I don't understand what I've done wrong

After more than a year of day-in, day-out operation of an HDC2450 in a CANbus system on my daughter's wheelchair, the chair started badly misbehaving even though there have been no circuit changes and no script changes in many months.

At first, it was just sometimes sluggish starting out.
Then it started going backwards no matter which way the joystick was moved, and going backwards at full power when the joystick was released.

I removed that entire system (joystick module, power distribution module, display module, HDC2450) from the chair and mounted my spare set. The mis-behaving unit was set up with dummy loads on my bench and I tried to find the fault, or damage. It appears that both Allegro motor current sensors have been damaged. When first tested, I saw ca. 1300 mV instead of 2500, and they are now giving only about 60 mV output - i.e. extremely strong reverse direction current which IR motor compensation then uses to bias throttle in that same direction.

The spare unit behaved OK for a week or so, and then it started to do the same thing - first just not accelerating well in the forward direction, then always going backwards.

The two sets of modules are not identical, the power distribution module in particular used automotive relays for seat lift and tilt in one, and STM VNH3SP30-E integrated motor drivers in the other. I have not been able to find any changes (shorts, opens) in board layouts or wiring, but the Roboteq 5V pin is not at 5V IF Dout 7 IS CONNECTED TO THE LEFT BRAKE DUMMY LOAD - it is at battery input voltage, or just a volt or two less, dropping sharply if motor current is flowing. If that brake load is disconnected, the 5V line is indeed at nearly 5V and rock steady. Having the right brake dummy load connected or disconnected makes no difference whatsoever.

Clearly I have done something VERY wrong recently, but haven't a clue what it was. The chair is dangerously unusable and I am completely baffled. I am attaching the configuration profile in case that might give you a clue. Any suggestions will be most welcome.

Sincerely,
Lenny Robbins

Clearly the voltage should never be anything other than 5V, but what in the world happened nearly simultaneously after over a year's use to TWO controllers.

BTW, the 5V pin goes to B+ if brake 1 (or a 2.2k dummy load) is connected, EVEN IF there is no connection at all to the 5V pin in the DB9. How could presence of a 24V pull up at Dout7 translate to an overvoltage at the output of a 5V regulator?

I assume that "RCB daughter card" refers to the smaller board sitting on top of the larger, high current board. I could certainly pull one of them and send it to you for diagnosis. As important as it is to know what failed, it is equally or more important to know why it failed.

Do you see anything problematic in the configuration profile?

Thanks,
Lenny

Dout 7 was connected, in the chair, to a 60 ohm motor parking brake and on the bench to a 2k2 resistor. The board has a Vishay SMBJ28CA-E3/52 600W 28V bidirectional TVS between Dout7 and ground.

Please send me instructions for shipping a complete unit to you.

Thanks for your prompt replies,
Lenny

Will do.