Nema34 - DM860A Driver & Mach3

[Clive S]

Today I tried to set the drivers to 1600 as someone suggested in this other post >>> LINK


While in the motor tuning section I put the values 320 steps 1500 speed and 40 acceleration.

Then I used the auto-calibration tool of Mach3 and it slightly changed value, like 326.5 ....

The axes with this configuration move quite well, maybe it's a little too fast.

What is still not good is the accuracy of the movement, in fact it is wrong about:
- 0.5 cm if I do G0 X20 F400
- 1 cm if I do G0 X40 F400
- 2 cm if I do G0 X80 F400

The positive thing is that when I do G0 X0 F400 from one of the commands above it is not wrong and returns correctly to the point of origin.

What could be the problem? The settings in the driver? Should I redo the mach3 motor-tuning?


For the Nema 23 engine, at the moment I used the same settings (different current values of course) but I have not yet checked if it is wrong during the movements.

First it is good you get consistent results.

How are you using the auto calibration? When testing use G1 not G0 (Go are rapid moves)

The greater the distance you measure over is better so try G1 X500 F400 and measure the distance moved then put that in the auto box and test again. Don't mix units up ie 0.5cm and 5mm generally we use mm or inches.

Another way is to work out the % difference in the distance moved and then use that % to change the step per.

[Channel]

Hi Clive,

Thanks for the tip on not mixing the units.

unfortunately this afternoon I did not have time to try to execute the command you suggested to me.

As a self-calibration I used that of the Mach3 software, in practice I have to enter a distance X for a given axis, the machine performs the movement and then I have to enter what has actually moved.
After that, the machine modifies the steps in the motor tuning section.


But, while I'm answering you, I'm in doubt ... the movement you suggested to me >> G1 X500 F400 in practice moves the x axis of 500 MM with a speed of 400 right?


Because in reality if I do, for example G0 X20 F400, the X axis moves 20 CM ...

So I might have set ten times bigger values ..

What do you think about it??

[cropwell]

I do not know your leadscrew pitches, but let us assume 10mm pitch and direct drive (no pulley ratio to consider).

One turn of the stepper motor (200 steps) will take you forward 10mm so your resolution (distance per step) is 10/200 i.e .05mm (which is OK for woodwork).

If you use microstepping, you will gain resolution - for example 16 microsteps will give you a resolution of .05/16 or .003125, which is over the top for woodworking and especially difficult to achieve that accuracy with a wood framed machine.

All that microstepping will do for you is make the motors sound smoother. I would suggest a value of 8 is practical, 16 at the most. Any higher and you will lose speed.

There is a common misconception that you gain accuracy with higher microstep values, but if you look at the way microstepping is done, you will see that for the majority of us amateurs, it has little practical advantage.

Have a look at this article https://hackaday.com/2016/08/29/how-...epping-really/ It is a simplification, but an easy read.

[Channel]

I took a quick look at the link you suggested, I read it better tonight.

What I knew is that if I increase the resolution too much then I will have problems with the engine torque. In fact, I imagined that 51200 steps / revs could not be used as a value.

My transmission of the X axis is made with:

- M12 screw
- a pulley 20 teeth 5 pitch on the motor axis (I have to check if it is 20 or 15 teeth)
- a 60-tooth pitch 5 tooth pulley on the M12 screw
- two 40-tooth pitch 5 pulleys connected to the ends of the screw
- two 40-tooth pulleys step 5 for "neutral"

I attach photos:


For the Y axis it's all the same except that the screw is an M10.

I attach photos:


For the Z axis I have:

- a pulley 15 teeth step 5 on the drive axle
- a pulley 20 teeth 5 pitch on M10 screw

I attach photos:



The information I miss is:

- How much should you set the driver for? (from what you say 8 could be an optimal value)

- Once the driver has been set how to calculate the "step for" parameter in the motor tuning section of mach3. (My idea is to do the theoretical calculation, and then use the "auto-calibration" tool of mach3 to adjust the value.

For speed / acceleration I thought I would try some values ​​until I found the optimal ones. (unless there is a formula to calculate these two values)

How can I get these values?

Thanks again for the advice and patience XD

[cropwell]

OK, so X and Y are belt drive and are a different problem.

On your Z axis you say M10 screw. Do you mean a piece of threaded rod with the normal coarse pitch of 1.5mm?

If this is so then on your Z axis 200 steps -> 15 (driving pulley)/20 (driven pulley)x 1.5 mm of movement = 1.125mm, so steps per mm is 200 / 1.125 = 225 (note this is steps per millimetre). If you then use microstepping, you will have to multiply that by the factor you set in the driver DIP switches. (other forum members - could you check this as I am tired today, I have done some work and I am not used to it )

I will get to the X and Y later.

A lot of information is here http://www.machsupport.com/wp-conten...all_Config.pdf

[Channel]

Hello,

yes, it is exactly a piece of M10 threaded bar.

So for the Z axis I was able to calculate it like this:

driving-pulley / driven pulley x 1.5mm (threaded bar) the result is divided by 200 (physical steps)
and in this case it would be just like you wrote:

15/20 * 1.5 = 1,125mm 1,125 * 200) 225.

Assuming to set the driver through the dipswitches to 1/4 (400) I have to make 225 * 4 = 900.
Or if set to 1/8 (800) 225 * 8 = 1800

This value is what I have to set in the motor tuning section of mach3 for the right Z axis?

Is this calculation valid even if the pulley on the threaded rod has not been tapped but is held in place by nut-locknut?
I attached a photo for a better clarity:

[cropwell]

Is this calculation valid even if the pulley on the threaded rod has not been tapped but is held in place by nut-locknut?

Only if the pulley does not slip, and that arrangement looks as if it will do so at some time.

You are only using threaded rod on X and Y as axles, so the pitch is irrelevant.

I have yet to get my head round the X and Y axis arrangements, but it looks like the belt attached to the axis will move 200mm per turn of the axle, but you are geared down 3: or 4:1 depending on the number of teeth on your drive pulley. So if you are using 15T drive then your resolution is 4 steps per mm or if 20T 3 steps/mm. So I can see why you are hoping to get some positional accuracy using microstepping to increase your resolution.

I have doubts that this current design will be very accurate. You should also look at some way of fixing your pulleys to the axles other than locking nuts, are there no grubscrew holes on the pulleys ?

[Channel]

OK, so X and Y are belt drive and are a different problem.

On your Z axis you say M10 screw. Do you mean a piece of threaded rod with the normal coarse pitch of 1.5mm?

If this is so then on your Z axis 200 steps -> 15 (driving pulley)/20 (driven pulley)x 1.5 mm of movement = 1.125mm, so steps per mm is 200 / 1.125 = 225 (note this is steps per millimetre). If you then use microstepping, you will have to multiply that by the factor you set in the driver DIP switches. (other forum members - could you check this as I am tired today, I have done some work and I am not used to it )

I think you make an error... because the result of 200/1.125 is not 255 but 177,77 ... BUT 200 X 1.125 is 225... so is the calculation or the result wrong?

Anyway I confirm the accuracy of the movements of the Z axis with the values set at 533.33 as described in the previous post ...

[cropwell]

I think you make an error... because the result of 200/1.125 is not 255 but 177,77 ... BUT 200 X 1.125 is 225... so is the calculation or the result wrong?

Anyway I confirm the accuracy of the movements of the Z axis with the values set at 533.33 as described in the previous post ...

It is 200/1.125 but I did say I was tired.