C1253/138 After Brand New Actuator

I recently entered into the Prius world by buying a 2005 with about 198000 miles on it. It came to me with a known bad brake actuator and my first order of business was to get it swapped out. After trying two different used units from junkyards, I gave up and bought a new Toyota unit. I installed it, bled the system, I'm pretty sure I did all the relearns and resets, and all was good. For about 3,000 miles or so.

This is the car I bought for my daughter to drive to college, so she's put most of the miles on it. She told me that after a hard braking event the three brake/ABS/VSC lights came on. I brought it back into the garage and found that it had a C1253 with INF code 138: MTT input is 3.5V or less for at least 1 sec (0.2 sec during initial check) when main relay 1 monitor is 9.5V or more and motor relay 2 is on.

Knowing that things get wonky with old batteries, I put a new battery in it and everything seemed okay again. I couldn't get it to fail on me so I sent it off. It happened again while she was away at school so, since she's a less-experienced driver, I hooked up the trailer and went and fetched it (and her) to get it back into my garage to work on it.

I followed the flowchart:

1. Active test of motor relay 1 and 2. Worked fine - relays clicked, pump ran.
2. Inspect ABS-1 and ABS-2 fuse. Checked resistance all along the fusible link assembly and had 0 Ohms.
3. Inspect brake actuator assy. 6 different resistance tests, and they were all (probably) good. Probably? Well, BM2 to MTT and MB1 to MTT were both 500 Ohms exactly, as opposed to the "about 33 Ohms" that the flowchart calls for. It's a brand new unit from Toyota and I saw another thread or two where others indicated that the 500 Ohm reading was expected. So I moved on to the next test.
4. Check harness and connector between the skid ECU and the brake actuator. Three resistance tests, and all three passed.
5. Use the scanner to measure the voltage from the accumulator pressure sender. This was also within spec

The only thing left after that is "replace skid control ecu". So I went out to the local junkyard this morning and pulled a unit with the same part number off of another '05. After figuring out how to get to all three bolts without removing the steering column, I bought it home, got it installed in my car and... same code showed up.

I fiddled around a bunch with trying to clear the memory, reset the yaw sensor, re-run the linear valve offset learn, to no avail. If I cleared out the codes, they would come back immediately if I ran the manual activation of the ABS MTR relays from my scanner. Eventually, I performed the memory clear, left the battery unplugged for a while, and installed the resistor that came with the new brake actuator. I reconnected the battery, then did the linear valve offset relearn and the yaw sensor relearn. No codes came back, so I then went through and manually activated the MTR relays about a dozen times. Still no codes. I gave it a quite drive for a couple miles, including some hard braking. Still no codes.

I don't know if I did anything or if it's just going to have the same problem again. The only thing that really changed is that I have the resistor replaced now, but I haven't read anything so far that indicates that the resistor would make any difference other than what it sounds like. I suppose it's certainly possible that I grabbed a used skid control ECU that is broken in the same way as my original one, but I haven't seen any way to test one - it seems like if you don't find any other problems, replace that.

Is there anything else that might not be on that chart which could cause this sort of problem? I have no other codes, but could a system that isn't totally bled get that sort of result? I hate to call it "fixed", because it's done this to me before, and I'm very open to suggestions!