Engine time. Nice color match, right?
The generator head is to be coupled to a 690 cc Honda v-twin engine, which produces up to 22 hp at full speed (3600 rpm). I'll be running at around 3000 rpm, before the torque falls off too much, and where I can still expect decent BSFCs. All that stuff is to be tested later.
The engine and generator head are joined using a flexible shaft coupling assembly, two steel pieces with a Hytrel solid spider:
To my surprise, there wasn't any need to raise one device or the other to get the right vertical shaft alignment. My buddy Chris and I worked together late one night to mount the two pieces together on a frame of strut channel and steel plate. While cutting some of that plate (half inch thickness), Chris ran his DeWalt Sawzall out of juice, so we hooked it up to a car battery to finish the job.
I mentioned it before, but the garage is exceedingly dusty with this ridiculously thick, black dust everywhere. Yukon tried to sleep on top of stuff, but the yellow Labrador gradually turned into a black lab by the end of the night.
We got the genset mounted on the frame and I was able to start it up later on.
This past week I redid the frame slightly to help with shaft alignment as well as add a bit of strength to the bracing. The engine sounds a lot better with a muffler attached as well.
The three power phases from the generator go to my 6-diode rectifier, which feeds DC to the caps.
The generator is self-exciting, and after wiring it up, I could measure open circuit phase-to-phase voltage as the engine idled. Here, the generator is in an "underspeed" condition, where the regulator cuts the excitation current to drop the voltage. Speeding the engine up would clear the error and provide 240 VAC phase-to-phase.
My hope was that the engine and generator would be able to handle the short-circuit type of condition that exists when the caps are mostly dead, but that was not the case. The generator's regulator was constantly demanding 266 VAC delivered at precisely 60 Hz, and the engine couldn't keep up. Faults kept tripping, and it was clear that some kind of external control would be required. Here is the short version of what's been done so far.
First, I eliminated the route for the regulator to "talk to" the exciter coil by removing two quick disconnect leads. Here, the blue wire is the exciter negative, and the yellow wire is the exciter positive.
As I mentioned to fechter in a previous post, the exciter is supposed to be run between 0.3 A (no load) up to 1.1 A (full load) excitation current. I used a current-controlled DC power supply to vary the field on my own.
To maintain a consistent engine speed, I've got a knob-type throttle installed now. The tachometer uses an optical sensor, with a patch of reflective tape placed on the shaft coupling. Idle speed is around 1850 - 1950 rpm.
This morning, Chris adjusted the engine speed while I supplied a few different excitation currents to the winding and wrote down data, while everything was at open circuit. The voltages were measured on the DC side of the rectifier, so this is not the true open-circuit delivered peak voltage. However, it was a good guideline.
So, using an excitation of 0.35 A at 3000 rpm, we charged the capacitor bank until it settled at 362 V. It turned out that the true peak voltage was somewhat lower than what appeared during our test, so we had to boost the excitation to about 0.45 A in order to top the bank off at the desired 380 V.
While I can certainly do the work related to figuring out the optimum engine speed for cruising, I do NOT want to be playing around with excitation currents manually to maintain a constant current-type charging of the capacitor. One option was to connect a current-controlling SCR, like you would use for a three phase heater, to the generator outputs.
I am positive this would have worked, but it would take up a bunch of space, would require cooling, and would cost about 3000 USD. First I'm going to try a cheaper method using a brushed DC Kelly motor controller at 12 V:
Here, a current transducer is placed on one of the incoming generator phases before the rectifier. Depending on the current it detects, it sends a 0-5 V signal (0 for no current, 5 for my choice of 10, 20 or 50 phase amps) to the Kelly Controller's throttle input. The throttle input is most commonly arranged for 0-5 V, that is, 0 V input causes zero throttle/current demand, and 5 V is full throttle/current demand. However, it may also be configured to 5-0 V input. Then, when the phase current is the prescribed maximum, it will be at +5 V, and asking for no "throttle" (zero current to exciter winding). If the current drops off, there will be more current sent to the winding, increasing the delivered voltage and therefore raising charging current. The Kellys are great in that they may be programmed for a current limit, which I'll set to about 0.5 A at 12 V. That should prevent overexcitation. Thanks, by the way, to Jeremy Harris for discussing this with me over a PM.
That system will have to be tried out when all the parts arrive. For now, things are coming together.