Christian's head - quadruple head

[Christian Thomasser]

Hello, it's me again !!

Based on your objections to a thread driven Z-axis a few months ago, I designed a quadruple head driven by belts.

Before I build this now, I ask you if I overlooked something important or did something wrong. I am open to any suggestions for improvement.

Note: I did not draw any screws or wiring boxes/mounts, that would have been too much work for me.

best regards
Christian

[John Plocher]

I like where your head is at 😎

Will the close spacing of the heads impact the size of components they can pick without interference?

E.g., if you pick large ICs on two of the heads, and need to rotate them, will they knock into each other?

  John

On Tue, Nov 5, 2019 at 4:13 PM Christian Thomasser <Christian...@gmx.at> wrote:

Before I build this now, I ask you if I overlooked something important...

[Christian Thomasser]

Hello John,

Yes they will knock into each other.

But this is not important for me because i don't use ICs in that size.

Should I ever have to place such ICs, I would place them one at a time.

The spacing between the nozzles is exactly 22mm

[Jason von Nieda]

Nice looking design Christian! Looks quite compact. Is the camera and camera mount one of the common "Mini Case" cameras with a CS mount lens?

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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/07f384bd-5bbb-46ca-bb12-917c2b8d44ae%40googlegroups.com.

[Christian Thomasser]

It is this camera:

Svpro 2,0 Megapixel 2,8-12 mm...
https://www.amazon.de/dp/B07R7JXV35?ref=ppx_pop_mob_ap_share

[Jason von Nieda]

Okay, thanks. That is similar to what I am thinking of. I recently picked up a similar one with global shutter and I was impressed with how solid the case was.

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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/2b787d8b-b578-4b88-9673-898d41ba02dc%40googlegroups.com.

[Pavlin Stoyanoff]

Here is my design - Click

The belts are missing and some of the bearings from behind. But I think it is ok to get the idea. I think it is simple.

[Maple_Dude]

nice design

[Sandra Carroll]

How do you plan to route the vacuum to head head?

If there’s 22mm spacing some if not all is going to be taken up by the vacuum leads assuming these are thru shaft motors on each head

 

Nice design it’s very simple which is great but packs allot into that small package

 

Sandra

--

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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/07f384bd-5bbb-46ca-bb12-917c2b8d44ae%40googlegroups.com.

[Christian Thomasser]

Thank you all for your comments!!
With my unsufficient talent in english it’s a bit complicated for me to explain the vacuum and air pressure routing of the seperate heads. But i will take some pictures of, and on of a cut through the rotating nozzle holder. There you can see how it must be working ... hopefully 😂

[Maple_Dude]

What size are the Up-Down motors at the top? Nema 17? They must be really thin to fit in that space. 

[Christian Thomasser]

That‘s right! They are Nema 17 0.4A 28Ncm. Hope they are strong enought. I eill test it the next days.

[Maple_Dude]

Yes, please keep us posted!

[Christian Thomasser]

Okay, now i've testet the strength of the stepper.

At this time I am sure it is strong enough. But i am not sure about the speed in the current designed.

I designed with 20T pulleyes. The have a diameter of about 12mm that is a circumference of 37,7mm.

Step frequency will be 1kHz (I've tested, this works fine) -> 200 Steps/rev -> this will be 5 Turns per second (start and stop ramps not considered)

After all calculations, about 188mm per second will result as maximum Z-Axis Speed. I think in practice (because of ramps) it will be abaout 150cm/sec.

But what i don't know is .... Is this slow, okay or fast?? Is it usefull to achieve more speed on the Z-Axis???

When the head is finished i will design the rest of the machine. For X-Axis i intend to use a 12Nm Nema34 closed loop stepper and about 5Nm for Y-Axis -
Does anyone know if this performance is sufficient? Maybe for a slightly heavier and stiffer design too??

As i promised today i will now show you the routing of the vacuum.

I don't use Nema 8 with hollow shafts. I've designed a construction where the stepper drives a partially "semi" hollow shaft capsulet in an airtight round body mounted under the stepper.

this system is almost absolutely airtight. the two holes in the body are for vacuum and compressed air. Maybe i will add a third hole for a smc vaccum sensor but i am afraid it will be damaged by a compressed air blast i will use to blow of the smt parts.

I still have to test how though these sensors are. --- Does anybode have a better solution??

[Iceturf]

Interesting. 

    My head currently moves Up/Down at 6000mm/min = 100 mm/s. I've seen nozzles on other machines run at double the speed my head runs at.  

--
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/BOmK72AsypY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/7b072db8-d690-469e-b061-9dd2080028fc%40googlegroups.com.

[Emil Johnsen]

You can probably run it at 1000 to 1200rpm before it starts loosing much torque.

[Christian Thomasser]

I tried that. I slowly increased the step frequency while the motor was running. at 1.2 kHz, the armature is no longer behind the rotating field

Maybe i have to optimise the stepper driver, or get better ones.

I think ther is anough power to use 40T pulleyes. This will double the speed.

For more power maybe I'll overload the stepper motors too. There is still room for a Nidec fan in the head. This keeps even overloaded engines cold.

In addition, I will work with the enable input to save power and avoid engine heat.

[Maple_Dude]

Without doing the math myself - I find most stepper motors cap out ~200 RPM for open loop. What is your stepper motor running at during MAX speed?

[Christian Thomasser]

At the moment 300rpm. This works pretty fine. The stepper also starts with good torque and that without a ramp.

[Michael Anton]

I have no problem driving the steppers on my 3D printer to 450rpm, and they are open loop.  The X and Y axis on that machine can run up to 300mm/s, with 20 tooth GT2 pulleys (40mm/rev, so 7.5rev/s * 60s/min = 450rpm).  That said, the drivers run at 24V, so if you were running at 12V, then 200rpm would be about right.

[Christian Thomasser]

Where is the best place to mount the valve block??
I intend to use hoses with an inner diameter of 2mm between the individual heads and the valve block. From the valve block to the compressed air and vacuum tanks, I use the next larger diameter.

I like to mount the valve block on the quadruple head unit. But i am not sure if the extra weight is bad for performance. Otherwise i dont know if it works well when the 2mm hosses are to long in that case i will mount the valves in the machine socket.

[Mike M.]

Hi Christian,

Have a look at this web - https://docs.mgrl.de/maschine:pickandplace:vacuumsystem

for your Qadrature head this is total of 8 valves + 4 vacuum readers - It would be good to have them as near of nozzles as possible,.. but if the space is the problem than better off the head...

For min size see here enclosed manifolds from SMC -- https://www.smcpneumatics.com/pdfs/VQD.pdf

Mike

[Mike M.]

2mm ID vacuum hose is OK

[Kylra]

Hi guys,

i added a schematic with four pressure sensors. Honeywell MPR series.

https://github.com/Kylra/Kylras-Tiny-Motion-Controller

Kylra

[Maple_Dude]

Kylra, 

    dumb question on my part - in your schematic you have 4 temperature sensors. What are these for? Perhaps to check the temperature of the stepper motors? 

thanks

[Kylra]

They are for the drivers.

Kylra

[Kylra]

To check the temperature of the steppers is a good idea.

Kylra

[Maple_Dude]

Oh, have you tried the Soft Polyurethane hoses? I found they are kind of nice to use on the PnP head. 

[Christian Thomasser]

I have not tested them yet, but I already have 50 meters of it at home because I want to use exactly these. 😁

[Mike M.]

Look for Silicine hose - abt ID 3mm is my preference

[Maple_Dude]

Do you use a barb on the Hollow Stepper motor or how do you manage an air seal? 

[Christian Thomasser]

I don't use hollow steppers.

Air pressure and vacuum are fed under the stepper motor

I hope you can see it in the sectional view:

[Christian Thomasser]

The air seal is done by the two sealed Bearings (2RS)

[Maple_Dude]

Sorry, you mentioned this before - I forgot. 

[Mike M.]

For manifold - valve block = 4 x10mm soleonid size 60x25 x 10mm

https://www.aliexpress.com/item/4000181743051.html?spm=a2g0o.productlist.0.0.2cc52b559vRgH9&algo_pvid=e2cca3fd-95cf-49b5-899d-481234701e45&algo_expid=e2cca3fd-95cf-49b5-899d-481234701e45-1&btsid=fa2b97f4-31ab-41ad-8dd0-fe237ad7d41e&ws_ab_test=searchweb0_0,searchweb201602_8,searchweb201603_55

see enclosed

Mike

On Tuesday, 12 November 2019 17:18:25 UTC+1, Christian Thomasser wrote:

[Iceturf]

What solenoids are compatible with these solenoid blocks?

--

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/BOmK72AsypY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/f50bc7f6-c131-4d0b-bb4b-5fcd98a033c3%40googlegroups.com.

[Mike M.]

Hi,

Google SMC- SY114 adat to your DC Voltage

-Google as exsample -https://www.ebay.de/itm/Smc-Corporation-SY114-5LOZ-3-Port-Direkt-Betriebenes-Magnetventil/293214737757?epid=10034148172&hash=item4444f5d55d:g:fCcAAOSwXFldIQ5p

- https://www.ebay.de/itm/Smc-10-sy114a-5gs-Ventil-Solenoid-3-Port/163570718934?hash=item26159318d6:g:BUAAAOSw241YWIZv

Mike

On Thursday, 21 November 2019 18:44:15 UTC+1, Iceturf wrote:

What solenoids are compatible with these solenoid blocks?

On Thu, Nov 21, 2019 at 10:06 AM Mike M. <mike...@gmail.com> wrote:

For manifold - valve block = 4 x10mm soleonid size 60x25 x 10mm

https://www.aliexpress.com/item/4000181743051.html?spm=a2g0o.productlist.0.0.2cc52b559vRgH9&algo_pvid=e2cca3fd-95cf-49b5-899d-481234701e45&algo_expid=e2cca3fd-95cf-49b5-899d-481234701e45-1&btsid=fa2b97f4-31ab-41ad-8dd0-fe237ad7d41e&ws_ab_test=searchweb0_0,searchweb201602_8,searchweb201603_55

see enclosed

Mike

On Tuesday, 12 November 2019 17:18:25 UTC+1, Christian Thomasser wrote:

Where is the best place to mount the valve block??
I intend to use hoses with an inner diameter of 2mm between the individual heads and the valve block. From the valve block to the compressed air and vacuum tanks, I use the next larger diameter.

I like to mount the valve block on the quadruple head unit. But i am not sure if the extra weight is bad for performance. Otherwise i dont know if it works well when the 2mm hosses are to long in that case i will mount the valves in the machine socket.

--
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/BOmK72AsypY/unsubscribe.

To unsubscribe from this group and all its topics, send an email to ope...@googlegroups.com.

[Mike M.]

And I drilled larger holes in manifold original was 1.1mm --- increaded to 1.7mm

Mike

On Thursday, 21 November 2019 18:44:15 UTC+1, Iceturf wrote:

What solenoids are compatible with these solenoid blocks?

On Thu, Nov 21, 2019 at 10:06 AM Mike M. <mike...@gmail.com> wrote:

For manifold - valve block = 4 x10mm soleonid size 60x25 x 10mm

https://www.aliexpress.com/item/4000181743051.html?spm=a2g0o.productlist.0.0.2cc52b559vRgH9&algo_pvid=e2cca3fd-95cf-49b5-899d-481234701e45&algo_expid=e2cca3fd-95cf-49b5-899d-481234701e45-1&btsid=fa2b97f4-31ab-41ad-8dd0-fe237ad7d41e&ws_ab_test=searchweb0_0,searchweb201602_8,searchweb201603_55

see enclosed

Mike

On Tuesday, 12 November 2019 17:18:25 UTC+1, Christian Thomasser wrote:

Where is the best place to mount the valve block??
I intend to use hoses with an inner diameter of 2mm between the individual heads and the valve block. From the valve block to the compressed air and vacuum tanks, I use the next larger diameter.

I like to mount the valve block on the quadruple head unit. But i am not sure if the extra weight is bad for performance. Otherwise i dont know if it works well when the 2mm hosses are to long in that case i will mount the valves in the machine socket.

--
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/BOmK72AsypY/unsubscribe.

To unsubscribe from this group and all its topics, send an email to ope...@googlegroups.com.

[Per Eklund]

Does any one know what thread it is on the manifold in the link below? Can't find that information.

I'm looking for a matching fitting.

https://www.aliexpress.com/item/32814345399.html?spm=2114.12010612.8148356.8.67516c74wuxAFR

--Per

[Mike M.]

Hi Per,

If you are refering to image of manifold above ( 4M_Dim_60x25x15.jpg) it has all M5 threads for quick connection fittings.

Mike

[Per Eklund]

Thanks!

Do you use 4 or 6mm hose?

[Javier Hdez.]

Awesome!

Thank you for share it. I like for my next machine.

El miércoles, 6 de noviembre de 2019, 0:13:34 (UTC), Christian Thomasser escribió:

Hello, it's me again !!

Based on your objections to a thread driven Z-axis a few months ago, I designed a quadruple head driven by belts.

Before I build this now, I ask you if I overlooked something important or did something wrong. I am open to any suggestions for improvement.

Note: I did not draw any screws or wiring boxes/mounts, that would have been too much work for me.

best regards
Christian

[Mike M.]

Per - I use OD= 4mm hose, Abt ID=2mm

Mike

[Mike M.]

Finaly all parts arrived and I got most of my head assembled last night

[Mike M.]

one more with juki

[Balázs buglyó]

Mike, is this design opensource?

--
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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/2645afd3-c4a2-4eab-b97f-079b2e0ec2f8%40googlegroups.com.

[Mike M.]

Hi,

I have no problem with sharing my desighn files and explanations of used parts but for now I have to test it first

I am not sure if Nema14 can handle Z motions

I am using this steppers but I could change this Z motion steppers - https://www.ebay.de/itm/0-9-Degree-NEMA14-2-phase-4-wire-35MM-Stepper-Motor-Copper-Pulley-CNC-3D-Printer/263942579499?epid=16021090588&hash=item3d7434252b:g:8rQAAOSwhmNcwrzw

 As you can see from pictures Vacuum is taken out from head from one side which eliminates use of Hollow shaft steppers.

Mike

On Wednesday, 19 February 2020 12:23:08 UTC+1, Balázs buglyó wrote:

Mike, is this design opensource?

Mike M. <mike...@gmail.com> (időpont: 2020. febr. 19., Sze, 11:49) ezt írta:

one more with juki

On Wednesday, 19 February 2020 11:42:34 UTC+1, Mike M. wrote:

Finaly all parts arrived and I got most of my head assembled last night

On Wednesday, 4 December 2019 21:44:14 UTC+1, Mike M. wrote:

Per - I use OD= 4mm hose, Abt ID=2mm

Mike

On Wednesday, 4 December 2019 17:47:37 UTC+1, Per Eklund wrote:

Thanks!

Do you use 4 or 6mm hose?

--
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 ope...@googlegroups.com.

[Marek T.]

Hi Mike,

I must have missed some discussion about it... so how do you deliver the vacuum to the nozzles?

[Mike M.]

See here:

[Mike M.]

This are the nozzle to nozzle dimensions - Head is 70 x 70mm size

[Marek T.]

 Looks really nice if will work good with hermeticity (tightness) good enough.

[Jim]

On Wednesday, February 19, 2020 at 3:05:44 PM UTC+1, Mike M. wrote:

This are the nozzle to nozzle dimensions - Head is 70 x 70mm size

Hello Mike,

Your new head looks cool - a light weight design.
Looking forward to see it in action with attached camera and coax light.
Are you planning to use your existing tip garage?

Jim

[Mike M.]

Marek - with Juki there is no 100% sealing (nozzle and nozzle vertical guide has gap (at spring location) and you can not avoid vacuum lost). But the good thing is vacuum lost is minimum & constant with all nozzles becouse the size of nozzle body and hole is for all nozzles the same.

@Jim - I am using this coaxial camera already now with two nozzles & I know this will work perfectly for me,...

Regarding Tip storage-changing - I do not think I will need any nozzle change - with 4 nozzles I can handle all my SMT parts for a Job. If needed manual change will do.

Mike 

[Dave B.]

Keep up the good work.  This is really interesting to see!

-Dave

--

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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/0e18a325-d758-4de0-bf69-6997b7e2bd8f%40googlegroups.com.

[Mike M.]

I opened a new page here: https://groups.google.com/d/msg/openpnp/FKnGRrXwaGU/cf9XhfOoAwAJ

Mike

[Marek T.]

I know this Mike, I have meant tightness no worse than with using hollowed shaft motors.

[Yevhenii Shcherbakov]

I don't really understand how to control such a quad head. 

I don't remember where, but on some of the resources I read that openpnp can control the X Y Z C1 C2 axes only through one controller / board. 

Smoothieboard only supports 5 axes, Sbase is similar to 5 axes.

So how do you manage this? 

Maybe I missed a simple solution?  

среда, 19 февраля 2020 г. в 20:35:00 UTC+2, Marek T.:

[Yevhenii Shcherbakov]

I had said that a smoothieboard can control only 5x axes, but in a PnP with a quad head you need to control 8x axes.

вс, 13 сент. 2020 г., 14:37 Yevhenii Shcherbakov <yevhenii.s...@gmail.com>:

--
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/BOmK72AsypY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/d760a16c-5850-4a11-8723-858e44a1c3fcn%40googlegroups.com.

[serge]

hi,  answer is in the documentation :-)  add more controller board 

[Arthur Wolf]

Do the non-XYZ steppers need to move at the same time, or will they always move one after the other ( never together/at the same time )? Or at least the last 3 steppers.

If that's the case, you can control 8 steppers with a single Smoothieboard, just ask me how and I'll explain.

Cheers.

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 view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/df2f70d8-85d9-4eef-8079-af6e186ab057n%40googlegroups.com.

--

勇気とユーモア

[Yevhenii Shcherbakov]

serge, wolf.arthur, I will be very grateful if you explain how to connect them.

вс, 13 сент. 2020 г., 18:26 Arthur Wolf <wolf....@gmail.com>:

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAMHZkm4zgkjcUxq4f5fR3RyJ3L%3DjE1MuObHvseSoo7jBggLi2w%40mail.gmail.com.

[Arthur Wolf]

You would essentially use the exact same technique used for beds that have more than one stepper motor ( typically 3 or four ).

Get a Smoothieboard 5XC, then if you need 3 more axes ( total of 8 ), get 3 TB6600 stepper drivers ( or neater ones like trinamic drivers in pololu quick-connect adapters ), wire them to any pins that are free, and start following the instructions at : http://smoothieware.org/switch?s[]=multi&s[]=axis&s[]=homing#homing-a-multi-motor-axis

Now you've got three stepper drivers all sharing the same axis ( that'd be C here ), but you also have M-codes that allow you to enable/disable any of those axes as you wish, which you can use to do anything you want from there.

If you run into any trouble don't hesitate to let me know.

Cheers!

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAHyyacEc2xYbKcXTism_MUMR%2BhNYD9Q6i_j5qvXahS3G0xR%3Diw%40mail.gmail.com.

--

勇気とユーモア

[ma...@makr.zone]

Hi

I just wanted to point out that I'm working on optimizations of motion control in OpenPnP that will not work with multiplexed axes. The idea is to blend the motion of axes together to increase speed, e.g. the travel time should be used to rotate all the axes (parts) at the same time.

So if you find another way i.e. a true 8-axis controller, that would be much better. Apparently the Duet3 controller can do that using a CAN-bus connected extension, though I haven't tried it!

https://duet3d.dozuki.com/c/Duet_3_Mainboard_6HC

https://duet3d.dozuki.com/c/Duet_3_Expansion_board_3HC

IMHO, aiming for a 4 nozzle machine is really ambitious. You should be prepared to spend some money and effort on the controller ;-) Otherwise it might be better to build a lighter and faster machine with less nozzles.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAMHZkm5sbr5pi6jSi%3D7Q6e1o4M_10Th_PcwOjxPQu_LdmrPX-g%40mail.gmail.com.

[Mike Menci]

Hello Mark

"Shared C" (for 4 nozzles)  should remain to work ?

Mike

[Arthur Wolf]

On Sun, Sep 13, 2020 at 7:24 PM ma...@makr.zone <ma...@makr.zone> wrote:

Hi

I just wanted to point out that I'm working on optimizations of motion control in OpenPnP that will not work with multiplexed axes. The idea is to blend the motion of axes together to increase speed, e.g. the travel time should be used to rotate all the axes (parts) at the same time.

So if you find another way i.e. a true 8-axis controller,

Note if somebody wants to make Smoothie actually natively 8 axis, that's likely less than 10 lines of code to change for basic 8-axis support.

If somebody wants to take a crack at it, I can help them, a lot.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/91da4e50-d839-63b4-9b19-1ad056b99c5a%40makr.zone.

--

勇気とユーモア

[Jason von Nieda]

Hi Mark,

Are you saying your changes will not support axes spread across multiple controllers at all, or just that it won’t be as well optimized?

Thanks,

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/91da4e50-d839-63b4-9b19-1ad056b99c5a%40makr.zone.

--

Sent from my BeOS enabled toaster

[ma...@makr.zone]

Yeah, shared axes are ok, but of course they can't benefit from that optimization (but it can still benefit from other optimizations.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/d5206245-0b45-471f-a0cf-b4c358381947n%40googlegroups.com.

[Yevhenii Shcherbakov]

Wolf arthur

But in this case, there will not be enough ports for the auxiliary periphery: vacuum and purge valves (in the case of using a set with Aliexpress).

In theory, it is easier to use a set of demultiplexers, for example 4051, and take the microcircuits according to the amount of required equipment. But I do not know how to send a code via Gcode to select the desired multiplexer output.

P.S. Now I am designing my first PnP mechanism, and I am very familiar with the possibility of openPnP. My task is to achieve a performance of 4000 components / hour, with a board size of 600x600.

вс, 13 сент. 2020 г., 19:28 Arthur Wolf <wolf....@gmail.com>:

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAMHZkm5sbr5pi6jSi%3D7Q6e1o4M_10Th_PcwOjxPQu_LdmrPX-g%40mail.gmail.com.

[Arthur Wolf]

Oh I'm sorry I don't know anything about the counterfeit boards on aliexpress. ( this page gives a bit of an explanation why we don't collaborate much with those: http://smoothieware.org/troubleshooting#what-is-wrong-with-mks )

If you were using a Smootheboard I'd be able to help you free additional pins as you need them, but on a clone/counterfeit board I don't think I can help you ( I doubt the seller will either ).

If you know how to drive a multiplexer, you can use the switch module ( smoothieware.org/switch ) to control any pin with any m-code, so that would work for you I expect.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAHyyacH%3DNBafHQyHudPUFg%3DLOLR6CJjq%2Bxtd5UV4EXNpRbQGog%40mail.gmail.com.

--

勇気とユーモア

[ma...@makr.zone]

Hi Jason,

> Are you saying your changes will not support axes spread across multiple controllers at all, or just that it won’t be as well optimized?

just that it won’t be as well optimized.

Basically, it will have to revert back to the old way of always waiting for each move to complete, before the next move can be started. For example, if the Z axis is on a different controller than X, Y and I want to do the typical moveToLocationAtSafeZ() move, I will need to move up Z first, then wait for completion, then start moving X, Y, then wait for completion, then move Z down again. Otherwise I could crash the nozzle, obviously. I call this "interlocking". It is only needed between moves that involve different controllers (but unfortunately, this will often be the case, if Z or C are distributed). These limitations mean, that the multi-threading and asynchronous operation that I implemented will be useless in these move sequences. Machines will incur the full communications "ping pong" lag and the will physically perform full stops between moves.

Having said that, all these machines will still benefit from all the other enhancements on single moves: like jerk control simulation, axis-centric rate and acceleration control etc. OpenPnP can also still perform a single move across as many controllers as you like (formerly just two could move simultaneously: the main driver and the first sub-driver).

On the other hand, if everything is on a single controller, all these moves can be queued and executed asynchronously and even be blended. I'm talking about the Z "overshoot" that can be used to avoid slowing to a stop in the corners.

For C it is the same thing, only less obvious! Frankly, the full ramification of C needing the same blending only occurred to me some days ago. :-/ Until then I thought that putting C axes on a second controller will be fine. It won't.

There is a slight hope that it may still be possible: if we determine that our prediction of the timing of controllers is 100% reliable (it is so far on my Smoothie tests), we could send dwell commands (G4) to synchronize intermittently non-moving controllers to the moving ones and therefore generate a continuous fully timed "punched tape" across controllers... It might be too dangerous for Z (collisions!) but acceptable for C. The worst case there is that a part would be "stirred" in the solder paste, because a C axis command took longer than predicted.

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jyZG2M-aADnVtvXP8SQL8df5TTA1qTumJ04HeL3KPR1Zg%40mail.gmail.com.

[Yevhenii Shcherbakov]

Arthur Wolf,  

When I mentioned aliexpress, I was referring to the use of head kits, as well as ejector systems and high pressure valves to capture and dump components.

Type of this

https://m.aliexpress.com/item/32807025833.html?spm=a2g0n.productlist.0.0.1c967ea67gDlPv

This system uses 2 valves per nozzle.

And the S moothieboard doesn't have a lot of free pins.

P.S. Sorry for my English.

вс, 13 сент. 2020 г., 21:29 ma...@makr.zone <ma...@makr.zone>:

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/be40f27e-a1f3-9474-2d06-0ab18966af94%40makr.zone.

[Arthur Wolf]

If you have a Smoothieboard and you are running out of free pins, there are several ways you can go to get more.

You can disable the 4 leds ( in config ) and use these as gpio, you can start using the play/pause button+pin leds, you can use the pins on the second spi port.

If you need even more there are a few less obvious ways, don't hesitate to email me directly.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAHyyacEyd_D0_ovpPd%2Bkx7O1Muj8L2tjgqbF%2Bmx6b8MHCGg5GQ%40mail.gmail.com.

--

勇気とユーモア

[Jason von Nieda]

Thanks Mark - just wanted to make sure that ability would not be removed. I think it's important that we maintain flexibility in controllers.

Thanks,

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/be40f27e-a1f3-9474-2d06-0ab18966af94%40makr.zone.

[Marek T.]

I think that 4000cph is mission impossible if you want to perform vision alignment for every parts.

W dniu niedziela, 13 września 2020 19:57:06 UTC+2 użytkownik Yevhenii Shcherbakov napisał:

Wolf arthur

But in this case, there will not be enough ports for the auxiliary periphery: vacuum and purge valves (in the case of using a set with Aliexpress).

In theory, it is easier to use a set of demultiplexers, for example 4051, and take the microcircuits according to the amount of required equipment. But I do not know how to send a code via Gcode to select the desired multiplexer output.

P.S. Now I am designing my first PnP mechanism, and I am very familiar with the possibility of openPnP. My task is to achieve a performance of 4000 components / hour, with a board size of 600x600.

вс, 13 сент. 2020 г., 19:28 Arthur Wolf <wolf....@gmail.com>:

You would essentially use the exact same technique used for beds that have more than one stepper motor ( typically 3 or four ).

Get a Smoothieboard 5XC, then if you need 3 more axes ( total of 8 ), get 3 TB6600 stepper drivers ( or neater ones like trinamic drivers in pololu quick-connect adapters ), wire them to any pins that are free, and start following the instructions at : http://smoothieware.org/switch?s[]=multi&s[]=axis&s[]=homing#homing-a-multi-motor-axis

Now you've got three stepper drivers all sharing the same axis ( that'd be C here ), but you also have M-codes that allow you to enable/disable any of those axes as you wish, which you can use to do anything you want from there.

If you run into any trouble don't hesitate to let me know.

Cheers!

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

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/df2f70d8-85d9-4eef-8079-af6e186ab057n%40googlegroups.com.

--

勇気とユーモア

--
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/BOmK72AsypY/unsubscribe.

To unsubscribe from this group and all its topics, send an email to ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAMHZkm4zgkjcUxq4f5fR3RyJ3L%3DjE1MuObHvseSoo7jBggLi2w%40mail.gmail.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 ope...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAHyyacEc2xYbKcXTism_MUMR%2BhNYD9Q6i_j5qvXahS3G0xR%3Diw%40mail.gmail.com.

--

勇気とユーモア

--
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/BOmK72AsypY/unsubscribe.

To unsubscribe from this group and all its topics, send an email to ope...@googlegroups.com.

[bert shivaan]

I think even without vision it will be a stretch. That is just over 1 part per second.

To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/92a5af9b-aa73-4e94-a327-8f2c973cf9e1o%40googlegroups.com.

[serge]

don't seem to be realistic, here is an fast 4 heads openpnp base machine with servo. https://www.youtube.com/watch?v=y14pdfjYsyo

 its 2218cph 

[Marek T.]

I have 3 pretty fast pneumatically driven nozzles and servomotors few hundred watts each.

And my machine reaches some 1500cph with vision (so this 2218 sounds still real) and moreless 3500cph without vision.

[Yevhenii Shcherbakov]

Thanks for the info.

пн, 14 сент. 2020 г., 15:02 Marek T. <marek.tw...@gmail.com>:

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/2603c419-55a4-4e40-acf4-040a0cf70bddo%40googlegroups.com.

[Christian...@gmx.at]

Hi guys,

can anyone of you give me a tip for a vacuum pump that generates enough vacuum for a quadruple head (speed about 2,000 pps).

Is one of these pumps sufficient? Or how much pressure and liters per minute do I need??

https://www.amazon.de/Vakuumpumpe-Elektrische-medizinische-Behahrung-Instrument/dp/B0786L1C3R/ref=sr_1_6?dchild=1&keywords=Mini+Vakuumpumpe&qid=1601489260&sr=8-6

or

https://www.amazon.de/Vakuumpumpe-Uunterdruck-Saugpumpen-Halter-500mA/dp/B07F5CQ1CD/ref=sr_1_11?__mk_de_DE=%C3%85M%C3%85%C5%BD%C3%95%C3%91&dchild = 1 & keywords = mini + vacuum pump & qid = 1601489460 & sr = 8-11

or

https://www.amazon.de/Vakuumpumpe-Uunterdruck-Saugpumpen-Halter-500mA/dp/B07H7CSXC5/ref=sr_1_11?__mk_de_DE=%C3%85M%C3%85%C5%BD%C3%95%C3%91&dchild = 1 & keywords = mini% 2Bvakuumpumpe & qid = 1601489460 & sr = 8-11 & th = 1

How durable are these things anyway?

[Mike Menci]

For start go with TWO of this: https://www.ebay.de/itm/DC-12V-15V-Double-Head-Suction-Pump-Mini-Diaphragm-Pump-Vacuum-Air-Pump-80kpa/122117403709?hash=item1c6ec3b43d:g:EB4AAOSwFdtXyXw8

Each nozzle uses 1 diafragm head.

Mike

[Christian...@gmx.at]

Hi Mike,

Thank you for youranswer!

I forgot to mention that I want to use a vacuum tank with a volume of about one liter.

And because of maybe no measurable loss of vacuum I also will use a air pressure tank for blow of option

Another important thing should be mentioned at this poin:

Because the two ball bearings let too much air through in the test, I redesigned the nozzle rotating device a bit

I added two shaft seals and a spacer sleeve.

This means that I should no longer have a measurable loss of vacuum in the nozzle turning device.

Because I have not tested it so far (I am stil waitign for the seals), I don't know yet whether the shaft seals have too much friction for the low torque of the stepper motors,...

... but I just think positively :)

[Mike Menci]

Christian,

Which nozzles and which bearings are you using? You have a link ?

Are the bearings pressed into 3D printed housing or ?? material?

With Nema8 you might have difficoulties to turn shaft if the seal will have to much friction...

Can you share your desighn or cross section view?

People here might help!

Mike

[Christian...@gmx.at]

Hi Mike,

I use juki nozzles and standart 8x14x4mm bearings.

the housing is not 3D printed, it's machined out of solid POM

Here is the desing of the turnig section:

[Michael Anton]

If those are rubber seals on the shaft, you might want to flip them around from what you have drawn, as they will seal better.  As drawn, the vacuum will try to bend the seals outwards, causing a leak, whereas the other way around they would be pulled towards the shaft, and seal better.

[Christian...@gmx.at]

You are of course right, but I try to achieve the exact level of vacuum where it is just kept in between and at the same time reduces the friction of the seals on the shaft. The seals are made of rubber, but it is a very stiff rubber. So enough vaccum should build up before the seal fails.

[Christian...@gmx.at]

Hello guys, there is news again!

I have now built and tested a single pipette unit.

The two shaft seals are braking less than I thought.

The stepper motor manages to move the nozzle holder at maximum speed without loosing steps.

But to be on the safe side, however, I will lathe off the nozzle holder by one to two tenths of a millimeter in the area of the seals to reduce friction even more.

By the way, the system is completely tight and I don't lose any vacuum at all, right up to the juki nozzle.

Here I have a few pictures of the individual pipette unit:

Next to the vacuum connection there is a second connection for system ventilation or optional blow of air.

This one is not connected yet, because I have to make a few more of the brass nipples

[Christian...@gmx.at]

Here is the latest design of my quadro head:

[Christian...@gmx.at]

Hi Guys,

I wrote to several companies and asked for an offer for the aluminum milled parts. The cheapest offer I received was this:

Quadro Head Ground plate: € 156.00 (one needed)
Quadro Head Side Plate: € 192.00 (two needed)

Shipping Europe-wide € 12.90

Anyone interested in the parts ?? I would then order several parts at once.

[Yevhenii Shcherbakov]

An expensive design turned out. 

Can rework the project? 

Aluminum parts for the head of my design cost about 40 € + welding (20cm).

пятница, 30 октября 2020 г. в 00:06:26 UTC+2, Christian...@gmx.at:

[Shai]

Very nice quad head design! If I may suggest, I would put a set screw on the bottom left of the camera where the lens goes into the hole. Otherwise your camera will swivel on the top axis and vibrate as the machine moves, causing inaccuracies. Looking forward to seeing it work!

[Shai]

Also - what are you using for the linear blocks there? Link?

And I would be against using bearings to prevent air leakage. I've tested it before and air does leak out. Jason had a simple solution to use an O-ring with some oil on it and it works great. Just have it slightly compressed (10% compression of O-ring height). See below how I've done it.

[Christian...@gmx.at]

Hello Shai,

Thanks for the hint with the adjusting screw at the bottom of the lens. I will come up with a nice solution.

I use the MGN9H linear carriages with MR9 Rails.
can be easily found on ebay:
E.g .:
https://www.ebay.at/itm/MGN9H-100mm-Mini-Linearfuhrung-Linearschienen-Linearfuhrungsschiene-Gleitblock/323553021161?hash=item4b55434ce9:g:lLUAAOSwabRb7qCq

8 posts above you will find a cutaway model of my rotary nozzles.

I also tested whether the sealed ball bearings were sufficient to hold the vacuum. As you correctly write out, this doesn't work at all. For this reason I also used shaft seals between the ball bearings.

To reduce friction between the hollow shaft and the shaft seals i lathed the shaft between the bearings to 7.7mm.
The systems are completely sealed and the small nema 8 stepper motors (length: 34mm) can turn the nozzles at 300 RPM without any problems. Apart from the Juki nozzles, I don't lose vacuum anywhere else. During the tests, I was surprised how extremely well it worked.

[Shai]

Yes I saw the model. I still prefer Jason's way after doing lots of tests :)

By the way, since you mentioned you will pick large ICs one at a time (so they don't interfere) since the nozzles are close together, how does this happen in OpenPNP? Does it automatically realize the nozzles are close and only picks one large IC at a time to avoid collision? 

[Shai]

By the way, looking at this again - I don't think you need 4 end stops. Shouldn't 1 endstaop for each motor be sufficient?

[Christian...@gmx.at]

Hey Shai,

Right now the largest ICs I use are 18f45k80 - so far this issue does not need to be fixed - I only scratch when it itches :)

No, two end stops for each motor ist very important.

If I only need the end stops for referencing, I would agree with you.

But I also need the limit switches to check whether both nozzles are really in Safe Z again before I move in X and Y.

I don't trust the linear bearings 100 percent. If one of the bearings got stuck and the nozzle didn't go up again, I would have a crash.

[betzt...@gmail.com]

Can you show a model (or just images) of the parts in question? Are you saying 540 Euro for 3 parts? Or 348 Euro for 3 parts? Are they flat parts that can me machined from @ 6mm aluminum plate? It is possible I can help....

Peter.

[Christian...@gmx.at]

Hello Peter and all other guys,

thank you for your help!!
I have now considered using laser cut parts instead of the very expensive machined ones.
I found a company nearby that sells all three pieces for a total of € 80 thank you for your help.

I have now considerd to use laser cuted parts instead of those very expensive machined one.

I have found a company nearby that offers all thre part for 80 €

I have also reduced the thickness of the plates to 5mm.

[Betz Technik Industries Ltd.]

 

Laser cut aluminum parts? That requires a very expensive fiber laser. Good for you if you have found someone to operate it at a reasonable price.

 

Regards,

Peter.

 

 

[jdlv]

Hi Christian,

laser cutting 10mm aluminum looks possible:

https://www.john-steel.com/en/

If you have dxf files you can have a instant quote.

joël

[Iceturf]

how square is the end of a laser cut part? 

--
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/BOmK72AsypY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to openpnp+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/8ade49c8-7756-4d0d-426e-40c9d470d913%40gmail.com.