With all the wiring checks you’ve done, I agree the box is most suspect. I’m curious to know whether, in all of those blown TCMs, was it always the one output driver that was blown? If it were always, say, the 1,3,5 solenoid driver, then that solenoid may have shorted coil turns or a short to its ca...

I have a feeling that the it is holding both solenoids on at the same time Now, this is a bit worrying. A short between the drive lines to the 1,3,5 and 2,4,R solenoids would of course cause them to energise simultaneously (but incorrectly) and it would also screw up any valid feedback current sens...

OK, things are a bit clearer now that you've provided a schematic; unfortunately, it doesn't make the solution any easier. Sorry I gave you such a simple "solution" first: using a zener. I had my suspicion about the source of that offset, but thought "why complicate things to start wi...

I'm not familiar with the specific output driver of the Izuzu TCM, but relay or solenoid drivers are typically open-collector transistors, or open-drain MOSFETs. The driver acts as a switch, i.e. it's either fully on or fully off. When ON, the full supply voltage is developed ACROSS the load. Conver...

The SSR's input current should be around the 10 - 15mA mark with a 12V supply. For a 3.3v zener, power dissipation works out at around 45mW, so a standard 400mW rated device will be more than enough. Yes, just in series. Be careful of polarity; positive current flow is into cathode of zener. You won...

Shouldn't need anything as elaborate as an opamp.

A low-voltage zener diode, say 3.3v, connected between each driver output and its respective SSR should block the residual DC but still allow the now slightly reduced PWM to switch the SSR correctly.

OK, good to know. The MPC6001 has rail-to-rail outputs and a common-mode input voltage range that exceeds the supply voltage. These would allow it to accept a 0% to 100% duty cycle PWM input range for a full 0v - 5v output. Sure beats a lowly LM324 or LM2904!

Just another thought: if the SMA 10/25V is a diode, then it would remove any input voltage below ground, and the input PWM signal would then be integrated by the 2.5uF cap to +5V rather than to ground. It would make no difference to the opamp, though. It would still buffer with a low output impedance.

I can't immediately identify the IC, but the SMA "diode" has markings that I would normally expect from a capacitor - in this case, 10uf/25v. If it is in fact a capacitor, then the IC appears to be configured as an opamp voltage follower (or buffer). A single opamp stage, e.g. a single sta...

The unexpected response could be due to the fact that you're sending $6D as the first header byte of a physically addressed message, when it should be $6C.