Mesa 7i48 problems - knackered 7i48?

It sounds like possibly one or more analog outputs were shorted to GND
or miss-connected in some way.

The 7I48 has only one +-10V analog out (AOUT)
so this should go to drive pin 9 and the adjacent 7I48 GND pin
to drive pin 8

I would try powering the 7I48 with no analog connections to
verify that its OK. Then perhaps add a say 1K series resistor in the AOUT
pin and try one drive at a time.

I'm not 100% clear here, because between the way the drive manual is worded and the way the 7i48 manual is worded, I find I can't be certain (so please forgive me if I check in detail!): first, I'm assuming that because the 7i48 says it's a 6-channel interface, when you say "The 7i48 has only one +-10V analog out (AOUT)" - you mean one per channel, 6 in total, using AOUT(0-5) and GND(0-5), yes?

Second: 7i48 manual says to use Up/Down PWM, i.e. type 2, and scoping the relevant pins (PWMGEN pin 0/1) on the 50-pin cable shows that:

for + voltages we get 0V on PWMGEN0 pin 0 and a 0 to 100% duty cycle on PWMGEN0 pin 1, and:

for - voltages we get 0V on PWMGEN0 pin 1 and a 0 to 100% duty cycle on PWMGEN0 pin 0 (or the other way around, but whichever, they swap roles when passing from - to +)...what the drive manual seems to call 'differential'.

That should translate to 0 to +10v on AOUT0 and 0 to -10v on GND0, with whichever pin isn't doing that, acting as 0V reference. Let's call this setup Version 1; I connect AOUT0 to drive pin 7, GND to drive pin 9, and cable shield to pin 8 with no bridging to 7 or 9.

(The Bosch manual says its command cable is set up as a differential input. This behaviour isn't what I understand to be called a differential signal, but looking at the scope output it kind of seems to make sense.)

Version 2 (single-ended?): AOUT goes from -10V to +10V and GND stays at GND, and the command cable is rigged as described on p8 of the drive manual i.e. with the shield pin and GND pin connected. This would seem to require that PWMGEN is set to type 0, but I tried that, and got an error saying that the 7i48 must be set for type 2. Plus, the drive goes from 0 speed to max speed with 0V-9V (as tested with a 9v battery through a potentiometer), but only in one direction, and the drive has no direction pin. So I don't think it's Version 2, but am a bit lost here.

Sorry for so many words! I will try what you suggest with the additional 1K resistor and go step by step through the wiring. If what I've written makes sense and you can say V1 or V2, please do. Thank you for your help!

GND is always GND. The AOUT pins swing from -10 to +10V
so version 2

If you GND any AOUT pin you are shorting out the analog outputs

Oh. In which case I don't understand at all...How is this consistent with what the scope shows (the pwm duty cycles changing from one pin to the other when crossing from + to -)? Or rather, how does the up/down pwmgen setting work?

The UP/DOWN PWM pins drive filters then a single ended output amplifier
so on the analog side GND is always GND ( and there is a single output per channel)

Ok, so: AOUT0 goes to amp pin 9, GND goes to amp pins 7&8, and shield...also to pin 8? or maybe to PC power GND signal, or case?

It looks like the most likely cause of the problem would have been mixing up wires 7 and 9 and ending up connecting AOUT to GND, as you say.

Yes, that looks correct

Also if the drives truly have differential inputs, you can
wire each AOUT to IN+ and AGND to IN- as long as
there is a common mode ground connection between the drives
differential inputs and the 7I48 GND.

Is that what seems to be described/illustrated on p8 and p9 of the drive manual - amp terminals 0v, 7, 8, shield and cnc command source GND (ie 7i48) all connected to the 0v busbar, and 9 to AOUT only? (there are no other command inputs on the drive)

Once again, sorry for so many words, and thanks for all your help!

Happy to report that it's all working now: "AOUT0 goes to amp pin 9, GND goes to amp pins 7&8, and shield goes to pin 8" was the right answer.

The 'weird scaling issue' I mentioned earlier on turns out to have been a mistaken about MAX_OUTPUT. I'd been thinking it was in volts for some reason, and left it set at 10 which made everything very slow. It seems to just refer to the OUTPUT_SCALE value: mine is set to 120 - max vel in machine units/sec - and MAX_OUTPUT being 10 meant I couldn't go above a very slow speed. Changing MAX_OUTPUT to 120 as well has fixed this and the motor speeds are much better.

Thanks again to PCW for the troubleshooting.

Time for servo tuning.