Need Help!- Building my first cnc & having problems.

[eddie3000]

Hello! This is a great forum, loads of interesting stuff to read.
I started building my first cnc out of mdf a few weeks ago, and I'm still at it. I bought a 4 axis stepper motor kit on ebay which seems to work, and I seem to have trouble tuning the motors.
As I've read, the slower you run the motors the more torque, and viceversa. I also read that, to check the max speed, you should test run the motors until they start to stall and then back off 10-20% or so.
I've tried many configurations on the driver board, full step, half step, 100% current, 75 %current, 50%current, etc.... But the motors seems to stall on ocassions even if they aren't running flat out, and miss quite a few steps during periodic test movements. They might even tend to stall at 8mm/s but go smoothly at 13mm/s. Accelerating from 0 to 15mm/s sometimes stalls, but mostly doesn't. On the other hand, running the motors at 25% current, 100mm/s2 accel and 25mm/s seems to work fine, though not much torque.
I can make them work at 1 or 2mm/s, loads of torque, and hardly no missed steps, but it would make machining an eternal task.
Could somebody point me in the right direction please? What am I missing? Are chinese drivers no good maybe? Should I set it to work dead slow and go off to drink some tea/beer in the meantime?
Thanks.

[RomanLini]

Your Windows setup may be chewing up chunks of time with interrupts etc. Try running the minimum software you can with everything (not CNC) turned off.

Some CNC software has test modes that let you check if Windows is causing glitching.

[MrWild]

Lots of people have problems with chinese drivers. One problem could be related to misalignment's. Extra drag really kills performance

[eddie3000]

Thanks 4 your replies guys, and sorry because I haven't specified what I have.

I am running EMC2 on ubuntu on an old pc using the onboard parallel port. The chinese driver & step motors are like these on ebay:

CNC Kit 4 Axis Driver Board 263 Oz-in Motor + PSU 220V en venta en eBay.es (finaliza el 15-dic-10 07:45:32 H.Esp)

My router is aprox. 900x700x290mm of workable area (the thing's quite a bit bigger), and the threaded rod is a standard steel 1,5mm/rev 10mm diameter.

EMC2 has a Max jitter setting, and I'm not sure what that is. I adjusted it to about 14000 using the jitter testing app while opening other applications, browsers, playing videos and all. I do not have a dedicated graphics card (could this be the cause?).

When I run the tests I have EMC2 running only. Nothing else.

It is very possible that I have slight alignment errors as it is my first cnc router and I haven't much experience. Most of it is cut by hand too. My machine is a bit tatty, but I think it runs fairly smoothly. And my intention is to learn as much as possible, get it going as best as possible, and then use it to make my second cnc, much better than the first. The thing is that the motors stall in a similar fashion even when not mounted on the machine, without mechanical load.

It may be all my fault for getting a chinese driver. Mmmmm....

So this will be my course of action, (unless advised differently by the nice people on this forum):

1) Change the pc for a more powerful one with dedicated graphics card. (Already have one, so it shouldn't take too long).

2) If after 1) the problem persists, I will check the driver PSU fisrt with an osc. to see if there is any noise.

3) Check whether the isolation between driver board and pc is ok.

4) Check driver output for noise and/or other possible problems with. Double check that what comes out from the parallel port is what comes out the driver output, and the optocouplers on the board too.

Any other suggestions?

Thanks again!

[dertsap]

try a new port cable , it seems that a lot of the cables that get sent out with this stuff is inferior , I bought a breakout board a while back and had a number of issues until i used my original cable , then it was problem solved

[eddie3000]

That sounds like a good idea. Thanks. It's true that the supplied cable looks a bit thin to have so many cables in it. I'll try and get hold of a good cable and see what happens. In the mean time I may try to mess around with the oscilloscope.

[EduardKieser]

eddy.
I bought that exact same driver board with similar motors and I experienced the exact same problems.
I also used EMC2 on a old PC using ubuntu 8 as a os.

I fear that the 2 of us may have under-invested in drivers. I don't think the driver boards are necessarily bad, but they are way to weak to operate the motors they are paired up with.
I for example was given 4 354ozin steppers with my TB-6560. After a lot if painstaking time wasting it turns out that my 9mH motors actually require about 60V to operate properly.

apparently a good rule of thumb is: (recommended voltage = sqrt(inductance)*20)

The kind gentlemen of this forum let me onto it. I have since bought 3 keling kl-6050 drivers but i'm also having trouble getting them to respond to my PC.

Please let me know if you get your TB-6560 to work effectively. It woult be good to know I could still use it for something

Eduard

[ger21]

apparently a good rule of thumb is: (recommended voltage = sqrt(inductance)*20)

For best performance, try voltage = sqrt(inductance)*32

[eddie3000]

Good Morning! Thanks for all these replies. So according to your formulas, with my motor with a 3.6mH coils, I would need anything between 40 and 60V to get adequate performance. GOSH! The PSU only supplies 24V!
No wonder I was so surprised how cool the motors were after intense test periods. Surprisingly the drivers hardly get warm either (except for a 7812 reg that gets pretty hot).
Looks like there's something in these chinese drivers that attract Edwards!
I'm gonna get my osc. out right away to see if I get to any conclusions.

[eddie3000]

OK! So I've checked the output of the driver. What comes in through the parallel port seems to come out from the optocouplers, and into the driver chip as well. But the motors still occasionally stall. So, either the driver chips themselves fail (which I doubt), or maybe it is true that the driver isn't powerful enough.
I am no expert in step motor drivers nor step motors. This should be kept in mind as my conclusions may be completely wrong.
I've checked the output to the motors. The output signal seems to be pretty noisy(apparently random noise), this might explain the noise the motors emit when idle. I have read in other threads that drivers are more or less silent or noisier than others. There is also visible crosstalk between channels (you can see the signal from the y axis on the x axis output, only attenuated). I don't think that is a good sign. The overall build quality isn't very good either in my opinion.
The ground is common throughout the entire driver board. This means there is no "real" electrical isolation at all. Why didn't I check this before?
In an intense moment of inspiration, I accidently shorted one of the driver outputs while testing, and I'm afraid I burnt one of the chips. Doesn't matter, still got 3 left on the board to play with.
The cable seems to work even though it's thin and cheap.
So, my quick diagnosis (I may be wrong) is that the driver seems to work, and it is quite possible (as Eduard pointed out) that the driver is underrated for the motors it was sold with. It's all my fault though, because the seller does clearly give motor specs (assuming they are correct). If I'd have known all this before I wouldn't have bought this kit. Too good to be true!
I can't say the driver is no good, but maybe not adequate for the motors I have.
So after my slight deception with the kit, the motors seem to be ok (I hope), what drivers would you recommend? Even though the board has 3 good chips left on it, I think I don't want to waste time and money cutting material and discovering the piece has come out wrong due to missed steps! I've glanced through the chip data sheet and raising the voltage would fry the three remaining chips.
Thank you all for your replies. Any suggestions?

[lukewarm]

Hi Eddie

Strangely enough I bought the same driver board and motor combination and also run it on EMC2. I played with settings endlessly trying to tweak all the performance I could out of these motors and came to the conclusion that the chinese driver simply wasn't up to the task. I came to this conclusion after exhaustively going through every setting possible, adjusting acceleration values in stepconf, countless hal file edits, even switching computers and as a last resort using windows and Mach3. All these efforts failed to realize any improvements.

This is IMPORTANT - The ONLY thing I was able to do that increased performance was to adjust the pot trimmer on the back of my PSU to increase the output voltage by 5 volts. I was able to get about a 25% increase in RPM by doing this. I was on the threshold of frying this driver though, so I don't suggest it as a solution.

My conclusion is these specific driver/motor combinations are mismatched and the motors aren't being supplied with enough voltage. I also found the driver to be extremely buggy and only semi-properly worked in 1/8th stepping mode. I am buying replacement drivers from gecko. 80volt 7amp for 114$ G203V Vampire 10 Microstep Drive

Hope this helps. These were just my conclusions based on trial and error, if you discover anything different please post!

[eddie3000]

Oh well! It could be worse. I'll have to look around for some good drivers then. In the mean time I will use the three axis chinese controller I have left at very reduced speed.
I hope this thread puts others off buying the kit I bought.