Brother tc-215 retrofit advice

Hello,

I recently purchased old brother tc-215 machine (it will take a while until it arrives) and I'm now collecting info. Original controller works and machine seems to be in good condition.

I intend to retrofit it with linuxcnc as original interface (Brother proprietary conversational cnc) is just abysmal and cannot be used with any cam software that I know of.

Things I already know (gathered from reading forum posts here and on Cnczone) :
1. All axes use servos with resolver,  but servo drive outputs quadrature signals to the controller. This signal can be used as position feedback after I disconnect controller.
2. Servos are +-10V analog.
3. Tool changer has some makeshift optical encoder that can be read with simple digital inputs
4. Spindle is also AC servo motor and has some kind of position feedback (I'm not sure what it is, but it has to be present, this cnc is capable of rigid tapping)
5. I know that there was at least few successful retrofits of tc-215 or tc-225 here.

I intend to use original servos with their drives, add 2 Mesa cards (7i77 for servos and 7i92m for digital IO and PC interface)
X and Y servos are rather anemic, but will use this machine mainly for cutting aluminium and plastic and maybe some light steel jobs (mainly drilling and tapping), not hogging out material at high speed. It was designed to be tapping center after all.

I collected some PC parts that I had laying around and assembled computer (Asus M5a97 + amd FX-8120), it has approx. 30us of worst case latency with preempt-rt kernel. Is this OK for linuxcnc with Mesa cards?

I read that there can be some tuning issues caused by low readout frequency by original drives.

My questions (mainly for those that have been playing with retrofitting this machine)
1.is it better to use for example 7i49 (resolver interface board), read resolver directly and not rely on signals generated by the servo drive (I don't know if that is possible, drive generates it's own resolver excitation signal, 7i49 does that too - it may not be possible to connect card AND drive to resolver) or just go with quadrature signals from servo drives and try to tune stuff in software?

2. Is spindle drive also analog +-10v controlled? I found few posts about some difficulty in direction control, but I don't know what's the deal with that.

3. How are stops and limits wired? Do they connect to the servo drives, safety relays or do I have to rig something up after disconnecting controller to avoid running over them - I know that there are three switches per axis, two overruns and one for homing.

4. How is spindle overload handled? Does drive switch off and signal some error code or do I have to install current and temperature sensors and monitor it externally?

I will try to document my progress with the retrofit here

Yup. This and github repo from linked post was the reason I decided to buy this machine (retrofit with original servos and drives is definitely possible, I just need to figure out how to do this )

Hi,

For TC-225;

This guy says he's using genuine drivers and motors.
Can you also let me know if you can make the original spindle driver?



 

Thats a TC225 which has quadrature encoders

I will get the machine hopefully next week and I will se for myself if it has quadrature outputs by poking at wires with oscilloscope (if it doesn't have quadrature outputs, I'm probably buying Mesa resolver interface board)

Hello,

Thank You all for replies, unfortunately my Brother TC-215 purchase didn't go well. I ended up getting something a bit bigger - MAHO MH600C with Philips 432 control. Control is 3D capable (it even does helical interpolation). Currently I don't see any benefit from retrofitting it (maybe apart from new LCD and potential rigid tapping, but I have no idea if this machine is capable of doing it - damn manual is in german, that makes it hard to read) 

The resolver thing is a red herring. it's just quadrature as far as linuxcnc is concerned. Shame you didnt manage to get it, the maho 600c is... uh, a much more involved project. (i had a maho hc5hs whihc is basically a horizontal variation of the same machine).

I'd need to know what year the maho is and the control version (502, 703 etc), but the words phillips suggests its before 1990. (they were rebranded heidenhain in 1991).

As for benifits of linuxcnc... speed and memory and easy of use (networking etc). The maho w/432 can certainly do a 3d path or hsm path, but it will do it extremely slowly, pausing every few lines of code. You need to drip feed larger programs which may require additional and expensive software and hardware.

Maho parts are brutally expensive used. so perhaps a strategy is - if it fits your needs as is - wait til something breaks on it, then retrofit.

For the sake of information, a retrofit would involve:

New servos (the stock ones don't have encoders) OR new linear scales (the philips ones are not readable by linuxcnc - later heidenhain ones can be).

Many io - like a hundred - for various pumps, solenoids, motors, gear changing, sensors etc. Mahos rely heavily on hydraulics.

Not an impossible project by any stretch, but in a different league than xyzs and a few switches on a brother.