Leland, I hate to hijack this thread but here's the response to your questions:

I think the radiator has 4 rows.  No overheating yet, even at 100F and no wind.  It's monitored by a Picaxe chip which will shut her down if it does happen, so I don't worry about it.  But I would go with a GEO sized radiator and a shrouded, exhaust powered (suction) airflow if I was to do it all over.  My "House of Lister" is too far from the house to be used as a heat source, so heat exchangers are out for me.

The interior walls of the "House of Lister" are drywall that was laminated with 1 mil aluminum foil before it was installed, and the seams were then foil taped.  I did this because I had a chemical injury in a new sick building and don't tolerate diesel/oil fumes well.  The foil adds about about R-5 or better insulation value, and is totally non-absorbent.  Because it was built late in the fall, I couldn't get epoxy on the floor in time for it to cure well before building, so the floor was covered with galvanized sheet metal.  When I spill oil or fuel or coolant, it cleans up without leaving a stink for me to cope with.  I ran 100% biodiesel until this fall; it was much nicer to be around.  I hope to get a regular supply of it someday.

The generator "frame" is wood, 4x6 frame with 2x6 planks that was a temporary test frame.  It's glued with Gorilla glue and all screwed, painted Hunter Green with ACE spray paint.  It's still doing just fine.  It's mounted to the floor with 3/4" soft rubber pads (McMaster.com) , scored 1/8" with a saw on both sides.  Gorilla glue holds the pads to the floor and frame.   I keep waiting for it to fail but after 2 years (while I've been building my shop and house) it hasn't yet.  I'll add rubber isolated through bolts to the slab (with epoxied-in-slab threaded sleeves) when it does.  The "House of Lister"slab is 8 inches thick under the engine, 5" everywhere else.  Not nearly as nice as an isolated slab, concrete bedded steel frame, but it's been working. 

 My stock machine would vibrate the whole building (a little), even with rubber pads. After balancing with the Mr. X method, it is now a "glass of water stays put on the valve cover" engine.
(The Mr X method is marking the running flywheels and adding clay opposite the mark, leading a few inches, then lead wheel weights.)  For the fine tuning, forget marking, just try moving 4 oz of clay around the flywheels to see where it helps or hurts, using your feet and the glass of water as indicators.  Expect to either add to or be opposite the flywheel counterweight.  My Metro needed over a pound opposite the flywheel counterweight, on each flywheel.

If I was going to do a concrete mounted (bedded) steel frame, I would first balance the Listeroid on a temporary wood frame or put some old carpet layers under the steel frame for balancing, then mount it to the concrete.  The Indians don't control piston and connecting rod weights, so you need to adjust the external counterbalance to better match what you've got.

Bruce M

To clarify for Stan, in an induced draft cooling system, the exhaust is directed inside (and in the direction of cooling air flow)  into the OUTPUT duct of the shrouded radiator.  The exhaust flow then pulls along a bunch of cooling (fresh) air through the radiator.

After Dail's reference to the Rumely Tractor, I remembered that Andy Hall was the guy whocame up with this as an idea for Listeroid cooling, along with the reference to the Rumely Oil Pull tractor.  Thanks Dail.

Lots of ways to skin the cooling cat, and other system design issues tend to make the choice for you. (Such as if you have to run a coolant pump anyway, or are sometimes pumping air with no AC being generated.) 

The other concern for a pressurized system is the O rings around the cylinder liner.  Made me nervous, so I opted for the no pressure approach.  But I know some guys have done fine with modest pressures.

Best Wishes,
Bruce

Brian, the peak volts out of the Harmonic are over 300V for the ST3, so consider a bridge diode rated for at least 600 volts.  No harm in going overboard, I use 800V rated or more.  I prefer the 25 amp integrated units with a metal body- they will conduct the heat better to your new heatsink...which can be homemade out of some aluminum.  They run about $7-8.  In my case, the bridge diode mounts to a steel case with some aluminum channel on the outside for additional cooling; the engine room is a warm place.  Don't forget to use some thermal grease for mounting it, too.  The thermal grease (very thin layer is best) goes between each piece.

Overkill on the bridge diode is the way to go, as you suspected.

Best Wishes,
Bruce M

My ST-3 made noise under load (from hum to low growl), increasing with load, until I replaced the bearings. Now it's quiet.
Bruce M

Neither of my two conrods will fit down through the cylinder sleeve, so I have to put the piston in the cylinder, then lower the whole assembly in place.  It is a two person job, and not fun. 

Bruce M

Cujet,
I agree the RC servo "tweaker" approach is probably the safest, easiest, cheapest solution.  If the spring attachment point was replaced with a long lever, with the spring attached very close to the pivot, then a cheap servo could be used to move the long end of the lever. Consider using a Picaxe chip, then others can use your (Basic) program and download their own Picaxe chip from a PC serial port with no development system cost.  They are very cheap (most under $10).   I have used the Picaxe chips on several projects, and do use one to control my Lister, including an RC servo to flip an air switch which controls my air compressor unloader valves.  My special thanks to MrLister for suggesting the Picaxe chips to me a couple years ago.

The load change related frequency variation is slow enough that any decent AVR will be able to hold a good constant voltage. Most AVRs even respond fast enough to greatly reduce the power stroke speed changes.  I expect this would solve virtually all your equipment problems except for possibly the UPS units which might also be watching frequency.  Most other digital gear won't give a hoot if the frequency of the AC varies, as long as voltage is held withing spec.


Bruce M

Wonderful job Ron! When I have time for some more Listeroid twiddling, I'll be using your idea to check timing. 

I was wondering...did you change your engine timing based on the strobe, or is 20 degrees BTDC spill really where it should be (at 2.5 degrees, really)?