Well, gents, I have made a couple more trial runs, and found some interesting things.

Yesterday, 10/28/2010, I visited a local repair shop to get another oil filler cap, so I can destroy one and still have a cap in case I need the generator. While there I chatted with the proprietor (an old acquaintance), who told me of measuring oil temps on his old BMW 1000 boxer. He said there was available an oil thermometer which replaced the OEM dipstick, and on which he observed oil temperatures approaching 300 F. He mentioned Klotz, saying it got just as hot, but seemed to give him perhaps one or two mpg improvement. He also said Klotz was the only oil he ever used which made his BMW leak, and when he went back to petroleum oil the leak went away.

Based on Verslagen's IR temps for oil, I bought a meat thermometer with a high number of 220 F. I drilled the oil filler plug to a press-fit for the stem of the thermometer, and put the whole works on with the thermometer stem down in the engine oil. I also placed a piece of tape on the generator housing and made a pointer which I attached to the throttle butterfly. I marked throttle wide-open as 100%; throttle closed (to the idle stop) as 0%, and roughly divided the rest of the tape in 10% increments. Bear in mind this is a governed engine with no user throttle control, which is designed to produce somewhere near 60 Hz power, 3600 rpm.

My idea was to start the engine, and as soon as it came up to speed plug both my electric heaters into it and start the stopwatch. When I plugged in the heaters, the throttle solidly pinned against 100% and the engine bogged a little. I didn't like that much, so removed one heater for a moment. Then I figured to see what would happen, since the generator has an overload breaker. I started the stopwatch and plugged the heater back in. At start, the 10W-30 oil temp was 64. When the oil temperature reached 120, in about three minutes, the throttle began to drop. At five minutes the oil was 198 and the throttle had dropped to about 75%. I walked away for a few minutes. When I returned at twelve minutes, the oil temperature was off-scale, at an estimated 240, and the throttle had dropped to about 70%. At 17 minutes the oil was at an estimated 270 and the throttle was about the same. When the engine ran out of fuel at 29 min 25 sec the oil was perhaps 280. Temps above 220 are estimates, since the pointer was off-scale.

Conjectures so far. The oil viscosity decreases as the engine warms, and this is quite significant in terms of throttle position with a fixed load. In fact, it is a heck of a lot more significant under my test conditions than I would ever have guessed. Without casting doubt on Verslagen's measurements, either his Savage engine is cooled a heck of a lot better than the forced-air cooling on my Briggs, or the Briggs is working a lot harder than the Savage. Be it noted the Savage holds about 2.4 quarts while the Briggs holds 20 ounces.

I did some more experimenting today, but I am still digesting results.

A motorcycle going down the road isn't under a constant  load like your generator is,Unless you were going up a 10 mile hill or pulling a heavy trailer.

My experiments yesterday, 10/29/2010, revealed some interesting data, although I didn't run the full tank of fuel. First, I determined my original voltage measurements were in error by about five volts. I had used an old analog meter; this time I used a digital and recalibrated the analog. The virtue of the analog meter is its AC plug, making connection easier.

I started by measuring the actual loads of my electric heaters. The one marked 1000 watts is actually closer to 750, depending of course on input voltage. The other was about 1050, but its rating label long since fell off. I did this by measuring current and voltage, then multiplying. For resistive loads this is OK with AC, although strictly speaking I should have called the result VA instead of watts.

On initial engine start, the 10W-30 oil was 70 F; the throttle settled at 38% with no load. Of course, the engine always has the load of its cooling fan and whatever windage load exists in the generator. Output  was 129 volts. I plugged in the larger heater; throttle went to 67%; voltage 116; current 9.4 A; wattage 1090. I unplugged the big heater and plugged in the smaller one. Throttle 45%; voltage 120; current 6.8; wattage 816. I plugged both in; throttle 95%; voltage 105; current 15.1; wattage 1586. This didn't last long, because the power strip into which I plugged the heaters has a circuit breaker that didn't like that much load.

I spent some time fooling around with the different heaters and the power strip. By the time eight minutes had passed the oil temp was over 200 F. A few minutes later, when the oil had reached 225 (est) I did some more checking. With no heaters, throttle was 20%.Small heater, throttle 38%. Large heater, throttle 45%. Both heaters, throttle 70%. Note particularly the differences in throttle position between cold and warm oil. No load cold, 38%; warm 20%. Full load cold, 95%; warm 70%.

I unloaded the generator at about 15 minutes, just to see whether the oil temperature would drop. It did, but quite slowly. I then shut off the unit and called it a day.

At this point I conclude I need to get a thermometer with a higher range to get better numbers. I also conclude the varying viscosity of the oil as it warms causes far more change in the engine loading than I would have imagined. You folks who noticed the idle speed changing as your motorcycles warmed up are not imagining it. Frankly, I am not sure whether my experiment as originally formulated will produce anything worthwhile, but I plan to keep playing with it. I have already discovered information I find interesting.