PNP machine build update

[Trampas Stern]

I have been slow building my PnP but worked some on it this weekend. 

I started with Peter' head design but redesigned it to use belts. 

The belt is not continuous but rather two cut pieces of belt. The plastic pieces are 6mm width so it allows the NEMA 8 z-plates to slide past pulleys with out interference. I 3D printed the NEMA 8 plates with an Envision Tech printer, but can be machined (requires some drilling and taping for screws in the edges). However it might be possible to make the Z axis plates where they can be laser cut using the nut slot trick that Peter did in his design.  

The Z axis plate design I put some slots in such that the belts can be looped and doweled, see image below. I also raised the stepper motor on the Z axis plate such that the wire chase can go under the stepper motor, I personally prefer this as it makes machine look neat and provides a track for the wire chase on the X-axis rail. 

The limit switches on the Z axis are just screwed into the z plate with M2 screws and contact the MGN9 carriage block. 

The basic XY mechanical design is based on Anthony's machine, but with some modifications, I redesign the blocks between the X and Y rails with slots for the belt loop, however I made the slots a bit too wide and had to add a wire tie to hold belt.  

I got the machine all wired up and currently working on configuring Ray's Remix board to control everything.  Once I get the machine up an running I will start adding the smart steppers to the design, however I wanted to baseline without the smart steppers first. 

I am still waiting on my downward camera to be delivered and I have not make lights for the camera yet.  I was thinking about just using LED strip lights for the moment. 

I also have not picked out a good reel feeder design yet. For the moment I will just use cut tape, might try using double stick tape to hold cut tape to bed. 

Also with the head design if I use a thinner stepper for the Z movement it should be possible to add a second set of heads on the back side of the rail, for a quad machine.  However right now I would be happy to have one head working....

Trampas

www.misfittech.net

[oliver jackson]

Looks great! Also looking forward to seeing how your rotational smart steppers experiment goes.

--
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/c20701e2-ab05-4a01-88ad-8c9347cbfa83%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jason von Nieda]

Nice work! 

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/3CDB1616-6594-4082-984C-E02C2BA89C3B%40gmail.com.

[Daniel]

What controller will you use ?

if is something like smoothie I am interested to find out how you will wire Z1 max and Z2 max endswitches and how will configure Z min switch.

I am waiting as well to see how performs intelligent driver !

Thank you,

Daniel

[Trampas Stern]

For the moment I am using the remix board with smoothie firmware, http://cohesion3d.com/cohesion3d-remix/. 

On my head design I have two limit switches Z1 max and Z2 max, Since Z2 max is the same as Z1 min, I just wired Z2 max to the Z min of the remix board.    

It ends up I have around 60mm of travel on my head, which divided between the two heads means each can move 30mm each.   In the config.txt I told smoothie firmware that Z min was -30mm.  This way when the machine homes it sets the Z height to -30mm when it hits the min switch.  In OpenPnP I thing changed the Gcodedriver such that for homing it sends G28 which moves axis to min switches, followed by G0 Z0 - this then moves both heads to the "Up position", ie mid way on both's travel. 

What I noticed with the Remix/smoothie and the A4988 step sticks is that the axis moves are really really slow to keep from missing steps. What I have found out is that:

1. I was using 0.9 degree motors with 16x microstepping which makes things slow.

2. I was using 12V for motors, which makes things slow. 

I switched to the Nano Zero Smart stepper (NZS) and 24V and the axis moves much faster, I still need to do more testing with Remix/smoothie as I could not get it run as fast when Remix was controlling the NZS.  I think it might be the microstepping again limiting the speed, but need to test more...  

So if you build a machine I recommend using a high voltage for the motors, this will add a lot more speed.  

I have more testing and set up to do. I also got my head controller PCBs in but have not populated or programmed them yet, but I will keep you guys updated. 

Trampas

www.misfittech.net

[Daniel Dumitru]

Trampas, 

Thank you for nice explanation on having one motor for 2 Z heads ! I was really need them since I am struggling to start again my old Zeva.

Like others I am highly expecting your results on testing driver with force feedback on dual Z heads !

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/4e24ee6e-f9ea-4416-82f0-f8681431833d%40googlegroups.com.

[Arie Lashansky]

Hi   Tampas   

I like your  mod of the Z stage  sliders can you send me the 3d model 

Thanks Arie 

[Trampas Stern]

Attached is the step files for the head... 

There is a base plate, left and right NEMA8 plates, belt clamp and then horizontal NEMA 8 mount. 

[Trampas Stern]

As far as an update on the machine goes here is a picture from last night. 

One thing I like about the head design is the drag chain rests on the extrusion, sounds silly until you try and figure out how use drag chain when it can not rest on extrusion. 

This was before I learned that the PCB and strip feeders need to be at same height. Speaking of which here are some lessons learned.

1. A4988 Step Sticks from china are limited to 1A not 2A! I kept wondering why I could not get speed out of my machine.
2. A4988 step sticks from china have high failure rate, took 8 to get 5 working ones. 
3. Mechanical design should be done such that PCB and strip feeders at same Z height such that camera focus is correct. 
4. The down looking camera needs to have lens calibration/alignment done before installing. 
5. Cameras need to mounted such that they are perpendicular to the build plane, and life is easier if they are square to mechanics. 
6. I kept having problems using camera with OpenCV driver, the webcam driver seems more stable for me.
7. Setting up OpenPnP takes some time, and even longer to become proficient, which I am not.  
8. Anything involving vision needs good contrast, white tape strips should be put on dark background, black tape strips on white background.
9. Cut tape strips should be put on a holder (even if just an MDF board), do not tape to bed else you will be sorry when you have to move them...
10. No software will help a bad mechanical design.
11. Get a good computer, with lots of independent USB ports for Cameras, get good camera. 
12. On the counterbalanced dual head design, going forward I will want a "home" switch. One concern I have is moving the head with a nozzle down. Jason might already have a check for "Head Safe before move" in the software. With the smart stepper this will be easy (no physical switch required). 
13. Your up looking camera should be placed between your parts that need alignment (QFN, TSSOP, etc) and the PCB, so that head moves less per part. Not sure mine is optimal above. 
14. Be prepared to make mistakes and rebuild several times. 
15. Test early and often, don't put down a lot of parts like, I did above, at once, test each one as you place strip down. 

Some things I have learned about OpenPnP which may be obvious to others (and me now that I know them)... 

OpenPnP coordinated system is based on the down looking camera. The Nozzles have an offset from this location. Using the vision system requires that the nozzles offsets be correct. What this means is that camera should have a good stable mounting to head as it will be referenced for all things vision.  Along the same lines, your usable area for the machine is where the space where the camera and both heads can reach.  So expect to loose real estate because of this, you can see the red marks where I have estimated my usable real estate.   If you did a 4 head system with another Z head on back for the top rail, the usable real estate would further decrease. 

Note for automatic reel feeders you might not need the camera so you might be able to place them in the "Dead zone" on one front side,  for design above. 

To configure the counterbalanced dual head system will require you editing the machine.xml file, so install a good text editor. 

Smoothieware documentation mentions the axis gcode being having an "Alt" and "B" axis, however I found no documentation (maybe poor googling) that mentioned "E" axis which was for extruder on 3D printers. Therefore it took some time to figure out how to rotate the B axis. Basically the 3D printer firmware for smoothie uses "E" axis for extruder, then to change to different extruder you send "T0" or "T1" to change "tools", this changes which axis the gcode "E" command moves.  

Double stick tape is your friend in getting started, put on PCBs to hold parts when placed. Use to hold down cut tape strips, etc. 

It is far better to copy a design than to build your own. That is the configuration of the steppers, setting up heads and such will take you hours (or days in my case). Thus it is far easier to copy someone else's design and configuration that works, but where is the fun in that... 

OpenPnP like all software requires you to have a certain work flow and mental model.  For beginners like me I often come biased to what I feel the work flow should be, usually this exposes bugs as that no one would have ever thought to test before.  Therefore it is best to take the time and learn the work flow of the software before starting. Read all of the wiki and watch all of Jason's videos to start getting a feel for it.  Even then you will run into problems... 

I would have never gotten this far without Jason's help. So I owe him several beers! 

Note to date I have only been able to place one part on the PCB. So I am not done with my learning.

Trampas

www.misfittech.net

   

[Anthony Webb]

You've come a long way my friend.  Keep up the good work!

--
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/aa99d7c7-b5a6-47ed-8caf-2615e48e134e%40googlegroups.com.