Hello Guest it is October 03, 2023, 08:41:22 PM

Author Topic: Missing steps running code  (Read 690 times)

0 Members and 1 Guest are viewing this topic.

Missing steps running code
« on: February 02, 2022, 09:59:41 PM »
I was able to get my mill running with Mach3 and ran several dry runs that seemed to work.  I tried a test on aluminum and my Y-axis lost steps, many steps, within a few minutes.  Several more tests gave the same results. I can hear the stepper whining whenever it starts missing steps.  I can jog around with both X and Y axes together and never have a problem,  but will lose steps running code sometimes on the first move of either X or Y.

I'm running 25KHz kernel speed and Driver Test says 'excellent' every time.  I have tried increasing step and direction pulses from 1 to 15 ┬Ásec with no results. I've tried active high and low for step and direction.  I noticed that if I leave the diagnostics screen up it will show the pulse count at around 25k, but it randomly drops to 392 or something crazy low then goes back to 25k.  That would seem to indicate a pulsing issue that doesn't show on Driver Test.  The computer is a Dell Dimension 4600 with a 2.66GHz CPU, 512 Mb RAM. OS is XP Pro 32 bit. It is set up as a Standard PC and per the optimization list I found here.  There are no other programs running other than Mach3. While running code CPU usage is 5-7% and Memory usage 180Mb.

If it is a pulsing problem wouldn't that ever show up while jogging too, or is that just dumb luck that it never seems to miss steps then?  If it is a pulsing issue is it better to just bite the bullet and invest in a UC100 or similar, or try another PC?

Any help appreciated.
Re: Missing steps running code
« Reply #1 on: February 02, 2022, 11:09:15 PM »
Hi,
when you are jogging or even just a dry run there are no cutting forces for the steppers to contend with.

Try reducing the max velocity of both X and Y axes to half, and try again.

All steppers lose torque the faster they go, so reducing the max speed will mean that the stepper has more torque to contend with cutting forces and acceleration.

Craig
'I enjoy sex at 73.....I live at 71 so its not too far to walk.'
Re: Missing steps running code
« Reply #2 on: February 02, 2022, 11:42:14 PM »
That is a good point, thank you.  I did try reducing speed by 10% or so but will try reducing more.  The speed is already pretty slow @28 in/min jogging and the 20 for the feed rate, but it's worth a try.
Re: Missing steps running code
« Reply #3 on: February 03, 2022, 12:04:06 AM »
I reduced the speed by 50% and left the acceleration at 3 and did a dry run of about 1200 lines.  It sounded different but no obvious misses and it returned to zero when I stopped the program.  I'll do some more testing tomorrow, I have to shut it down for tonight.
Re: Missing steps running code
« Reply #4 on: February 03, 2022, 01:10:19 AM »
Hi,
really the speed of the machine is immaterial, what matters is that the steppers have sufficient torque at the speed of operation to NOT
miss steps....whatever that speed may be.

You have not said what your steppers are, they could be high inductance 'first time buyer traps', nor have you said what voltage drivers and supply
you are using. If its 24V or even 36V then its no bloody wonder the things won't go.

First thing to establish is 'is there a speed below which these steppers behave'. Second step is to assess whether the speed obtained is just the physical
limits of your hardware or is it some fault, be it hardware or software.

Craig
'I enjoy sex at 73.....I live at 71 so its not too far to walk.'
Re: Missing steps running code
« Reply #5 on: February 03, 2022, 01:13:54 AM »
I don't know what the steppers are and can't find it in the scarce documentation.  It's a Taig/Microproto 2000 that I bought used.  It was barely used and I don't have any reason to think it's not original.  I don't think they are low enough quality to be considered suspect, but maybe I'm wrong about it.
Re: Missing steps running code
« Reply #6 on: February 03, 2022, 01:37:45 AM »
Hi,
you will never know until you measure the inductance.

Do you have an LCR meter?

High inductance does not mean low or lower quality, but it does mean severe speed limitations which is of particular importance in CNC machines.
The same high inductance motor which misses and stalls at the slightest provocation but could still be a top candidate for a telescope positioning
motor say.

All steppers lose torque the faster they go, that's the physics of how they work. Inductance is a good measure of how bad that torque degradation
will be. A 500oz.in stepper but 8mH will be lucky to have 5% (25oz) of its torque at 1000rpm while a 300oz.in stepper but only 1.2mH will
have 40% (120oz.in) at 1000rpm. First time buyers are likely to go for the higher torque unit, but at speed the lower torque unit, with very low
inductance, will far outperforms it.

The classic way to mitigate the loss of torque with speed is to use the highest voltage drivers and power supply you can. At the current time there
is a good selection of good to excellent quality 80VDC drivers at fair prices. If you want to get your steppers to perform at speed, doesn't matter
what steppers, then get some 80VDC drives and a 80V supply.

Craig
'I enjoy sex at 73.....I live at 71 so its not too far to walk.'
Re: Missing steps running code
« Reply #7 on: February 03, 2022, 12:41:29 PM »
Hi Craig.  Sadly I do not have a way to measure inductance.  I understand that could be a problem but for the purpose of exploring the question of pulse stream I'm willing to give the steppers and drivers the benefit of the doubt for now. I'm going to assume (dangerous sometimes) that the factory gave it steppers powerful enough to drive the table under no machining loads, and this problem can happen on the very first 'rapid' move the machine makes in a program, and usually happens when I am running above the table with no tool in the spindle.  I'll revisit the steppers.

I'm curious about the pulse stream since that is something I can 'see'.  When I run DriverTest the pulse count at the top is always 25400-25401.  When I look at the diagnostics screen I always see this, or something very close.  It cycles through counts like this all the time, several times per minute.  Is this just a display anomaly since the test program is always solid, or is this something that could be a cause of the problem?
Thanks again.

Re: Missing steps running code
« Reply #8 on: February 03, 2022, 01:32:32 PM »
Hi,
it could be to do wit the parallel port. Small variations like you have described are normal.

I upgraded to an Ethernet SmoothStepper eight years ago, and is light years ahead of the parallel port. Not that I ever
had that much trouble with the parallel port, but the ESS is just that much faster, smoother and more reliable.
I've also been using Mach4 for seven years, and it too is light years ahead of Mach3.

I still think your problem lies in the tuning of the steppers, but without some extra diagnostic equipment we can't prove or disprove it.

Craig
'I enjoy sex at 73.....I live at 71 so its not too far to walk.'
Re: Missing steps running code
« Reply #9 on: February 03, 2022, 03:15:40 PM »
Hi Craig.  I've been concerned about the pulse stream looking like it almost shuts off but I can't imagine that it really does.  I've been dropping acceleration and so far the results have been good, so it's probably more due to stepper tuning as you say.

I did find info on the steppers.  Underneath the machine I found a label on one that says they are from Lin Engineering, model 5718L-10D-01, 1.00A.  I've seen on other forums that they were made to spec for Taig.  I don't see a 1.00A in the standard models.  These are also said to be 200 oz-in torque but I have yet to find a torque curve for them.  I also found that the PSU is 48V, 7.5A.  Maybe I'll find more details.  Thanks for the help.