[dingo117]

Hello every one, sorry if this thread is in the wrong place but i think it’s the right category, so I made my own machine and got in to the hobby and with the hope to actually produce some commercial parts later on.

So the basic details of my machine are as follows:
axis motors: 3.1Nm Nema 24 Stepper Motor 3.5A 4-wire (2 motors on x, 1 motor for y and z)
stepper controllers: m335 driver, unknown/no brand, pictures attached
generic 24v 400watt switch mode power supply
control board: generic parallel board with optical isolation for inputs ( I think…) and amplified signals, schmitt triggers I think, can’t remember what the item listing said… pictures attached
5mm pitch ball screw and linear bearings for all axis
Spindle is controlled via 0-10 analogue output from control board
80 steps per mm on half step is how I drive it
Running mach3 using pcs inbuilt parallel port

the issue is it seems to lose steps, I did not notice it when running simple motion paths, I saw a little bit of lost steps but thought it was just losing steps from too aggressive a cut or moving to fast, but I recently tried a more 3D motion path (pictures attached) and its lost a lot of steps on all axis and pretty equally as well, when I return to home its offset by like 2-10mm in the same direction each time, so…
I tried all manner of setting changes in mach3 and on the motor drivers, I separated all the components to help avoid EM interference, I ended up disabling the spindle and all but one axis, unmounted the stepper so it drives nothing and free spins, I put tape on the shaft and body and mark out the home location and run the program and note each time I return to home the marks never match up. Even when just doing a few jogs from the keyboard it doesn’t match up. I tried driving the motor controller without the breakout board as well, just jumped wired directly from the parallel port and same deal. On a side note, mach3 has my motor speed limits at about 1700mm/min the motion paths should have cuts limited to 1000mm/min in fusion 360, but when I run the program in mach3 I sometimes see the feed peak to over 3000mm/min for a short period and I can hear it rev that high as well, the heck? Isn’t mach3 meant to always limit the top speed to what you set the motor to under tuning? I set the motors to 100mm/min and still lose steps when running my motion path as well, and I tried overriding feed speed to 40% in mach3 and still no luck. I ran mach3s driver test thing and its shows the timing all fine for my pc.

The crux: So I believe the issue is either the parallel port motion control method, mach3 or the motor drivers, unless any one knows something I may have missed, I think I am in for a new motion controller and 4 new motor drivers.
So my questions are:
1. any ideas on the best control combination for my setup that won’t lose steps and is the cheapest?
2. is it possible to not lose steps if your cutting rate and axis doesn’t bog anything down? Or is lost steps a fact of life for open loop stepper motors? In which case I will need to update to a closed loop system
3. Also I notice if the motors spin to fast It gets an oscillation issue and it stops spinning and just whines until the speed changes, I hear gecko drivers have resonance damping, does that help stop this issue and gives faster top motor rpm?


UPDATE: jog spamming in one direction results in no step loss on my unloaded test motor, but if i jog spam in both directions i think it loses a step on each direction change, returning to home never match up with my marks if i do this.

[CitizenOfDreams]

Try inverting your STEP signal polarity. (I'm afraid you have more serious problems, but it does not hurt to try).

[ger21]

Try cutting your acceleration in half. That's what usually causes lost steps.
Also, try changing the microstepping to 1/8 step (1600steps/rev). It should run smoother, and reduce resonance.

Is the machine screw driven? If so, what pitch?
What you really need are better drives, and a higher voltage power supply.
Something like these, https://www.omc-stepperonline.com/di...or-dm542t.html
Or a Gecko G540.
Use a 48V power supply.

[KH0UJ]

Try inspecting your ball screws, seems like all the symptoms on your machine points to a semi stuck ball screws (XYZ)

[dingo117]

i tried different accelerations, 100 mm/s/s was my default, i tried 25 but the same deal, the pitch is 5mm, i will try some lower step divisions tomorrow, i stayed away from it because i thought it would slow down my already pretty slow speed but i dont know this for sure, https://www.geckodrive.com/g250x-dig...per-drive.html i was thinking of these drivers, seems to be the same ones in the g540 but i could use any other motion controller if my parallel port output ends up being the cause of my problems, i have no fear of getting dirty with wiring and soldering things myself. but thanks for showing me the DM542T, seems to have the same functions and is a bit cheaper.

[ger21]

With 5mm pitch screws, you need high rpm's to get speed, which requires low inductance motors, and/or high voltage.

[wizard]

Lost steps are NOT A FACT OF LIFE with steppers

Beyond that if I understand your Post correctly, you are loosing position with no load attached. That is really bad. I’d take a long look at all of those generic parts to try to determine if electrical noise is an issue, or part quality is an issue. In a nut shell you should be able drive an uload d stepper back and forth and not loose position. The exception might happen hitting a bad resonance spot.

[dingo117]

Try inverting your STEP signal polarity. (I'm afraid you have more serious problems, but it does not hurt to try).

Hello, i gave this a try just now, inverted it under the motor ports and pins tab in mach3 but no difference, just inverts the direction, i did find that i only lose steps when changing direction of travel how ever, i dont lose them if just jogging in one direction

[dingo117]

Lost steps are NOT A FACT OF LIFE with steppers

Beyond that if I understand your Post correctly, you are loosing position with no load attached. That is really bad. I’d take a long look at all of those generic parts to try to determine if electrical noise is an issue, or part quality is an issue. In a nut shell you should be able drive an uload d stepper back and forth and not loose position. The exception might happen hitting a bad resonance spot.

i suspected as much, and yes the motor is unloaded free spinning shaft with tape on it and markings i match up when i set home and i check to see if they line up after i do some moves and return home. i found out if i just jog in one direction i dont lose steps, but if i move in both directions i think i lose a step for each direction change

[ger21]

Hello, i gave this a try just now, inverted it under the motor ports and pins tab in mach3 but no difference, just inverts the direction,

Change the Active Low for the Step, NOT the Direction. If it changed the direction of travel, then you changed the wrong one.

[dingo117]

Change the Active Low for the Step, NOT the Direction. If it changed the direction of travel, then you changed the wrong one.

ops brain slipped there, but yea i also tried changing step, all that did was make the stepper move very roughly and much slower

[CitizenOfDreams]

yea i also tried changing step, all that did was make the stepper move very roughly and much slower

OK, so that was not the problem.

[dingo117]

i dont know if i sated that i had fixed this, a while ago, just replaced the motor drivers, but i think the original drivers would have worked properly if i had set the signal timing right but i dont think i can do a custom signal in mach 3

[ger21]

You can change the pulse widths.

[dingo117]

thanks for the reply, i might have tried that at the time but the new drivers are quiet have given me 0 problems throughout a few years of work now so i am happy!

[rodw]

I can't believe nobody has pointed out to you that your choice of drivers and power supply is the problem. Your motors are unable to generate enough torque at 24 volts. I learnt the hard way too. Roughly doubling volts will double torque. But in your case you need to replace both your drivers and your power supply to do that. Whilst 48 volts will make an enormous difference, you would be far better off getting up to 70-80 volts. I'd look at finding a toroid power supply in the 70-80 volt range.

I'm in the process of upgrading my 48 volt Switch mode power supply with a 64 volt AC toroid supply and AC stepper drives which will give me about 90 volts at the motors

[dazp1976]

I have my Nema 24's running on 60vdc.
It's a BIG difference and I used to run on 36vdc.

24v is laughable. You may as well have put as little as 180oz/in motors on the machine at that voltage.