[tfins]

Hello,

I have been struggling over the past few weeks on an intermittent problem with the router that I built about 6 years ago. I don't believe the problem has always been present but it has been highlighted recently by more use due to an uptick in product demand.

The problem: The machine will randomly stop while cutting occasionally it will studder a bit prior to stopping. This has happed in the x and y axis.


Details on the machine:
- Frame is 8020 constructions with aluminum tooling plate uprights
- Nema 23 motors (425 oz-in holding torque, 3 A rated current)
- DM-556D stepper drivers
- Dual motor Y axis
- 5 mm pitch ballscrews
- Limit switches, e-stops, & motor wire all runs in shielded twisted pair that is grounded to a common point in the control box (using mechanical limits and only use them for homing as there have been false triggers from vibration)
- Motor acceleration is set pretty slow at 6 in/s^2
- Typical cutting speed 60 - 100 ipm
- Tooling: Carbide roughing 3 flute 3/8" or 1/2"
- Spindle: 2.2kw water cooled
- Controlled from Mach 3 through a UC100 into a C10 Breakoutboard

Initially there was only one switching power supply (non-branded 48V 12A) that fed the 4 motors. It was fine for a while but over time has started to show the stuttering issue more frequently.

Second power supply was added (Meanwell 48V 12A), this seemed to help but did not fully resolve the issue on heavier cuts.

The non-branded powersupply was replaced with another Meanwell of the same specs. Each of the power supplies are wired to a single point 120v connection. The same goes for each stepper driver.

With the two meanwell switching power supplies the machine is still showing the stuttering behavior somewhat randomly. I started up the machine today to run some plywood and two times in a row had the machine stall on me while jogging between the holes it was cutting at a constant speed. This stall was a few minutes after I started the machine up.

Below is a link to a video that show the machine stalling while cutting out a skateboard blank. This cut, using a 1/2" 3 flute carbide rougher, will pull 3.8-4A on the spindle readout.

Theory on what is happening:

The stepper motors pulling enough additional current, in a fast manner, when the machine goes into a heavier cut causing the protection circuitry to think that there is a short circuit causing the power supply to foldback voltage or current or both causing the stall.

When there was only one power supply driving all 4 motors they would all stop once this stall occurred. Now that there are 2 power supplies (1 has both Y motors, the other has X and Z) X will keep going when the Y stall occurs.

Today I tried to prove what was happening by connecting the multimeter to the outputs of the Y axis power supply and watch the voltage. Naturally the machine did not show any issues while the multimeter was connected. I plan to continue to do this to see if I can catch something happening during a stall event.

I have done the math on the 5mm pitch ball screw at 50 RPM increments up to 200 ipm following the torque curve of the motor. I have a hard time believing the motor is stalling on the axis that has two motors vs one (There are heavy cuts in both directions the video just happened to capture Y which is the failure the is most frequently seen). At 100 ipm I think I should be able to stall the spindle prior to stalling the motors.

I have run the machine off of a different computer to prove that windows isnt doing something fishy in the background stealing resources from Mach.

Am I just running into the limits of a protected switching power supply?

https://photos.app.goo.gl/3d1hPXAWR7jp3ERT9

I posted this in the electronics category as I believe the power supplies are the culprit here. Please move to DIY CNC if it fits better there.

[Jim Dawson]

When was the last time you cleaned and lubricated the rails and screws?

An unregulated linear power supply is normally better for driving steppers. Something like this https://www.automationtechnologiesin...4820-with-5vdc

[tfins]

Rails and screws are lubricated from the outside (not a great method I realize) about every other time the machine gets turned on. A lightweight oil is what I typically use. I have considered that the gantry could have resistance due to a lack of lubrication inside the ball nuts and trucks but I have yet to disconnect the ball nuts from the gantry to see how the rails feel.

The linear rail has zirk fittings and the ball screws have threads for a fitting to be added. I am planning to pick up a grease gun that could hold oil today to try and lubricate them from the inside out.

I originally wanted to do 10mm pitch ball screws on the machine knowing how the torque curve on stepper motors falls off at higher RPMs. Looking back at those calculations maybe it is time to swap out the 5mm pitch for 10mm pitch. Once I can gain confidence in the machine running consistently again, I will likely bring the feed rates up and will get more force out of the 10mm pitch beyond 100 ipm.

Back to electronics though... The power supply seems like the likely culprit to me since it only seems to happen frequently when the machine is first switched on and warming up.

Is there a capacitor I can add to the motor drivers to give some smoothing of the quick high amp draw from the stepper motor during a heavy cut?

With the linear power supply will one unit be able to handle all 4 motors? (Edit: At 20A rated current it should not be a problem)

[Jim Dawson]

You might get away with large capacitors across the output of your existing power supplies. Just guessing here, but maybe 5000mF, 100V would be helpful. But doing this I would expect the switching power supplies to foldback on startup due to the large inrush current while charging the large cap, but this should not really be a problem once they are powered up and stable. I only mention this because it might shorten the life of your power supplies.

[tfins]

You might get away with large capacitors across the output of your existing power supplies. Just guessing here, but maybe 5000mF, 100V would be helpful. But doing this I would expect the switching power supplies to foldback on startup due to the large inrush current while charging the large cap, but this should not really be a problem once they are powered up and stable. I only mention this because it might shorten the life of your power supplies.

Thanks, I think the unregulated linear supply is the way to go. I appreciate the help