[Buzz9075]

Just finished building a new CNC table using the Sosylva Plans. Spent many hours aligning it make sure the axis are all moving freely and nothing is binding but I keep coming back to the same issue and starting to wonder if it has something do with the electronics. Here is what is happening.

I appear to have it calibrated everything moves very smoothly on all axis. Wrote a bunch of code to work each axis individually and they all work fanastically at 52"/min and acceration of 4". Now put the code together for multiple axis movements and all appears fine for a while, then all of sudden on one or more axis it will make this lovely grinding/growling sound and they stop moving then or shortly afterward. I have tried tons of stuff, some but not limited to: reduce speed to 30" (just takes longer for it to happen), reduce acceleration, shut down driver board and move axis at point of issue and it feels very smooth with no drag. In order to see what it would take to stall the machine I applied pressure one time on gantry till the motor did the same thing this time due to my fault... it took significant force to cause this.. no way any of my axis are out this much of.
As well I can be jogging along at 50"/min and change direction or something random and poof it does it. For example press the left button for 10" or so then reverse direction and then back again poff it looses it. You apparent reason.
Normally you never know when it is going to happen, most often it is when one axis is moving slower than the others, for example:
g00 x0 y0 z0
g00 x1 y1 z-6
or
g00 x0 y0 z0
g00 x1 y5 z-1

and the step slope appears to shut down one or more of the axis (usually only one but at times two), then try
g00 x0 y0 z0
goo x5y1z-1

and it could happen again. But if you just run all the axis by them self they work perfectly. Two axis I have running so smooth now you can barely hear them running (y and z). X axis is noisey because of the longer threads and the rod wonder a little.

Many times I can tell when it is going to happen. The machine is instructed to change direction and you can hear one or more of the axis groaning as it moves along, unless it is a very short movement you know it is going to crash.

This happens to any of the axis (normally y and z) and it is random. Run the same pattern 10 times it might happen 3 times.

Starting to believe it has something to do with the electronics portion.
I have the 4 axis prepackage from Xylotex. Suppose to be plug and play, no need to set vRef or install and hardware.

Any one got any advice for me?

Thanks in advance for your help.

[paulC]

Have you worked out the rpm of the motors at this speed?
Steppers loose torque with speed so without knowing all your hardware setup its my first question.
Paul

[Buzz9075]

RPM would be 52"/in * 18TPI(RerPerInch) = 936RPM. The failure occurs at all different speeds. Often the stepper slops will cause the crash... on a stepper slope some axis might only be doing 10"/minute while others are doing 52".

One thing I have noticed if I left the motors free floating it happens a less (not completely gone but less occurances of the crashing/stalling). What I mean by free floating is the loosen of the nuts the hold the motor base plate to the machine. At first I thought the alignment of the motor to axis might be of, but looking it over I don't see anything missaligned here.

[ger21]

What software? What kind of PC and how fast?

From what I've read, you're going faster than other Solsylva users have been able to go.

[Riceburner98]

I had / have a very similar problem with my Taig milling machine. 4 axis Xylotex board with thier big honkin' 269 oz motors. The machine runs *great* with my Laptop (HP / Pentium, 1.6Ghz), but I built a 3Ghz machine to dedicate to the setup; I can run my CAD programs 10x quicker on the new machine, but for some reason the new machine randomly crashes the motors. Very similar symptoms to what you have going on. I've tried 2 different motherboards thinking it was the Parallel port drivers not quite hitting +5v (I've heard that can be a problem with newer computers) and also tried 2 add-on Parallel port cards with the same results. Go back to the Laptop, and it's fine. I haven't put a 'scope on the outputs of the new desktop machine, but I suspect it's something to do with the parallel port drivers. It certainly can't be the processing power not keeping up with the driver software, it's a fresh install, and I've tried Windows 2000 and XP, running MachII and MachIII with the same results. Can't be the Xylotex board because it works on the Laptop. Got to be the new Parallel hardware..... Anyway, just a thought!

[soapy]

It could be that your driver board is crashing because the hardware in that can't keep up with the data rate from the faster computer. It shouldn't happen due to the bottleneck of the parallel port, but your new parallel port might well be running in a new mode, like ECP or EPP, and so winds up being too fast for the board. Try reading through http://www.lvr.com/jansfaq.htm for some clues. Best I found was an issue with Toshiba's causing issues by adhering to the specs while others don't.

As for the stalling issue, it is likely to be a cross-feed issue. The slight error in one axis means that the slight error in the other axis causes it to bind. If it is at the reverse, it might be due to too tight a tolerance on the backlash. I'd run a few tests, try doing a reverse at one side, then repeat it at the other, and in the middle, on both axes. (Or all three, but I assume you ruled out the Z axis changing and messing things up already?) If it happens more at one side than the other, and not in thte middle, or whatever, you might learn a bit more about the problem. Then run it in reverse, and see if the problem is the same. Finally, have it approach the turns from the other side, i.e. if the machine stalled going from +x to -x in the bottom left of the table, have it do a -x to +x in the same place, with the same y axis move.

You could probably work around the problem by stopping (or nearly stopping) one axis while you do the reverse, then start up again.

This is all half-recalled from years ago, so others might want to chip in.

[scheek]

Check your fan in power supply to make sure the drivers are not over heating after continuous use.

I had a problem similar to this once and it drove me crazy. I finally opened my powersupply and put my hair drier (on cool) directly over the the driver(s) that gave me trouble. That corrected the problem, so I moved my enternal fan to blow closer to the drivers to keep them cooler. NO more problems.

good luck.

[Buzz9075]

Awesome support guys, thanks for all your help. Think I might be on to something, see the following in another forum:
http://www.cnczone.com/forums/showpo...&postcount=505

I am going to work on this first then gather up a list of all items identified here and the other forum and pick them of one at a time.

Cooling one, yesterday I would have been behind that one 100%. This morning though I ran the following twice, second one caused the stall then I loosened the motor up and it stopped (see link above for details).
g00 x0 y0 z0
g00 x1 y3 z-6.5
m30
Would love to test more today but work calls.


I have tried three different computers thinking it might be communictions: 350Mhz with 512MB RAM, 1.2Ghz with 512MB RAM, 2.8GHZ with 1GB RAM (got lots of machines to play with). The ports on the last two are probably set to and advanced setting, the 350Mhz one I am not sure what it would be set at. Will review this tonight.


Going faster than others, problem is occuring when I reduce the speed as well just takes longer for what ever is happening to happen. For example set speed to 30ipm and you might have to wait 4 or 5 passes and it does it again. At 52ipm axis work fine for long periods of times with no issues, more axis equals issues.

Oh ya software is Mach 3. Just installed latest version (see attached link on findings).

[scheek]

What happens once you stop, shut down and wait 20 minutes and then crank back up? Does everything work for a few times and there it goes again?

[Buzz9075]

First thing I did this morning was run the following code:
g00 x0 y0 z0
g00 x1 y3 z-6.5
m30
First pass OK
Second pass crash.

This was with the latest version of the software. And it was the Z axis that crashed with no groaning prior to stalling. Yup the controller was powered down.

When you experience over heating, how long did it take for the over heating condition to take effect.

[scheek]

It depended. I clocked it many times and it was always within 15 minutes. Jogging would take longer sometimes. However, it always failed at somepoint.

Cooling down period always let it jump start again and work properly for another 8-15 minutes. Then crash.

Note, my CNC groans too, especially when XY are both working at the same time. I think that is the nature of the beast. Jogging is much quieter when jogging one axis at a time. I don't think that is your problem based on my experience.

Did you build your powersupply? I'm a novice but I built my own CNC and powersupply and incountered every problem imaginable. Your similar problem was up there at the top. I replaced motors, drivers, and even the break out board. It was a process of elimination...but it came back to the simple fact the drivers were not cool enough.

Hope yours is something that simple.

[Buzz9075]

Scheek, the power supply was included in the complete package from Xylotex.

[thkoutsidthebox]

Firstly, I know absolutely nothing! Now that that is cleared up.......

....you said that its still happening, but just taking longer when you slow the machine down, that would lead one to think that it is happening at a specific line of G-Code that it just takes longer to get to when the speeds are slower.Can you identify a specific line of G-Code for each axis where it is happening for other more experienced persons than me to have a look at? Also, is that same line anywhere in the program before the line that causes the stalling? If not then maybe its related. If it is then Im probably way off track.

If it happens, and then you switch everything off and on, then run the same code, does it happen again at the same point after the same length of time. If not, and it happens immediately, then maybe its a hardware problem such as something shaking itself loose, or as mentioned overheating that hasn't cooled down yet.

Just because the package came assembled from Xylotex doesn't mean its infallible. You could be the unlucky 1 in whatever customer that had some tiny component damaged in shipping. But hopefully not.

Sorry if these are no use. Just trying to help.

[Buzz9075]

Not really it does not take longer to get to the same point... except for a few paths (which don't always create the issue but do more often than the rest) the problem just takes longer to exhibit itself sooner or later. TO cause it faster I need to put together more code with angles that have motors running at different speeds. Normally the problem occurs when one axis is going slow and another is fast. For example you would NEVER see it happen with the following which has all axis running at 52"/min
g00x0y0z0
g00x6.5y6.5z-6.5
m30
I can repeat the one above for ever and it will NEVER stall.

But something like this will sooner or later cause a crash.
g00x0y0z0
g00x1y3z-6.5
m30
This exact code crashed on the second run this morning for the z axis going from 0,0,0 to the -6.5 location. As well it crashed a little later when I entered a G00x0y0z0 from the location of x1,y3z-6.5. This type of path causes stalls anywhere on the board. I have moved it around the board (allong with others that cause movements of axis at different speeds). Testing went on for MANY hours yesterday....

The second path the z motor running at 52"/min and y at say 28"/min and x at probably 10"/min. And the z axis will stop.

Another that I would try that might stall the Y axis (have not tested it but this is the type of stuff that causes the low groan yesterday... soon as you hear that you know you are going to die):
G00x0y0z0
G00x21.5y2y-4

another one that I would try.
G00x0y0z0
G00x1y21.5z-6

While my code is all coming from 0,0,0 for this documentation the same type of angles on the machine anywhere will cause the problem.

From what I am seeing there is not a specific location on the table that the problem happens, more what is happening at the time with the motors (motors running at different speeds/velocities).

I hope all this made sense.

[paulC]

If the problem only occurs when all axis are moving it could be the power supply can't handle the load. Have you looked at that as a possible cause?
I have also seen some pretty wierd things happen when the motor power cables are not screened and running in a tight bundle or parrallel with mains power.

[Buzz9075]

I would not say it only happens when multiple axis are going, but the majority of the time this is when it happens. Z axis NEVER binds when running alone. Y axis once in a blue moon (maybe once an hour even two, usually right after I run a full machine test run)... but for sure when regularly going on multiple axis.

I don't know how to confirm the power supply is adequete, I would figure the solution supplied as a full package would have enough power to drive 4 axis and at present I am only using 3.

This is what I got: http://www.xylotex.com/Econo4AxSpecial.htm

[ronic]

I am not sure if it's the case here but I have some ploblems with motors impedance on 3977 driver which I built. To be more specific they run very well at high speed 1500 rpm but al low speed they lose steps somehow randomly. I changed the motors with a higher impedance ones and the problem gone.

[paulC]

It would be difficult for you to try testing the power supply initially. I think we have to assume the kit is ok for the moment
Ive been trying to follow the threads of your build.
I notice you have threaded rod butt jointed to the stepper in one picture. I would think with the inacuracies of allthread you would need some sort of coupler.
The lack of a flexable coupler could be the source of your binding.
If you are using plastic tubing as the coupler I would think you would need a reasonable gap to allow the plastic to flex in taking up the missalignments. I would look at the instructions of the kit, but if its not mentioned I would try about a half inch gap.
A picture of your couplers may help.
Paul

[swagman]

instead of doing a rapid move, try it with a g1 move and put a feed rate at the end of the line. g1 x2 y2 z-2 f50

[rgammage]

I was amazed to see that this software is running on Windows. Windows spends so much time sorting out its rubbish to find time to control your 3 axis at high speed. If you can try LESS memory ... closing down every application and as many of those little icons you see in the bottom right hand as you can, disconnect the internet and all superflous peripherals, printers etc. The less time it spends looking at other computers and external devices the better. Good Luck. I am looking at Linux for this task!