[Ricc]

Hi all, I'm getting insane after this. So i have a 6040z chinese machine. The machine overall has always run decently well, however sometimes the Y axis stalled, which is indeed the axis with more torque on it as it carries the entire bridge.
Recently I have mounted a 3d printing head on the machine, and on my first big prints it is clear that the prints turn out inclined along the Y axis by a substantial amount on the order of the mm. Meaning that printing layers are shifted from their expected position the more the printing proceeds. I have made all what possible on the machine hardware side to reduce excess torque: cleaning all components up the single ball bearing, motor alignment, greasing and oiling, assembling the machine from scratch with perfect alignment. Also now there is no mandrel mounted which is normally very heavy. I am also cooling the steppers with fans. Software side also everything seems fine, the 3 motors all have the same drive model and are tuned the same. So to myself I can explain this only in one way: step loss due to electronics. Keep in mind that on 3d prints, the job is much longer and the paths too of course, so I assume that systematic step loss becomes more apparent. A slight amount of inclination seems to appear also along the X axis but much more lightly.

My machine mounts the attached TB6560 based drives and chinese breakout board. The drives are set to 1/8 microstepping. I know something about electronics but stepper drivers and breakouts are tad complex for me. Steppers are nema23 57BYGH76, 3A, 1,5Nm holding torque.
Anyone had a similar problem(step loss and occasional stalling) and know how to fix it?

[joeavaerage]

Hi,
those steppers have a 5mH inductance, and that is too high if you want to run them reasonably fast.

All steppers lose torque the faster they go, its physics and no getting around it. Inductance of the windings determines how bad that torque degradation will be, the lower the better.

For example a 23 size stepper of 1.5mH inductance (driven by a good driver) will retain approx 40% of its holding torque at 1000rpm.
A 23 size stepper of 5mH (driven by a good driver) will have....maybe....5% of its torque at 1000rpm....if it has not stalled already.

You may have to consider getting low inductance steppers. The manufacturers of steppes know that first time buyer look for maximum holding torque, and they can
deliver, but at high inductance. Only then does the hapless buyer learn about inductance. Low inductance motors tend to have fewer windings and therefore require more current for
the same torque....and be more expensive.

The classic method to overcome the loss of torque with increasing speed is to use a high voltage driver and power supply.
Your TB6560 is limited to 36V, and the power supply is possibly even lower than that, say 24V. That's a joke! You should get some decent
drivers that can handle 80VDC and use an 80VDC power supply. That will make your steppers sit up and take notice!

Craig

[Ricc]

Ok, I am honestly not sure that the inductance is 5mH. A few years back I had a datasheet from the manufacturer(I will try to find it back), but I couldn't find it again online. Online I only found a few random datasheets but not from the manufacturer and some are discordant one to another, on torque and amperage too. For one I found 2.2 mH. I agree however it's 5mH to a certain probability.

Btw, I am not turning very rapidly honestly, I'd say with a max of 300 rpm right now, considering thread pitch 5mm and 1500mm/min feedrate. Despite this, it stalled in one of my last works. Also I tested running very slowly or with half acceleration (150 mm/s^2 instead of 300) and this didn't seem to have an effect on step loss. What's strange is that while Y seems be the most affected, decreasing acceleration doesn't seem to improve the step loss, which made me think of defective or unstable drives. Also the randomness of the stalling events, apparently independent of load, makes me think of unstable drives, So, since I am in a hurry, I followed my instinct before your message and I ordered new and apparently better drives since I read about many issues on similar drives to those that I have. I will get them today, if they work I'll be happy, otherwise I'll resort to better steppers as per your suggestion, which btw were in the plans either way.

Thank you for your very interesting insight.

[servtech]

Hi all, I'm getting insane after this. So i have a 6040z chinese machine. The machine overall has always run decently well, however sometimes the Y axis stalled, which is indeed the axis with more torque on it as it carries the entire bridge.
Recently I have mounted a 3d printing head on the machine, and on my first big prints it is clear that the prints turn out inclined along the Y axis by a substantial amount on the order of the mm. Meaning that printing layers are shifted from their expected position the more the printing proceeds. I have made all what possible on the machine hardware side to reduce excess torque: cleaning all components up the single ball bearing, motor alignment, greasing and oiling, assembling the machine from scratch with perfect alignment. Also now there is no mandrel mounted which is normally very heavy. I am also cooling the steppers with fans. Software side also everything seems fine, the 3 motors all have the same drive model and are tuned the same. So to myself I can explain this only in one way: step loss due to electronics. Keep in mind that on 3d prints, the job is much longer and the paths too of course, so I assume that systematic step loss becomes more apparent. A slight amount of inclination seems to appear also along the X axis but much more lightly.

My machine mounts the attached TB6560 based drives and chinese breakout board. The drives are set to 1/8 microstepping. I know something about electronics but stepper drivers and breakouts are tad complex for me. Steppers are nema23 57BYGH76, 3A, 1,5Nm holding torque.
Anyone had a similar problem(step loss and occasional stalling) and know how to fix it?

Steppers are not closed loop, there is no feedback encoder or scale to hold a tight position control. Steppers are solely reliant on calibration, a few microns off on a short program is no problem, but the longer you run without near perfect calibration, losses by drift, through heat, or inbuilt inaccuracies in the motor increments, the more cumulative errors increase.

The only solution I can recommend is to have axis homing switches and home the axis from time to time during the program to re-zero each axis if at all feasible.

[joeavaerage]

Hi,
my mini-mill used open loop steppers that were well tuned and did not miss steps and with an accurate calculation as to the calibration it would not gain or lose
more than a few microns in 10 hours of machining.

Open loop steppers can when well tuned be perfectly accurate.

OPs problem is that its missing steps somehow or other......if not excessive inductance and therefore drooping torque then a low voltage (24V) driver. He should be running 48V at the very least and
70V-80 preferably.

Craig

[Ricc]

Steppers are not closed loop, there is no feedback encoder or scale to hold a tight position control. Steppers are solely reliant on calibration, a few microns off on a short program is no problem, but the longer you run without near perfect calibration, losses by drift, through heat, or inbuilt inaccuracies in the motor increments, the more cumulative errors increase.

The only solution I can recommend is to have axis homing switches and home the axis from time to time during the program to re-zero each axis if at all feasible.

Hi, yes, apart from the stalling events(the main reason why I decided to buy new drives), I thought that not enough accurate "steps per mm" setting could be the issue too on long jobs of many hours. I tried calibrating with a dial indicator but I really don't think it's enough. As I have limit switches I quickly thought about your suggestion too, ultimately leaving it as a last resort or so, because whilst on milling/engraving jobs stopping them to zero the axis wouldn't be too much of a pain, on 3d printing jobs this would disrupt the regularity of the polymer extrusion, ultimately altering the quality of the printing, apart from the fact that it would require quite some work to code a gcode postprocessing routine, and adding further preparation steps to my jobs. I would highly prefer to find a way to calibrate the "steps per mm" in an extremely accurate manner for long jobs, however for now I am not aware of a method to do so. I wonder how commercial 3d printing manufacturers calibrate this setting. Also I am not sure whether not enough accurate "steps per mm" would result in tilt of the 3d print, If I had to guess I would say it would result in inaccurate dimensions of the part but I am not sure. I have to make some more research on this topic.

Hi,
my mini-mill used open loop steppers that were well tuned and did not miss steps and with an accurate calculation as to the calibration it would not gain or lose
more than a few microns in 10 hours of machining.

Open loop steppers can when well tuned be perfectly accurate.

OPs problem is that its missing steps somehow or other......if not excessive inductance and therefore drooping torque then a low voltage (24V) driver. He should be running 48V at the very least and
70V-80 preferably.

Craig

Do you have any suggestion on how to calibrate the "steps per mm"? They are now actually set to 320, which arises from basic calculation knowing 5mm pitch, 1/8 microstepping, 1,8° step angle. The calibration attempt with the dial indicator clearly returns a similar number, however in the end I decided to leave 320 because I thought that "if the X axis runs decently on 320, why shouldn't the Y axis too?". However I now realize that this assumption is possibly incorrect, even in the case that the motor is not losing steps(which honestly I don't have a proof of but is very likely). Also i presume that backlash calibration could have something to do in this. However I think my backlash, at least, should be calibrated correctly. Anyway I am now setting up a gcode for step calibration on long runs(until now I only used 1,5cm run with the dial indicator, which is clearly not enough), aside from testing new drives today.

I get what you are saying on voltage/inductance. Have to be honest I'm not too much on the verge of spending money for new steppers and power supply as it's not the easiest of times for me economically, but I will probably make the expense if I the new drives or better steps per mm calibration don't resolve the thing.

[joeavaerage]

Hi,
the steps per mm depend on the mechanics of the axis, nothing else.

A two phase stepper (1.8 degree per full step) rotates once in 200 full steps, not 199.8 or 200.4, 200 exactly. That same stepper driven by a driver with microstepping of 8
micro steps per full step takes 1600 steps to rotate once, again that is exact, its not 1599 or 1601 but 1600. With a 5mm pitch ballscrew the steps per mm is 1600/5 =320.
Note this is exact, its not 320.1 or 319.8 but 320.

I have seen one mistake that has been made in the past. Ballscrew made to inch measurements commonly have a pitch of 0.2 inch which corresponds to 5.08 mm. It would be very easy
to mistake the two if you had an imperial screw but thought it to be metric but is not. The correct way to measure this is to have the machine rotate the ballscrew say 50 times, and measure exactly how far
it actually goes. Thus, if OPs machine is set to 320 steps/mm then a G0 X250 will cause the ballscrew to rotate 50 times. Then measure how far it actually went. If it went 250mm then the ballscrew is indeed
5mm pitch however if it is an imperial ballscrew (0.2"=5.08mm) then it will have moved 254mm. That should be an easily distinguishable measurement.

I suspect that over a period of time you will have to upgrade both the steppers, drivers and power supply to satisfy. Given the poor reputation on the Toshiba driver IC, probably rather unfairly, it's just such limited voltage
rather than anything wrong with it, I would start there. Get drivers capable of handling 80VDC and then a supply of 80VDC. If you get more capable drivers but don't increase the voltage you'll get no
improvement.

Craig

[Ricc]

Hi, ok I see, it makes sense. Again thank you for the interesting info. I thought that maybe a low quality leadscrew for example might not be precisely 5mm. I think I would have noticed, but I think it will be better to check what you said on the 5,08 mm pitch. But btw, I installed the new drives (took me quite some re-wiring time) and launched a 3d print. It's almost 11 hours in and I don't see that tilt artifact anymore. Nor it stalled. I used the drives for X,Y, E for now since I bought only a pack of three for 30 bucks and Z is used very little in this print. Hopefully they will continue to be stable. It is proving to be a decent solution as these drives are also much more flexible as I can change current an microstepping on the fly.
A possible reason for the step loss was also that the extruder, which should run on 1.5A was actually running on 3A as the old drives were all set to run on bigger motors. Then the PSU has only 15W of capacitance so it's maybe possible that getting near to that limit with a motor not absorbing the correct power might have brought some problem. The fan too was emitting strange noises that now seem to be disappeared. But overall I think the drives were not good, bad quality, many say that. If it works I'll stay like this for some time, until I can upgrade the machine.

[Ricc]

Ehm I meant 15A max for the PSU, not 15W of course.