I see the MAX4995 has a minimum supply voltage requirement of 1.7V volts so I will take it out of the circuit.
The TMS570 has some complex power up conditions in voltage and timing. If not correct the chip can go into a constant reseting mode like I was seeing. Something that may need more study. I can see where this would all work fine in a prototype but be on the edge and when in space operating at hot of cold have an issue on startup.
Initially I checked out the supplies with no CPU and all looks fine and constant. I measured them many times with a volt meter but I could not see this reset glitch happening. I finally did look ant the error output pin on the CPU and this was the point I realized there was an issue.
There is about 2 uf of bypass and 22 uf of filter/loop capacitance on the 1.2 volt supply that all has to get up and stabilized. In some systems this would all happen before the power is handed over to the TMS570. The power up delays in the TMS570 may be enough to take care of this if it does not mind some instability while in the delay which is in the order of a millisecond. The TMS570 will not allow the 1.2 volt signal to fall below 0.9 volts typically and above 2.1 volts typically. On a worst case bases this could be squeezed to 1.13 volts min 1.40 volts max.
I took out all the extra resistors I penciled into the JTAG interface. May be able to take out more. I think EMUn needs to be pulled down low not high. I have connected them this way because I was seeing EMU interrupts in the startup error messages earlier. They are not there on the 10 pin version of the JTAG programmer. I will test it early next week but I think it will work fine and be simpler to deal with. I see not startup error messages now when loading code.
Bob