Brother tc-215 retrofit advice

ihavenofish wrote:

(...)
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.
(...)

 

That is exactly what I have in mind for this machine.

Though, I must admit - retrofitting it looks daunting. I assumed I can forget about HSM toolpaths. I need it to do mostly simple stuff.

It has Indramat servos and spindle, plus philips linear scales (the bad type not readable by Linuxcnc). Spindle is of newer type - not round induction motor, but this black square servo-looking thing (With matching and rather huge Indramat driver electronics box mounted in the rear cabinet). It looks like it has some sort of position feedback - according to the manual it can do oriented spindle stop, though I did not try it yet.

Control version seems to be one generation older than last Philips generation before it was sold as Heidenhain (I think it was designated as 432/9, but I may be wrong here. I thought that 432/9 cannot do 3d paths and is 2.5D not 3D because of weak CPU, but seller claimed that it does and It happily does helical paths and 3D lines). Software inside was definitely upgraded by someone - version numbers do not match these in manual, but control panel is of older type without top row of soft buttons (maybe controller broke and was replaced before? It's hard to know now - I cannot take a look in control cabinet right now as it's almost flat against a wall until we repair the floor and move it to it's final spot. We dragged it across the floor on the rollers, but this thing is so heavy that it busted few holes in the ceramic tiles on the floor.

I know that it works (for now at least) and I hope it will continue to work until I recover my money (with jobs I bought it for). Then if it breaks I will try to do a retrofit.

spindle is field wound dc with a tachometer. i assume with trk6 drive.
the encode for the spindle is on the spindle itself. orientation is mechanical with sensors.

432/9 is useful enough, the main meaningful differrnces were graphics, and memory if i remember. the /9 isnt an intel 8088, the /10 was a 286 i think.

the software version is more important. v703, v603, then v502. 703 being the newest.

You are right about trk6 - I have one.
Machine is from 1984 and I cannot reliably check software version because crt monitor doesn't start fast enough to see it (by the time image appears machine is already booted up). I have 2 programming manuals, one for V600 (but it looks like I got it by accident, it's for 432/10 control) and one for 302 software version - I think that I"m running this version, at least constant list matches the one in manual.

I managed to get machine running with some hiccups - Z axis started homing in wrong direction and tripped estop which dropped hydraulics and locked up everything, I had to undo belt cover and manually turn the motor to get it unstuck.
Hydraulics are a bit flakey - after boot up pump turns on for split second every 30s or so and light behind the button turns off while it's running, but it works good after it warms up. I found relevant page in the manual with description of this exact phenomenon - it blames check valve and it makes perfect sense, but I have no idea where this valve is.

Apart from this I took first chips (yay!) in aluminum as I'm waiting for coolant to do some steel.

Old crt screen is a bit wobbly so I bought cga-VGA converter and regular lcd monitor, but there some difference in the viring: converter expects r, g, b and h, V sync, but maho only has h, V and video (monochromatic signal I guess?). If I wire the same video signal to all 3 RGB and adjust gain for each one on the converter (it has little pots to do this) it should work and give me white colored signal, right?

The most serious thing is I couldn't get RS232 to work - no matter what the settings (Baud rate, flow control etc.) it just does nothing. (I'm using RS232 to usb converter with isolation based on ft232 and some analog devices isolation chips + nullmodem cable and db9 to db25 adapter).
Computer outputs data just fine (TX led on converter blinks, if I short rx and TX it comes back) but control does not see anything (yes I added eot at the end and %PM at the beginning ).
When control outputs data, nothing gets through (if I used wrong Baud or something I should be able to get anything, garbled bytes, some trash on the terminal - but there is nothing)
I will pull out oscilloscope and start poking at the wires, it's probably wired incorrectly)
 

fun fun.
mine did the backward homing as well, it happens when the machine was left with the axis in overtravel. its gets confused what direction its meant to go.

i dont know much else sadly. my machine had a bad trk6 and a bad Z scale so the control never fully booted (i could hold the hydraulics on and jog and home, but not much else).

on the control version, it will be written on the chips on the cpu board. my machine was upgraded at least twice, so you may well have a much newer control than the 1984 date suggests. my machine had 1984 written on it as well, but the sales receipt was 1987 so it sat around, and was likely updated before sale and the control had 1991 on some parts that they upgraded later on.

Hello,

I got RS232 working, I can download and upload programs, tool data and constants and it all works great - my null modem cable (sold to me as null-modem) wasn't null modem after all but after little surgery it works .

Sadly my software version is too old for DNC it doesn't have BTR option. Something was definitely changed inside,  as my schematic has comments and some changes done in pen, but control software and hardware match perfectly with manual. The biggest clue is machine constant list that ends on 214 and has constant for 'new faceplate' set - it seems to be the updated version of 432/9.
I tested out movements and all canned cycles and it all works beautifully including helical interpolation and automatic corner chamfering. I bought this machine with 4th axis add on (the dividing head that you can mount on the table) and that is my next target to get it running.

Another thing cropped up - when Y (vertical) moves fast (G0) the malfunction light comes on with error code 16 on the led display (it means Y axis slip clutch issue) but control does not throw any errors, machine works OK, does not stop and completes movement, keeps position to the micron. I did not notice any error in operation of the machine, but this little light is bugging me.

The slip clutch is in the motor coupling. It is for crash protection. If the clutch slips, the sensor tells the control. However, since the machine uses glass scales, this doesn't necessarily equate to lost position on G0. On G1 it would probably fault.

Somewhere in the manual it will give you instructions on how to calibrate it. You set it to the max torque the servo puts out in normal conditions (1.5-2x rated for example). This way it should only slip if you slam into a vise or whatnot. Seems likely that after years of use the clutch is soft and slipping at normal working torque, and needs to be tightened.