Knob fingers

Started off OK, wiring up the servo power transformer (previously modified to give 100VAC). Transformer is back in its box, cable glands fitted, IEC plug and socket fitted to flying leads etc. These will plug directly into the back of the console box.

No pictures, but tried out the soft start circuit for the servo power supply. Flashed up the AC input, enabled the Mean Well PSUs and checked out the sequencing of the time delay relay and soft start relay. Looks good. The soft start relay remains open for several seconds (can be varied), then closes, allowing time for the DC bus caps to charge up through the precharge resistor.

Then I looked at the 12V power supply connections for the servo drives. The drives conveniently use RJ45 sockets for both the "encoder" and "main" external connections. Obviously this presents a great opportunity to bung 12V into the encoder circuit. To be precise, if you stick the "main" plug into the encoder socket, it's the B+ input that gets to see 12V. I spotted this opportunity to let expensive smoke out when I first read the manual, so naturally I wasted no time getting down to it at the first opportunity ie today. Bollocks. The closest I can come to an excuse is "parallalax" ie it looked OK from where I was standing when I did a quick eye up before powering it up.

Thought I'd better rip the offending drive out and look for black bits that weren't supposed to be black. The identical connectors are visible at the bottom of the pic. The first device in the firing line is the SO-16 package at the bottom right. It's an AM26LS32A, which is a differential line receiver. Usually it likes to see 5V supply and sub-5V inputs.

Luckily, the input circuit is tolerant to a certain amount of idiocy:

And the spec seems to support that view:

Well, TFFT!! So it looks as if the CNC gods smile on me again. I'll wop it back in tomorrow. I feel happier at the thought of trying the circuit out if there is at least some basis for believing it won't shaft the encoder in the process.

FYI, here's what the whole drive looks like:

PIC micro at bottom RHS, 4 switch FET bridge (devices mounted under board, using baseplate as heatsink - not ideal but good enough here), 4 drive optos, current sense resistors, a big-ish electrolytic for the DC bus. Encoder line receiver at bottom LHS of picture. Not what you'd call industrial quaality but clearly they have spent a fair while developing the product. It's clearly capable of generating a 3-phase output with different components fitted, so I assume their "digital" servos use the same setup.