I recently did some high elevation and high load testing. Here is the data and analysis. Setup: NGK CR7HS plug and Wen 55201 magnetic dipstick. 5 hrs of break-in with 10W-40 high-mileage (with ZDDP zinc) mineral oil (not synthetic), alternating 1500 and 3000 watt loads for 10 secs working up over an hour to about 90 seconds, with 30 to 45 second cool-downs between each load. Oil changed after every hour, which went from very dark to fairly clean. Then allowed to cool overnight, completing one thermal cycle. (Note this is a "hard" thermal cycling break-in process to seat the rings, based on engine builders using dynamometers to show what actually works. Good ring seating produces more power with less throttle, saving gas, keeping the oil cleaner, reducing wear, and increasing lifespan of the engine.) MXR3500 sitting on “Air Jade Rubber Anti-Vibration Pads” on motorcycle rack at rear of RV 7500 ft elevation, factory jetting for sea level, so power is down about 23% (-3% per 1000 ft of elevation gain), E10 86 octane. (Note higher octane would be pointless in a low compression engine like this, but E0 (ethanol free) would provide a couple percent more power.) 82F in shade, 86F inside RV, roof vents off, 20 dB on dining table with phone mic to rear Roof air conditioner: Coleman Mach 3+ 8333B SERIES (2006 model year), EER of 7.5 for 1800 watt load: very inefficient by modern A/C standards, but there is no efficiency requirement for mobile roof air conditioners. (Note this is a relatively cool RV indoor temperature. As the RV gets into the mid-90s, I have measured loads on the shoreline of 1900 watts. I would guess it would hit 2000+ watts in AZ 110F temps. In 2019 I ran the unit on a campground shoreline when the local weather said 115F, but did not make any load measurements. No shade. I ran it full blast for about 18 hrs per day.) 4p: Hour meter at 7.1 hrs, gen in sun, started easy w/o choke, Eco OFF Loosened ratchet strap around gen, did not seem to make much vibration difference. The pads seem good. Now 40 dB on dining table, so up 20 dB with generator on Hour meter at 7.2 hr, so 6 min warmup done 410p: ActiveStart app connected to soft start unit: Configured unit for a 3000W generator Coleman turned to LOW, middle temp setting on the dial 412p ActiveStart app: started out around 1400W, now pulling 1600W as the compressor pressure rises 51 dB on table, so up another 11 dB with A/C now on. Up 31 dB from ambient now. 417p 1700 watts 423p ActiveStart app: mid 1700s, 15A, very steady at 120 volts, 60 hz (Not sure what the resolution of the ActiveStart soft start device is, but the generator seems steady.) 430p Low 1700s, saw 1785 watts briefly 430p Turned on fridge A/C heating element to add more load. Now around 2050 watts. Sitting on the sofa, I can hear the generator as a slight background rumble “underneath” the roaring air conditioner. This is significantly quieter than the Onan 4000 running under the floor. In 2022, with another phone, I measured the Onan alone at 57 dB, then 68 dB with the same AC on low and the compressor on. So if these numbers are comparable, the MXR3500 on the rack is 17 dB quieter than the Onan and 17 dB quieter than Onan plus A/C. Either way, this is a lot of noise in a small space. Sitting at the table right under the A/C discharge, I have found that Bose 700 headphones on maximum noise cancellation remove almost all generator and A/C noise. So somewhere around 30 dB is eliminated. Expensive but I can do things in comfort for hours. Ear plugs also work and are cheap. Outside 12 ft pointing directly at exhaust: 60 dB 15 ft from front of RV at the entrance to campsite: 44 dB About 50 ft from front of RV = "neighbor noise": 40 dB The MXR3500 definitely has a louder exhaust buzz than Honda EU's. There is probably lots of air induction noise also, because the EU's have a sophisticated "organ pipe" silencer before the air filter. Honda probably spent hundreds of hours perfecting this. But for now the MXR3500 is quite satisfactory for 21% of the cost of an EU3000i. When I am convinced the EU3000i is at a Honda level of perfection in all weather conditions, then I will buy one. Looking at the Spectrum analyzer (Spectroid app), there is a very steady peak at about 50 hertz (we can hear down to around 20 hz). This might be a harmonic created by other frequencies. Then there is another peak around 60 hertz which is probably the inverter. Then there is a constantly churning zone centered around 120 Hz that might be the exhaust and induction noise. 503p Inside temp is now about 83F. A/C has dropped it 3 degrees, but RV is in the sun, and is poorly insulated. Turned off the fridge load. A/C switched to HIGH cool, max cold setting on the dial Now 52 dB so HIGH is not much louder even with the compressor on. It is pulling about 1750 watts, because the inside temp is 3 degrees lower and the A/C already has pressure built up. Starting right away on HIGH is a much higher load, as noted above. 520p all loads off, turn on Eco for cool-down 36 dB on the table. Sitting on the couch, the generator now seems quite loud without the white noise of the A/C unit. 530p Generator off 20 dB on table: quiet background The MXR3500 used about 53 US fluid ounces (0.41 US gallon) over 1.5 hrs. So that is about 0.28 gallons per hour for an average load of maybe 1700 watts. Using an inverter and running at high revs (4850 rpm), it consumed about half the gas of my ancient Onan 4KW. At $3/gallon, that is 85 cents per hour. I should have put on the Kill-A-Watt meter, but let’s guess the MXR3500 made 2500 Watt-Hours of energy (1700 watts * 1.5 hrs). That would be an efficiency of around 50 Watt-hours per ounce of gas (2500 / 53). So if the fuel tank actually holds 4.7 Liters that is 1.24 US gallons. At 0.28 gallons per hour (as calculated above), it should run over 4 hrs. There are rumors of an auxiliary fuel tank from the vendor. The fuel cap is much larger than the Honda EU series, which most 3rd party auxiliary tanks are sold for. So far I like the MXR3500. For a product with a cost at the factory door of perhaps US$150, it is pretty impressive. First I will shoot for 500 hrs. If it can last for 2000 hrs, I calculate it may achieve the same cost per hour of an EU3000i: around 25 to 30 cents per hour, not including fuel of course.