Hi Ray, Thanks for the kind words. I haven't done a tear-down, or even seen one. But I have an analysis of the PIP-4048GE based on the eBay ads
and the manual
1. Yes, if they are 60 cell panels (as 250 W panels would usually be), you will need to put all ten in series for this inverter.
2. It is not obvious that its battery-side MOSFETs or capacitors would be different from those in the PIP-4048MS (where they are a known weakness). But I note that they have slightly improved the voltage rating of these parts in recent MS deliveries (but not the lifetime or impedance of the capacitors).
3. We do not have access to firmware for the GE and so cannot fix the charging bug, if it still exists.
4. The array voltage is lethal and yet they do not have any insulation monitoring (if designed to have a floating PV array) or any earth-leakage/residual-current detection (if not). And it is unclear which of those two types of shock protection will work (if any).
5. It appears to not have any cover over its 230 Vac terminals.
6. Its AC output cannot be paralleled with others of its kind. Although the manual makes no mention of this either way, the old ad linked above states in plain chinglish: "**Note: this PIP4048GE model is NOT with Parallel able design."
7. Its maximum total
charge current is limited to 80 A (unlike the 140 A of an MS) despite the fact that it can separately charge at 60 A from the AC, and 80 A from the solar, just like an MS. This suggests that the GE's SCC (Solar Charge Controller) connects to the 400 V side of the DC-DC converter stage (instead of the battery side as it does in the MS), and so the total battery charge current is limited by what can be pushed backwards through the DC-DC converter stage.
Further to 4 above: If the array is floating, it could be made safe by an insulation monitor such as the Bender isoPV425, or some other isoPV models shown on this page
However, if I'm right in 7 above, then it will not work with a Bender unit, because the PV array is not floating, nor can it be earthed, because it would connect to the DC bus that the 230 Vac output sine wave is produced from, with no intervening transformer. And it seems unlikely there would be an isolating transformer in the SCC itself. So the PV array will have a common-mode AC voltage relative to earth, possibly at 30 kHz or so. This is the same as what happens with a transformerless grid-feed inverter. This will not work with any of those bender units because they assume an array that is floating with respect to earth (or functionally-earthed via a high-valued resistor).
If I'm right, you would instead need to use a Type-B RCD (Residual Current Device) on the AC output, to detect earth leakage from the array. Type-B RCDs can detect a DC component in the AC signal, where an ordinary RCD can't.
Either way, to know whether you were protected, you would need a means of testing, by deliberately applying some leakage from each of the array terminals to earth. The protection devices should have a "Test" button that does this job. Failing that, three 10 kΩ 5 watt resistors connected in series would probably do it. But you'd be dealing with deadly voltages, so you should have some training in electrical work.
If anyone already has a PIP-4048GE, and the required training in electrical work, and wants to tell us whether the array is floating or not, or you want to find out before you go and buy a Bender or a type-B RCD: You can use those same resistors to soft-earth one side of the array and then measure the AC voltage across the resistors. If it's floating, the AC voltage should not be more than a few volts. If it's transformerless then there will be more than 100 V of AC. For these tests, the PIP itself needs to be earthed, as it normally would be. And its neutral output also needs to be connected to earth, as it normally would be via the MEN link in your switchboard, if it was powering household loads.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).