Hello Guest it is March 28, 2024, 05:40:49 AM

Author Topic: Only X-axis motor is working  (Read 33567 times)

0 Members and 1 Guest are viewing this topic.

Offline rcaffin

*
  •  1,052 1,052
    • View Profile
Re: Only X-axis motor is working
« Reply #130 on: April 01, 2018, 02:18:08 AM »
Yes, you will get better performance by putting them in parallel.
Yes, the stepper motor will get warm. More advanced drivers not only have anti-resonance circuitry, they also have the ability to reduced the holding current (and hence the heat) when the motor is not turning.

Unipolar vs bipolar: that is more a matter for the driver than the motor. All stepper motors have to be 'bipolar': the field has to reverse in each winding. Stepper motor primers are on the web.

Cheers
Roger

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Re: Only X-axis motor is working
« Reply #131 on: April 01, 2018, 03:26:30 AM »
I would NOT go with Nema 34's - they are big, slow and need a serious setup of power supply and drives (high voltage/high power) to make them dance. If i were calculating I needed a Nema34 I would switch to AC servo's without further thought, a small AC servo can out perform a Nema34 easily.

IIRC the average "good" setting for micro steps is 1000 but it can be anywhere you want really, just remember that the higher you go, the faster Mach3 has to generate pulses - theres a limit (if you are using M3 to generate pulses that is)

Definitely belts as mentioned, another good profile is HTD5, very common and no backlash, its the shape of the teeth that matter, anti backlash belts generally have round profile teeth whereas the old ones had trapezoidal shapes.

Getting the balance between motor power and machine build is pretty much impossible first time - the calculations are horrendous so most builders think they need oodles of power and look to Nema34's etc - this then opens up a world of pain because you now have speed issues and your machine, although it can easily break a 25mm tool off, cannot move very fast when needed ;)

On most machine builds, It seems the most common solution now is for Nema23/4Nm on something like a 70v power supply with good anti-resonance digital drives. I have these on my plasma table, and even though I'm only running 48v it can certainly shift when needed, One on X and two on Y, Z has a smaller 3Nm motor. This is gear on rack drive so not good for milling etc but adequate for plasma which has zero axis load.
Re: Only X-axis motor is working
« Reply #132 on: April 01, 2018, 12:41:38 PM »
I THOUGHT I HAD IT!!!

I had made a lot of different configuration changes. Finally I found one where the X-axis (working with only one axis til I found something that worked) was working flawlessly.
I had it set to single step mode and ran a test G-code file and after three times running the file I checked and the X-axis hadn't lost a single step.

Also, during this configuration searching I realized I wasn't dropping as many steps as I had thought. My rotational indicator (a zip tie on the shaft of the motor) was slipping as it went around the table side of the arc.

Anyhow, after the test, verifying that the X-axis hadn't missed a step, I hooked up the other 3 motors with the same configuration. Then I ran the test file again.

The test file is just a spiral design. Nothing fancy.

After a single run of the test file I found that all three active axis had dropped a LOT of steps. As many as 60, based on the angle of the indicators.

So, I am thinking that it's very likely that the problem is that the computer can't communicate properly with that many steps at the same time. Possibly the 3.3v on the parallel port issue.

Without an oscilloscope I won't be able to "properly" diagnose this. It's going to be some trial and error.

The other possibility is that the power supply can't handle 3 motors but my better judgment tells me this power supply ran (checked again) SEVEN of these motors in their original machine so it should be able to handle three of them.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

On the other side of things, I do believe I will stick with a NEMA 23 setup. There are a lot more options available in the 23 and even the dual shaft that I'd like to have for the X-axis.  Plus they are a lot less expensive than the 34's.
If I need more torque I can always gear them up/down a little.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I'm not ignoring the advice on the timing belts. It made it into the noggin. I've been toying with belt timing, lead screws, gear drive or ballscrews. My mill will definitely have ballscrews (bought two axis of them already) but the CNC router will probably be run on timing belts or gear rails.



Now I'm going to see about increasing the parallel output to 5v if there's any possibility. I can't help but wonder if a PCI card would deliver 5v.

I'm also going to try running the file again with the network controller unplugged (it's a USB type).

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Re: Only X-axis motor is working
« Reply #133 on: April 01, 2018, 04:22:33 PM »
Sounds like there is still some work to do there ;)

Even with ball-screws, belt connection is common and helps a lot with resonance issues etc, it also allows you to easily adjust ratio up/down without major alteration.
Re: Only X-axis motor is working
« Reply #134 on: April 01, 2018, 04:57:49 PM »
Went into the BIOS on the computer. It was selected as a PS/2 port. I changed it to EPP and measured the voltage. Still about 3.3
Tried to run code. No change in the performance. Still drops steps.

I checked the voltage at the drivers, coming from the BoB and it measured 4.25v so the BoB is allowing the USB power from the puter to switch based on the signal from the PP. Basically there's USB power coming through to the drivers.

I'm considering hard wiring directly from the computer power supply to the BoB. It would allow more amperage to be supplied to the BoB than the USB can supply.


Good point on the timing belt on the ballscrew. I had considered that the single step on the stepper on a 5:1 ballscrew allows for a .001 precision which would be plenty for my needs. Having a gear reduction and microstepping on the motor could make that smaller (not needed) but also a lot quieter.
« Last Edit: April 01, 2018, 04:59:42 PM by Inferno »

Offline rcaffin

*
  •  1,052 1,052
    • View Profile
Re: Only X-axis motor is working
« Reply #135 on: April 01, 2018, 05:21:36 PM »
Hi Inferno

Oh, things are not that bad. There is usually more than one bug, and you have dealt with the first one.

Zip tie - hilarious! Gowachin Law: first exonerate the judge. Ah well.

likely that the problem is that the computer can't communicate properly with that many steps at the same time.
No, I do not think that will be the solution, not at all. That is not how Mach works. I think you should assume that the right pulses are coming out the PP. From memory, EPP is the right setting. Also remember that when you exercise the X axis only, it works perfectly.

These are all steppers? If so we can discount encoder noise.

I see two main possibilities: pulses are being lost going through the Bob, or the drivers are missing some pulses. But the loss is most likely due to the extra load when you go multi-axis. Something is causing the X axis drive chain to go wonky.

Power supplies are where I would be looking. Without a CRO this gets a bit tricky. Set up your system with that spiral test program (a good idea, btw) and monitor the various power supplies on the BoB and on the drivers as it runs.
Btw, that 4.25 V - that sounds wrong. Either 3.3 V or 5.0 V would be OK, but not 4.25 V. You may need to try an external PS on the BoB.

Cheers
Roger

« Last Edit: April 01, 2018, 05:23:24 PM by rcaffin »
Re: Only X-axis motor is working
« Reply #136 on: April 01, 2018, 05:37:44 PM »
Hi Inferno

Oh, things are not that bad. There is usually more than one bug, and you have dealt with the first one.

Zip tie - hilarious! Gowachin Law: first exonerate the judge. Ah well.

likely that the problem is that the computer can't communicate properly with that many steps at the same time.
No, I do not think that will be the solution, not at all. That is not how Mach works. I think you should assume that the right pulses are coming out the PP. From memory, EPP is the right setting. Also remember that when you exercise the X axis only, it works perfectly.

These are all steppers? If so we can discount encoder noise.

I see two main possibilities: pulses are being lost going through the Bob, or the drivers are missing some pulses. But the loss is most likely due to the extra load when you go multi-axis. Something is causing the X axis drive chain to go wonky.

Power supplies are where I would be looking. Without a CRO this gets a bit tricky. Set up your system with that spiral test program (a good idea, btw) and monitor the various power supplies on the BoB and on the drivers as it runs.
Btw, that 4.25 V - that sounds wrong. Either 3.3 V or 5.0 V would be OK, but not 4.25 V. You may need to try an external PS on the BoB.

Cheers
Roger


I was mistaken about the 4.25
It's actually 4.45
That's splitting hairs.

I checked the USB on the BoB and it's reporting the 4.45 so that's where the drivers are getting their voltage.
I checked the power supply on the computer and I can get 12.08 or 5.1v if I hard wire into the power supply.

I'm leaning that direction now. Hard wire.

I'm also wondering if I should switch the drivers from a common positive to a common ground and then set the Mach3 for that in the port settings. I suppose it can't hurt.

I'm really leaning to the USB not being able to supply enough power for the drivers all at once. Especially considering it's 4.45v. Theoretically, if all lugs are going at the same time I could be drawing current on 12 devices. 4 drivers with 3 "devices" each.

Offline rcaffin

*
  •  1,052 1,052
    • View Profile
Re: Only X-axis motor is working
« Reply #137 on: April 01, 2018, 06:06:58 PM »
Common Ground - DEFINITELY! I would never run on a common positive, ever. Huge room for disaster there.

Getting a good solid 0/+5.0 supply: YES.  Hard-wire - YES.

Now, a thought. The X axis now runs fine by itself, yes? Can you get the Y axis running properly *by itself*? If so, same for Z axis - *by itself*. If so, then the odds of there being a PS problem rocket.

Btw, if you are checking the PS voltage, check it at the PS output AND at the BoB PCB AND at the the driver PCB as well. Do not assume the wired connections are perfect. Advice for the day.

Cheers
Roger
Re: Only X-axis motor is working
« Reply #138 on: April 01, 2018, 06:56:34 PM »
Common Ground - DEFINITELY! I would never run on a common positive, ever. Huge room for disaster there.

Getting a good solid 0/+5.0 supply: YES.  Hard-wire - YES.

Now, a thought. The X axis now runs fine by itself, yes? Can you get the Y axis running properly *by itself*? If so, same for Z axis - *by itself*. If so, then the odds of there being a PS problem rocket.

Btw, if you are checking the PS voltage, check it at the PS output AND at the BoB PCB AND at the the driver PCB as well. Do not assume the wired connections are perfect. Advice for the day.


Cheers
Roger

Good thoughts. I will go there next.

I just swapped out to a common ground and flipped the standard in the port setup.
I also hard wired to the Computer power supply.
I checked voltages and now I get +4.5v at the driver (direction pin and enable pin) when it's engaged. That should be more than enough to ensure a "good" signal.

Same results, if not a little worse.

During all the different tests, I get weird voltages all over. At one point I found 2.2v 

I think it's time I mapped this out.

I will isolate the X then Y then Z to see if one of the steppers is giving a weird feedback. 
Re: Only X-axis motor is working
« Reply #139 on: April 01, 2018, 07:27:05 PM »
X isn't working right now when isolated. There goes THAT test.

It's odd because when I jog the stepper I can see that the direction will switch back and forth when jogging without me touching the other direction. For instance, I jog clockwise a few times and it might make three jumps clockwise and one jump counterclockwise. ONLY when I jog it though. If I hold down the key to move it it will sometimes jog backwards for about 5 steps (7 degrees) and then go forward.

The Y axis and Z axis displayed the same symptoms.

As part of the above test, I switched around the coils and rain on just a single set of A/B. It didn't change the symptoms
I also flipped the polarity of one coil to see if it changed anything and it didn't.

There's something weird here.