Help debugging hardware gone bad!

northbear
Posts: 255
Joined: Mon Apr 09, 2012 9:21 pm

Help debugging hardware gone bad!

Post by northbear » Tue Jan 19, 2016 2:49 am

Looking for some help on how to best debug what I believe is some piece of hardware that went bad.

I was in the middle of running a job, and then the CNC simply stopped (with the spindle on, vacuum running, but no movement in any direction.). I stopped the program which should have turned off the spindle and vacuum but didn't. I knew something was wrong and then cut power to the machine. I powered it back up and now I have issues even when just trying to jog. I can move in the x axis no problem (both directions). I can move in the negative y direction, but when I try to move in the positive y direction it makes some bad noises and tries to twist the gantry. The z-axis refuses to jog up or down. When powered on all the stepper motors seem enabled in that they are energized and stay in position. Also as soon as I turn on the power to the machine the spindle and vacuum are enabled rather then automatically turning on when I starting a job.

My setup is a bit different from a standard ShapeOKO as I am using LinuxCNC with a breakout board and individual TB6560 stepper drivers (see http://www.shapeoko.com/forum/viewtopic ... 317#p17593 for more info on my setup)

I am thinking it is the breakout board that went bad as that seems to be the one thing in common to everything that is going wrong. Either that or three stepper drivers and a relay all went bad at the same time?1? Again, everything was working great before this happened so it shouldn't be a software or wiring issue.

Any ideas on what/how to debug and figure out exactly what what is wrong?
My buildlog is here

AnonymousPerson
Posts: 758
Joined: Sun Apr 26, 2015 1:16 pm
Location: 3753 Cruithne

Re: Help debugging hardware gone bad!

Post by AnonymousPerson » Tue Jan 19, 2016 3:04 am

If no-else gets back to you, it's not a bad idea to email support@carbide3d.com. Just in case. :)
Shapeoko 3 #516

WillAdams
Posts: 8596
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Help debugging hardware gone bad!

Post by WillAdams » Tue Jan 19, 2016 3:07 am

There have been instances of corrupted drivers and firmware installations causing weird problems.

Start w/ the basics — check wiring continuity, then check driver and firmware functionality, then swap around parts / connections to isolate the bad component?

Do you have an oscilloscope?
Shapeoko 3XL #0006 w/ Carbide Compact Router w/0.125″ and ¼″ Carbide 3D precision collets
Nomad 883 Pro #596 (bamboo)

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

Re: Help debugging hardware gone bad!

Post by cvoinescu » Tue Jan 19, 2016 3:17 am

I say bad breakout board. They tend to have two or three buffer ICs, each handling six or eight signals. What you see is consistent with one of them having gone bad. It could be as simple as a single bad solder point, or the IC could be fried.
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk

northbear
Posts: 255
Joined: Mon Apr 09, 2012 9:21 pm

Re: Help debugging hardware gone bad!

Post by northbear » Tue Jan 19, 2016 3:53 am

Thanks for the quick responses guys!

AnonymousPerson - I could contact them but my CNC is not an official ShapeOKO, but instead a "Shapeoko inspired" design of my own. I know the guys at Carbide are helpful but I feel it may be asking a bit too much to help me debug my own machine :lol:

WillAdams - I had considered swapping parts / wires but haven't had the time yet to methodically debug things and also am not sure how to test if the breakout board is bad (other then process of elimination)

cvoinescu - Thanks for your views on what went bad and letting me know that my symptoms would be consistent with a bad breakout board.


Is there a quick and easy way to test the breakout board? If not I may just spend the $15-$20 and buy a new board, swap it out and see if that fixes the problem.

Thanks again!
My buildlog is here

AnonymousPerson
Posts: 758
Joined: Sun Apr 26, 2015 1:16 pm
Location: 3753 Cruithne

Re: Help debugging hardware gone bad!

Post by AnonymousPerson » Tue Jan 19, 2016 5:46 am

northbear wrote:AnonymousPerson - I could contact them but my CNC is not an official ShapeOKO, but instead a "Shapeoko inspired" design of my own. I know the guys at Carbide are helpful but I feel it may be asking a bit too much to help me debug my own machine :lol:
Good point. :D
Shapeoko 3 #516

WillAdams
Posts: 8596
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Help debugging hardware gone bad!

Post by WillAdams » Tue Jan 19, 2016 11:55 am

Yeah, if one is self-sourcing, I think it’s best to do a modular design, and have spares of any parts which are not duplicated w/in the design for this sort of diagnostic.
Shapeoko 3XL #0006 w/ Carbide Compact Router w/0.125″ and ¼″ Carbide 3D precision collets
Nomad 883 Pro #596 (bamboo)

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

Re: Help debugging hardware gone bad!

Post by cvoinescu » Tue Jan 19, 2016 2:22 pm

Find the Y that goes the wrong way, and measure the voltage on the direction output (with respect to ground) while executing a slow, long move in one direction, then while moving in the other direction. They should be very different (about 0 V and about 5 V), but I suspect they'll be the same, because of the broken chip. Do the test with the axis that works, too, just to check. You can disconnect the motors, or remove the belts, so that you don't hurt the machine.
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk

northbear
Posts: 255
Joined: Mon Apr 09, 2012 9:21 pm

Re: Help debugging hardware gone bad!

Post by northbear » Tue Jan 19, 2016 10:54 pm

cvoinescu wrote:Find the Y that goes the wrong way, and measure the voltage on the direction output (with respect to ground) while executing a slow, long move in one direction, then while moving in the other direction. They should be very different (about 0 V and about 5 V), but I suspect they'll be the same, because of the broken chip. Do the test with the axis that works, too, just to check. You can disconnect the motors, or remove the belts, so that you don't hurt the machine.
Thanks for the directions on how to test. In the meantime I also ordered another breakout board. If that is not the problem, then at least I will have a spare!
My buildlog is here

northbear
Posts: 255
Joined: Mon Apr 09, 2012 9:21 pm

Re: Help debugging hardware gone bad!

Post by northbear » Wed Jan 27, 2016 9:52 pm

I replaced the breakout board last night and now it is back up and running! Hurray! But leads me to the question:

Why did it fail? My current thoughts:

1) Cheap electronics. I purchased the first board off of e-bay direct from China (buyer "goodlucksell"). It was cheap, but once I received the board, the documentation looked like it was from another seller, and the board was not exactly like the pictures in the listing. I probably should have known, but at the time it didn't seem to matter and worked. The board I just installed was via e-bay again but from a US seller in California (savebase). The board matched their pictures in the listing, has a better silkscreening (much clearer), looks like thicker wire traces, and the components are on a bit "straighter" I am sure the board was still made in China, but hopefully better quality as it was also more expensive.

2) Old fluorescent shop light causing an issue. I have a very old (20+ years?) shop light plugged into the same power strip as the computer and CNC which will flicker a bit every so often. Light was installed about 2 feet directly above the CNC. Not sure how this could cause an issue, but I may just find a different light for use over the CNC just to be safe.

3) Static? Not sure how it could cause a problem as everything is well grounded, but I do notice some static buildup on the vacuum as fine wood dust can cling to the outside of the vacuum hose. I could use tinsel or a grounding wire, but I am not very convinced this is the issue.

My best guess just cheap electronics, but would be interested if anyone else has any thoughts on why it failed.

Thanks again for the help!
My buildlog is here

Post Reply