I’m using the standard X/Y axis CNCRP Pro nema 34 rack and motor mount assemblies in my build. PRO Rack and Pinion Drive, NEMA 34 | Avid CNC - that part, but I"m using a clear path servo.
I’m struggling with my servo tune and the parts i’m running. If I cut above about 100 IPM, I get some overshoot in my corners and some noticeable deviation in high acceleration moves, visible in the cut. I’m wondering if anyone else has done this, and if so, did you experience any tuning issues with the belt system? I don’t think I have any mechanical backlash, but I do have a very heavy gantry (same 3x6" 8020 extrusion, but 72" long.
Maybe I assembled it wrong? anyone have any good ideas here?
I have the avid electronics kit. I run over 200ipm with no problems. I have a couple hundred hours of operation with no issues don’t know if that helps other tha. A comparison
Are you getting overshoot on both the X and Y axis? When I went through the auto tune the X and Z axis motors performed well but the Y axis with the two motors was difficult to get right. I finally ended up having Teknic manually tune the motors. They can do that through an internet connection with an application running on your control computer. After the tuning I have no overshoot and am able to cut up to 600 IPM and have rapids set at 900 IPM. Make sure you have the spring tension set correctly for your X and Y motors.
I did have them tune it that way, and I got a good but very jerky result, and then tuned it a different way with them a separate time which was smoother but spongy.
I’m going to approach this as a mechanical problem first and see if I can improve rigidity in my assembly first and then come back to tuning.
I’m the meantime, can anyone explain Mach 4’s CV wizards? At operating speeds it’s impossible to “see” a difference.
What acceleration setting were you using in Mach 4 when they manually tuned it? If the motors are sized for the gantry weight, I am at a loss for a explanation. I use the Clearpath CPM-SDSK-3421S-ELN and have power to spare. Do you use the SmoothStepper with Mach 4 as the motion controller?
I am not familiar with that wizard but I never have had a problem tuning the motors that couldn’t be solved by just adjusting the acceleration. Then further refining it with servo tuning.
As a side note: I have the same exact motor as you - would you mind sending me the config file from you Y axis (dual drive) motor? I’d love to take a look at the servo parameters you ended up with.
I will go to the shop this afternoon and get the configuration files but I only have a standard 4 foot gantry on my PRO4848. I have been running at 100 for acceleration so 60 should be OK. Avid’s default setting for the NEMA 34 steppers is 50 but I have heard that some people run them at 30. Lowering the acceleration should help with you overshoot issue though, easy to try.
Have you added weight to the base for stability? I have the leg kit which I reinforced with 1" MDF and added two additional shelves also out of 1" MDF. Then I used the bottom shelf for sheet good and lumber storage. All tolled I have probably added 300 pounds to the base. I can tell a big difference in the performance from when I was running with no added weight. With your larger gantry it would be even more important.
How you run your CAM can also be a contributor to overshoot. If you are making several passes for a profile cut make sure your plunge rate is the same as your feed rate. If you use a slower plunge rate with a ramp you will be slowing down and speeding up at each end of the ramp. That can make a big difference for small circles. The safest way to go if you have Vcarve Pro is to use the spiral option for profile ramps with the plunge rate set equal to the feed rate. That will give you nice steady feed all the way to the bottom of the cut. If it is a small circle for a countersink or screw hole an inside profile cut will be more accurate than a pocket cut.
So, the problem is fixed and it’s because I am an idiot. I had assembled the rack and pinion drive units with that nylon hex nut in place. I see now that the instructions call out that you should remove that. Instead, I had assumed that that part was a necessary spacer. Anyway, it turns out that a nylon nut has interesting dynamic properties. I replaced that with an aluminum spacer and it all works fine now.
Glad you found the problem. Now that you have everything assembled correctly I would still have Teknic manually tune the X and Y motors. Auto tune works fine for Z but I have found that manually tuning definitely benefits X and Y. Are you using the hard stop homing or the proximity sensors? I have found that hard stop homing with precision homing activated to be extremely accurate and dependable.
There might be a better way to do it but I let the motors home by themselves when they are enabled and then do a home in-place in Mach4. Works fine after you get use to it.
I am a Teknic OEM and my best advice to you is turn them to your system. I didn’t even know the factory would try to tune then for you. Learn something every day!
You will find it a bit challenging to turn the two Y axis motors but everything else should be straight forwards.
Also, do not exceed the machine specs as AvidCNC defined it even though your motors can do better!
Why?
Because the limitation of your system wasn’t the original NEMA 34s. Its that rack an pinion drive.
Not only can you rip the teeth off the pinion gear but the accuracy and repeatability AvidCNC publishes is the rack and pinion, not the electronics.
Now the one function that the Teknic motors give you that matters more than anything else is their patented g-stop function. That would greatly benefit the AvidCNC machines! It removes all of the vibration and whiplash in the system when tuned correctly. It would make the AvidCNC machines really shine.
That said, I believe you would have to use their SC line of motors and the Meridian controllers to access that. Now, they are releasing new stuff all the time so my knowledge might be dated but that is a really expensive system and your would be coding all of your own motion control as they do not provide that.
Anyways, YES!! Retune now. Your machine isn’t performing to spec until your do!
Best of luck. I will try to help if I can. The Teknics guys and gals are always great though. Even though your are not an OEM they will still return your calls.
I know this may be a bit late but I read your post about that nylon nut on the rack and pinion, did you really leave it on? Couldn’t tell if you were joking?
I’ve been working on my build using the standard X/Y axis CNCRP Pro NEMA 34 rack and motor mount assemblies, along with the Avid CNC PRO Rack and Pinion Drive. However, I’ve opted for a clear path servo which adds its own set of intricacies.
I’ve gone through my setup and I’m confident there’s minimal mechanical backlash. It’s worth noting that my gantry is on the heavier side, comprising the same 3x6" 8020 extrusion but spanning 72" in length. Could there have been a misstep during assembly? I’m open to any constructive suggestions or insights from those who’ve treaded this path before. Your input could potentially help me iron out these kinks and achieve smoother operations.
I’m still intermittently troublshooting my setup as well. One thing I’m thinking of trying is messing with the gear reduction for the avid unit. I’m thinking of shrinking the large pulley so that I get less mechanical advantage, and increasing the drive sprocket so I get more engagement with the rack. The clear path’s have so much torque that it’s somewhat hard to modulate under very short accelerations.
I can say, though, that my worst situation is jogging where there is no CV mode.
As I said earlier, more RAS on the clear paths helps, but the tradeoff seems to be some loss of roundness in small bores and sharp corners under high speeds/accelerations.
I have Clearpaths on my PRO 4848 and they work very well. I did have to have Teknic tune my Y axis via an internet connection. Have you done that? Just using the auto tune results from the X axis for Y did not work for me. I also do not set the acceleration in Mach4 any higher than 50.
On mine too, although I’m using custom servos, not clearpath. The tuning values for Y were way different because (1) the gantry wieghs far more than just the spindle, and (2) you have two motors.
When I watched Clough42 set up his, he was careful to weigh the gantry, then add mass to the spindle equal to half the gantry, tune X, copy X params to Y, remove mass, retune X. It seemed to work for him.