motion controller for pnp ? multiple axis needed !!

1,551 views
Skip to first unread message

Alex Angel

unread,
Apr 10, 2017, 11:58:08 PM4/10/17
to OpenPnP
Hello all,

I am looking to build a pnp machine with following requirements but can not find a all in one solution controller.



plenty additional inputs and outputs for other sensors 



So I need a board that allows me to connect AT LEAST 16 stepper motor inputs
plenty of additional inputs and outputs for sensors 
 at least 8/16 solenoid valve relay and perhaps even pressure sensors to match 

Does anybody here know a board or multiple boards that I can use and is compatible with OPEN PNP software?

Because I require so many stepper motor inputs I can not use smoothstepper or similar products. 

Please suggest something as cheap and as simple as possible. 

greets Alex

Daniel Dumitru

unread,
Apr 11, 2017, 4:53:09 AM4/11/17
to ope...@googlegroups.com
It has been discussed here in this group about a controller based on smoothie with configurable/pluggable connection boards. That may fit your requirements



EVen so , having 16 motors , 16 inputs  and so on I think that it's much over something free..
You are going to professional area of machines.
Are you willing to retrofit a machine ? 

Br,
Daniel

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ddad2689-9698-4e8e-82c8-abf53b1ccfb3%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Arthur Wolf

unread,
Apr 11, 2017, 5:23:19 AM4/11/17
to ope...@googlegroups.com
Do you need the "extra" axes ( other than XYZ ) to move at the same time ? Or do they always move alone ?
Because if they don't have to move together, you can just use a multiplexer to select which motor is active at any given time ( receives step/dir instructions ).
It's easy to setup Smoothie to control the multiplexer using Gcode ( smoothieware.org/switch ).
I think this is by far the easiest path to what you want.
( I'm assuming here you can make a PCB to fit your needs, this is open-pnp after all :) )

Also, Smoothie is going to be able to do what you want out of the box in the future : we are working on a system that allows to add any number ( 16, 50, whatever ) of axes to the board.
But it's not ready yet : https://docs.google.com/document/d/1yJnCG1fGhp64-zeYBpuORSVDRx_d7kL48Qw4BJoBYs4/edit#


--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/ddad2689-9698-4e8e-82c8-abf53b1ccfb3%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
Courage et bonne humeur.

oliver jackson

unread,
Apr 11, 2017, 6:15:53 AM4/11/17
to ope...@googlegroups.com
Great idea Arthur thanks. Multiplexing the rotation axes makes a lot of sense. Don't even need to worry about the position of the motors not currently in use as vision will find the position when it's time to rotate 
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

Alex Angel

unread,
Apr 11, 2017, 2:37:12 PM4/11/17
to OpenPnP
I am either looking for 1 single board or multiple boards of the same type combined together. 
Does OPEN PNP software support multiple boards at the same time ? For example I connect 2 boards through usb , will OPEN PNP use them both at same time ? Is this possible ?

I saw this, looking for something similar to this. But Im not familiar with this product so I don't know if it can support 16 axis with 1 controller board or if I need to buy 2 of them. It also have different breakout boards for sensors etc ..

https://www.poscope.com/products/pokeys-devices/

You guys only seem to talk about the smoothieboard. Is there a  list of supported devices for this software ? Who can I ask directley about the supported hardware and it's limitations or add ons.. ?

For more details about what I am trying to build:

I need the following :

4 motors for X axis ( meaning 1 single stepper motor port will do)
2 motors for y axis ( 1 single stepper motor port will do)

8 motors for the nozzle rotations ( 8 nozzles = 8 seperate nozzle rotation motors so I need 8 stepper motor ports for this part )
4 motors for the nozzle z movement  (1 motor for 2 nozzles  - motor turning to the left will push down the 1st nozzle , motor turning to the right will push down the second nozzle  ---- so for this I need 4 stepper motor ports )

so In total I got 14 motor ports that I need , than I would of course like some extra motor ports for let's say maybe a automatic focusing lens for the vision camera's?

*So now you know why I need 16 or more motor control inputs that run together.

*than I need at least 8 relais for the 8 nozzles to use with the vacuum valves ? ( Or do I need 16 suction and pressure ? )

*8 pressure sensors

*6 or 9 sensor inputs for end stop and homing ( 2 or 3 for each axis  - 2 at each axis end and 1 for the homing)

What else am I missing ? maybe some more relais for the vacuum pump itself ? controlling the leds for the vision camera etc ..

I need to start working with a board or multiple boards that are EASY to implement into the  OPEN PNP software and that are very reliable 

For the feeder system I will make an automatic independent system so it will not be attached to the software what so ever.

Greets Alex



Jason von Nieda

unread,
Apr 11, 2017, 2:42:03 PM4/11/17
to OpenPnP
Alex,

OpenPnP can communicate with multiple boards, but there is currently a limitation that all primary motion, i.e. movement in X, Y, Z or C (Rotation) must happen via a single serial port. Other things like relays, lighting, solenoids, pumps, etc. can all be on other boards and ports.

So, if you need that many steppers, it will be best if you can get them to all communicate over the same serial port. Whether that is via one board or multiple boards that talk to each other doesn't matter, but OpenPnP must only need to talk to one.

This is only a limitation of the GcodeDriver, which is the main driver in OpenPnP that most people use to talk to their controllers. If you are willing to write or modify your own driver (using Java) you can do anything you like.

Jason


--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

Alex Angel

unread,
Apr 11, 2017, 5:02:08 PM4/11/17
to OpenPnP
So what about smooth stepper for example. This is one board but uses 2 serial ports or 3 if not mistaken ?

Does this means that if somebody using smooth stepper will not fully benefit of all the serial ports on the board ?
The board is connected to the pc through usb or lan and probably emulates multiple serial ports on hardware level.

What boards can somebody suggest to me that uses only 1 serial port and allows me to connect 16 motors ? or perhaps multiple breakout boards to 1 main board ?


greets Alex

Arthur Wolf

unread,
Apr 11, 2017, 5:08:21 PM4/11/17
to ope...@googlegroups.com
You can have 6 axes with one Smoothieboard, so assuming my multiplexing solution won't work for you ( I didn't see an answer on that ? ), then you could use 3 Smoothieboards.
Either via an OpenPNP driver than can talk to 3 boards, or via a small arduino nano that would dispatch the commands to the various boards ( this is probably like 30 lines of C++ ) over serial.
Another way of implementing this, is with one Main Smoothieboard, and two slave Smoothieboards ( being talked to over serial ), we have a feature planned to implement this but it's not fully ready yet.

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Alex Angel

unread,
Apr 11, 2017, 5:55:48 PM4/11/17
to OpenPnP
Hello,

Ok this sounds more promising. I am very newbie at this . I can only understand it once I can make it work and see it myself. At the moment is all a bit too much.
Basically I need to know with what boards I should go and most important boards that are EASY to use and have full support on OPEN PNP software.

I am not familiar with smoothie board , and definitely not familiar with C++ or any programming language. I am more a electrical, hardware and mechanical guy with some particular understanding of electronics like microchips.
But I don't have very deep knowledge or hands on experience on how to do stuff right away ..

What about pokeys ? They have pokeys mainboard and also pookers extension boards that seem very user friendly and easy to work with as well as compact in size.

Further I know a little bit about smoothstepper , and also I saw something about raspberry

https://www.poscope.com/product/poextbusoc16cnc/
https://www.antratek.be/automation-hat-for-raspberry-pi

 There are so many boards out there .. I will be using 0.9 deg stepper motors on dsp drivers
and I need or love to work with something that is easy upgradable and tons of addons like arduino ... just an example. 

The smoothie board seems to be all in 1 product and not modular ..

But if not choice ... than it will do ..

Greets Alex

Op dinsdag 11 april 2017 23:08:21 UTC+2 schreef Arthur Wolf:
You can have 6 axes with one Smoothieboard, so assuming my multiplexing solution won't work for you ( I didn't see an answer on that ? ), then you could use 3 Smoothieboards.
Either via an OpenPNP driver than can talk to 3 boards, or via a small arduino nano that would dispatch the commands to the various boards ( this is probably like 30 lines of C++ ) over serial.
Another way of implementing this, is with one Main Smoothieboard, and two slave Smoothieboards ( being talked to over serial ), we have a feature planned to implement this but it's not fully ready yet.
On Tue, Apr 11, 2017 at 11:02 PM, Alex Angel <angelf...@hotmail.com> wrote:
So what about smooth stepper for example. This is one board but uses 2 serial ports or 3 if not mistaken ?

Does this means that if somebody using smooth stepper will not fully benefit of all the serial ports on the board ?
The board is connected to the pc through usb or lan and probably emulates multiple serial ports on hardware level.

What boards can somebody suggest to me that uses only 1 serial port and allows me to connect 16 motors ? or perhaps multiple breakout boards to 1 main board ?


greets Alex

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

Arthur Wolf

unread,
Apr 11, 2017, 6:00:56 PM4/11/17
to ope...@googlegroups.com
Hey.

On Tue, Apr 11, 2017 at 11:55 PM, Alex Angel <angelf...@hotmail.com> wrote:
Hello,

Ok this sounds more promising. I am very newbie at this . I can only understand it once I can make it work and see it myself. At the moment is all a bit too much.
Basically I need to know with what boards I should go and most important boards that are EASY to use and have full support on OPEN PNP software.

As far as I know, Smoothieboard is by far the best supported board for OpenPNP.
 

I am not familiar with smoothie board , and definitely not familiar with C++ or any programming language. I am more a electrical, hardware and mechanical guy with some particular understanding of electronics like microchips.
But I don't have very deep knowledge or hands on experience on how to do stuff right away ..

What about pokeys ? They have pokeys mainboard and also pookers extension boards that seem very user friendly and easy to work with as well as compact in size.

Further I know a little bit about smoothstepper , and also I saw something about raspberry

https://www.poscope.com/product/poextbusoc16cnc/
https://www.antratek.be/automation-hat-for-raspberry-pi

 There are so many boards out there .. I will be using 0.9 deg stepper motors on dsp drivers
and I need or love to work with something that is easy upgradable and tons of addons like arduino ... just an example. 

The smoothie board seems to be all in 1 product and not modular ..

The Smoothie firmware that runs on the smoothieboard is by far the most modular CNC firmware you will find anywhere.
It's been designed that way, we've been making it even more modular for 5 years, and there's nothing like it around.

The board itself breaks out all pins, and the firmware allows you to do *a lot* of different things with those pins.
I can tell you for sure Smoothie is the main choice of most people who want to do something "weird"/unusual ( which seems to be your case ), because of the extreme modularity.
Maybe read smoothieware.org ?

I don't think you answered the question I asked twice :
What level of concurent motion do you need ?
You have 8 nozzles, do they need to work at the same time, or are you fine with them moving one after the other ?

Cheers.

 

But if not choice ... than it will do ..

Greets Alex

Op dinsdag 11 april 2017 23:08:21 UTC+2 schreef Arthur Wolf:
You can have 6 axes with one Smoothieboard, so assuming my multiplexing solution won't work for you ( I didn't see an answer on that ? ), then you could use 3 Smoothieboards.
Either via an OpenPNP driver than can talk to 3 boards, or via a small arduino nano that would dispatch the commands to the various boards ( this is probably like 30 lines of C++ ) over serial.
Another way of implementing this, is with one Main Smoothieboard, and two slave Smoothieboards ( being talked to over serial ), we have a feature planned to implement this but it's not fully ready yet.

On Tue, Apr 11, 2017 at 11:02 PM, Alex Angel <angelf...@hotmail.com> wrote:
So what about smooth stepper for example. This is one board but uses 2 serial ports or 3 if not mistaken ?

Does this means that if somebody using smooth stepper will not fully benefit of all the serial ports on the board ?
The board is connected to the pc through usb or lan and probably emulates multiple serial ports on hardware level.

What boards can somebody suggest to me that uses only 1 serial port and allows me to connect 16 motors ? or perhaps multiple breakout boards to 1 main board ?


greets Alex

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/a89bc8ca-803d-41d3-936d-9823a115f0e4%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.



--
Courage et bonne humeur.

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Alex Angel

unread,
Apr 11, 2017, 6:35:42 PM4/11/17
to OpenPnP
hy,

The nozzles will of course work one after another , they can not work in the same time because its almost impossible to pick at the same time .. and to place at the same time because of the space between the nozzles.
So there will always have to be some motion in between. The same goes for the rotation of the nozzles right ? The rotation will be perhaps on the go I don't exactly see how this can happen all at once.

Like the x and Y axis movement this can happen at once yes for curved movement etc .. But the other movement I see them pretty linear .
The use of 8 nozzles will reduce the time needed to pick and place an item so it will just carry more items at once closer to the board location but it will never place more than 1 item at the same time , unless there is a bug in the software and the position of 2 nozzles correspond with the place position and the software will try to place them both at once . But even so , that will be when the head assembly will be standing still ..

Further the Z axis movement uses 1 motor for 2 linear guides if the same motor turns left than 1 guide moves down , the other guide moves up. If it the motor turns right than the other rail carriage will move down and the other up.
When they are neutral the motor is somewhere in the middle. This is something I came up with and have no idea how to implement this yet or what kind of motion or programming I need to do in the software.

I also need allot of sensors ... use of plenty of optical sensors to make sure of perfect calibration and I would DEFINITELY want in the near future to implement encoders on the all axis including rotation , and Z movements so there are allot of encoders do to. 

Can you back up your main suggestion ? What would be your second choice ? And your third ?
What should I avoid ? IF you can argument this I would very much appreciate . No need of saying I'm trying to make a cheap as possible pick and place machine that would run in a local shop pretty good.

greets Alex

Paul Kelly

unread,
Apr 11, 2017, 6:41:04 PM4/11/17
to ope...@googlegroups.com

Alex, I’m assuming that you want all the extra steppers for feeders?

If so, you would be best off making a small board with a driver for each stepper and using some kind of serial bus to control them. If you search back you’ll find that this has been discussed previously on the list. IIRC the differential signalling protocols (RS485, RS422) came out on top because they were cheap and noise immune.. SoC stepper drivers (for small motors) are ubiquitous simple to implement, and cheap. 

 

Then you just end up with a small control board plugging into the PC and enumerating as a serial port. It takes messages from and OpenPnP driver or sub driver and sends commands to the appropriate address on your serial bus..

 

PK

--

Arthur Wolf

unread,
Apr 11, 2017, 6:41:39 PM4/11/17
to ope...@googlegroups.com
On Wed, Apr 12, 2017 at 12:35 AM, Alex Angel <angelf...@hotmail.com> wrote:
hy,

The nozzles will of course work one after another , they can not work in the same time because its almost impossible to pick at the same time .. and to place at the same time because of the space between the nozzles.
So there will always have to be some motion in between. The same goes for the rotation of the nozzles right ? The rotation will be perhaps on the go I don't exactly see how this can happen all at once.

Then this is pretty easy to do with a Smoothieboard, a multiplexer ( or even without one, just a bunch of transistors would work ), and the switch module ( smoothieware.org/switch ) to map pins to gcodes.
Simply do the following for your 8 nozzle axes :
* Configure Smoothie so that the A axis's step/dir pins are any given pins you choose
* Wire those two pins in parralel to the step/dir inputs of 8 stepper drivers
* Cut the step pins of each with a transistor
* Wire the input of each transistor to a pin on the smoothieboard ( or use a multiplexer to use 3 pins instead of 8 )
* Configure switch modules ( in smoothie config, this is where smoothie's modularity is neat ) so you can switch each pin on/off with gcode
* Configure openpnp to output those gcodes whenever needed ( not sure how configurable that is right now )
* Now with gcodes you can select which stepper driver receives the "step" pulses at any given time and you can control the rotation of all 8 no problem.

Makes sense to you ?

Like the x and Y axis movement this can happen at once yes for curved movement etc .. But the other movement I see them pretty linear .
The use of 8 nozzles will reduce the time needed to pick and place an item so it will just carry more items at once closer to the board location but it will never place more than 1 item at the same time , unless there is a bug in the software and the position of 2 nozzles correspond with the place position and the software will try to place them both at once . But even so , that will be when the head assembly will be standing still ..

Further the Z axis movement uses 1 motor for 2 linear guides if the same motor turns left than 1 guide moves down , the other guide moves up. If it the motor turns right than the other rail carriage will move down and the other up.
When they are neutral the motor is somewhere in the middle. This is something I came up with and have no idea how to implement this yet or what kind of motion or programming I need to do in the software.

I also need allot of sensors ... use of plenty of optical sensors to make sure of perfect calibration and I would DEFINITELY want in the near future to implement encoders on the all axis including rotation , and Z movements so there are allot of encoders do to. 

Can you back up your main suggestion ? What would be your second choice ? And your third ?
What should I avoid ? IF you can argument this I would very much appreciate . No need of saying I'm trying to make a cheap as possible pick and place machine that would run in a local shop pretty good.

greets Alex

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Alex Angel

unread,
Apr 11, 2017, 8:21:46 PM4/11/17
to OpenPnP
hello,

Yes I understand the part with the transistor , this acts like a that chip they used on nintendo for getting  the different input signals through one port or something like that forgot the name is kinda late here ...
I am looking for plug and play modules because if I need to make custom hardware it will take even more time to complete and debug everything and it will also be probably more expensive.

But will the stepper motors work perfect with no interruption or delay? Like I mentioned the stepper are 0.9 deg stepper so they get like min 400 pulses/RPM
They operate smooth when connected to a dsp driver but I have no idea if this will have any impact on the resolution as losing steps is definitely out of the equation.

For PK :
No this steppers are used for nozzles z axis , nozzles rotation and motion of the x and Y axis , for the feeder system I will use an independent system that will automatically trigger when a nozzle picks something.

Greets Alex

Op woensdag 12 april 2017 00:41:39 UTC+2 schreef Arthur Wolf:


To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

SMdude

unread,
Apr 11, 2017, 8:22:15 PM4/11/17
to OpenPnP
Hi Arthur,
It is my understanding that the smoothie needs to keep track of the position of each axis including the nozzles.

If we keep moving say A axis the positions will all end up wrong as we need to treat each nozzle as an axis of its own and smoothie needs to remember the position of each nozzle so when openpnp commands say nozzle c to rotate to 90 degrees it ends up in the correct position.

I am liking where this is going...

Cheers

Paul Kelly

unread,
Apr 11, 2017, 10:15:19 PM4/11/17
to ope...@googlegroups.com

 

I must admit, I’m still at a bit of a loss as to why everyone wants to use steppers/servo’s for z.  As I see it; it just adds quite a bit more complexity and creates problems (like having to program part heights) where there needn’t be problems.

Re R axis, you’ll only be placing one part at a time, so you can just slave all the rotary axii yes?

 

PK

Juha Kuusama

unread,
Apr 11, 2017, 10:31:32 PM4/11/17
to OpenPnP
It is not here yet, but www.plug.ee is building a smoothieware compatible system with up to 15 slots. One board with all slots filled with stepper drivers, another for switches and the rest.

Alex Angel

unread,
Apr 12, 2017, 12:15:06 AM4/12/17
to OpenPnP
since it only place 1 part at the time then yes I don't need everything working at the same time togheter.

HOWEVER when you guys talk about slave or separate boards I wonder what is the delay on this ? are we talkin in milliseconds ? seconds? is there a synchronisation between the boards ? does one board waits for the other to complete?
I can go with multiple boards but I need to make sure everything runs smoothly and does not have glitches like sudden stops for performing something unless the nozzle is directly above the place point

greets

Op woensdag 12 april 2017 04:15:19 UTC+2 schreef PK:

Alex Angel

unread,
Apr 12, 2017, 1:18:51 AM4/12/17
to OpenPnP
What about a FPGA motion controller? Any with 16 or more stepper outputs ? As it seems FPGA could provide easily multiple axis control and tons of programmable inputs and outputs 
This seems to be a nextgen thing in motion controllers maybe next step for diy ...


greets Alex

Op woensdag 12 april 2017 04:31:32 UTC+2 schreef Juha Kuusama:

Daniel Dumitru

unread,
Apr 12, 2017, 1:36:00 AM4/12/17
to ope...@googlegroups.com
What a discussion, started by a question.

Angel , please tell us what machine would you like to make .
Do you have already one PnP machine now ?  

If not , my advice is to make first a simple machine with one head and strip feeders. 
After that go further to more complicated features.

Best choice for controller it's Smoothie.

Br,
Daniel
 

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

Juha Kuusama

unread,
Apr 12, 2017, 7:42:40 AM4/12/17
to OpenPnP
Sorry, it is https://plugg.ee/ (with one g, it goes to some Estonian site).Their campaign is at https://www.crowdsupply.com/plugg-ee-labs/juicyboard.

On Wednesday, April 12, 2017 at 5:31:32 AM UTC+3, Juha Kuusama wrote:

Arthur Wolf

unread,
Apr 12, 2017, 7:54:45 AM4/12/17
to ope...@googlegroups.com
Hey.

On Wed, Apr 12, 2017 at 2:21 AM, Alex Angel <angelf...@hotmail.com> wrote:
hello,

Yes I understand the part with the transistor , this acts like a that chip they used on nintendo for getting  the different input signals through one port or something like that forgot the name is kinda late here ...
I am looking for plug and play modules because if I need to make custom hardware it will take even more time to complete and debug everything and it will also be probably more expensive.

This is pretty trivial hardware to make, I'm pretty sure if you search a bit you can even find existing hardware that can be adapted to do this ( logic gates breakouts for example )

But will the stepper motors work perfect with no interruption or delay?

Yes.
 
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Trampas Stern

unread,
Apr 12, 2017, 11:15:33 AM4/12/17
to OpenPnP
Alex,

How big will your machine be? 

One of the issue you will find is that the more heads you have the less effective working area you have.  For example on my machine I used 600mm x 600mm rails for the X and Y thinking this would provide a huge working space. However when I added the camera and two heads it turned out the the distance between heads (and camera) was around 60mm. What this means that if I have feeder I want too use with both heads and camera it must be inside my effective working area which was reduced 60mm on each side so my 600mm machine became 480mm x 480mm.  Then if you include room for end stops and such becomes closer to 450mm x 450mm.  I was thinking of doing a 4 head system in which case the working space that all heads and camera could see would reduce even further. 

Unless you do something cleaver with the 8 heads you will end up with even further reduction of effective working space or a big machine... 

I personally made the mistake in thinking that building a PnP would be simple, and I ran into problems I could never have predicted.  I have taken my machine apart and reassembled multiple times trying to address issues, and I wish I had just went and purchased some openbuild rails and made a disposable prototype first.  That is I have found so many issues with camera setups, head setups, mechanics Z height issues, configuring software, software bugs, and so forth that by having built the straw man prototype first I would have learned these problems and done a far better design on the final machine.

Another mistake I made was in assuming OpenPnP was turn key and production ready. That is I figured I would start up OpenPnP run through the manual setting things up and be placing parts in a day after my mechanics were done, I was wrong. There is a lot of learning and debugging you will do when you start trying to use OpenPnP, do not under estimate this.   For example simple things like using strip feeders can be a problem, if you use vision alignment on strip feeders OpenPnP can get confused if strips are too close.  These little things are things you want to learn before doing a larger design.  Note the whole vision aspect is hard as every machine is different as is the lighting where your machine is, so don't expect vision to work out of the box. 

OpenPnP has a massive amount of options and configurations to support a wide variety of machines, this means it can be difficult to figure out the correct XML foo to make it work. Additionally if you have any mechanical issues with your machines, like backlash, squareness, missing steps, mm/step settings wrong, etc it will compound the setting up of OpenPnP as you will not know if the software is broken or if the hardware is the issue.  So you need to make sure your mechanics are rock solid before starting OpenPnP, that is measure repeatably, squareness, etc before ever starting OpenPnP.   Even little things like how to setup the lights for the camera makes a big difference in machine performance, so it is best to learn these before tackling larger problems.  

I still have not gotten OpenPnP running on my hardware, I have stopped trying for the moment and started writing some Python scripts for testing out the hardware reliability/repeatably before attempting OpenPnP again.  When something does not work, I want to be able to run a "test mechanics" script to make sure it is not a mechanical problem before starting to debug software. 

Do not make your first design complicated and push the envelop, get a simple machine running with one or two heads, this will save you time and money on your final build.



Hey.

Bernd Walter

unread,
Apr 12, 2017, 11:57:04 AM4/12/17
to OpenPnP


On Wednesday, April 12, 2017 at 5:15:33 PM UTC+2, Trampas Stern wrote:
Alex,

How big will your machine be? 

One of the issue you will find is that the more heads you have the less effective working area you have.  For example on my machine I used 600mm x 600mm rails for the X and Y thinking this would provide a huge working space. However when I added the camera and two heads it turned out the the distance between heads (and camera) was around 60mm. What this means that if I have feeder I want too use with both heads and camera it must be inside my effective working area which was reduced 60mm on each side so my 600mm machine became 480mm x 480mm.  Then if you include room for end stops and such becomes closer to 450mm x 450mm.  I was thinking of doing a 4 head system in which case the working space that all heads and camera could see would reduce even further. 

I'm building with 4 picker, because I'd wanted to have 4 different nozzle tips mounted.
Seemed easier than to go with a nozzle changer.
What I didn't though about is the influence it has on feeders.
If you need to pick something with the right nozzle on a left feeder bank your feeder needs to have enough room for the left 3 nozzles (and maybe components already hanging on them).
I have yet to verify, but it seems that left/right feeder banks seems to be impossible with the Yamaha feeders, so I might have to go with a single front feeder bank instead.
The back side is also impossible to use because I can't go far enough to the back that the feeders wouldn't collide with the frame.
Fortunately I have >800x800mm movement, because I'd started with 1000mm rails.
However I'm using 12mm supported rods, but they are already bending a bit with the heavy head and I might need to add more support.
The height added by using big round rods instead of flat rails added some other challenges because my picker had only has a short Z path.

So many things to consider.
In the end I've already decided to build this machine machine to completion for the learning effect and start a new one with the new knowledge.

Trampas Stern

unread,
Apr 12, 2017, 2:43:34 PM4/12/17
to OpenPnP
One thing to consider on the Z height is the camera focus.  That is you want everything that uses the camera to be at the same Z height, ie PCB strip feeders, tray feeders, etc.  Also if you do double sided boards this working Z height has to be high enough to accommodate components on the back side board being populated. 

Also if you use vision the head camera needs a good mount, that is error between nozzle and camera causes errors in placement when using vision. 

On the 4 heads things you can assume the end reels are only for the end heads but it is limitation. I have also assumed for reel feeders the camera does not need to see the feeders.. 

Yes I keep trying to get mine working, and I will get there but I already see a need for another machine.  My dream is to have cheap reel feeders and a really large (8' wide) machine such that I never take parts off machine, however that is not really practical.  I do a lot of prototype builds, so I have a database of parts I work from, and by having all on the machine I can do one off builds easy (assuming a solder paste dispenser too). Again not practical but it is nice to dream.   It would be nice to have another machine for large volume production as well. 

Trampas 

Alex Angel

unread,
Apr 12, 2017, 9:45:39 PM4/12/17
to OpenPnP
hello Trampas,

Thanks for the reply. I have somehow experience with failures and errors so I do not underestimate this pnp setup.

The machine size will be 1500mm x 1000mm I have already made the design a little in autocad . Off course there will be plenty of unknown factors and adjustments.

The most important thing is choosing the right controller. That's why I started this forum question.

I am searching for a cheap motion controller that is modular and uses open source software, easy to integrate easy to upgrade etc.
Must have as many stepper or servo motor ports available , or even both. I need at least 16 steppers. They don't have to run all at once, but they do have to have a quick reaction time in milliseconds.
I saw smoothie , I saw pookeys , arduino .. I saw smoothstepper , KFLOP and than I started looking for FPGA motion controllers they usually allow you to have many gpios and motor ports .

What I am NOT looking for is for example motion controller boards from GALIL or similar. They are probably the best and have 16 axis but are extremely expensive. I have an old ISA slot card from galil 4 axis card I once bought it for cheap to integrate it into my cnc but never happend since I also needed special scsi cable 100pin and special breakout board etc.. which were more expensive then the card itself

Im looking for something between 50 - 350 usd , 400 would be too much just for the motion controller only.
If it would be 350 it would have to be an extremely user friendly design with pro features and easy part swapping etc .. just for example

greets Alex

Daniel

unread,
Apr 13, 2017, 7:00:18 AM4/13/17
to OpenPnP
Arthur , on this application there is not needed LAN. 

On RAy's board LAN pins are available as external connector.

Just one question for you :

In smoothie it's possible to declare/use 6 extruders ?

IF possible than we need to check with JAson if we can use extruder type outputs instead of aditional Z

Daniel

PS : but I don't understand why are needed 4 nozzles. My machine stays unused 98% of time ...

Arthur Wolf

unread,
Apr 13, 2017, 7:07:56 AM4/13/17
to ope...@googlegroups.com
On Thu, Apr 13, 2017 at 1:00 PM, Daniel <dand...@gmail.com> wrote:
Arthur , on this application there is not needed LAN. 

On RAy's board LAN pins are available as external connector.

Just one question for you :

In smoothie it's possible to declare/use 6 extruders ?

Yes, by modifying the makefile and being careful of RAM usage.

But if you are doing that many extruders, I strongly recommend using the multiplexer solution I mentionned, it's just a few $ worth of electronics available anywhere, and it's much simpler.

IF possible than we need to check with JAson if we can use extruder type outputs instead of aditional Z

Daniel

PS : but I don't understand why are needed 4 nozzles. My machine stays unused 98% of time ...


On Wednesday, April 12, 2017 at 12:08:21 AM UTC+3, Arthur Wolf wrote:
You can have 6 axes with one Smoothieboard, so assuming my multiplexing solution won't work for you ( I didn't see an answer on that ? ), then you could use 3 Smoothieboards.
Either via an OpenPNP driver than can talk to 3 boards, or via a small arduino nano that would dispatch the commands to the various boards ( this is probably like 30 lines of C++ ) over serial.
Another way of implementing this, is with one Main Smoothieboard, and two slave Smoothieboards ( being talked to over serial ), we have a feature planned to implement this but it's not fully ready yet.

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Daniel Dumitru

unread,
Apr 13, 2017, 7:23:38 AM4/13/17
to ope...@googlegroups.com
In smoothie it's possible to declare/use 6 extruders ?

Yes, by modifying the makefile and being careful of RAM usage.

But if you are doing that many extruders, I strongly recommend using the multiplexer solution I mentionned, it's just a few $ worth of electronics available anywhere, and it's much simpler.


In the very first second I considered the multiplexing idea brilliant. And I was thinking to multiply nozzle rotation.

However , during operation without uplooking camera , nozzle it's rotating until placement.
And in this case would be needed to be able to rotate all nozzles on same time...

BR,
Daniel 

Arthur Wolf

unread,
Apr 13, 2017, 7:28:05 AM4/13/17
to ope...@googlegroups.com
Rotation is extremely fast, you could just rotate them in sequence before moving, or at the target, it'd have an extremely small time cost.

BR,
Daniel 

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Trampas Stern

unread,
Apr 13, 2017, 7:48:49 AM4/13/17
to OpenPnP

If you are good at Java programming just change OpenPnP to work with multiple controller board boards.  This is on my personal wish list as I would like to use nano zero steppers without a controller. Hence each motor has it's own motor controller and own serial port. 



On Thursday, April 13, 2017 at 7:28:05 AM UTC-4, Arthur Wolf wrote:
On Thu, Apr 13, 2017 at 1:23 PM, Daniel Dumitru <dand...@gmail.com> wrote:
In smoothie it's possible to declare/use 6 extruders ?

Yes, by modifying the makefile and being careful of RAM usage.

But if you are doing that many extruders, I strongly recommend using the multiplexer solution I mentionned, it's just a few $ worth of electronics available anywhere, and it's much simpler.


In the very first second I considered the multiplexing idea brilliant. And I was thinking to multiply nozzle rotation.

However , during operation without uplooking camera , nozzle it's rotating until placement.
And in this case would be needed to be able to rotate all nozzles on same time...

Rotation is extremely fast, you could just rotate them in sequence before moving, or at the target, it'd have an extremely small time cost.

BR,
Daniel 

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

Matt Brocklehurst

unread,
Apr 13, 2017, 7:58:02 AM4/13/17
to ope...@googlegroups.com
I personally don't like the idea of multiple simple controllers - how can you guarantee everything stays in sync, you don't want for example want the z dropping whilst the x y is still moving (potential for collision)

Sent from my iPhone

Trampas Stern

unread,
Apr 13, 2017, 8:55:42 AM4/13/17
to OpenPnP
This is one of the reasons I want multiple controllers.... 

So basically with a smoothie boards if you ask to move from (0,0) to (100,100) it will move at the rate for the slowest axis ensuring a linear line path. That is it assume the X,Y, and Z axis are dependent on each other, for CNC machines even the A,B,etc are  all dependent with X,Y, and Z. For a Pick and place the axis are not dependent, but rather independent.  That is as long as the Z is at safe height you can  take any path you want to get to desired location, just do it as fast as possible (up to point part fly off nozzle) 

For staying in sync this is the job of OpenPnP.  For example with a nano zero stepper I can query the z controller and get it's current height before moving in order to know it is safe to move. Then I can tell X axis to move, then Y axis. Then I can poll the X and Y controllers to know when the movement is complete (low error in position). Then drop Z axis to pick part, poll controller to know when done, raise Z to safe height and poll when done, then move X an Y again.  
Currently OpenPnP already has to maintain this synchronization on the Z axis. However OpenPnP sends a move command with X,Y,and Z which actually increases the probability that Z is moving during X and Y.   I actually found a bug where this was happening a few weeks ago. 

Currently as I OpenPnP does not have the X and Y move commands independent, if you sent command to move X axis, then sent command to move Y the controller would not move the Y axis until the X move was complete. So it has to send the X and Y together so both axis are moving at same time, for fastest moves.  With independent controller this limitation is removed. 

Also in OpenPnP currently a move command is usually followed by M400 command for smoothie. The 'M400' means it waits until queue is empty (assumption OpenPnP makes is this is wait until move is done). This again is a problem as the move is not finished when M400 returns, I actually ran into this problem with doing testing in python.  The only way around it was to put in a delay wait after the M400 returns, and assume that xx milliseconds latter the move has actually completed, if you have fast machine this delay has to be larger.  This does not affect the pick and place as that if you follow the XY move by a Z move it will complete the X and Y move before the Z, but it does mean there is a wait after move before you can use vision.  So again with independent controllers and feedback you do not need the delay between moves and use of vision as you know when the move is done. 

Again OpenPnP currently implements the synchronization between axis, and with multiple controllers it still would have this requirement. 
Using independent controllers with  feedback (and changes to OpenPnP) can reduce the possibility of head collisions and improve machine speed. 
Message has been deleted

Alex Angel

unread,
Apr 13, 2017, 12:36:04 PM4/13/17
to OpenPnP
Op donderdag 13 april 2017 14:55:42 UTC+2 schreef Trampas Stern:

Lisandro B

unread,
Apr 13, 2017, 12:42:51 PM4/13/17
to OpenPnP
Take a look at the Mesa boards, have one in my CNC for years now and never let me down, cheap, powerful and open source

Arthur Wolf

unread,
Apr 13, 2017, 12:46:29 PM4/13/17
to ope...@googlegroups.com
Are you sure those are Open-Source and not just "compatible with Open-Source LinuxCNC" ?

On Thu, Apr 13, 2017 at 6:42 PM, 'Lisandro B' via OpenPnP <ope...@googlegroups.com> wrote:
Take a look at the Mesa boards,  have one in my CNC for years now and never let me down,  cheap, powerful and open source
--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Alex Angel

unread,
Apr 13, 2017, 10:48:02 PM4/13/17
to OpenPnP
hello,

Ok does anybody have other suggestions ? and please state why

Something besides smoothieboard. Something that uses FPGA ? 
Anybody experience with pookeys ? arduino ? raspberry ?

greets

Op dinsdag 11 april 2017 05:58:08 UTC+2 schreef Alex Angel:
Hello all,

I am looking to build a pnp machine with following requirements but can not find a all in one solution controller.



plenty additional inputs and outputs for other sensors 



So I need a board that allows me to connect AT LEAST 16 stepper motor inputs
plenty of additional inputs and outputs for sensors 
 at least 8/16 solenoid valve relay and perhaps even pressure sensors to match 

Does anybody here know a board or multiple boards that I can use and is compatible with OPEN PNP software?

Because I require so many stepper motor inputs I can not use smoothstepper or similar products. 

Please suggest something as cheap and as simple as possible. 

greets Alex

Paul Kelly

unread,
Apr 14, 2017, 12:09:09 AM4/14/17
to ope...@googlegroups.com

My machines have 3 heads, 80 feeders and use 3 axis (ok 4 on the one with the nozzle changer) they do about 2000cph.

Even if the heads were driven by steppers, that’d only be 6 axii..

 

I really think that that you are looking at the problem in a way that just makes it harder than it needs to be.

PK

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.


To post to this group, send email to ope...@googlegroups.com.

Cri S

unread,
Apr 14, 2017, 12:23:21 AM4/14/17
to OpenPnP

Something besides smoothieboard. Something that uses FPGA ? 

Mesa was already mentioned.
 
Anybody experience with pookeys ? arduino ? raspberry ?

yes, but this configuration are not supported officially be OpenPnP, same like FPGA. short answer, don't think about it.
longer answer, if you want it you need to have the skills for doing that, know someone that do the work for you, ...

Use any board you are able to buy and program.
You are able to drive 100 rotation nozzle with one motion controller output, you just need at least 50 separate drivers for that.
For multiplexing, Z multiplexing works, but Rotational axis multiplexing require Code modification to OpenPnP.

There exist nice I2C motor driver including motion controller that can be connected on the same bus,
http://www.ulrichradig.de/home/index.php/projekte/tmc222-controller

For rotation there are ideal and can be connected to smoothie/... board, again, you need to be able to modify you'r selected motion controller, in order to implement the M code to i2c command sequence.


DAniel Dumitru

unread,
Apr 14, 2017, 12:24:44 AM4/14/17
to ope...@googlegroups.com
Alex,
Again start with 2 heads and optimize it later.

IT would be a good idea to read OpenPNP documentation (I don't do it enough either). Motion controller board has to be able to interpret Gcodes directly and execute them.

POKyes- I haven't looked to that lately , but was working with a Plugin for MAch3 and MAch4  -as far as I know didn't interpret g-codes directly.
raspberry - there is no distribution real time for ARM that would be able to drive motors + RPI has very few output pins, no way to control "plethora" of in/out that you have requested
arduino - smoothie it's better than all versions of arduino motion controllers , even cortex M3 versions

But if you like to know the market :
- look on KFlop  - nice board but does not interpret directly G-Codes , has powerful MCU + fpga, up to 8 axes
- beagleboard black cnc  only 4 axis but high speed

Unless you have very good skills on programming embedded  and you are willing to spend a considerable amount of time on this, my suggestion it's to stay to Smoothie.

Br,
Daniel
--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/5d12f820-ab00-4a74-89e1-0fe025c174c1%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



Virus-free. www.avast.com

Arthur Wolf

unread,
Apr 14, 2017, 8:13:45 AM4/14/17
to ope...@googlegroups.com
On Fri, Apr 14, 2017 at 4:48 AM, Alex Angel <angelf...@hotmail.com> wrote:
hello,

Ok does anybody have other suggestions ? and please state why

Something besides smoothieboard.

I'm not sure why this is the third time you ask this, people keep telling you "just use smoothie" and you don't listen.
Smoothie is the official board for openpnp, it's the best supported, it's the easiest to adapt to "weird" things ( like you are doing ), and it does everything you want with just a bit of configuration and a few $ worth of electronics ...
Not sure what the obsession with an alternative is about. Maybe just try it, no ?
 
Something that uses FPGA ? 

Smoothieboard v2 will use a FPGA.
 
Anybody experience with pookeys ? arduino ? raspberry ?

greets

Op dinsdag 11 april 2017 05:58:08 UTC+2 schreef Alex Angel:
Hello all,

I am looking to build a pnp machine with following requirements but can not find a all in one solution controller.



plenty additional inputs and outputs for other sensors 



So I need a board that allows me to connect AT LEAST 16 stepper motor inputs
plenty of additional inputs and outputs for sensors 
 at least 8/16 solenoid valve relay and perhaps even pressure sensors to match 

Does anybody here know a board or multiple boards that I can use and is compatible with OPEN PNP software?

Because I require so many stepper motor inputs I can not use smoothstepper or similar products. 

Please suggest something as cheap and as simple as possible. 

greets Alex

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

Cri S

unread,
Apr 14, 2017, 10:36:58 AM4/14/17
to ope...@googlegroups.com
@Daniel, Kflop can execute Gcode, dll that can parse gcode and
translate it to kflop native
commands is available for free including visual basic examples.
>> email to openpnp+u...@googlegroups.com.
>> To post to this group, send email to ope...@googlegroups.com.
>> To view this discussion on the web visit https://groups.google.com/d/
>> msgid/openpnp/5d12f820-ab00-4a74-89e1-0fe025c174c1%40googlegroups.com
>> <https://groups.google.com/d/msgid/openpnp/5d12f820-ab00-4a74-89e1-0fe025c174c1%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>
>
> --
> Courage et bonne humeur.
>
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "OpenPnP" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/openpnp/LeOoxE6kHrI/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> openpnp+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/openpnp/CAMHZkm54ztAzO%2BoEFK9ncRb1-q9iS-ZPnWuD1_%2B0Tmp1KB7VEA%40mail.gmail.com.

Alex Angel

unread,
Apr 14, 2017, 4:55:38 PM4/14/17
to OpenPnP
So when will smoothie V2 come out ?

I can not test and build a pnp with 1 or 2 nozzles because :
I have recentley started a company for small line production.
I must build a PNP machine because it's much smarter and better and also allot cheaper than to buy one poor build desktop pnp machines.
It will also allow me for future upgrades or modify there were necessary.

Do I need to use 2 or 3 smoothie boards ? Or do I need 1 smoothieboard with different external boards ?
Personaly I think using 3 smoothieboards is just crazy plus all else that is needed we are than going to achieve prices that are simmilar to some poffesional motion controller cards

I see some in pretty good shape motion controllers with 16 axis and with few months warranty from mitsubishy , fanuc, baldor, allen bradley, siemens etc for under 700 usd.
As I am myself an open source supporter, the quality, build and research & development of such devices are probably much higher than any other hobbyst boards. As well as the stable and proven support that comes with it.

Again I must find the best solution for a small business, limited on budget but trying to get most of it. I can outsource parts from everywhere new or second hand doesnt matter.

Please advice..

Greets Alex


Jason von Nieda

unread,
Apr 14, 2017, 5:01:30 PM4/14/17
to OpenPnP
Alex,

I think you are asking for information that does not exist. No one, to my knowledge, has used OpenPnP with a machine as large as you are describing, so there is no advice to be given. If you want to use that many steppers with OpenPnP you will more than likely need to design your own motion controller and learn how to integrate it yourself.

This is why people keep recommending starting with a smaller machine. You are unaware of OpenPnP's features and limitations. Without understanding this, you may build a machine that simply won't work without extensive changes. If you are willing to make those changes, of course it's fine, but you should know going into the project that it will not be "plug and play".

The recommendations given to you so far take into account what will be needed to work with OpenPnP. It's good advice, and you should listen to it.

Jason


--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

Paul Kelly

unread,
Apr 14, 2017, 6:04:17 PM4/14/17
to ope...@googlegroups.com

When we bought our first pick and place machine we were a small business with only about AU$500K turn over. 

We spent AU$15K on a CSM84vz with about 60 feeders. We had to ship parts from around the world to get it.

I actually only paid about AU$1500 for the base machine, $10K on feeders and the rest on shipping…

This got as going with a machine we didn’t have to build, for a price we could afford.

 

Subsequently I bought the CSM 60 as a turn-key system, with feeders but a broken vision board for about $5K landed.

This machine worked fine for a year and is the one we are upgrading to OpenPnP now..

 

I would strongly recommend against starting a business that will rely on the first pick and place machine you ever build…

PK

 

From: ope...@googlegroups.com [mailto:ope...@googlegroups.com] On Behalf Of Alex Angel


Sent: Saturday, 15 April 2017 4:56 AM
To: OpenPnP <ope...@googlegroups.com>

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

fšk

unread,
Apr 14, 2017, 6:27:20 PM4/14/17
to OpenPnP
you could try arduino due with a custom breakout board for drivers and use aprinter firmware. im not 100% sure but i think it can drive  the number of axes you need. try asking the author on IRC. 

fšk

unread,
Apr 14, 2017, 6:40:23 PM4/14/17
to OpenPnP
Ups nope i was wrong it can not drive 16 for sure. Due doenst have enought timers for 16 axes... sorry.

i need to ask him about teensy 3.6. that could probably do what you want.

SMdude

unread,
Apr 14, 2017, 7:24:38 PM4/14/17
to OpenPnP
Spot on PK!

If you are going to be relying on the pnp for your income, buy one. You will pay it off in no time!
I chose to build mine because none of the commercial desktop machines could do pin pitch down to .3mm. I need this, so I decided rather than buy a machine that was not capable and make it capable(maybe??) I would build my own!
The amount of time I could have better spent designing my products and building things, I would have probably paid off a half decent machine by now!~
Luckily, I am not reliant  on this to make an income, but still it hurts to think about it!

Since I started looking into PnPs, the commercially available desktop machines have come a fair way.
If I were to buy one, I would personally go for one of the Smallsmt machines as their backup support and willingness to fix software(and produce good software to begin with) seems to be really good compared to Neoden, Charmhigh etc.

Its still pretty cool to have built a pnp machine! I will find out in the coming weeks if it can get those .3mm BGA's into position reliably.

That's my 2 bobs worth anyway....

Alex Angel

unread,
Apr 14, 2017, 7:57:56 PM4/14/17
to OpenPnP
I can't buy a new one

I have already made some hardware designs in autocad.
In the past I have made some cnc routers. The hardware is pretty modular and reliable. I do not use extruded aluminium profiles I am using cnc machined profiles and mounts etc..

I think my design is very solid and the precision will be very good. 
The only think I'm considering right now is the main controller board.
Like I said I even have a solid idea and plan for the feeders. That too won't be a problem.

The only problem I have to deal right now is the main controller board.
I will learn how to deal with it when the machine will be ready, for that I need first to make a BOM list. That's why I need to choose the right controller for my needs.
I explain in previous message that I find using 3 controllers a bit too much even that I'm not familiar with this smoothie board.
 I asked if there was a possibility of working with 1 single smoothie board and having other cheaper boards to extend so I can connect my number of axes needed.
Also when does the smoothie board V2 come out ? I have 3 months time to get this done.

What about the other controllers I mentioned ? The fanuc, siemens , mitsubishi , galil  etc .. 

Like I mentioned I was looking for spending about 350 top for something that would make me able to run 16 steppers or even more.
Those stepper do not need to work simultaneously. For the arc motion is only 2 axis , X and Y. The Z don't even have to run together because I rather wait until the machine moves to it's right position in order to avoid collisions with other parts if moving to quickly. 


greets

Paul Kelly

unread,
Apr 14, 2017, 8:25:07 PM4/14/17
to ope...@googlegroups.com

SMdude, you make some good points. There is a reason you can buy a 1990’s pnp machine for AU$1500, that’s because it can’t place anything smaller than 0603 and 0.5mm

 

Indeed one of the reasons (but not all of them) that we are retrofitting the CSM 60 is in the hopes that we’ll get better placement of smaller parts.  We’re not banking on it, but it’s an aspirational goal.

 

Back to “A PnP machine is going to be a key part of my business, how much should I spend?”

 

Folks, I know and respect that this is irrelevant stuff for those who want the challenge of building a pick and place machine, or like SMdude and I, are chasing an additional capability that doesn’t fit the model I present. Or you are trying to ‘open up’ PnP technology space.

If that’s you, then read no further.

 

The OP stated that this machine was to be part of his business. This is the (most common) way a business looks at capex.

 

Let’s say you are just starting out so your machine will be running 20% of the time. Ie 44 days a year.

Medium sized boards (say 50 components)

Slow placement rate (1000cph)

 

This works out to 44 X 8 x 1000 / 50 = about 7000 boards a year. Western world assembly costs for this size of board for short runs would be about $5   so your machine is going to earn you $35000 per year.

 

Point #1, there isn’t a lot of money in assembling electronics.. What you get by doing it in house is flexibility and more control of cash flow via smaller batches..

 

Pick and place machines last quite a while at these low utilisation rates so let’s say it will run for 5 years without major expense.

 

You just need to decide, how much of that $35000 will go into paying for your PnP hardware (I’m ignoring ovens and stencil printers etc because, at that rate of production, you can just make these)

I’m going to say 10%, but you may choose a different number.

 

So you have a budget of $3500 a year for your machine.

At 10% interest on a 5 year business loan you can afford $13500 for your machine

Alternatively, a $13500 machine would pay for itself in 5 months

 

So….

Point #2 you can get you a machine like ours with lots of feeders. And it will either cost you < $300 per month over 5 years, or pay for itself in 5 months.

 

Sure you might have to make some design compromises because your machine can’t place 0201 components.  But you were going to have to make those compromises anyway because nothing you can build cheaply will place those parts..

 

Go spend $13500 and move onto the really difficult parts of the problem like ‘how am I going to find customers and convince them to give me money’

PK

Paul Kelly

unread,
Apr 14, 2017, 9:09:25 PM4/14/17
to ope...@googlegroups.com

Enough philosophy for one morning, I had to come in to work so I grabbed a pic of where we are at with our retrofit project.

 

panel progress.jpg

down cam.jpg

 

Downward camera mount is done.

Power supplies (80V, 24V, 5V) are done.

Servo voltage clamp is done.

Isolated IO board (the blue one) is done.

A small fraction of the wiring is done.

 

Waiting parts for:

Servo drives (the Kelling 160V drives went to el Salvador for reasons that aren’t clear to me).

Boards and  parts for the modified Tarocco drive will be here Tuesday

RGB LED lighting for the up camera and it’s controller board.

 

 

image001.jpg
image002.jpg

Alex Angel

unread,
Apr 14, 2017, 9:14:09 PM4/14/17
to OpenPnP
Ok,

I just explained what I need and trying to do and what my question were.
Just to give you again more specifics because you are assuming wrong.

This business is in partnership , there is a budget allocated for this. I CAN NOT buy a new pick and place machine. I can buy a cheap chinese one but I'm extremely sure I can build one better myself.
What I'm trying to do is to build one as complex as possible just to be safe. If things don't work out I can always make it  LESS complex.

board assembly is about 150 -170 smd's with many of them being 0402 and 0.5  fine pitch Ic's
I can not give you a sales estimation of the boards but it's safe to say that I will need to be able to make about 50/day with one machine
So I'm not even calculating the time for adding the solder paste, reflow time etc .. in other words 4 hours time I need to place 7500 components .

We do not want to make use of external companies for this. This has a significant impact on the profit. I don't even know why I am explaining this details. All I asked was questions about VARIOUS boards that I could use.
I only got answer about the smoothie board but nothing specific. I also can not even find a decent website where they sell this along other components that would be compatible with.
For example arduino is sold on 1000+ website with 10000+ shields maybe a little exaggerating but you get my point.

I would like to go the full 100% with this board but so far nobody has convinced me to.

On wiki all I see is '' please stop asking when board will be ready'' This by itself gave me a bad vibe because I have to take my business serious and do not have to rely on luck or chance. So first impression is a big maybe.
On the other hand if I don't have a choice than I don't have a choice.

Like other have said, If I would use other boards than I would have the write some code, and MAYBE I could even hire somebody to do it for me. Taking this into account I am thinking what is best ? Working with this and do it all by myself and maybe some help from the community or go with other boards that have more opportunities or they are hardware more powerful would be cheaper than 3 smoothies together + other extra boards ...

Then I hear somebody tell me go ahead and make your own hardware .. LOL . I have made a motion controller and I can tell you that, that wasn't a joke .. This isnt something you can build in 3 months what kind of '' suggestion '' is that ?

Anyway .. in few days time I will need to establish what hardware I will use for this pnp...

greets 

Op zaterdag 15 april 2017 02:25:07 UTC+2 schreef PK:

Michael Anton

unread,
Apr 14, 2017, 10:44:12 PM4/14/17
to OpenPnP
A couple of people were suggesting you buy a used PnP machine, rather than a new one...  I would say this is a valid option.  I would buy a used machine in a heartbeat, if I could find a reasonable one that fit my needs and that I could get into my office.  The only reason I'm considering building one, is finding one with those constraints doesn't appear to be too easy.

Alex Angel

unread,
Apr 14, 2017, 11:08:28 PM4/14/17
to OpenPnP

Ok people ..

Can somebody just tell me what I should use ?  I wrote this several times I will not buy a used machine nor a new one.
I already bought allot of parts and already have a design I am looking for a board to control 16 steppers

Liteplacer is 800 usd , chinese desktop is 4000 usd

I am building one myself budget is 4000 usd and it should be allot better than the chinese one.
I don't understand why I don't get specific answers to the questions I asked.
Isn't this group suppose to encourage new things ?

I am very much wondering why nobody has used or made an IO board for the nucleo F4 or F7 and used that for OPEN PNP ? I thought this was open source or am I wrong ?
I had similar issues with some other open source software that was actually made for specific hardware and no room for others.
That's why I started and created this motion controller I will put for sale in the next few months.
I hope OPEN PNP is truly open source and not just claim to be while forcing other to use particular hardware. Most of the times this thing are only supported and developed by few persons which have allot to gain.

Does anybody know when that V2 smoothie will come out ? 
And can somebody give me some schematics or pe very specific how I could get those 16 steppers running with 1 smoothie board and '' other '' extension boards ? 
I prefer to use something I can buy and connect to this smoothieboard and not make hardware myself it takes too long. I will have more than enough problems to solve and to adjust on the machine even IF I get the right hardware.

This is a motion controller with 8 axis from 1987 ... now it's 2017 , we have so many different types of development boards and all I see are wrapped up motion controller boards like a swiss army knife.
I am looking for a truly open source board that is '' truly modular '' . I only see different versions of the same thing. All the same with different mcu's or different pinout.

You guys also say that open pnp is very open source and has a modular design. Does open pnp have support for the F4 mcu or the F7 ? They are pin and function compatible either way.
So why on earth did nobody build it on a nucleo board for example ? and since it's supposed to be open source it should have been ported to other mcu's as well


greets Alex

Bernd Walter

unread,
Apr 15, 2017, 12:25:21 AM4/15/17
to OpenPnP


On Saturday, April 15, 2017 at 5:08:28 AM UTC+2, Alex Angel wrote:

Ok people ..

Can somebody just tell me what I should use ?  I wrote this several times I will not buy a used machine nor a new one.
I already bought allot of parts and already have a design I am looking for a board to control 16 steppers

Liteplacer is 800 usd , chinese desktop is 4000 usd

I am building one myself budget is 4000 usd and it should be allot better than the chinese one.
I don't understand why I don't get specific answers to the questions I asked.
Isn't this group suppose to encourage new things ?

You are not encouraging at all to help you.
 

I am very much wondering why nobody has used or made an IO board for the nucleo F4 or F7 and used that for OPEN PNP ? I thought this was open source or am I wrong ?

Well - the current open control designs all come from the small CNC milling/laser machines and usually just do 2-3 axis.
The 3D printer world added the requirement for extruder motors and since no cartesian axis control seems to be cool for some people it became
hard to run on those old 8-bit AVR controllers.
So recently some boards became available to use modern ARM processors from Atmel or STM.
Those processor need support, so not every popular available source runs on them.
The cheap 3D printer boards available have 5 stepper drivers (or slots for driver modules) to run 3 main axis plus 2 extruders.
Some boards have support for 6 drivers.
An exception is the DUET board, which has 5 onboard drivers plus an easy to plug expansion additional 4 drivers, making 9.
However - I wouldn't say that the DUET is a cheap board.
16 Axis is nothing anyone ever needs so much that it justified to build and produce a special board for it.

I had similar issues with some other open source software that was actually made for specific hardware and no room for others.

Well -  the Source is there.
The driver Interface is very clean and simple.
People do run different hardware, but it seems the Smoothie with the Smoothiefirmware board is the most common to be used.
Nevertheless there is existing support for Marlin and others - Marlin itself supports plenty of different boards and is opensource as well.
 
That's why I started and created this motion controller I will put for sale in the next few months.
I hope OPEN PNP is truly open source and not just claim to be while forcing other to use particular hardware. Most of the times this thing are only supported and developed by few persons which have allot to gain.

Really - I don't see what Jason has to gain from that.
He even offered "send me your machine configuration, I'll edit it for you" support for free.

Does anybody know when that V2 smoothie will come out ? 
And can somebody give me some schematics or pe very specific how I could get those 16 steppers running with 1 smoothie board and '' other '' extension boards ? 

You are basicly asking for a turnkey solution for your custom requirement.
Most of the 3D printer firmware is using GPIO to hook up step/dir driver chips.
The firmware is rarely limited to the number of steppers they support unless you need to run the simultaniously.
E.g. for a 6 axis milling machine you need that, but for an SMT PnP machine it is usually ok to not rotate all pickers at the same time.
It is quite easy to take any free GPIO and hook it up to an external stepper driver.
For example you can get an Arduino expansion board with 4 slots for driver modules from china.
You can't directly plug it in and need to wire it manualy to free GPIO pins.
Each additional stepper driver just needs 2 GPIO (maybe 3 if you want to disable the driver).
You can also hook up beefy chunky drivers - for example my CNC mill is running Leadshine drivers wired up to a 3D printer board.
All you need to care about when using big drivers is that some of them want 5V signals, which ARM systems don't create.
You also need to rebuild the board firmware defining on which GPIO additional motors are hooked up.
And if the selected board has enough free GPIO you can hook up the additional drivers when you actually need them.

I personally go the multi board way, because it works and looks cleaner.
Since I need 8 drivers for the base axis, which OpenPnP has not support to spread on multiple boards it's not turnkey for me either.
I could connect all 4 rotation axis to one driver and a single 5 driver board would be ok however.
Everything else ( I plan to have to steppers for feeders) can go to individual boards.

I prefer to use something I can buy and connect to this smoothieboard and not make hardware myself it takes too long. I will have more than enough problems to solve and to adjust on the machine even IF I get the right hardware.

This is a motion controller with 8 axis from 1987 ... now it's 2017 , we have so many different types of development boards and all I see are wrapped up motion controller boards like a swiss army knife.
I am looking for a truly open source board that is '' truly modular '' . I only see different versions of the same thing. All the same with different mcu's or different pinout.

Yes - this is an axis controller - nothing more.
It has neither support for onboard drivers or onboard driver modules.
This is just a board you hooks up external drivers too.
In other words: any kind of bigger plain Arduino with e.g. Marlin firmware on it.
The only reason why people are using onboard drivers/modules is because it is easy to handle and the standard firmware configuration knows to which GPIO they drivers are hooked up to.


You guys also say that open pnp is very open source and has a modular design. Does open pnp have support for the F4 mcu or the F7 ? They are pin and function compatible either way.

This doesn't even make sense.
It has support for any board running a "firmware" on it, which understands some basic interactive G-code.
I was even suprised how many people just run stock firmware and OpenPnP without any sourcechanges at all.
I'm not aware about any opensource for STM ARM processors.
Marlin and Repetier Firmware runs on Atmel Cortex-M3 ARM processors.
Smoothieware runs on NXP Cortex-M3 ARM processor.


So why on earth did nobody build it on a nucleo board for example ? and since it's supposed to be open source it should have been ported to other mcu's as well

Because you are too blind to see that those exist.
Maybe not for the Nucleos, but on other processors in the same ballpark.

Jason von Nieda

unread,
Apr 15, 2017, 12:47:19 AM4/15/17
to ope...@googlegroups.com
Looking very neat Paul! A few questions:

* What is that massive power resistor (?) for?
* How are you intending to handle RGB lighting? What controller, in particular?

Jason


--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

Alex Angel

unread,
Apr 15, 2017, 1:16:49 AM4/15/17
to OpenPnP
I thought I saw somewhere that smoothie runs on M3 cortex from ST that's why I was talking about the M4 and M7,

Further you say that openpnp does NOT support multiple boards? Somebody earlier said it does . I just didn't want to buy 3 or 4 of them.
Does this smoothieboard have extension boards for  extra IO's ? The stepper are not the only problems. I also need allot more sensor inputs that can trigger events in the software.

Working with a board that uses shift registers so I can connect for example 16 steppers but is connected to one input on the smoothie ? On a serial or I2C port ?? Does something like this exists ?
Or perhaps using multiple I2C stepper drivers  ? would that work ? IF somebody can give me links ? facts ?

greets

Cristian Nicola

unread,
Apr 15, 2017, 1:45:59 AM4/15/17
to OpenPnP
http://www.st.com/en/evaluation-tools/steval-3dp001v1.html 

i think this runs marlin, but i could remember it wrong.

Chris

Paul Kelly

unread,
Apr 15, 2017, 1:50:57 AM4/15/17
to ope...@googlegroups.com

Looking very neat Paul! A few questions:

Neatness is goodness. When it breaks mid run we need to be able to dive in and fix it before the solder paste dries out..

 

* What is that massive power resistor (?) for?

Apologies if I’m telling you how to suck eggs (as we say in my country).

When an axis decelerates the motors turn into generators and the supply voltage bulk storage caps start to charge.

For tiny little machines, the output caps on the PSU or input filter caps on the drives just soak this up and the supply rail goes up by a volt or two. If one axis is accelerating whilst others are decelerating then the accelerating drive will just consume the extra energy.

 

The heads on our machines weigh 5-10Kg and we have half a kilowatt of servo motors driving them, mostly on diagonal moves where X and Y will decelerate simultaneously.

The nominal supply voltage is 80V and the red board has 20mF of 100V caps on it.

There is also a circuit that monitors the supply rail and, should it rise to 99V turns on a FET that puts the aforementioned resistor (10ohms, 250W) across the rails to suck the excess energy out.  The FET will stay on until the voltage drops below 92V.

In addition, there is an Estop input that will just turn the FET on until the voltage drops to about 20V and the Aux PSU drops out. The estop simultaneously kills power to the transformers..

 

Nb, we used toroidal transformers rather than a SMPS because a SMPS will just freak out if it sees its output voltage rise.  You have to use a diode into a big cap bank to stop back feeding it. Then you have to use an inductor to limit inrush into that cap bank.. then…. You get the idea, it’s simpler and cheaper to use transformers.

 

* How are you intending to handle RGB lighting? What controller, in particular?

J So, after the overwhelming number of responses to my “How much light, what colour?” request, I decided that stupid overkill is the best way forward.  I have about 64 WS2182a’s (we have reels of the things) on some PCB’s that will solder together to form the ring light, and a micro and 5V buck reg  (the thing will pull 3.5A @ 5V) on a separate control board. I have the option of an isolated USB uart interface, or a simple signal line from the IO board.  I suspect we’ll muck around with the former then move to the latter..

 

PK

Jason

 

 

On Fri, Apr 14, 2017 at 8:09 PM Paul Kelly <te...@caswa.com> wrote:

Enough philosophy for one morning, I had to come in to work so I grabbed a pic of where we are at with our retrofit project.

 

panel progress.jpg

down cam.jpg

 

Downward camera mount is done.

Power supplies (80V, 24V, 5V) are done.

Servo voltage clamp is done.

Isolated IO board (the blue one) is done.

A small fraction of the wiring is done.

 

Waiting parts for:

Servo drives (the Kelling 160V drives went to el Salvador for reasons that aren’t clear to me).

Boards and  parts for the modified Tarocco drive will be here Tuesday

RGB LED lighting for the up camera and it’s controller board.

 

 

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/004701d2b584%24d86edb60%24894c9220%24%40com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

Jason von Nieda

unread,
Apr 15, 2017, 2:36:17 AM4/15/17
to ope...@googlegroups.com
Thanks Paul, very useful stuff for me since all my experience is with small systems and steppers. Have not had a chance to play with big hardware like that :)

Before you spend too much time with those LEDs, try them with your camera. I used similar ones and it caused horrible banding in the image. PWM frequency is too low. That looks like an ELP camera in your image, which is what I use, so I suspect you'll see the same thing.

Jason


Paul Kelly

unread,
Apr 15, 2017, 2:45:15 AM4/15/17
to ope...@googlegroups.com

J ta for the tip. Did you try at 100% PWM?  One option I have is to just set them to white and dial up intensity by changing the buck converter output.

 

PK

Downward camera mount is done.

Michael Anton

unread,
Apr 15, 2017, 3:06:12 AM4/15/17
to OpenPnP
Except WS2812 parts need 5V to run, so the variation that you will have available on the buck converter may be limited.  This would be the downside to using LEDs that have a built in driver.  You could look at the AP102 parts as an alternative, as they have a higher PWM frequency.

BZHEX

unread,
Apr 15, 2017, 3:49:42 AM4/15/17
to OpenPnP

Both WS2812 and APA102  have 400Hz PWM freq , with is NOT enough for camera view. It will cause flickering images. PWM must be at least 1000Hz for changing LED colors.

Paul Kelly

unread,
Apr 15, 2017, 4:12:05 AM4/15/17
to ope...@googlegroups.com

It would have been nice if these topics had come up in response to my question about lighting… 

As I keep saying to my wife, “I’m not angry, just disappointed.”

 

Are you sure about the APA102?

I’m finding vague references to it having a high pwm freq…

The datasheet only mentions a 1MHz clock.  1000000/256=4 ish KHz…

 

I have boards arriving on Tuesday anyhow. I’ll try 100% duty cycle and just changing the supply voltage.

PK

From: ope...@googlegroups.com [mailto:ope...@googlegroups.com] On Behalf Of BZHEX
Sent: Saturday, 15 April 2017 3:50 PM
To: OpenPnP <ope...@googlegroups.com>
Subject: Re: [OpenPnP] Build progress

 

Both WS2812 and APA102  have 400Hz PWM freq , with is NOT enough for camera view. It will cause flickering images. PWM must be at least 1000Hz for changing LED colors.

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

Paul Kelly

unread,
Apr 15, 2017, 4:33:24 AM4/15/17
to ope...@googlegroups.com

Can somebody just tell me what I should use?

You asked!

 

http://www.ebay.com/itm/Philips-Assembleon-Sapphire-Pick-and-Place-High-Speed-SMT-Chip-Shooter-2000/322230174609?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D2%26asc%3D40130%26meid%3D69841eb827de4874bff9cd68348879e7%26pid%3D100005%26rk%3D6%26rkt%3D6%26sd%3D112344577300

 

Doesn’t say anything about feeders but I seem to recall that these use the pneumatic ones that are activated by a hole in one of the mounting pins, but it has two heads, 8 nozzles per head. 4 cameras and can do two boards at once. Should be good for 5000CPH .

 

http://www.ebay.com/itm/Samsung-CP30V-3-Head-SMT-Pick-and-Place-Machine-in-Good-Working-Order/181891440565?_trksid=p2047675.c100009.m1982&_trkparms=aid%3D888007%26algo%3DDISC.MBE%26ao%3D1%26asc%3D42404%26meid%3Ddc67f6fc2b7a4cbba9c197f59cf3dd00%26pid%3D100009%26rk%3D1%26rkt%3D1%26sd%3D351940992067

Again, with vision, but the feeders may be harder to source. Price is good though.

 

http://www.used-line.com/semiconductor-and-pcb/pcb-pick-and-place-machines/quad-ii-c/item-7624512

from a dealer

 

 

Buying from a dealer will cost 10-20% more, but they will come and set it up and offer service if it breaks.

 

From: ope...@googlegroups.com [mailto:ope...@googlegroups.com] On Behalf Of Alex Angel
Sent: Saturday, 15 April 2017 11:08 AM
To: OpenPnP <ope...@googlegroups.com>
Subject: Re: [OpenPnP] Re: motion controller for pnp ? multiple axis needed !!

 

https://lh3.googleusercontent.com/-qS9DsDPG0uQ/WPGNMgxY66I/AAAAAAAAAAs/JUOzVHs8XKME3xVCUi031MbsOTCMKFHSgCLcB/s320/54458.jpg

Ok people ..

 

Can somebody just tell me what I should use ?  I wrote this several times I will not buy a used machine nor a new one.

I already bought allot of parts and already have a design I am looking for a board to control 16 steppers

 

Liteplacer is 800 usd , chinese desktop is 4000 usd

 

I am building one myself budget is 4000 usd and it should be allot better than the chinese one.

I don't understand why I don't get specific answers to the questions I asked.
Isn't this group suppose to encourage new things ?

I am very much wondering why nobody has used or made an IO board for the nucleo F4 or F7 and used that for OPEN PNP ? I thought this was open source or am I wrong ?

I had similar issues with some other open source software that was actually made for specific hardware and no room for others.
That's why I started and created this motion controller I will put for sale in the next few months.
I hope OPEN PNP is truly open source and not just claim to be while forcing other to use particular hardware. Most of the times this thing are only supported and developed by few persons which have allot to gain.

 

Does anybody know when that V2 smoothie will come out ? 

And can somebody give me some schematics or pe very specific how I could get those 16 steppers running with 1 smoothie board and '' other '' extension boards ? 

I prefer to use something I can buy and connect to this smoothieboard and not make hardware myself it takes too long. I will have more than enough problems to solve and to adjust on the machine even IF I get the right hardware.

https://lh3.googleusercontent.com/-qS9DsDPG0uQ/WPGNMgxY66I/AAAAAAAAAAs/JUOzVHs8XKME3xVCUi031MbsOTCMKFHSgCLcB/s320/54458.jpg

This is a motion controller with 8 axis from 1987 ... now it's 2017 , we have so many different types of development boards and all I see are wrapped up motion controller boards like a swiss army knife.
I am looking for a truly open source board that is '' truly modular '' . I only see different versions of the same thing. All the same with different mcu's or different pinout.

 

You guys also say that open pnp is very open source and has a modular design. Does open pnp have support for the F4 mcu or the F7 ? They are pin and function compatible either way.
So why on earth did nobody build it on a nucleo board for example ? and since it's supposed to be open source it should have been ported to other mcu's as well

 

 

greets Alex

 

--

You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To post to this group, send email to ope...@googlegroups.com.

Arthur Wolf

unread,
Apr 15, 2017, 4:34:14 AM4/15/17
to ope...@googlegroups.com
I think you posted this in the wrong thread.

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Arthur Wolf

unread,
Apr 15, 2017, 4:38:20 AM4/15/17
to ope...@googlegroups.com
On Sat, Apr 15, 2017 at 1:57 AM, Alex Angel <angelf...@hotmail.com> wrote:
I can't buy a new one

I have already made some hardware designs in autocad.
In the past I have made some cnc routers. The hardware is pretty modular and reliable. I do not use extruded aluminium profiles I am using cnc machined profiles and mounts etc..

I think my design is very solid and the precision will be very good. 
The only think I'm considering right now is the main controller board.
Like I said I even have a solid idea and plan for the feeders. That too won't be a problem.

The only problem I have to deal right now is the main controller board.
I will learn how to deal with it when the machine will be ready, for that I need first to make a BOM list. That's why I need to choose the right controller for my needs.
I explain in previous message that I find using 3 controllers a bit too much even that I'm not familiar with this smoothie board.
 I asked if there was a possibility of working with 1 single smoothie board and having other cheaper boards to extend so I can connect my number of axes needed.

As I said three times now : you can do this, using the method I described, a smoothieboard, a $3 logic gate chip breakout and a $3 multiplexer breakout
 
Also when does the smoothie board V2 come out ? I have 3 months time to get this done.

Smoothie v2 won't be out 3 months from now, for sure. But you don't need v2, v1 does what you want, I was just mentionning it because you mentionned FPGAs and v2 uses them.

What about the other controllers I mentioned ? The fanuc, siemens , mitsubishi , galil  etc .. 

Like I mentioned I was looking for spending about 350 top for something that would make me able to run 16 steppers or even more.
Those stepper do not need to work simultaneously. For the arc motion is only 2 axis , X and Y. The Z don't even have to run together because I rather wait until the machine moves to it's right position in order to avoid collisions with other parts if moving to quickly. 


greets

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Michael Anton

unread,
Apr 15, 2017, 4:43:07 AM4/15/17
to OpenPnP
I think the datasheet on them is wrong, as everything I've read indicates the PWM rate is much higher than the WS2812 parts.  This seems to claim that as well: https://www.adafruit.com/product/2343, and states that it is 20kHz.  According to this guy: https://cpldcpu.com/2014/08/27/apa102/, the global brightness is modulated at 580Hz, but the individual color brightness is modulated at 19.2kHz.

Arthur Wolf

unread,
Apr 15, 2017, 4:45:25 AM4/15/17
to ope...@googlegroups.com
On Sat, Apr 15, 2017 at 3:14 AM, Alex Angel <angelf...@hotmail.com> wrote:
Ok,

I just explained what I need and trying to do and what my question were.
Just to give you again more specifics because you are assuming wrong.

This business is in partnership , there is a budget allocated for this. I CAN NOT buy a new pick and place machine. I can buy a cheap chinese one but I'm extremely sure I can build one better myself.
What I'm trying to do is to build one as complex as possible just to be safe. If things don't work out I can always make it  LESS complex.

board assembly is about 150 -170 smd's with many of them being 0402 and 0.5  fine pitch Ic's
I can not give you a sales estimation of the boards but it's safe to say that I will need to be able to make about 50/day with one machine
So I'm not even calculating the time for adding the solder paste, reflow time etc .. in other words 4 hours time I need to place 7500 components .

We do not want to make use of external companies for this. This has a significant impact on the profit. I don't even know why I am explaining this details. All I asked was questions about VARIOUS boards that I could use.
I only got answer about the smoothie board but nothing specific.

Yes you did ! I explained to you exactly how to implement what you want. If you combine what I told you, with reading the smoothie documentation, you have everything you need to do what you want !
 
I also can not even find a decent website where they sell this
 
What ???
smoothieware.org, -> Big blue button says « Get Smoothieboard here » -> http://smoothieware.org/getting-smoothieboard with a list of 10+ resellers all around the world !

 
along other components that would be compatible with.
For example arduino is sold on 1000+ website with 10000+ shields maybe a little exaggerating but you get my point.

If you had read the smoothie documentation you'd know smoothie is designed to be modular so there is very little limit on what kind of hardware it's compatible with.

I would like to go the full 100% with this board but so far nobody has convinced me to.

This is not anyone's job to convince you, it's your job to do research and determine what's best for you.
Everyone here has done their research, and many of them has chosen Smoothie, that should help you, but if it doesn't then that's your problem.

On wiki all I see is '' please stop asking when board will be ready''

This is only in regards to the v2 system, and it's a tongue-in-cheek answer to the *very* strong community enthusiasm around the project ...
 
This by itself gave me a bad vibe because I have to take my business serious and do not have to rely on luck or chance. So first impression is a big maybe.
On the other hand if I don't have a choice than I don't have a choice.

Like other have said, If I would use other boards than I would have the write some code, and MAYBE I could even hire somebody to do it for me. Taking this into account I am thinking what is best ? Working with this and do it all by myself and maybe some help from the community or go with other boards that have more opportunities or they are hardware more powerful would be cheaper than 3 smoothies together + other extra boards ...

Then I hear somebody tell me go ahead and make your own hardware .. LOL . I have made a motion controller and I can tell you that, that wasn't a joke .. This isnt something you can build in 3 months what kind of '' suggestion '' is that ?

I think you are being way too adversarial for this sort of volunteer-based community-based mailing list ...
Remember you are talking to volunteers and you didn't pay anyone for anything ... take it or leave it ... 
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.



--
Courage et bonne humeur.

Michael Anton

unread,
Apr 15, 2017, 4:47:43 AM4/15/17
to OpenPnP
Pretty sure about the APA102, and posted links in the above post.

I probably didn't know about these when you were discussing lighting.  I went looking for them when discussing a POV type project with a friend, and knew at the time that the WS2812s were slow, so I sought out something different.  The guys on the FirePick Delta forum figured out long ago that the WS2812s caused problems...

Paul Kelly

unread,
Apr 15, 2017, 5:24:38 AM4/15/17
to ope...@googlegroups.com

Except WS2812 parts need 5V to run, so the variation that you will have available on the buck converter may be limited.

We run 2812a’s at 3.3V,  I’ve run them on 2 AA batteries

 

 You could look at the AP102. parts as an alternative, as they have a higher PWM frequency.

I’m confused, I thought I said that before and you disagreed??

Michael Anton

unread,
Apr 15, 2017, 5:47:40 AM4/15/17
to OpenPnP


On Saturday, April 15, 2017 at 3:24:38 AM UTC-6, PK wrote:

Except WS2812 parts need 5V to run, so the variation that you will have available on the buck converter may be limited.

We run 2812a’s at 3.3V,  I’ve run them on 2 AA batteries

 

 You could look at the AP102. parts as an alternative, as they have a higher PWM frequency.

I’m confused, I thought I said that before and you disagreed??


I don't think I disagreed with you, and that it was BZHEX that disagreed.  I don't remember you suggesting the APA102 parts, as I thought I brought them up (though I spelled the partnumber incorrectly).  It sounds like we are on the same page now though... :-)

Trampas Stern

unread,
Apr 15, 2017, 6:50:26 AM4/15/17
to OpenPnP
On the LEDs I had the PWM problem. I turned PWM pin to be digital output and then put a power resistor in series until I got the light level I wanted. 

If you need a LED driver that does programmable current limiting with no flicker please contact me off list. 

Trampas

Trampas Stern

unread,
Apr 15, 2017, 7:15:23 AM4/15/17
to OpenPnP
Alex,

If you want a system that controls 18 stepper motors from one serial port, please contact me off list and I can provide you with a solution.  Specifically I can quote you a box with 3 smoothie boards, and an interface board with interpreter to control the three smoothie boards for you.  

However if you need to build 50 boards a day, please consider an option of building several PnP machines. Then each machine only has to build a fraction of the total production. 

For example assume the LitePlacer is 1000 cph, then you might need 5 of these machines.  By getting one of these machines you can be up and running at slower volumes quickly, and when sales increase you can always buy more machines.  This is a very safe bet as that if sales do not come you have not spent a fortune on a PnP.  Additionally  as you change hardware designs over time or add different boards you can bring one machine down and use it for prototype builds, rather than trying to do change overs on the main production machine. 

One of the issues in having one machine to rule them all, is that when that machine is down, so is your business.  

Again if you want to go with 18 steppers, I can provide you with a hardware solution!   

Trampas

BZHEX

unread,
Apr 15, 2017, 7:51:22 AM4/15/17
to OpenPnP
I study the data sheet a little and yes, PWM for each color is at high freq - 19.2Khz - flicker free. But do NOT use global brightness , witch has only 500-600Hz , this will cause serious flicker on camera. Put global brightenss at maximum 32, and only set PWM for each color. This will work.

Alex Angel

unread,
Apr 15, 2017, 12:45:27 PM4/15/17
to OpenPnP
hello,

Yes, but I can not find any shop that sells those extention boards .. Can you give me some links? Where do I find this multiplexer ? 
I need to see some examples or the boards you are talking about, where will I connect it to on the smoothie ? The smoothie only provides inputs , outputs and some outputs for the motors right ?
Does the smoothieboard provide ports ? Like I2C , serial ? etc .. ?

The next question is something I would feel more comfortable working with.
IS it possible to use I2C stepper drivers daisy chained ? Would this work for C1, C2 , Cn .... and Z1, Z2, Zn ... ?
The only thing is that OPEN PNP would have to support that kind of code would that work ? OPEN PNP would have to adress the specific I2C driver each time sincer they can be so many of them 128 ? not sure ...

Regarding the LED driver you can also use a led driver through I2C or PWM right ? There are many small boards with different drivers for it and you would need to write a small code for it. Not sure if OPEN PNP can take care of it ...


greets


Op zaterdag 15 april 2017 10:38:20 UTC+2 schreef Arthur Wolf:


On Sat, Apr 15, 2017 at 1:57 AM, Alex Angel <angelf...@hotmail.com> wrote:
I can't buy a new one

I have already made some hardware designs in autocad.
In the past I have made some cnc routers. The hardware is pretty modular and reliable. I do not use extruded aluminium profiles I am using cnc machined profiles and mounts etc..

I think my design is very solid and the precision will be very good. 
The only think I'm considering right now is the main controller board.
Like I said I even have a solid idea and plan for the feeders. That too won't be a problem.

The only problem I have to deal right now is the main controller board.
I will learn how to deal with it when the machine will be ready, for that I need first to make a BOM list. That's why I need to choose the right controller for my needs.
I explain in previous message that I find using 3 controllers a bit too much even that I'm not familiar with this smoothie board.
 I asked if there was a possibility of working with 1 single smoothie board and having other cheaper boards to extend so I can connect my number of axes needed.

As I said three times now : you can do this, using the method I described, a smoothieboard, a $3 logic gate chip breakout and a $3 multiplexer breakout
 
Also when does the smoothie board V2 come out ? I have 3 months time to get this done.

Smoothie v2 won't be out 3 months from now, for sure. But you don't need v2, v1 does what you want, I was just mentionning it because you mentionned FPGAs and v2 uses them.

What about the other controllers I mentioned ? The fanuc, siemens , mitsubishi , galil  etc .. 

Like I mentioned I was looking for spending about 350 top for something that would make me able to run 16 steppers or even more.
Those stepper do not need to work simultaneously. For the arc motion is only 2 axis , X and Y. The Z don't even have to run together because I rather wait until the machine moves to it's right position in order to avoid collisions with other parts if moving to quickly. 


greets

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

Arthur Wolf

unread,
Apr 15, 2017, 12:50:22 PM4/15/17
to ope...@googlegroups.com
On Sat, Apr 15, 2017 at 6:45 PM, Alex Angel <angelf...@hotmail.com> wrote:
hello,

Yes, but I can not find any shop that sells those extention boards .. Can you give me some links? Where do I find this multiplexer ? 
I need to see some examples or the boards you are talking about, where will I connect it to on the smoothie ? The smoothie only provides inputs , outputs and some outputs for the motors right ?
Does the smoothieboard provide ports ? Like I2C , serial ? etc .. ?

Now you are getting to the point of what we have written documentation for.
Please read the documentation, and once you have, if you have any kind of question that is not answered there, I'll be very glad to help you.

To give you a short version : most pins on smoothie can be configured as either inputs or outputs, and actions/gcodes mapped to them, simply by editing the config file. It has plenty of ports, including i2c, spi, serial etc.
Here you would mostly use GPIOs and connect those to a multiplexer ( any would do ) which in turn controls a logic gate chip ( any would do ) which in turn switches whether stepper drivers receive step signals or not.

The next question is something I would feel more comfortable working with.
IS it possible to use I2C stepper drivers daisy chained ? Would this work for C1, C2 , Cn .... and Z1, Z2, Zn ... ?

Nope, that doesn't exist, and i2c would be a terrible choice for it.

I did link you 10 mails back to a project the smoothie project is working on that does something very similar ( using rs485 ), but you obviously don't read what you are pointed to ...
 
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Alex Angel

unread,
Apr 15, 2017, 1:04:36 PM4/15/17
to OpenPnP
Yes, I saw it but there is only example fo the code,

I wanted to see the board.

I need to work with something that I can buy ready made . Something nicely made like for example pokeys have extention boards for extra sensors or for steppers, relais etc .. :





Op zaterdag 15 april 2017 18:50:22 UTC+2 schreef Arthur Wolf:


Arthur Wolf

unread,
Apr 15, 2017, 1:06:42 PM4/15/17
to ope...@googlegroups.com
We have those planned for smoothieboard v2, but right now you'll have to instead use breakout boards from somewhere like adafruit or sparkfun.
This is the best option available to you right now, there are "all integrated" options in planning, but right now you *have no other option* than using brekout boards and doing a bit of wiring yourself.

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Trampas Stern

unread,
Apr 15, 2017, 1:27:00 PM4/15/17
to OpenPnP


Alex, 


I have a board I designed where we needed to test some firmware running on hardware. The hardware had to be electrically isolated during testing. So I made a relay board that is a 1 to 8 JTAG mux. This board will switch 14 pins to one of 8 outputs. 


So for example if you have step/dir/enable for each stepper motor it could switch 4 motor outputs from a controller to 32 different motors.   




Alex Angel

unread,
Apr 16, 2017, 1:11:18 AM4/16/17
to OpenPnP
Ok , I'll keep all this in mind ...

Further anybody idea how to get multiple sensor inputs ? For example I would LIKE to use as many sensors as possible.
2 mechanical end stops for each axis for safety in parallel with  optical sensors , 2 soft stop sensor near the end of the axis end ? or can I choose this in the software for example that my X axis moves max rate and slows down at last 10% of the axis ? Then I would also need a homing sensor. So 3 or 5 sensors for each axis.

Anybody have an idea what to use to connect a capacitive type encoder to a stepper so I can get it closed loop ? In this case I would not miss a step due to fast acceleration or sudden stops.
I'm using only nema 17 and nema 14 motors max2ah each motor. I plan using 4 nema17 motor for the Y , 2 nema17 for the X 0.9 deg step angle motors with double pulleys on each one of them .
Im hoping to get maximum resolution by forcing 4 motors to synchronize in parallel to work together for each step so I'll get less step errors. Im looking for a mini cheap controller to place on each motor so I can connect to the encoder and driver to get the closed loop also.


Then is there a way to let OPEN PNP get the encoder input ? So they also can act as a sort of feedback measurement device ? and then by some wonder code that it calibrate itself along with the sensors etc ..

Also what cpu's or mcu's does OPEN PNP support ? smoothie uses M3 from atmel right ? Would the code gain more from using an M4 ?

greets

Daniel Dumitru

unread,
Apr 16, 2017, 1:23:17 AM4/16/17
to ope...@googlegroups.com
Please , please , take 15-30 minutes to read from OpenPnP documentation !
(every thing related to motion happens in motion controller, no code from open pnp runs on microcontoller )



Cristos a inviat !

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.
Message has been deleted

Alex Angel

unread,
Apr 16, 2017, 1:45:10 AM4/16/17
to OpenPnP
Adevarat a inviat ! :D

But motion controller is the MCU or micro control unit ..   This uses an atmel M3 , somebody mentioned the EVAL board from ST with an M4 6 axis , almost everything from ST is well supported and open source.
The smoothie board might have better support as is now for the open pnp but I'm trying to make a decision because either way I will need somebody to take a look at the config file and maybe write some code.
Some will say smoothie is easy and probably is , as is now .. but in my case ..  just don't know . If I could I would even prefer to run on a standalone controller , but you have the vision most probably that would be very hard for someone to program anyway there is allot of mumbo jumbo in my head right now when I will know enough I will probably gamble on my decision .. and hoping it will at least get me where I need to be. And in worse case changing the controller is not a big thing, the bigger issue is the time I will have to put into it understanding it and make it work with specific hardware, If I will be forced to change controller than I would have to start all over again

greerts

Cri S

unread,
Apr 16, 2017, 2:14:38 AM4/16/17
to ope...@googlegroups.com
On you'r requirements, don't loose time on small controllers, use
LinuxCnC/EMC(2) .
That can do all what you want and OpenPnP can interface with EMC.
But read the documentation first, othwise you need things differently
very often.
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "OpenPnP" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/openpnp/LeOoxE6kHrI/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> openpnp+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/openpnp/3fb7ac26-f7b4-46f4-a4a1-a8329aa78234%40googlegroups.com.

Alex Angel

unread,
Apr 16, 2017, 3:24:00 AM4/16/17
to OpenPnP
hy,

I need a modular motion controller not software ... if you are referring to the standard breakout boards then I rather not use them, I need something decent with a powerful micro controller buffer etc
 a real motion controller but something that will make me able to run 16 steppers or more if needed. Only 2-3 axis need to run together in interpolation  the others not .. 

greets

Op zondag 16 april 2017 08:14:38 UTC+2 schreef Cri S:

TheCunningFellow

unread,
Apr 16, 2017, 3:46:14 AM4/16/17
to OpenPnP
a clue

BZHEX

unread,
Apr 16, 2017, 4:00:10 AM4/16/17
to OpenPnP
this one is modular for example: https://plugg.ee/

write me on bogdan...@gmail.com ( in romana sa ne intelegem ) . If you are able to wait 2 - 3 months from now, i think i can give you a solution.

Cri S

unread,
Apr 16, 2017, 8:48:31 AM4/16/17
to OpenPnP
If you ask about ONE controller, extensible or not, then all axis are synchronized for that controller. As example if you drive the nozzle and a feeder and a tape peeler (tm220 like), 3 separate controllers are needed because the actions need to be asycnchron in order you don't need to wait on feeder to completed feed before beginning head motion. Big controllers like emc2, fanuc, ... allows asynchronous motion .

Bernd Walter

unread,
Apr 16, 2017, 9:51:47 AM4/16/17
to OpenPnP


On Sunday, April 16, 2017 at 7:11:18 AM UTC+2, Alex Angel wrote:
Ok , I'll keep all this in mind ...

Further anybody idea how to get multiple sensor inputs ? For example I would LIKE to use as many sensors as possible.
2 mechanical end stops for each axis for safety in parallel with  optical sensors , 2 soft stop sensor near the end of the axis end ? or can I choose this in the software for example that my X axis moves max rate and slows down at last 10% of the axis ? Then I would also need a homing sensor. So 3 or 5 sensors for each axis.

At this point it really would be the best I you've just installed OpenPnP and played with it.
OpenPnP comes with a simulated PnP machine, which isn't perfect but good enough to get an idea how some of the things work.
The problem is that it is not possible to tell you where you are wrong, because you aren't, but you have an over complicated view on many of the parts.
At best I would suggest you to go and buy a RAMPS 1.4 kit with Arduino and stepper drivers, plus 2 steppers and a 12V power supply, then install Marlin on it.
I understand that you just want to plan right now and not buy, but on the other hand nothing beats playing with stuff and a RAMPS kit is the cheapest option and
you even might use some or all of the hardware in your system.

For a PnP machine none of the axis need to be synchronized and theoretically you can use an individual board for each motor, each sensor and each actor.
In many cases however it is reasonable to control the XY path, so having the X and the Y motor moving synchronized is something you want.
OpenPnP sets the restriction that it wants to have the main axis on one controller, this is X, Y, all Z and all C (nozzle rotation).
This is an OpenPnP restriction, which can be handled by custom code in OpenPnP, or some kind of aggregation driver, which takes all the commands
from OpenPnP via one channel and then sends it to subboards - non of this has existing code AFAIK and needs to be done by yourself.
Everything else - feeders, valve outputs, sensors, ... are free to be done with individual boards.

But also be carefull about what you read and havn't tested yourself.
Many of the people here, including myself, are just building their own machines and may be wrong on some points.
The main difference is that we are prepared to tweak for wrong decisions.

Anybody have an idea what to use to connect a capacitive type encoder to a stepper so I can get it closed loop ? In this case I would not miss a step due to fast acceleration or sudden stops.
I'm using only nema 17 and nema 14 motors max2ah each motor. I plan using 4 nema17 motor for the Y , 2 nema17 for the X 0.9 deg step angle motors with double pulleys on each one of them .
Im hoping to get maximum resolution by forcing 4 motors to synchronize in parallel to work together for each step so I'll get less step errors. Im looking for a mini cheap controller to place on each motor so I can connect to the encoder and driver to get the closed loop also.


Then is there a way to let OPEN PNP get the encoder input ? So they also can act as a sort of feedback measurement device ? and then by some wonder code that it calibrate itself along with the sensors etc ..

Also what cpu's or mcu's does OPEN PNP support ? smoothie uses M3 from atmel right ? Would the code gain more from using an M4 ?

Wrong questions and I've already tried to explain that.
OpenPnP doesn't care - it most likely runs on a beefy modern PC and can't run on a microcontroller at all.
It communicates via serial channels sending ASCII commands to connected devices and parses ASCII responses.
How the boards implement that is not the job of OpenPnP - it just needs to know which board to send which command.
It really becomes obvious once you start playing with OpenPnP.


greets

Alex Angel

unread,
Apr 16, 2017, 5:47:32 PM4/16/17
to OpenPnP
hy,

 so since open pnp have nothing to do with the smoothieboard mcu then why should I go with the smoothie at all ? They say it's supported the most but in what ?
If we are just talking about assigning pins to the config file or parameters than this should work in any other bob board as well ..
What about that '' supported g code '' for what is that good for ? I asked if a M4 board would be better for open pnp than the current M3 from atmel that OPEN PNP is using and I got as answer that it does not matter ...



greets

Arthur Wolf

unread,
Apr 16, 2017, 6:00:46 PM4/16/17
to ope...@googlegroups.com
You very obviously do not read anything that is sent to you. This thread has several dozen demonstrations of this.
I spend half my days providing support for smoothie, I'm very happy doing it, but once a year or so, I have to say to someone : please don't become a Smoothieboard user, you are not capable of listening or reading documentation well enough to use it, it will make you unhappy, and trying to support you will make us unhappy.

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+unsubscribe@googlegroups.com.

To post to this group, send email to ope...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Alex Angel

unread,
Apr 16, 2017, 6:06:58 PM4/16/17
to OpenPnP
Arthur,

I asked several times what the benefits are for open pnp software using M3 mcu or other type of mcu's.
I got answers as you can also read .. saying that this does not matter. But how can this not matter when a mcu is everything on a board.
You told me that smoothie is mostly supported by open pnp. That means that the mcu is supported right ?
WHERE is the code for ALL the hardware ? IS it in the smoothieboard firmware itself  or is it in the OPEN PNP ? Does open pnp just recognize the board configuration file and pin assignment or does it do much more ?

greets
Reply all
Reply to author
Forward
This conversation is locked
You cannot reply and perform actions on locked conversations.
0 new messages