I'm planning out a project that will use an Arduino to control a high voltage servo (24-75V.) The power supply offered by the servo's manufacturer is very expensive. Also, the manufacturer's power supply is a 75VDC unit. I haven't been able to find any motor shields / drivers for the Arduino that can handle voltages much higher than the minimum servo voltage.

The manufacturer advices against using switching mode power supplies because of the reverse voltages caused by a suddenly decelerating motor. I've read that adding a capacitor at the tail end of a switching power supply can guard against reverse voltages (I hope I'm not misusing the term 'reverse voltage.') Looking at pictures of the manufacturer's power supply, I'd guess that it is not a linear supply.

Here's my options:
1. Use a 25-30V linear power supply with a large capacitor used to deal with voltage drops and handle peak current.

2. Use a 75V supply (save up for the servo manufacture's) and make a circuit external to a motor shield to allow for higher voltages (remember, I can't find a motor shield that can handle 75VDC.)
I'm not sure how to do that. I briefly read something about using a mosfet--they said something about another quadrant. I'd have to learn more about it.

I've attached a schematic pulled from elmomc.com. I could also try to mimic something like that.

I'd like to make my own power supply because thatd be the cheapest option, but I obviously don't know much about it.

What direction do you think I should take?

 

You can look at motor drivers or design a snubber circuit that can damp the ringing current generated by the inductive load.

http://www.cde.com/resources/technical-papers/design.pdf
http://www.cde.com/resources/catalogs/igbtAPPguide.pdf

The snubber works the same way a flyback diode connected across a DC relay coil does. It damps the reverse current generated by the collapsing magnetic field. Do some research on snubbers.

 

What are the plate details from the servo motor?
Are you planing on servo action, i.e. closed loop? Or just simple bi-directional control?
Max.

 

You can use MOSFETs in an H-Bridge. Use heatsinks.

 

What are the plate details from the servo motor?
Are you planing on servo action, i.e. closed loop? Or just simple bi-directional control?
Max.

Here is a link for the servo I'm considering: https://www.teknic.com/model-info/CPM-SDSK-2310S-RLN/

I plan on using it for simple bi-directional control.

 

This stuff is expensive for a reason. If it is the application that is important, then don't screw around learning to build the components you need. You have a pretty high mountain to climb from your present level. If on the other hand time is not a factor, then divide and conquer. You need to establish what kind of supply the manufacturer is using. The reason they do not recommend SMPS is that most of them have very poor performance in the presence of load transients. These happen when the motor accelerates and also when it decelerates. You can compensate for this weakness, but if you have never designed one before I would not recommend this application as a learning exercise in the design of power supplies.

I think you should consider designing the shield that you need to handle the voltage you want. If you buy the vendors power supply and design your own shield, then you can design a power supply as a secondary project. You might want to ask them if they will sample you the supply so you can design the shield, and maybe you'll buy it or return it after some period of time. The worst that can happen is that they say no.

 

Certainly appears to be similar to a stepper in torque characteristics, the torque drops drastically from 500rpm.
Unlike a servo where the torque curve is essentially flat from zero to max rpm.
For stepper and servo supplies unregulated linear supplies are adequate without going to the frailty of the SMPS.
Max.

 

Hi Max,
I suspect it is based on a stepper motor with an encoder to check and correct for missed steps. It is quoted with a resolution of 0.45 degrees which is 800 steps per revolution. (A normal 200 steps per rev stepper using quarter stepping.) It could also be a system like the UHU servo that uses a PM DC motor with an encoder and electronics so that it behaves like a stepper motor. (Just step and direction input signals.

Les.

 

Here is a link for the servo I'm considering: https://www.teknic.com/model-info/CPM-SDSK-2310S-RLN/

I plan on using it for simple bi-directional control.

I don't see a current rating for the motor. Strange.
But here is a discussion on a driver that @cmartinez has built.
https://forum.allaboutcircuits.com/threads/h-bridge-driver-for-a-90vdc-motor.123808/
Tell us some more about what you expect it to do? Reversals can be very tricky-braking not so bad.

 

I don't see a current rating for the motor. Strange.
But here is a discussion on a driver that @cmartinez has built.
https://forum.allaboutcircuits.com/threads/h-bridge-driver-for-a-90vdc-motor.123808/
Tell us some more about what you expect it to do? Reversals can be very tricky-braking not so bad.

I'm making a CNC router. Will use the Arduino to interpret G-Code and control the servos. If I can get it within .002" accuracy, I'll be happy.

 

I just studied the uhu servo document that you posted. And I'm not sure it has the capability of doing what you want. That is, make a dc motor behave like a stepper.

First off, there is no holding torque. And the controller only attempts to compensate when there's a deviation from a static position. Or at least that's what I understood from the document.

And second, considering the previous assertion, although the motor will reach its intended destination with good accuracy, it won't do it synchronously. That would make it impossible to perform motion in a coordinated way between axes, which is an indispensable characteristic in all cnc routing machines.

 

. If I can get it within .002" accuracy, I'll be happy.

Which will depend largely on the mechanics, ball screws etc.
Instead of re-inventing the wheel, any reason why not using Mach etc, etc?
Max.

 

Which will depend largely on the mechanics, ball screws etc.
Instead of re-inventing the wheel, any reason why not using Mach etc, etc?
Max.

Isn't Mach comparatively expensive?
I'm trying to keep the electronics under a grand. I'm making the mechanics myself.

I've seen YouTube videos of people using the ClearPath and Arduino. The ClearPath website claims that the motor I posted can be used for CNC applications.

I'll look at Mach products again, though.

 

Mach3 is free up to so many lines of code.
I have not used the Arduino, but I would hesitate to implement interpolated servo motion from scratch when there is also off the shelf, such as Kmotion and Linux CNC.
Max.

 

I just studied the uhu servo document that you posted. And I'm not sure it has the capability of doing what you want. That is, make a dc motor behave like a stepper.

First off, there is no holding torque. And the controller only attempts to compensate when there's a deviation from a static position. Or at least that's what I understood from the document.

And second, considering the previous assertion, although the motor will reach its intended destination with good accuracy, it won't do it synchronously. That would make it impossible to perform motion in a coordinated way between axes, which is an indispensable characteristic in all cnc routing machines.

Here's a video of a CNC made using these motors. I don't know what he's using to drive it though.

 

Here's a video of a CNC made using these motors. I don't know what he's using to drive it though.

Just watched the video, and I have to say I really liked those motors. But after seeing their performance curve, I think I'll stick to the slightly more expensive Oriental Motor type. For 30% more the price (about $600 dlls for the CPM-SDSK-2311S-RLN with power supply and cables, Vs $780.00 Oriental Motor) you get far more torque and much better performance.

 

Clear Path CPM-SDSK-2311S-ELN Torque/Speed Curve:

​

Oriental Motor AR911AA-3 Torque/Speed Curve:

​

 

I guess I got spoiled with using Servo's off the bat and their basically flat torque curve when I got involved in motion control.
I have never really developed an affinity to steppers, in spite of trying them on occasion.

Typical servo torque curve.
Max.

 

I guess I got spoiled with using Servo's off the bat and their basically flat torque curve when I got involved in motion control.
I have never really developed an affinity to steppers, in spite of trying them on occasion.

Typical servo torque curve.
Max.

Yeah... much more powerful... and much more expensive too!

 

Yeah... much more powerful... and much more expensive too!

Over the years I have acquired many deals, Ebay etc, From DC brushed to high end BLDC.
I admit bargains are getting harder now than early years.
I still have a few I am trying to find homes for.
Max.