Anyone with a working TinyG?

[Jason von Nieda]

Hi all,

In another thread I'm helping Chris try to get his TinyG based machine working. Does anyone have a working TinyG configuration they can share? Especially if you are using the GcodeDriver. Alternately, if you are using the TinygDriver, can you share the version of TinyG you are using, and whether you had to make any modifications?

Thanks,

Jason

[Daniel]

I am working somehow with TinyG.

But it's a custom board and even some changes on firmware.

If is possible I would try to help Chris to start his machine.

BR

Daniel

[obiwanke...@gmail.com]

I'm not all setup yet, but I am using the tinyG and it's driver and can move X, Y, Z, A. Do you just need the machine.xml ?

Chris have you looked at the tinyG setup instructions ? Flow control was my problem when setting up tinyG.
https://github.com/openpnp/openpnp/wiki/TinyG

[Jason von Nieda]

Obi,

You may run into the same problem that Chris is having eventually. The machine will move, but when you try to actually do a job things will happen out of order. The issue is that TinyG changed their code a while back to process dwell commands asynchronously and the driver has not been updated to handle this. I'm trying to push people to use the GcodeDriver instead, as this driver is much more flexible. 

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.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/977302c2-0bc2-4754-8e8d-642f0a06a475%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[obiwanke...@gmail.com]

I'm working on  a version of Peter's head, and rotation of the second head would be a problem for the V8 TinyG, so a different driver will be needed, we've got a Deskcnc, Kflop, tinyG, and a smoothie around here from various projects. This morning I got an email from ST about a new 3D printer demo pcb, http://www.st.com/content/st_com/en/products/evaluation-tools/solution-evaluation-tools/computer-and-peripherals-solution-eval-boards/steval-3dp001v1.html?ecmp=tt3670_gl_enews_may2016 which looks interesting. Will the GcodeDriver even care which driver pcb ? Will it just issue Gcode commands and expect a Gcode response and is indifferent to the hardware ?

[Graeme Bridge]

thats a nice looking board, good price too

[Mark Harris]

Yup, i got that email too. Its only using L6474 drivers, but those are still pretty nice :)

It runs Marlin for the firmware, or rather, the ST port of Marlin. Has anyone used OpenPNP with Marlin?

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/d4ffd077-6abe-4ee0-814e-4c811a527471%40googlegroups.com.

[Jason von Nieda]

Obi,

Should work fine. In general, if the board speaks Gcode the GcodeDriver will work it. The Gcodes are all configurable and customizable.

More info at https://github.com/openpnp/openpnp/wiki/GcodeDriver

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/5fac90a2-1bf0-4aab-9fa8-45c8bf5d7de5%40googlegroups.com.

[Jason von Nieda]

Marlin has historically not been great for OpenPnP because it did not have per axis acceleration but maybe it's improved. In any case, Marlin should work fine with the GcodeDriver.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAJMBTySGur31Vy5QHd5%2B5nFNqPWpiGX%3Dn1bR_uenXtfv-zcmsw%40mail.gmail.com.

[Mark Harris]

Well:

#define DEFAULT_MAX_FEEDRATE {200, 200, 5, 35} // (mm/sec) #define DEFAULT_MAX_ACCELERATION {2000,2000,100,2000} // X, Y, Z, E maximum start speed for accelerated moves. E default values are good for Skeinforge 40+, for older versions raise them a lot.

https://github.com/St3dPrinter/Marlin4ST/blob/master/stm32_cube/Middlewares/Third_Party/Marlin/example_configurations/SCARA/Configuration_adv.h

seems to have individual accelerations, maybe it got better... or maybe it is just better on non-8bit platforms with tiny clocks.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jzo0RVaVUoPaaO%2BDvc0ObZOgpwwbXAH3N239xdWaX3Uxg%40mail.gmail.com.

[Jason von Nieda]

It's been a while, but if I remember correctly the issue was that it would limit acceleration, or possibly velocity, to the slowest axis in a given move. This made moves that included the Z or C axes unbearably slow. As I said, it's been a long time since I messed with Marlin and it's possible it's changed a lot since then. It's definitely worth a try.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAJMBTySFn22U9KkoyW92steFnYHs5jf1gTLvouiQN%2BUcqdHKeA%40mail.gmail.com.

[Michael Anton]

I suspect most 3D printing and CNC firmware is designed for coordinated motion, so it has to slow down for the slowest axis.  For PnP, we don't really need, that, so having one axis moving faster than another is not a problem, and probably desirable. 

[Red Davies]

If we can identify the errant commit we can fork at that point and
potentially stay current with that change removed.

> https://groups.google.com/d/msgid/openpnp/CA%2BQw0jzDr6b3Y5CJxP6N1sZaC0GVFMCiEVyONj74XH8ms24yQQ%40mail.gmail.com.

[Jason von Nieda]

Hi Red,

If you mean in TinyG, it's not an errant commit, so much as the way the TinyG folks want it to work. And it's not really a big deal to fix it in the OpenPnP TinyG driver. The reasons I am pushing people towards the GcodeDriver are:

1. Far more flexible. Inevitably once people get their machine moving they want to add things like lighting control, feeder control, etc. and with the controller specific, fixed drivers, this means code changes, which many people can't do. The GcodeDriver has the flexibility to do almost anything OpenPnP can do without code changes.

2. It's very difficult for me to support the many drivers we currently have, since I am usually only using one at a time. I have asked other people to take ownership of the drivers they use in the past, but that has not materialized. With a single codebase to support I can focus on making the driver work very well in more cases, and support changes through configuration.

So, if someone wants to fix the TinygDriver to work with the way TinyG works now, I'm all for it, but I'm going to focus my efforts on the GcodeDriver.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAFGw4%3DAJ6D6YW1ED5wRW7xONQ4v2zFpT%2BT1Jnmq8m7JBbG%2Bc%3DQ%40mail.gmail.com.

[Red Davies]

Well, to make sure I fully understand. I was under the impression
that tinyg users like myself couldn't use the generic gcode because
the firmware optimized behviour that we needed away. So, if we can
locate the patch to tinyg where this behaviour was added we may be
able to patch it out.

Alternatively, it wouldn't surprise me if if we were to enumerate the
issue directly the tinyg folks would add a configuration option to
disable that specific optimization.

<<-- Wants to be on the same driver as the project owner for selfish reasons :-)

> https://groups.google.com/d/msgid/openpnp/CA%2BQw0jy5_rbQuUWJ9_FiOozWefDBHGOcbkxmE5SYfW9xXkAiMw%40mail.gmail.com.

[Jason von Nieda]

Red, TinyG users can use the GcodeDriver as far as I know. The capabilities are there, but I haven't had a chance to get it yet. I'll be doing that this week.

I had already filed issues with TinyG related to this:

https://github.com/synthetos/g2/issues/138

https://github.com/synthetos/g2/issues/139

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAFGw4%3DB-%3DGdMp3OTLz15%2BpsVObjm25yc9s6XKUbvvaX5Xv7-Jg%40mail.gmail.com.

[Red Davies]

Looks like you're 20 steps ahead of me :-)

Oddly, I don't seem to be able to get any connectivity with the
straight GCODE driver as of yet - I'll enable debugging to try and
work out why.

> https://groups.google.com/d/msgid/openpnp/CA%2BQw0jx8z4f6R-8LZNUcT88pw_Hp7MxHhkDQ8XLVgSfGx9%3D7-g%40mail.gmail.com.

[Jason von Nieda]

You need to set dtr and rts I believ. Search the list for those terms and you will find a thread about it.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAFGw4%3DBmDErW7wZH4%3D6%2BjdZk9JAueOhh0fXBBbzXGs7p84rP-g%40mail.gmail.com.

[Chris Orita]

I've taken another shot at making this work, and I think I know where the problem is in the TinyG driver. The TinyG driver tries to find out if the machine has stopped by looking for the status code "60". The problem is that the status codes have changed, and 60 is not even used anymore (making me think that OpenPnP is just sending commands as space in the queue opens, which is a problem for commands like pick and place which don't seem to use the queue). The status code that the driver is looking for has been renamed to 201, though, I think something like status code 4 (operation complete) would be less of a hack. See: https://github.com/synthetos/TinyG/wiki/TinyG-Status-Codes. 

I would rather use the Gcode driver, but I haven't been able to figure out what command-confirm-regex response to use. All of my efforts have resulted in a timeout waiting for a response at startup. I am certain that the first line of the connect command is being sent properly (changing it to something like G0 Z10 results in the expected movement at startup), but I don't know what message (if any at all) is being sent by the TinyG after the completion of a command.

[Jason von Nieda]

Chris,

TinyG should be sending something like [mm]> ok after commands. Easiest thing to do is just connect to TinyG with a terminal program, or turn on trace logging in OpenPnP and look at what the response is and then model the regex after that. To start, something like .*ok.* should work fine, but using something a bit more precise would be better to avoid false positives.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/a39789b2-3dd2-4c89-be9c-961611f133fc%40googlegroups.com.

[Chris Orita]

here is an image of the log info that I get before receive a timeout error. I've tried using .*ok.*  ,  ^ok.*  ,  and variations on that. I've also tried just using the line "tinyg [mm] ok>", with no luck. Is there something i'm missing?

[Red Davies]

So I'm away from home currently so can't give you my exact configuration bit i battled the same demon a few weeks ago.

There were a number of settings that i needed to change.   From memory,  crlf.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/97638de2-7f25-4018-a4a5-3762d0b15fe4%40googlegroups.com.

[Jason von Nieda]

Hi Chris,

.*ok.* should have worked, as well as the full line that you mentioned. It's a little weird that you got the G21 back, I suspect maybe local echo is enabled for the TinyG, but that shouldn't cause any problems.

Ah, and I see Red's response now. Yes, CRLF might be an issue, especially if it's currently only set to CR. The driver is looking for a newline / LF to complete a command.

If that doesn't get it sorted, please send me your machine.xml and the OpenPnP.log file so I can investigate further.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/97638de2-7f25-4018-a4a5-3762d0b15fe4%40googlegroups.com.

[Chris Orita]

Thanks Red, 

Switching to LF instead of CRLF fixed the regex problem. There is still the problem where pick/place commands bypass motion commands, but now each pick/place command only jumps ahead of the subsequent motion command (instead of all of them like before). 

I think this has to do with the TinyG sending two "ok" messages per command, maybe it is sending one when it receives an instruction, and the second when it completes a command.

I have also been trying to use the move-to-complete-regex command to deal with this problem, but I haven't found any success yet.

[Jason von Nieda]

Chris,

The move-to-complete-regex should be something like: <move-to-complete-regex>.*vel:0.00.*</move-to-complete-regex>

You should see the vel:0.00 at the end of a move command. If you don't, make sure you have status reports turned on. TinyG docs cover that. There is some information about it here, as well: https://github.com/synthetos/g2/issues/139

Also, I beg of you, when you get things working properly please post your config. It sounds like we have several people using the driver with TinyG successfully, but no one has posted their working config. If someone would do this then others would not have to make all the same mistakes.

Thanks,

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/7d5fc887-8513-4e44-a140-832a8aa598cb%40googlegroups.com.

[Chris Orita]

Jason,

I'll be sure to post my config and whatever else someone would need to know to get a TinyG working with OpenPnP once it's running smoothly.

for the move-to-complete-regex command, i've tried .*vel:0.00.*, but I ended up getting a timeout error when giving successive move and pick commands. For the TinyG, the output "stat:3" indicates completed commands explicitly, but that also didn't work.

I have, however found a workaround. If you put a "null" move (e.g. G0 X0) as a preamble to the pick and place command, the null move will be skipped instead, effectively (and rather crudely) eliminating the problem.

[Chris Orita]

So that nulll move thing ended up not working the way I thought it was working. I've also started to run into unexpected errors when using move-to-complete-regex (sometimes the tinyg doesn't send a status report when I expect it to).

Attached is my machine.xml file and my TinyG config if anyone else want's to try to figure this out.

[Jason von Nieda]

Chris,

Can you also send OpenPnP.log? That will be the major thing that helps me see what is going on. Also, just to make sure I have all the info, can you tell me what hardware you are using for TinyG? v7, v8, G2, Due, etc.?

Thanks,

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/170e6bfd-49aa-4776-a0f8-1fa9439af224%40googlegroups.com.

[Chris Orita]

Here's the log file. I'm using the TinyG v8.

[Jason von Nieda]

Hi Chris,

Thanks for sending the log. As far as I can tell from looking at your machine.xml and your log, simply adding <move-to-complete-regex>.*vel:0.00.*</move-to-complete-regex> should get things going. TinyG is sending the vel:0.00 after each move completes and unless there is a bug in the move-to-complete-regex code this should work fine.

Here is what I'd like you to do:

1. Add <move-to-complete-regex>.*vel:0.00.*</move-to-complete-regex> to your machine.xml and make no other changes.

2. Delete your current OpenPnP.log so we have a clean slate.

3. Start the machine, home it, and do a test pick using the method you prefer. Either the Parts panel pick button, or the Feeders panel pick button will work. Try not to do anything else such as jogging the machine, pressing other buttons, etc. We want as little as possible in the log file.

4. Shut the machine down and send me the resulting machine.xml and OpenPnP.log.

Note: It would be best if step 3 is a fairly long move, so if you have to jog the machine away from the feeder first, that's fine.

This should clear things up. If it doesn't, I will be back to my desk on Sunday and will get my TinyG and TinyG2 out and figure out the right config for both so we can put this to rest for everyone.

Thanks,

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/5121c047-0a45-4122-840e-36dd7e06770c%40googlegroups.com.

[Red Davies]

Iirc, the setting 'strict json' results in the use of double quotes or not in those fields.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jwM5DUSUZLLYCKx%3DG_rh27eq7U9xtg9tsGrzSwwGQd%2Bkw%40mail.gmail.com.

[Jason von Nieda]

Red,

We aren't using the JSON reports in this case since we're sending Gcode. The status reports look like:

posx:56.133,posy:347.275,posz:-3.083,posa:0.000,feed:1000.00,vel:1000.00,unit:1,coor:0,dist:1,frmo:1,stat:9

Basically JSON without the braces, but we're just comparing regexs to the raw text without doing any kind of parsing.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAFGw4%3DCfC8at8KVy78Ezm5b%3DkLfw0ycdp8XVS8X-MMaCba_yPQ%40mail.gmail.com.

[Chris Orita]

Here's the log file. I did the test pick using the attached script.

To view this discussion on the web visit <a href="https://groups.google.com/d/msgid/openpnp/170e6bfd-49aa-4776-a0f8-1fa9439af224%40googlegroups.com?utm_medium=email&utm_source=foote

[Jason von Nieda]

Hey Chris, this all looks very good to me. Did you experience any problems? As far as I can tell it worked perfectly in this case.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/7a3e0ba6-d197-4100-89f5-042b0049413f%40googlegroups.com.

[Chris Orita]

I didn't notice anything weird when doing the test, but when I have my script place after the pick, the valve opens and then closes before the nozzle reaches the table. Attached is my log file of the same test, but the slightly modified script.

<blockquote class="gmail_quote" style="

[Jason von Nieda]

Okay, I think I know what's up. It's actually a script bug. You are calling task() for the move, but then calling pick and place outside of the task. So, while the task is running the pick then gets called immediately. Try putting the pick() and place() calls in a task() call like the moveTo, and I bet it all works fine. You still need the move-to-complete-regex that you added - that is working correctly.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/7f11396f-1ef0-4618-93b9-560e5cb6d1ab%40googlegroups.com.

[Chris Orita]

Thank you! Things are working exactly as expected now.

[Jason von Nieda]

Glad to hear it!

Can you please now post your working TinyG config and machine.xml? I'll post these to the TinyG Wiki page for future users.

Thanks,

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/1ae6b801-f0c7-4209-bc27-50b1026e53d1%40googlegroups.com.

[Chris Orita]

Here's the machine.xml and config that are working for me

>> >> >> For more options, visit <a href="https://groups.google.com/d/optout" rel="nofo

[Jason von Nieda]

Thank you Chris! I will get these into the Wiki in the next day or two.

Jason

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/e526646a-61b2-4886-ace9-b587bf8bfdf5%40googlegroups.com.

[Jason von Nieda]

Thanks again Chris! The TinyG wiki page has now been updated with this info: https://github.com/openpnp/openpnp/wiki/TinyG

Jason

[Red Davies]

I do have one issue but it is definitely in the firmware, not openpnp and i'll mention it here for archive purposes.

Tiny moves will timeout and there won't be any indication why in the debug logs.

If you make the same tiny move in json mode it will provide you with an error message.  202 MINIMUM_TIME_MOVE.

Reducing the speed to a ridiculous crawl "mitigates" this a certain amount but doesn't fix it.

It typically only presents itself when doing tiny jogs or setting up the feeder.

During a job I've yet to see a tiny move,  mainly because my head and camera are so far apart.

I don't forsee it stopping me doing any jobs,  i just need to remember to change jog speed before and after setup.

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/e526646a-61b2-4886-ace9-b587bf8bfdf5%40googlegroups.com.
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+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/CA%2BQw0jyFAutX12ncfW72Hr2r%2BSxZ%3DtWpaJJc0T0NU24p%3Dn_YOQ%40mail.gmail.com.

[Jason von Nieda]

I noticed that LitePlacer uses different Gcode for small and long moves with TinyG. I imagine we need to do something similar. Although, to me, this seems like a bug. I'd expect the controller to handle any move I send it in an intelligent way. Perhaps this a question for the TinyG folks?

Liteplacer code for small moves: https://github.com/jkuusama/LitePlacer-DEV/blob/master/LitePlacer/CNC.cs#L265

Jason

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/e526646a-61b2-4886-ace9-b587bf8bfdf5%40googlegroups.com.
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.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CA%2BQw0jyFAutX12ncfW72Hr2r%2BSxZ%3DtWpaJJc0T0NU24p%3Dn_YOQ%40mail.gmail.com.

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.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAFGw4%3DC%2BtF2kfe19dY2uL%2BvSV%3DzQz-ATEQ5gL4yMk%2B0-UffMDg%40mail.gmail.com.

[Red Davies]

Funny, I need to get it as low as F5 before it will move reliably.  No idea what the delta is... my guess would be the jerk.

Anyways, while pondering how to 'solve' my backlash issue I realized that if I apply the 'always arrive from the same direction' technique it could potentially guarantee that my moves are always long enough.  Can I implement it without a code-change, that is the question:

Potential solutions:

i) If it's possible to put a mathematical expression in the <move-to-command>, put two in.  Like:

  G1 {X:(X+5)%.4f}

  G1 {X:X%.4f}

ii) Use different frame of reference G5[45] and run the movement twice, once in each frame.

The obvious concern is that the move-to-complete-regex will trigger at the end of the first move.  A potential solution to this is to change said regex to be something like "end of my move yo" and then make the last command in that section that string which is then echoed back to OpenPNP.

I'll take a look at this over lunch and see how feasible it is.

Thanks,

Red

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/e526646a-61b2-4886-ace9-b587bf8bfdf5%40googlegroups.com.
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+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/CA%2BQw0jyFAutX12ncfW72Hr2r%2BSxZ%3DtWpaJJc0T0NU24p%3Dn_YOQ%40mail.gmail.com.

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+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/CAFGw4%3DC%2BtF2kfe19dY2uL%2BvSV%3DzQz-ATEQ5gL4yMk%2B0-UffMDg%40mail.gmail.com.
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+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/CA%2BQw0jyKdJxuDrLCTtx0P%2B0SL9FxwKZyfr266Why4E7xoALJ8g%40mail.gmail.com.

[SMdude]

G'day Red,

Is there any reason you can't adjust the backlash out of your system?
Backlash is a mechanical issue that should not really be tolerated in the X and Y system. Make sure your belts are sufficiently tight.

If your backlash is from your guide rails, it it affecting your static positioning?

Some things just should not be resolved in software IMHO

[Red Davies]

I completely agree.

When I'm doing video / sound design, nothing frustrates me more than
hearing a DP say 'he can fix that in post' (meaning me).

My belts are so tight that my steppers are fighting to drive the axes
and I'm having to actively cool my driver-board to keep it running.
This is not perfection - this is my 'learning build'.

(Un)fortunately I have a job I have a run in about 2 weeks so don't
really have time (or budget currently) to change from belt to
lead-screw which is absolutely my next job.

My 'business case' is a single batch of a couple of hundred boards
every year. Worst case I can do it by hand but I'd hate my life.

Open to other suggestions.

Thanks,



Red
PS:
(The 'crazy' part of me is wanting to build a PNP in a suitcase - tiny
axes, lead screw driven... I promise I am crazy)

> --
> 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/0ebaa682-6a34-4ef4-a72c-e4253952799b%40googlegroups.com.

[Anthony Webb]

It pains me to hear people having trouble getting the mechanical side of things for their PNP nailed down.  I can appreciate the frustration, and it is in many ways why I set out to create a reference build that if people would duplicate they would not have to worry about the mechanical side of basic PNP. 

There are a lot of smart people walking through the various mechanical pieces and doing a great job of lowering the cost and increasing the quality of the PNP experience.  It's not exactly "turn key" today, but we are beginning to arrive at what I think is a great stack and starting point to be confident enough to accomplish the task.  Hang in there Red!

-Anthony

> 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/0ebaa682-6a34-4ef4-a72c-e4253952799b%40googlegroups.com.
>
> 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+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/CAFGw4%3DD8hfNJbrqyfNY31-RNcwdGoEWcEyM0qPvy-HsDf5T5Yg%40mail.gmail.com.

[Red Davies]

Do you have a link to hand for your reference build?  I'd be curious to look at it for after this project cycle.

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CALsNZy1QzDvPgim-UaUCCytECsS578UnRrrwGoaHGM6q47RwLg%40mail.gmail.com.

[Anthony Webb]

The current reference is here, but there are some nice enhancements in the works that will be posted to the group soon too: https://hackaday.io/project/9319-diy-pick-and-place

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/CAFGw4%3DAb%3D_RXg_7ummpSRBo6%2BKV6QzGuuge8LgbW1MFk7mPWag%40mail.gmail.com.

[Red Davies]

I completely dismantled my machine this evening. 3 years to build,
about 30 minutes to get back to component atoms.

Going to re-build my motion platform from scratch and try again.
Fighting with makerslide and belts was never going to be 'tweaked'
into a reliable form.

>>>> > 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/0ebaa682-6a34-4ef4-a72c-e4253952799b%40googlegroups.com.
>>>> >
>>>> > 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.
>>>> To view this discussion on the web visit
>>>> https://groups.google.com/d/msgid/openpnp/CAFGw4%3DD8hfNJbrqyfNY31-RNcwdGoEWcEyM0qPvy-HsDf5T5Yg%40mail.gmail.com.
>>>> 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.
>>> To view this discussion on the web visit
>>> https://groups.google.com/d/msgid/openpnp/CALsNZy1QzDvPgim-UaUCCytECsS578UnRrrwGoaHGM6q47RwLg%40mail.gmail.com.
>>> 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.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/openpnp/CAFGw4%3DAb%3D_RXg_7ummpSRBo6%2BKV6QzGuuge8LgbW1MFk7mPWag%40mail.gmail.com.
>>
>> 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.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/openpnp/CALsNZy3EyZuCwQ0ihTVzHfQwdOdo%2BQ%2BiOdWTFpqd78DdbwktuA%40mail.gmail.com.

[Paul Jones]

I can vouch for Anthony's reference design being a good one. I based my machine on it and I'm very happy with it. The linear rails make backlash virtually non-existent. The hardest part was getting all the extrusions square and parallel, but nothing a bit of patience and digital callipers can't overcome!

Paul.

> >>> https://groups.google.com/d/msgid/openpnp/CALsNZy1QzDvPgim-
> UaUCCytECsS578UnRrrwGoaHGM6q47RwLg%40mail.gmail.com.

> >>> 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.
> >> To view this discussion on the web visit
> >>

> https://groups.google.com/d/msgid/openpnp/CAFGw4%3DAb%3D_RXg_7u
> mmpSRBo6%2BKV6QzGuuge8LgbW1MFk7mPWag%40mail.gmail.com.

> >>
> >> 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.
> > To view this discussion on the web visit
> >

> https://groups.google.com/d/msgid/openpnp/CALsNZy3EyZuCwQ0ihTVzHf
> QwdOdo%2BQ%2BiOdWTFpqd78DdbwktuA%40mail.gmail.com.

> >
> > 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.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/openpnp/CAFGw4%3DBs-dEHBAsM-
> PKYVGi67RqLRjO80fvC6gFRn-KvphAtLg%40mail.gmail.com.

[Red Davies]

Anthony,

Moving from makerslide to Linear Rail has adjusted pretty much all of
the backlash out of the system. However, when I use belts I don't have
sufficient resolution per-step in order to perfectly align YC164
(basic, 0603 packages).

The combination of 1/8 stepping and 1.8 degree steppers doesn't seem
to be cutting it (hence why I moved to screw and that's complicated
construction considerably).

Can you confirm for me that you're using 1/16 microsteps and 1.8
degree steppers?

I guess I"m going to double-check my tinyg firmware again that I have
it set to 1/8 and failing that, consider moving to 0.9 degree steppers
and/or a board that's capable of 1/16 stepping.

Thanks,



Red

>>>> > 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/0ebaa682-6a34-4ef4-a72c-e4253952799b%40googlegroups.com.
>>>> >
>>>> > 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.
>>>> To view this discussion on the web visit
>>>> https://groups.google.com/d/msgid/openpnp/CAFGw4%3DD8hfNJbrqyfNY31-RNcwdGoEWcEyM0qPvy-HsDf5T5Yg%40mail.gmail.com.
>>>> 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.
>>> To view this discussion on the web visit
>>> https://groups.google.com/d/msgid/openpnp/CALsNZy1QzDvPgim-UaUCCytECsS578UnRrrwGoaHGM6q47RwLg%40mail.gmail.com.
>>> 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.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/openpnp/CAFGw4%3DAb%3D_RXg_7ummpSRBo6%2BKV6QzGuuge8LgbW1MFk7mPWag%40mail.gmail.com.
>>
>> 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.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/openpnp/CALsNZy3EyZuCwQ0ihTVzHfQwdOdo%2BQ%2BiOdWTFpqd78DdbwktuA%40mail.gmail.com.

[Cri S]

What belt pitch and pulley spoket number you have?
12 and/or 16 with gt2 works.

[Red Davies]

Currently, GT2 - 20 tooth.

Maybe a move to 12 would be beneficial... I guess I didn't start
there since Anthony's machine speced 50 and the larger the number of
teeth, the greater the diameter so less resolution -=- right?

On Mon, Aug 22, 2016 at 8:09 AM, Cri S <phon...@gmail.com> wrote:
> What belt pitch and pulley spoket number you have?
> 12 and/or 16 with gt2 works.
>

> --
> 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/2dc75da7-8c66-4528-8245-3f3287b2e219%40googlegroups.com.

[Cri S]

Resolution of 20 theet 2mm pitch is 0.05mm
16 theet. Is 0.04 and 12theet  0.03mm .
This with 16 microstepping. There must be another problem because 0603 should be placeable without problems having 20 theet, and 8x microstepping should go too.

Il 22/ago/2016 14:19 "Red Davies" <noid...@gmail.com> ha scritto:

Currently, GT2 - 20 tooth.

Maybe a move to 12 would be beneficial...  I guess I didn't start
there since Anthony's machine speced 50 and the larger the number of
teeth, the greater the diameter so less resolution -=- right?

On Mon, Aug 22, 2016 at 8:09 AM, Cri S <phon...@gmail.com> wrote:
> What belt pitch and pulley spoket number you have?
> 12 and/or 16 with gt2 works.
>
> --
> 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/2dc75da7-8c66-4528-8245-3f3287b2e219%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

--
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/j-TAyyZ9XQ0/unsubscribe.
To unsubscribe from this group and all its topics, 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/CAFGw4%3DCLNd8XipTStCznqWj9XUSho4qA45pEBk9OhzqGkKZFcQ%40mail.gmail.com.