Just an educated guess. The microwave is an "unfriendly" load for the generator, with its current draw not directly following the voltage waveform. The circuit breaker reacts to time-averaged current, while there might in reality be very short current spikes many times the breaker value that don't result in a trip. There could be a weak link anywhere in the generator circuitry that can't handle the short current spikes, followed by a protection mechanism detecting it and shutting the generator down. Could also be a problem with the voltage regulator being confused by the unfriendly load, too. Not knowing the Onan internals in any detail, I'd have to say the problem could be with almost anything in the generator part. My first shot in the dark would be to make sure all the electrical connections are tight, clean all of the connectors, and reseat plug-in circuit boards if there are any.
You didn't say what the load for the test was. If was a resistive load like a portable heater, try it with a more unfriendly load like an air conditioner or power converter with a somewhat heavy load. If the waveform still looks good with those loads you have found yourself an above average conventional generator. For comparison, here's a picture of an Onan Microquiet running a fully loaded PD9160A converter (voltage trace on top, current on bottom):
Good chance it's a gummed-up carburetor. In that situation they usually just run rough with the choke off, but I could envision it dying completely with a severe gumming-up. Sometimes running a while with Seafoam in the gas will clear it up, other times the carb has to be taken apart and cleaned. See the first sticky thread in Tech Issues for more info.
I want to second Salvo's comment about the microcomputers. If you don't have the resources to do custom chip and/or circuit board design, this is where the action is. I have a good RV example. I wanted to add a trans temp gauge, which has little or no availability for the Sprinter chassis. The trans has an internal temp sensor with a voltage output proportional to the temperature. The sensor voltage can be converted to a voltage for direct voltmeter temperature readout with a linear equation, a perfect job for a simple analog computer op amp circuit. I tried and failed to get it to work reliably. I think it had to do with needing highly stable resistors, the physical layout, and good shielding. Embarrassed (still) I went to plan B, an ATmega development board for $20 plus a programmer device for $35. Added a precision voltage source for the on-chip A-D converter (a magic chip), an rs232 level converter (for the display, another magic chip), and a bit of software. Works like a champ with none of the analog circuit idiosyncrasies to fight with. Adding a couple of push buttons for mode and brightness control didn't require ripping it up and starting over. For many things, this is the only way to go these days.
I'll jump on the bandwagon of saying there's more to this than meets the eye. The way the GFCI outlets are mass produced as identical devices, I think the odds of one being able to reverse the hot and ground are astronomical. Coupled with the previous observation that the ground goes 'to', rather than 'through' the outlet, the odds are astronomical-squared.
The outlet tester gives a possible interpretation of the pattern of lights, not a definitive answer. With 2 lights lit in the 'hot-ground reversed' pattern, it's indicating voltage between ground-neutral and ground-hot. Another cause of this indication would be 'neutral connected to hot', which I believe would be consistent with your measurements and observations. While I still find it unlikely that the outlet would do this, the odds are IMO somewhat less than astronomical. There are probably other causes of the outlet tester's indication, too.
Ham radio, ~50 Heathkits, minicomputer research assistant in college, EE degree, 40+ years work in electronics, destroyed 10% of broken things attempting to fix them, destroyed 20% of working things trying to modify or enhance them. The lessons learned from the sparks and smoke are the ones best remembered.
Things are different for a younger person now. Devices are so miniaturized that they're close to impossible to work on. You can't look at something to get an idea of how it works. No more looking at a radio or tv and discovering things like the rf section, if section, deflection yoke, high voltage supply, etc. It's all in magic chips now. For the most part, you're a chip designer or you're clueless.
I have 2 of the P4400s and have found both to be surprisingly accurate after cross-checking them with a few other measuring devices including a revenue meter. Many light bulbs and heaters don't come very close to their advertised wattage, plus you have to know what voltage they're rated at to even know what to expect. As an example, while the standard rating voltage is 115, almost all hair dryers are rated at 125 volts. Because of this "nobody" will ever get the rated performance, but the higher wattage looks good in the advertising. The only places I've found the P400 lacking are the frequency and power factor readings with "modified sine wave" power. Even with MSW, the other functions give what looks to be good true RMS readings.
Never mind how today's Cummins's "Onan wannabe" generators are cooled, or oiled, or filtered, or how many cylinders they have ... how RELIABLE are they?Mine broke both valve springs at about 20 hours. There were hundreds, perhaps thousands of KY series generators that suffered the same fate at various Hobbs readings. I was lucky that my damage was limited to the springs. A lot of them sucked a valve and required total replacement.
I say exactly the same thing about my factory installed micro"quiet". Others say they are really quiet and the difference is the installation. When I'm boondocking I carry an eu2000i to charge the battery. For short times like running the microwave I just put up with the noise. It is pathetic IMO that the current built-ins haven't caught up to 10+ year old consumer technology.
I'm fairly confident that the strikeout algorithm is in the thermostat's software, so you're not going to change it. You could hot-wire at the thermostat output to keep it running, but then you'd lose temperature control. Another option would be to find another thermostat, but then you'd lose the gas heat/electric heat integration. You could do it with an additional thermostat, moving the electric heat control wire over to it.
I'm annoyed by the thermostat's insistence of running the gas heat when the spread between the current temp and desired temp gets too large. Sometimes I don't want the dang gas heat running and want to wait for the heat pump to do the job. The workaround of course is to chase the temp setting up as the heat pump runs, a pain in the neck.
As to "why", it's just somebody's bad idea of how it should work.
I'm really surprised to hear the first disc was so big, capacity-wise. The first one I worked with was in the 60s, from Digital Equipment, 32k 12-bit words. When we went to a megabyte platter, that felt like sci-fi. My college prof talked in terms of terabyte capacity. I thought he was a fruitcake, but he got the last laugh.
The manual for the motorhome should specify how much you can tow Do some motorhome manuals actually give useful information? Mine is just a bunch of generic info, less than half of which applies to my specific model. I have to figure out the tow rating myself by subtracting the actual motorhome loaded weight from the chassis GCWR, subject to the maximum in the motorhome's specs (which aren't in the manual).
Everything I write here assumes your RVP/Coleman heat pump is the same as my 15 kbtu RVP/Coleman heat pump. Mine came from the factory with the aux heat strip installed. If the creators had executed the controls differently it would be a good option. As they did it, it's IMO not worth dealing with.
The aux heat strip provides 5,600 btuh, equivalent of a 1,640 watt heater. By comparison, the specs say the heat pump produces 12,700 btuh in standard test conditions with the outside temp 47 degrees.
The control of the aux heat strip is done completely in the roof unit. The thermostat doesn't know it exists. The roof unit has an outside temp sensor that cuts the heat pump off at about 40 degrees. If the heat strip is installed, the power that is cut off from the compressor is routed to the heat strip. With that background, here are the unintended consequences:
1) The roof unit forces the fan to run on high whenever heat is selected. High fan is IMO unnecessarily fast for use with the heat strip and they should be running it on low so the outlet air feels warmer and the noise level is lower.
2) Since there's no way to force the heat strip on, you can't "de-stink" it at the beginning of the season or before a trip. The result is that when you finally hit a cold day and the strip activates, it stinks like heck from not getting used for a long time, and you want to open the windows to get the smell out. Catch-22!
Having too much time on my hands I modified my upper unit, adding a switch to force the heat strip on (for pre-season de-stinking), and at the same time force the fan to low speed. Without this change, I'd never want to use the darn thing.
How do they get 3000-3500 watts from only 5.5 horsepower? Honda and Yamaha barely get 1600-2000 watts from their 5 horse motors. 5.5 horsepower is equal to 4,103 watts. At a 3,000 watt output, that's 73% efficiency. Other generators tend to be in the 50% efficiency ballpark so they've got some great new technology or there's a mistake in the specs. Since the GEN3500i's displacement is 50% larger than the eu2000i and double that of the EF2000is, my first guess is that the horsepower is under-stated in the specs.
i see 0 reason for it not work, im pretty sure the honda will work with any generator inverter or not.
I'll disagree with half of this. The Honda eu-series generators have a very small frequency range they will operate in. They vary their frequency within this range to maintain sync. Sync difference is determined by comparing the phase of the voltage and current output. I can't imagine any small non-inverter genset being able to maintain the frequency range and stability required by the Honda to originate or maintain the synchronization.
Never read a post of connecting different brands. Here's one: Paralleled Honda and Kipor post
There have been a lot of Honda inverter generator "facts" promoted on the forum that have proven to be false. Some offhand examples:
1) Connecting one of the AC output wires to chassis ground will burn out the generator. (i.e. "neutral-ground bond").
2) Can't parallel different wattage eu-series generators.
3) Can't parallel more than 2 generators.
4) Can't parallel generators without additional synchronization apparatus.
There are outstanding assertions that can't be proven or disproven because they need volumes of long-term data we probably won't ever have:
1) Running air conditioner from eu2000i will burn out the air conditioner.
2) Running air conditioner from the eu2000i will burn out the generator.
3) Running a generator near its rated load will shorten its life.
I'm not encouraging experimentation without the full acceptance that the risk is totally with whoever does it. I do encourage being careful to separate opinion from actual experience or established fact.