I think I've solved the GSU power up problem after some rather careful testing. In all these tests, the GSU is plugged directly into the ECU as standard.

The first test was to separate the signal wire from the GSU and plug it in with just power (no data). The result is the same, when using the plane's power system the display is dark until you unplug it, then plug it back in. Obviously, with the data wire attached, it doesn't work on power up as that is the default setup. Unplug and plug it back in and it works fine.

The 2nd test is to just measure the voltage at various points from the battery to the receiver. It's all the same, around 6.5V with maybe .1 volt loss. Not the cause.

The 3rd test is to just plug the GSU into battery separately, then it works (no data but it lights up).

The 4rd test is to run a wire directly from the battery to the two receivers, and the ECU via throttle. Bypass all the wire runs in case resistance is causing brownout. My "switch" is just plugging it into the receiver fast. THE GSU now works immediately when everything is powered up. So, what is the difference. Switches and a voltage monitor. It dawned on me.

I use the Fromeco Cricket DC-Up power monitor. It has a bank of LEDs that tell you how low the voltage got during a flight. It also has a brownout protection built in, in the form of caps to smooth out voltage drops from heavy servo action, etc. For this to work, it has to charge the capacitors (you can see them in my picture), making the overall voltage come up more slowly than it would When you power up, and especially down, you can see the LED banks slowly climb and fall to the real voltage.

The 5th test was to turn the power off, and on (standard wiring setup) repeatedly, allowing the voltage to fall a bit more each time until the GSU won't boot on powerup. So guess what, there IS a certain low voltage level caused by the Cricket's capacitor array charge state, below which the GSU won't power up with the GSU.

So, this is why, in the original problem, just unplugging then plugging the GSU fixed it (made it light up). The caps were already charged on the second plugging.

So why go thru all this when it's easy to just wait to plug in the GSU? Well, the problem must be tracked down because the alternate to this cause could have been brownout to the receivers, causing a crash. It needed to be understood.

I do think maybe there is circuitry inside the GSU that is making it sensitive to slow power rise, that shouldn't be happening. There should be an input FET that waits to pass power to the rest of the screen until full voltage is attained. it's not a huge problem, but other customers using CAP arrays could waste some time on this, or stress out.