APU No Start

As a note.... this problem has been solved, but we had three frustrating days previous to the fix.

A recently replaced (supposedly overhauled APU) would not accelerate above 30%. Between days On/Off and shift changes, this APU had 50+ attempted starts over 4 days. EGT temps during start are around 750C. This drops quickly to around 350C-400C after a completed start. All of our starts failed. The faulty APU went from 750C to ambient every failed start. This type of thermal shock is not conducive to long term service of the unit.

A list of avenues traveled while troubleshooting this issue.....

All wiring. Continuity, megger , and load test .
Replaced APU battery.
Replaced starter.
Replaced MEC (Main Engine Control). )
Replaced IGV actuator.
Replaced generator. (Thinking the gen was loading down the APU.)

The ECU (Engine Control Unit) was not throwing any fault lights and there were no mini-flag indications either.

We faked the inlet door open signal. We eliminated an air input issue by removing an inspection plate at the APU plenum forward of the firewall and tried a start.
  
     
     
  
We disconnected the fuel line to the MEC and puked fuel all over the place with both the APU Start Pump and Main Engine Pump.

We verified the IGV Actuator operation by pulling it out before start. (It should close during start as soon as fuel pressure from MEC is available.) Shown in video below.

The battery was replaced  because of the thought that it didn't have enough balls to push the starter. It would drop to 17VDC and climb from there. (This is about normal for most APU starts.) Shown in video below.

We also verified the Surge Valve was remaining open though-out the entire start. (I initially thought it should close, keeping all air for starting - I was wrong.)
  
     
     
  
The Surge Valve should remain open. This relieves all compressor loads during start.

I even disconnected the plug from the MEC. The ECU sends a torque motor signal to the MEC during start. The APU made 18%. (This proved the torque motor signal wiring was good from the ECU to MEC as 33% could not be attained if this signal circuit was open.)

All of the mechanics involved were completely out of ideas. A decision was made to try another APU..... it started on the first attempt!!!  All that work (and parts) for nothing.