Designing a better Gamma chamber

[VincentG]

The displacer is shown here after 5 minutes of on and off heating with a MAP gas torch. The cold end was only 10F over ambient temperature with no active cooling other than my ceiling fan being on. The Tmax reached by the hot side was over 800F at times.

Gamma v1.00 initial displacer test high temp.jpg (228.4 KiB) Viewed 6847 times

[VincentG]

After testing several engines, including the Essex, it is clear that the displacer temperature itself plays a most crucial role in the operation of these engines. The standard metal displacer becomes quickly heat soaked and greatly hinders the running of the engine by counteracting the cold side. So much so that in the Essex, once heat soaked, cooling the cold end with water has little effect on its running. The only way to restore full power is to let the whole engine cool down and start again.

Imo this is why Tom and others have had good success with "regenerator" displacers. Not from regeneration, but from preventing the displacer from heat soak as a solid body displacer would. The unfortunate consequence of the, usually ss wool displacer, is a reduction in effective compression ratio.

So keeping the displacer at two distinct temperatures should be greatly beneficial.........so long as there is a way to block (and control) convection between either side of the chamber.

[VincentG]

https://web.archive.org/web/20071017030 ... rling.html

This is a good read with some excellent thoughts on design considerations for hot air engines.

[VincentG]

The chamber has finally been assembled for initial rough testing. No sense spending more time on this if it's a non-starter. Fortunately, it was extremely encouraging. I'll take more pictures of the individual parts at a later date, but for now know that the original goal of no welding or special tools is maintained. The metal assembly rings were ordered from an online laser cutting service.

The chamber will have adjustable displacement, but for now it sits around 65cc swept volume. I have not yet fashioned a displacer guide rod bushing so no pressure and volume tests for now, but it moves a lot of air! I expect it to expand more than 1 to 1, or in other words, it should push the piston out more than 65cc. This so far has been a personal benchmark I would like to reach with these chambers.

Most importantly for now, the thermal performance is excellent. Even with the chamber only partially submerged in water(this is actually the hot end but serves as the cold end for now) the chamber can maintain its full temperature differential sustainably. I expect actual results to be near PV=nRT, similar my little epoxy test chamber, but even at Tmax well over 600k.

Gamma v1.00 initial testing.jpg (200.24 KiB) Viewed 4683 times

An observation I did make here and also with my 150cc pancake chamber, is that the hot expanding air serves to drive the displacer upwards and minimize energy lost to lifting it. This only works when the hot side is on the bottom of course. With the addition of a larger displacer shaft and the subsequent ringbom effect, as well as the the displacer falling back down from gravity, the displacer may even contribute to net shaft power if this is gamed well.

[VincentG]

After making the displacer rod guide bushing, I had some disappointing results, only to remember that the displacer itself was never sealed to be airtight. In fact it leaks like a sieve, and it's a wonder this thing works at all. Even at that, I recorded a contraction of 30cc. The response time of this chamber seems incredibly rapid, owing to its extreme surface area to displacement ratio.

The thermal mass of the thin stainless steel cups is so low that it makes bench testing difficult, as temperature is lost quickly to open air cooling. I will have to fashion a proper propane burner and water jacket for meaningful testing.

[VincentG]

I’ve sealed the displacer a bit better but it still leaks badly. Looks like I can update the design to ditch the threaded rod and subsequently any holes that then need sealing.

Even still, I saw 40cc of expansion now, compared to just under 30 before. Contraction is always a bit more than expansion, surely due to the expanding air losing temperature in the syringe.

I’ll rebuild the displacer and test again, I have a hunch this chamber will easily expand over 65cc once airtight.

[Fool]

Excellent.

[VincentG]

So keeping the displacer at two distinct temperatures should be greatly beneficial.........so long as there is a way to block (and control) convection between either side of the chamber.

On the subject of convection, I had briefly tested the displacer without the internal convection barrier. Heat moving to the cold end was significantly increased. No surprise really but it was more than expected.

One thing I do notice with this chamber, response time is so rapid that it would really benefit from solenoid control at the expected sub 1000rpm operation speed. Any extra transit time of the displacer is just wasting heat into the cold end.

Any recommendations for a fast acting 12vdc solenoid with roughly .5” stroke?

[Fool]

Car starter solenoid?

Two car starter solenoids?

Modify a relay, or two?

Push pull?

[VincentG]

Good point, I could make something like that work, but I’d prefer an off the shelf solution. I’ll do some research on that. Push pull would be ideal but push only with spring return will suffice.

Solenoid control is appealing also because during periods of maximum load, the mechanical loss of the displacer can be removed completely.

[VincentG]

The displacer has been better sealed, and the silicone and rubber gaskets were removed altogether for an unrelated issue at the moment. The displacer is airtight, but the chamber still leaks a bit here and there. 40cc expansion was again recorded, with 65cc displacement. Expansion is rapid enough to cause pressure in overcoming the inertia of the piston and so air leaks out before it can act upon it.

But an interesting observation was made that has been talked about here and there by Tom. Does the buoyancy of air play a role inside the chamber?

Well now I can definitively say yes. When the chamber is held with the hot end on the bottom, the air expelled from the centrally located port(in the epoxy ring) is significantly hotter than if the hot end is help at the top. This surely will lead to more real-world power and efficiency.

[VincentG]

This chamber was designed to be directly inserted into a wood stove and contact the hot gas. But for other applications, an efficient way to apply heat should be a high temperature liquid in a highly insulated container. Any suggestions for a readily available liquid that can handle 700 degrees F or so?

[Fool]

Sulfer.

[matt brown]

Matt, humor me and run a graphic of "hot pp essex" with 400cc pp and 200cc dp 300-600k. But adjust starting charge pressure according to planet Bob.

So starting charge pressure at 200cc @300k would be roughly 19.5psi.

And or run a 400cc pp and 400cc dp 300-900k cycle where in this case starting charge pressure is only 11psi at 400cc and 300k.

While messing with Essex variations, I came up with this comparison which is amusing...

Essex options.png (19.58 KiB) Viewed 4019 times

Consider 3 Essex where all have the same 100cc DP and 200cc PP.

Essex A has 300-600k isothermal helium cycle similar cold gamma Stirling where regen is paramount

Essex B has 300-600k adiabatic helium Green cycle where a 1:6 expansion pressure ratio coincides with a 2:1 thermal ratio (and 1:3 volume ratio) whereby ideally, regen is not req'd, but added here to compensate for unideal input (short) during expansion.

Essex C has 300-450k adiabatic air Green cycle where a 1:4.5 expansion pressure ratio coincides with a 1:1.5 thermal ratio (and 1:3 volume ratio) whereby, like Essex B, regen is added to compensate for unideal input during expansion.

Note Essex B and C have (volume ratio) x (thermal ratio) = pressure ratio, but that the interplay between these values are constrained by whether the gas is monatomic (helium) or diatomic (air). B and C make these values appear linear, but they're not...air would require a 1:6 volume expansion for it's adiabatic expansion to cover 1:2 thermal ratio AND this would entail a 12x pressure swing during expansion.

Now consider C mod with 1/3 more charge pressure then C, whereby C mod has equal Pmax as B, yet C mod has more output then B (C mod has P=1.33 bar after expansion). So, this C mod will produce more power from the same size DP and PP, but from only 450k input vs B 600k input.

[matt brown]

Now consider C mod with 1/3 more charge pressure then C, whereby C mod has equal Pmax as B, yet C mod has more output then B (C mod has P=1.33 bar after expansion). So, this C mod will produce more power from the same size DP and PP, but from only 450k input vs B 600k input.

I botched this last part while trying to avoid timeout. The last line appears as if C mod cycle produces more power than B cycle, but I meant C mod expansion process produces more power than B expansion process. To clarify, here's a graphic...

Essex eff update.png (23.42 KiB) Viewed 4000 times

Essex D is previous "C mod" wherein D 3-4 expansion output does indeed exceed B 3-4 expansion output. However, D 1-2 compression input also exceeds B 1-2 compression input. Despite throwing around 6 bar DIY values, the backwork ratio is still relatively high for all 3 cycles, whereby each cycle efficiency is fairly low. The bottom line here is that the 300-600k B Green helium cycle has ideal eff= ~.30 whereas a similar Stirling has ideal eff=.50 (Carnot) thereby ideal B is .60 Carnot. Meanwhile, both 300-450k C and D Green air cycles have ideal eff=~.18 which is only .55 Carnot.

Gaming isothermal vs adiabatic processes into cycles can occupy endless hours before adding regen to the mix. However, there's no other way to scheme this stuff...