Retrofit servo drives for Brother TC211

More
07 Dec 2016 17:50 #83748 by JohnW
Hi All,

I have bought one of these recently, intending to do a full Linux CNC conversion and replacement of all electronics as I bought my machine without any electronics, but complete with original servo and spindle motors. Microswitches still in place too.

I would like to get the servo motors running first, they are removed already. From reading another TC211 thread on here it appears they are AC servo motors with Resolvers. It seems that (for example) a Granite Argon drive is needed, one per servo axis, at €489 per axis. The machine has a tool changer, and I might (later on) want to add a 4th axis, and even do more CNC conversions in future, so things look expensive going forward from a drive perspective alone.

I have been thinking about using a small microprocessor to control each motor with some Solid State relays to control the motor fields, and some analogue inputs to read the resolver outputs, etc. I can't seem to find existing projects out there doing this.

Would Allen Bradley motor drives off ebay be a cheaper/viable option? I am wondering if a drive like the Bradley would be possible to configure for these old servo motors with resolvers that might not conform to what is expected by the Allen Bradley driver.

Please Log in or Create an account to join the conversation.

More
07 Dec 2016 20:15 #83753 by andypugh

I have been thinking about using a small microprocessor to control each motor with some Solid State relays to control the motor fields, and some analogue inputs to read the resolver outputs, etc. I can't seem to find existing projects out there doing this.


LinuxCNC has the ability to convert many types of commutation signal.

For example you could use a Mesa 7i49 card to read the resolver angle and then convert that in HAL to the conventional 3 Hall sensor signals using the bldc component (or to any of a number of other formats).

My lathe and mill use the 7i49 resolver card and the 8i20 drives.

Pico also have a Resolver convertor, though that converts to conventional encoder signals. If the drives do not use the resolvers for motor commutation then that is an option too. (it is possible to commutate with an incremental encoder, but it isn't elegant)
The following user(s) said Thank You: JohnW

Please Log in or Create an account to join the conversation.

More
15 Dec 2016 21:07 #84221 by JohnW
Andypugh I see now googling more that you made an Arduino board to read resolvers once upon a time? looked good! Did that work well in the end? Also wondering if I could use resolver ICs you mention.

Once resolved, is it pissible to do commutation on a Mesa host pci card using the FPGA and an external amplifier?

Please Log in or Create an account to join the conversation.

More
15 Dec 2016 23:59 #84232 by andypugh

Andypugh I see now googling more that you made an Arduino board to read resolvers once upon a time? looked good! Did that work well in the end?


No, it never really worked very well. The Arduino analogue reads were not fast enough, really.

However, that was then and this is now, and perhaps with one Arduino Nano per channel at <$10 it could be made to work. It relaly needs a clever tracking-filter / dpll like Mesa use in the 7i49.

Once resolved, is it pissible to do commutation on a Mesa host pci card using the FPGA and an external amplifier?


Potentially. But the resolver-to-quadrature options are not great for commutation use, you need to rotate the motor to find the rotor alignment, and to do that before you know rotor alignment you need to run the motor "blind" at first.

If you are using resolvers for commutation then the 7i49 seems, to me, to be the best choice. Which is why I have two of them and don't use the Arduino approach any more.

Please Log in or Create an account to join the conversation.

Moderators: cncbasher
Time to create page: 0.077 seconds
Powered by Kunena Forum