metal in the rocket mass heater core

Erica Wisner wrote:...we do know a bit about Peter's work from the ProBoards forums. It is my understanding that he's skipped over the masonry designs and done most of his testing on all-metal prototypes. We wish he would try a masonry version, since he has some very cool toys for testing and we'd love to see similar numbers on our preferred designs.

The most recent prototype happened to utilize a cast refractory burn tunnel. Only the top end of the feed tube is made out of metal because it could be cooled more effectively. The lower end of the riser isn't made out of metal as well. See the picture from 2011 below. When this is mentioned already somewhere else, I do apologize.

Peter,
Thanks so much for posting this picture to clarify where you've been using metal in current designs.

I should probably apologize for quoting vague impressions. I was remembering conversations about a number of experiments that were all happening around the same time - a testing proposal involving a water jacket to quantify heat transfer, welders trying to emulate your 'turbulator' with modified steel burn tunnels, and other experiments that modified the original heaters so much they were basically irrelevant to actual home heating performance in a long-term or high-temp application.
More recently I read a secondhand reference to metal parts for 'the top of the burn tunnel' in your designs. That, coupled with a picture I saw of a sheet of metal with two holes from one of the DragonHeater kits, created a very alarming picture in my mind. The top of the feed opening is a different situation altogether.

I still prefer to minimize metal-to-masonry joins: I think the roundness and low temperatures of the barrel and ducting are part of what allows this hybrid heater to work well even when built by inexperienced novices.
I see you've used rounded corners and imagine there's expansion jointing where I can't see it, at the metal-to-masonry join. Ernie reported early metal parts literally 'crawling' out of place with repeated heating and cooling cycles.

We've been pursuing a similar line of reasoning about cooling the feed area in masonry installations: using less insulation around the top of the feed / feed-end face of the burn tunnel, allowing the casing thermal mass to bleed off excess heat, and then resolving the expansion issues that this creates with less-insulative materials and methods.
Curious to see how it works out in extended firings (we and some clients may run a heater 6 to 8 hours or more at need, at which point we tend to rely on skill/attention to keep the fire settling properly so the top of the feed never over-heats).

It looks like your mold may also have a fabrication joint along the top of the burn tunnel. Does that help with expansion cracking in the feed-to-tunnel corner area?
Does it require a secondary seal when installed, to cover any crack that may develop?

We particularly enjoyed seeing the ProBoards post of your thermal performance tracking charts. Ernie wants that kind of data for a full, standard, installed system (no supplementary air, simplest firebox layout (no turbulator, no ash pit, no tapers, etc.). From what we could gather the system being tested was a modified model of the combustion area alone, not a full, thermal-mass installation.

The drag and temperature drops of the thermal mass do affect firebox draft and performance; getting the whole system to balance well with neither excess nor insufficient draft is one of our primary concerns.

Can you tell us more about whether you've tracked long-term performance with full thermal-mass installations? Do you give specific limits for how much thermal mass, how long the heat-exchange channels can be, or how to calibrate for taller/shorter chimneys with your models?

Yours,
Erica W

Erica,
No need to apologize, it's easy to draw a different conclusion with only a few pictures to go by.
I'd think we are on the same path regarding steel in whatever form inside a rocket stove. In short: it will be eaten by the fire, in fact it is a very rapid corrosion process. Mainly due to the high temperatures combined with a relatively high oxygen and low carbon environment. Even stainless steel will buckle and give up in time.

The metal end in my design is there solely because of the opportunity to cool this actively using the incoming air for the secondary air intake. By cooling the feed, the secondary intake got fed hot air which will work much better. It's no use to go through several potentially troublesome construction details when there's no active cooling. By the way, only one of the extra openings of the feed tube is in use, the one facing the riser. This is done like this to create the opportunity to turn it around or even upside down when one side is worn out by the fire.

And yes there's an expansion joint between the steel and masonry. I fully agree with Ernie, the steel tube will crawl out of place without that. In the picture you'll see a white line between the left and right halves of the casting, that's an aluminum oxyde felt strip to seal both halves together. Casting the thing in one piece will result in major cracks. The same alu-oxy felt is around the steel tube to separate it from the casting.

The longest test run I've ever done with the small prototype lasted about 5 hours. I've let it run in the highest possible mode, needing constant attention and feeding, seriously trying to overheat the little critter. It still kept working as intended, coupled to a proper vertical chimney which had a too wide diameter, though. When moved outside in the garden, equipped with about two yards of bare vertical stove pipe, the top mode couldn't be reached and after an hour or so the fire crept up the feed. This prototype wasn't coupled to a bench or whatever, the high heat was taken away by means of two 16 gallon drums stacked on top of each other. The exhaust out the side of the steel litter bin which was used as the base, otherwise my testing equipment wasn't able to withstand the temperatures!

Can you tell us more about whether you've tracked long-term performance with full thermal-mass installations? Do you give specific limits for how much thermal mass, how long the heat-exchange channels can be, or how to calibrate for taller/shorter chimneys with your models?

I'm sorry, rocket mass heaters are few and far between in the Netherlands. I've conducted a single test run on only two of those, not enough data to draw conclusions except this: those full blown stoves had both the same spiky behaviour as the unalterated little model. And yes, the influence of chimney draft, i.e. gas velocity inside the stove, do have a large impact on performance. Virtually all woodstoves are made with large tolerances, the closer to the top range of performance the more difficulties to overcome.