Hi Buddy,
Expensive is a relative term. If your competition is doing the same for less cost, your method just got expensive, no matter how efficient it might otherwise appear.
Machining a crankcase from billet is going to be expensive when the other guy gets his cast enbloc engine from a foundry. He no longer needs to hog the whole thing out, bore holes and so on. His machining is now just clean up...and he only paid for the metal that was actually poured plus the cleanup chips. Machining from billet produces a lot of waste. This doesn't address the other costs. Well designed engine castings come from the foundry with motor mounts, auxiliary hardware brackets, coolant or lubricant passageways, skeletonizing to reduce weight and all sorts of other things built right in. If you start out with round stock and turn it, you can't even get those shapes and have to add a number of braces, brackets, fittings and such. Besides adding cost, all that stuff is more problematic in operation. I've had older cars with bracketry and newer ones with all the basic fittings cast in, the cast in components never vibrated lose or wiggled out of alignment.
The three cylinders in the WHE appear to be there for two reasons: to provide a self-starting capability and to preserve the 'trademark' Cyclone radial engine layout. Realistically, there have been any number of small diesel engines producing the same horsepower that have only a single cylinder. The self starting is nice, but probably not really relevant. Small electric starters are dirt cheap and all the rest of the competition use them, there is no competitive penalty to simplifying the design and sticking in a starter.
Thee are two simple ways to build cranks. One, you can use a built up crank that is made out of separate pieces all fitted together with accurately machined tapered joints on the mains and pins. Hardly radical, most two stroke engines with needle bearings go this route. The other is to NOT give the work to a job shop and go to people who are in the crank business. There are NC machines with conversational programming just for cranks as well as specialized crank grinders and polishers. Any billet crank is going to be expensive, but if you cast it out of high grade nodular (good enough for millions of Chevy small blocks from '55 until recently) or forge it first, the cost does come down. Reducing that sucker to two throws will also make a huge difference.
Let's face it though, gearing up to build 300 engines over 3 years, and of three different generations at that, sure isn't going to do much for profits.
The problem with dyno runs is that there are two different types. The first is what you described, a few minute run to measure power output and maybe economy, although economy is probably a poor fit because that usually improves as any powertrain heats up. That type of testing is what Cyclone appears to have been doing for years. I'll even grant them the possiblity that they achieved 1000 plus ft-lbs. Of couse, I've seen V-8s pulling well in excess of 2,000 horsepower as well. All very impressive, but no one is sticking a 100,000 mile warranty on it. That is exactly the type of testing Cyclone does not appear to have done, stick it on the dyno and run it 'til it breaks. Of course, you have to do that over and over if you want to build an engine you can sell and stand behind. If Cyclone really had a product for all these years, they would have taken a gen set to a public place and run it for a week or month while attached to a load just to prove they had the real goods. Instead they claimed they had a game changing product, but couldn't demonstrate it for whatever cheesy reason.
You're right, much of the stuff describes their heat regenerative engines. These are supposedly run on supercritical steam for maximum power density and economy. However, most of the engineering that has been made public looks highly questionable....like the spider bearing. The only reason for the very high torque specified is to overcome a very high starting load. Steam engines typically admit steam at the very top of the stroke and then cut off steam admission, forcing the small amount of steam trapped in the cylinder to fully expand, giving up its energy in the process. Naturally, this doesn't produce unusual amounts of torque. You get that by using 'long cutoff', admitting steam through most or all the stroke. Yes, you get enormous torque, but you use a lot of steam to do it....and every little bit of steam reflects the cost of the fuel burned to vaporize it. It is unlikely that they can develop this torque at high rpm, not for any period of time, the steam generator coils don't look anything like large enough to produce the kinds of steam needed. For that matter, the engine itself looks too light to handle that torque for very long. So the benefit may be less valuable than it appears.
I got into selling bridges after I realized there were too many competitors in alternate energy schemes...