random erratic behavior?

JoeyD
Posts: 18
Joined: Sun Nov 18, 2012 11:38 pm

random erratic behavior?

Post by JoeyD » Tue Jan 01, 2013 8:07 pm

I believe my precision issues have been ironed out but now I'm getting another strange occurrence. After various periods of running, the spindle is sent way off into a seemingly random direction. The first pic shows a pocketed job (Misfits logo) that was on it's 3rd and final depth increment. About 1.5 hours into it and it was humming along beautifully until I hear something awful and see the spindle running off the reservation. Crap.
misfits_fail.jpg
Misfit's logo failure
misfits_fail.jpg (172.24 KiB) Viewed 1607 times

That was my last piece of good wood so I decided to try some acrylic profile cuts. You can see the first run came out great. It took maybe 5 minutes to complete. On my 2nd run, about 3 minutes into it, the spindle took off in the other direction this time.
acrylic_fail.jpg
Acrylic failure
acrylic_fail.jpg (271.65 KiB) Viewed 1607 times
I'm not sure how to go about troubleshooting this. I have noticed that from time to time I'll have to reseat the USB cable on the laptop and restart gcode sender because it stops responding. This is usually between jobs though.
Last edited by JoeyD on Wed Jan 02, 2013 12:49 am, edited 1 time in total.

droo
Posts: 18
Joined: Sun Nov 25, 2012 6:32 pm

Re: random erratic behavior?

Post by droo » Tue Jan 01, 2013 10:03 pm

Looks similar to a problem I had - controller chips overheating ? I put heat sinks and fans on mine.

Droo

cvoinescu
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK
Contact:

Re: random erratic behavior?

Post by cvoinescu » Tue Jan 01, 2013 10:16 pm

It does look like missed steps, possibly due to overheating. Can't be missed commands, unless your G-code has relative coordinates (absolute is much more common, AFAIK) -- and the USB connection, once dropped, does not recover on its own anyway.
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk

JoeyD
Posts: 18
Joined: Sun Nov 18, 2012 11:38 pm

Re: random erratic behavior?

Post by JoeyD » Wed Jan 02, 2013 1:10 am

I have heatsinks on each of the motor drivers along with a small 24v fan blowing directly onto the X and Y heatsinks. I also have another fan blowing between the grbl shield and Arduino. Perhaps the fans aren't putting out enough air. I'll get some faster ones and try them out.

Thanks,
-Joey

edwardrford
Posts: 1250
Joined: Mon Apr 09, 2012 5:40 pm
Location: Dixon, IL
Contact:

Re: random erratic behavior?

Post by edwardrford » Wed Jan 02, 2013 3:37 am

Joey,

This same thing happened on a job I was running a few weeks ago.

Totally. Random.

The machine I was using had sinks and a fan. The best I could make of the situation was just PFM, in a bad way.

I tried just re-running the job (it was a short one, maybe 25 minutes) but the exact same thing happened in roughly the exact same place. I looked through the g-code and even ran it through a simulator again, all seemed well. No idea why that happened.

Not that this information helps you any, but misery likes company so I thought I'd chime in and let you know that I've been there too.

-Edward
Shapeoko 1 #0 - a couple of upgrades.
Shapeoko 2 #0 - a couple of upgrades.
Shapeoko 3 #2 - Stock

levlandau2
Posts: 187
Joined: Tue Apr 17, 2012 1:52 am
Location: Minneapolis, MN, USA

Re: random erratic behavior?

Post by levlandau2 » Wed Jan 02, 2013 4:12 am

Yes I can confirm exactly the same thing as described here has been happening to me as well. Most noticeably on the z-axis, although I have had it happen on the x-y axes as well.

I agree with everyone else that it is most likely missed steps, after which the tool looses its absolute coordinates and everything goes to hell :-). This has been causing most trouble for me when doing 3D profiling. I noticed, slowing down the z-axis feed rate makes the problem occur less often, but it is still there once in a while... I have improved my heatsink and fan setup to try to fix this issue, but am uncertain that I ahve a robust fix in place (see image of fan/heatsinks below).

This problem is particularly annoying, because it can cause the tool to crash into various objects (stock, rails, etc..), so it makes me paranoid to leave the tool running on its own.... It would be nice to have an error detection scheme where if steps are missed, the tool would shut down. Unfortunately this would require encoders on the steppers or something along those lines...

It would be really cool to add some temperature sensors close to the stepper drivers and correlate the failure with an increase in temperature. Then we can be sure that the hotter stepper drivers are causing this issue.

The long term solution would be to install a feedback mechanism on the steppers, the short term is better heatsinks, fans, and running at lower feed rates (does this actually keep the drivers cooler?)
Attachments
CropperCapture[15].jpg
CropperCapture[15].jpg (131.76 KiB) Viewed 1568 times

JoeyD
Posts: 18
Joined: Sun Nov 18, 2012 11:38 pm

Re: random erratic behavior?

Post by JoeyD » Mon Jan 07, 2013 7:21 pm

I upgraded to a larger 60mm fan and so far so good. I was never able to track down the exact CFMs of the old fans (gotta love surplus stores) but the new on is rated at over 23CFM which I suspect is 3x-4x more than the smaller ones. I need to run a few more cuts to make sure, but so far everything's looking great. This took about 1.5 hours last night.
image.jpeg
image.jpeg (421.39 KiB) Viewed 1518 times
I was toying around w/ the idea of getting some thermocouples to use with a spare Arduino to monitor stepper driver temps. I haven't looked into it too much as time is in short supply and the larger fan seems to be working currently.

-Joey

deejayspinz
Posts: 96
Joined: Wed Oct 17, 2012 9:33 am
Location: Burlington Ontario, Canada

Re: random erratic behavior?

Post by deejayspinz » Mon Jan 07, 2013 7:30 pm

FWIW.. When I had my MyDIYCNC I had this happen a few times to me. Here is what I found on the various occurrences:
- I had the X&Y speeds too high for the machine. I think the motors got ahead of the cut, and the negative forces caused them to skip, then bam, off it went.
- I have a fairly recent machine, and was running various things in the background. I had YawCam running and think that the machine may have gotten bogged down while processing and also caused another skip. When I removed the software and turned off any unnecessary services, the problem went away.

I never had an issue with the drivers overheating.

JoeyD
Posts: 18
Joined: Sun Nov 18, 2012 11:38 pm

Re: random erratic behavior?

Post by JoeyD » Mon Jan 07, 2013 7:33 pm

levlandau2 wrote: I have improved my heatsink and fan setup to try to fix this issue, but am uncertain that I ahve a robust fix in place (see image of fan/heatsinks below)
Levlandau2 - I've noticed that there's a pocket of slack air immediately in front of my fan. I had it sitting right on top of the heatsinks similar to yours. Pulling it back away from the heatsinks an inch or so allowed for more of the air to move across them which keeps them cooler. Might be worth trying?

cvoinescu
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK
Contact:

Re: random erratic behavior?

Post by cvoinescu » Mon Jan 07, 2013 9:52 pm

It may be worth trying to get some airflow under the driver boards. The driver chip is designed to transfer most of the heat through its bottom tab into the PCB, and the PCB has thermal vias that spread this to both the top and bottom copper layers. The top layer has all the connection to the pins, so those areas can't be used to carry heat away from the bottom of the chip, but the bottom layer has copper extending almost the entire surface of the board (small as it is). So cooling the underside of the Pololu board is the most important. Comparatively less heat is transferred to the tops of the chips and the heatsinks; the heatsinks do help, but not as much as one would hope for. A strategically placed baffle or two may be able to create a decent airflow under the boards despite the narrow space between the Arduino headers and the edges of the boards.
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk

Post Reply