Development of a compact batchrocket core.

Years ago, I tried a tapered riser Travis Armstrong posted about it. donkey32.proboards.com/thread/3379/tapered-riser-batch-box-core

The thing seemed to work quite well, work on this slipped into the DSR2 concept eventually. But the notion of a shorter riser stayed in my mind. Now that the DSR3 design is sort of established, one of its important properties is the maximised burn rate. Plus no steel in the firebox itself and a low profile. A bit complicated design to build first time, I admit that.

In line with the tapered riser, why not a riser that would be as wide and deep as the width of the firebox? And much lower, actually it leans on the notion that the volume of the normal riser is implemented in a compact way. I have been busy with building, changing and testing since November 6th and I have to say, it looks like it is going to work.

The firebox itself is according to the established recommendations, bar one aspect: the port. This is quite a bit narrower, it's a 60% one, formula 2.1B high and 0.4B wide. B being the base number that is used to calculate most of the meaurements of a batchrocket. The port has been a number of sizes, this size seems to work best. Riser is as wide as the firebox as I mentioned, and has a square shaped floor. Now this didn't work as intended, there was just a straight flame from the port and nothing else. No double vortex nor even a fire fountain. So in the small development model I introduced a liner on four sides but not any higher than the port. Everything above that is wide again. Not much above that, mind you, the total height of the riser is 5B instead of the recommended 10B.

The air inlet is entirely through the door frame, in fact it's the same door as I used for the DSR3 development model albeit I modified it to fit the new core. So no P-channel, floor channel or whatever. So far, with a riser top that's closed off partly with two strips of potter's shelf leaving a gap in the middle of just 110% of the system size csa. The system is one with a 130 mm diameter chimney pipe, just a little bit larger than 5". So it worked, but making the end port wider or narrower only resulted in more instability as it came dangerously close to overfuelling every time it was run.

Today I solved this, just by shifting the gap in the middle of the riser's top to the front side of it. Now it's not a converging thingy anymore, unlike a De Laval nozzle is doing. As I see it, there are two competing streams now, one that is able to go straight up and one that is rising at the riser's back wall. That last one is arriving at the gap slightly later and is interfering with the straight stream. The net result is a burn that won't race through, exactly like what the DSR3 is doing. It's still early days, the end gap is initially set on 100% now. It might be a bit tame for my liking so at some point I'll try a larger gap.

Pictures are on their way, will be implemented in due course. Another event: my trusted Testo 330-2 died quietly some time ago, repairing would be very costly with more repairs in the not-too-distant future. I pondered over giving testing up altogether and let this Shorty idea untried. But oh well, a pensioner has to have a hobby and hobbies are costing money so I bought a new Testo gas analizer, a T-300 this time around. It's getting used to, works a little different as compared to the old one but the new software could be run on the old small laptop I used for years for this purpose.

First a couple of pictures, just to show what has been done so far.
The setup is of two slightly conical barrels, with the open tops connected to each other. Both barrels were old ones and quite battered and didn't fit together well. So every time I wanted to change something new aluminum tape had to be applied to close it up again.

The core is built in a steel box which is fixed to the lower barrel.



The exhaust hole is situated below the core. Higher up the steel rod is visable where the steel box is welded to.



There's a second bar on the top front and also welded to the box. The core is built partly out of old 30 mm refractory pieces, some new 15 mm pottery shelves and leftovers of 10 mm shelves.



All around is 25 mm superwool, even under the firebox' floor. Next picture shows the riser box, insulated as well.



The door frame is bolted to the steel core box, any crevices between steel and refractory are sealed with superwool. Later on, I added roughly 30 mm to the height of the riser box, that's it. As is clear, the riser is very short.



And here's the complete setup, very ugly but functional. The door and door frame are the same as used for the DSR3 development, only the top door has been cut off. Air is supplied through the single inlet in the door, feeding into the door frame and streaming into the firebox through the slits in the top half of the door frame.



Here's the liner inside the riser box, done in 25 mm superwool. It shows a bit higher than the port, I took the surplus off later. There wasn't a difference that the Testo could measure so I am confident upto the port's top will be enough. Of course it could be done in any refractory material, the wool was the easiest since the hired wet saw was returned at the time.

I have been following a guy who built a rocket stove for his tiny house....



It does seem getting the amount of mass right is very key.


Thanks again Peter for all your info.

Jack Dody could have been my brother. Older, younger? Not sure about that. He's the better speaker of us two, that's for sure.

The height of the riser box has been at 4.5B, 5.5B and is now on 5B. The centered end port has been at 80%, 90%, 100% 110% and 120% of chimney csa.



The results of running the shorty with all those openings, one by one, made clear that instability was higher with the smaller ones. Sometimes a really good one came along and most of the runs would pass the EU conformation test for the parameters I am able to measure. The diagram from November 11th was notably nice, maybe due to the fact most of the refractory parts weren't dry yet.



Some others were downright not good, racing through and producing a large CO peak. Oxygen level went very low in a number of instances, although not all the time accompanied by the CO peak. The one from November 20th was remarkable in that aspect, oxygen went down to 4.5%, a sure recipe for disaster in the past.



Most of the test runs were done with a 60% riser port and a 110% end port, centered still. Until I got the notion that the centered end port would probably act as a convergent opening, getting higher speed and lower pressure. The narrower the port, the more speed. So I tried something similar to the DSR3 end port, the effect of that one is completely opposite, the narrower it is, the slower the burn. To this end, I shoved the opening towards the front of the core and tried a 100% opening.



The first testrun with this configuration was done on a very windy day with many gusts, it shows. But on the other hand, it looked quite good, the door was closed all the time and I did nothing to interrupt the way it behaved. End port 100% of chimney csa.



This went very well, despite the wind gusts it was all within EU parameters. This happened to be also a warm start and bottom lit, unlike the other runs which were all top lit. Next day, November 24, a calm day, firebox loaded with 2.8 kg of soft wood, door closed at the 11 minutes mark. This happened to be the best run to date, ran very calm with nice numbers.



Next goal: trying slightly larger and smaller openings, I am confident this is the way forward.

Hi Peter,

Exciting development! Clear report. We are following the process with much curiosity.

Good luck,

De Vrijlemerij

Yesterday night a test run with a slightly larger end port, 105% instead of 100%. The results are very much alike the before published diagram. O² went a bit lower, 7.5%, instead of 8.7%. Same fuel, same load, more or less the same wind conditions. I'll try to repeat this with hardwood the next test run.
This is the tuning phase, I am getting there albeit slowly. But I won't be hurried.

Sometimes, results tend to be very surprising, leaving me puzzled about the possible cause. Three days ago, November 27, I planned to do a testrun using hardwood because I wanted to see the differences as opposed to using softwood. I had a small crate of old oak scraps lying around, so I decided to use that. Very old oak, very grey on the outside.



The resulting diagram turned out to be very different, even disappointing, to say the least. Oxygen level came down very late in the burn, CO came up high and early. In general, to the eye it didn't look like a bad burn at all. But the Testo analyzer registered lots of CO, a mediocre burn. Nothing was changed in the configuration so I suspected something might be wrong with the fuel.



I let it sink in for a day or two and decided to use some healthy, unused, beech logs for a reference burn. It turned out to be necessary to split those logs up, since they were too large for the tiny firebox of the 130 mm (5.12") system. I could cram in a mere 4 kg (8.82 lbs) of bone dry, nice hardwood.



The burn progressed quite good, not as slow as the other one. This happened to be a cold start and top lit, as standard just to eliminate temperature and lighting differences.



This diagram looked much better, although the CO start peak wasn't very nice. Surprisingly, although the CO was on the high side of things, with both runs, the white filter of the Testo came out as just light grey. Is there a trade-off between soot and CO? I've seen it before, not as pronounced though.

In order to check the condition of the old oak planks, I'd burn another load in our own red bell heater yesterday night. Just the regular evening fire, uptil now we didn't have fires in the morning. Now this burn turned out as very sluggish as well, but no smoke stench at all outside.
So, the not-so-good results are probably due to the second rate fuel.
Next stage: enlarging the end port to 110% of chimney csa and see what happens. Softwood species first, those are free, hardwood have to be bought. Just one 4 kg load left of the better species.

Still eliminating port configurations in order to find out what the stable ones are. As I mentioned, I enlarged the end port to 110%, firebox port still at 60%. Fuel 3.3 kg (7.28 lbs) softwood, wind force 2, door closed at 10 minutes. The results were not good, the core failed to limit the burn rate sufficiently. To make matters worse, the white Testo filter turned to pitch black.



New day, new chance. I reasoned I crammed the firebox too full last time so I tried another load. Softwood again, 2.5 kg (5.51 lbs), cold start, wind force 2, top lit, door closed at 11 minutes. I kept loads of air at the top and the sides, maybe I could prevent overfuelling that way.



And yes! No overfuelling although not a good burn either. Now the burn refused to come up to speed at all and the diagram looked stable but shitty in the real numbers department.



After a day of grumbling and brooding over it, waking up at night a couple of times, I concluded this wasn't the way forward. The darn thing had to do its work well with a plethora of fuel arrangements and species otherwise I'd better ditch the whole idea. So, the end port was rearranged to 90% of chimney csa, and I mimicked the fuel load of the day before. Most of the other configuration details being the same, even the weather gods played ball.



This went much better, bar one spot. Where the CO went up in a sharp smallish peak I added two pieces of fuel. With the idea I might be able to provoke it into an overfuel situation. But to my pleasant surprise it didn't, it corrected itself within two minutes and went on the exact right way.



So, in order to repeat this latter behaviour without the peak, today I loaded the thing again with a bit more fuel 2.9 kg (6.39 lbs) and somewhat more densily packed.



Still lots of room above the fuel and behind it. This time the burn progressed very calmly, I refrained from tampering with it and it payed off. CO went down and stayed there for quite an extended period. O² went down nice as well, efficiency comfortably high as a consequence. The Testo filter happened to be just lightish grey, not very dark as the day before.



Progress has been made, next run I'll take my chances with the last portion of (very dry) hardwood species available in-house.

Can you clarify please what the %eta datapoint (red) represents? I'm assuming overall burn efficiency but I'd like to be sure. Is there a standardized formula for calculation or does your tester output this point itself?

Ezra Beaton wrote:Can you clarify please what the %eta datapoint (red) represents? I'm assuming overall burn efficiency but I'd like to be sure. Is there a standardized formula for calculation or does your tester output this point itself?

Yes, it represents overall burn efficiency, lower heating value. No standardized formula for it, the gas analizer is coupled to a computer and the Testo software generates a diagram and a spreadsheet. Only three of the values are measured directly, being temperature, carbon monoxide and oxygen. Other numbers are calculated from those three. The data is saved with an interval of 10 seconds so the spreadsheet is quite extensive.

Peter van den Berg wrote:

Ezra Beaton wrote:Can you clarify please what the %eta datapoint (red) represents? I'm assuming overall burn efficiency but I'd like to be sure. Is there a standardized formula for calculation or does your tester output this point itself?

Yes, it represents overall burn efficiency, lower heating value. No standardized formula for it, the gas analizer is coupled to a computer and the Testo software generates a diagram and a spreadsheet. Only three of the values are measured directly, being temperature, carbon monoxide and oxygen. Other numbers are calculated from those three. The data is saved with an interval of 10 seconds so the spreadsheet is quite extensive.

Thanks, now I can feel more comfortable commenting on the hardwood burn vs the softwood burn. If you could oblige one more test on the matter, I would suggest a burn where the bottom half of the charge is hardwood and the top half is softwood. I understand the end goal is to be able to burn a range of different charge types in one batch, and I think understanding the differences might help. Possibly it will only matter where in the charge you load different types.

Due to their physical properties, hardwoods naturally resist heating up quickly. Softwoods are more like insulation, and hardwoods are more like thermal mass. They will not burn quickly and will take some time to get to temperatures necessary for efficient burn rates. My solution for my own primitive woodstove is to split up a good core charge of kindling roughly 1/4"x1/4" (6mmx6mm) to make sure flames are hot enough to heat the firebox enough to get to secondary gas ignition as quickly as possible. While they resist temperature change, they do contain more energy (BTU's per volume) than softwoods, so it would follow that using hardwoods for heating should be a natural thing to do. I feel like the data charts speak to these differences pretty well. The hardwood burn did do quite well, but only later in time after everything has heated up.

If you do not have softwoods available for a particular charge, I would suggest splitting the hardwoods down much further than you would think necessary, at least for the top half of the charge, to a maximum size of 1"x1" (25mmx25mm) and see how those results pan out.

You are doing great work, please keep us updated. I'm sure I'm not the only one that finds this exciting here.

Ezra Beaton wrote:Thanks, now I can feel more comfortable commenting on the hardwood burn vs the softwood burn. If you could oblige one more test on the matter, I would suggest a burn where the bottom half of the charge is hardwood and the top half is softwood. I understand the end goal is to be able to burn a range of different charge types in one batch, and I think understanding the differences might help. Possibly it will only matter where in the charge you load different types.

I pondered over this idea, there might be something in it. By the way, all the burns are done with top lit and some split softwood on top.

Ezra Beaton wrote:Due to their physical properties, hardwoods naturally resist heating up quickly. Softwoods are more like insulation, and hardwoods are more like thermal mass. They will not burn quickly and will take some time to get to temperatures necessary for efficient burn rates. My solution for my own primitive woodstove is to split up a good core charge of kindling roughly 1/4"x1/4" (6mmx6mm) to make sure flames are hot enough to heat the firebox enough to get to secondary gas ignition as quickly as possible. While they resist temperature change, they do contain more energy (BTU's per volume) than softwoods, so it would follow that using hardwoods for heating should be a natural thing to do. I feel like the data charts speak to these differences pretty well. The hardwood burn did do quite well, but only later in time after everything has heated up.

Ahhh... Now you are trying to compare apples and pears. This development model is by no means a primitive woodstove, in fact the secondary gas ignition usually kicks in between 5 and 10 minutes into the burn. Both audible and visual, without very high temperatures at all. The batchrocket heater runs very well on softwood species. Most people think coniferous species are second-rate as firewood, but in my opinion they aren't.

Ezra Beaton wrote:If you do not have softwoods available for a particular charge, I would suggest splitting the hardwoods down much further than you would think necessary, at least for the top half of the charge, to a maximum size of 1"x1" (25mmx25mm) and see how those results pan out.

I mentioned some days ago that I have one last load of very dry hardwood left. But, being a curious human being, I gave it a go with two burns today. The first one your suggestion of hardwood with softwood on top. This is a cold start, windforce 3, 2 kg of hardwood, 1.5 kg of softwood on top and top lit.



The resulting graph is nice but not outstanding. The oxygen did not come down sufficiently which had a negative effect on efficiency level. In fact, somewhere in the burn the softwood at the top was almost burnt away but the underlying hardwood wasn't ignited fully yet. Between 20 and 25 minutes into the burn the pile collapsed a couple of times, visable as little hiccups in all of the lines in the diagram. This was the morning burn.



In late afternoon, I emptied the firebox, most of the glowing coals into a steel container and placed outside. The remaining hardwood consisted of one relatively big chunk split up and some clean hardwood scraps. Again with half of the volume in softwood on top. Top lit again, same wind conditions as this morning.



My guess was that the off-gassing would go too fast this time, due to a hot firebox and the finer split fuel. And it did, in a most spectacular way. It took the fire 11 minutes in an important part of the burn to stabilize itself again. The enormous peak of CO means lots of burnable stuff went up the chimney and wasted outside. The Testo filter came out real black, it's a pity.



Heated up from cold with fine split kindling is definitely a no-no, it was heated up now and ran out of hand. In conclusion: I won't go down the route you suggested. This method might be good for your stove but definitely not for this one. I'll revert to what I know best, not to mention the fuel that I have at hand.

Thank you for those results, they are wonderfully contradictory to my hypothesis. It appears hardwood is almost entirely improper to burn in this stove design in any way, and my mind struggled to grasp a pre-heated start being LESS efficient until seeing the results on paper. The first graph shows the burn going actually spectacularly well until the softwoods were gone from the charge, and the second graph is just an awful mess all around.

Eagerly awaiting further experimentation and results. Will you be making distributable plans once the design and testing phases are done?

Ezra, your conclusion that "It appears hardwood is almost entirely improper to burn in this stove design in any way" is all yours. I won't agree with that particular conclusion.

In the mean time, I did some other testruns.
On December 5 and 6 I did one testrun each day, both with a different idea in mind.
Let's start with the first, that one was to find out how compact the fuel could be loaded specifically. And as much as reasonably could be crammed in without blocking the air stream passing over the fuel.



This did go very well, the oxygen went down to 4.5% being close to the critical level of 4%, CO stayed down the 500 ppm mark for 40 minutes, while the burn itself lasted for 64 minutes. It also generated excellent average numbers, O² 10,75%, eff 89.03%, CO 265.5 ppm, TR 151.03 ºC. Much better than the minimum requirements of the EU norm. Maybe the fuel at the bottom of the firebox could have been somewhat bigger in order to extend the burn somewhat.



The second burn, in a cold heater, was done to find out what would happen with an open core. To be honest, I was using a spark screen that blocked the air stream to some extend. Probably to about 40%, so more than half of the front was completely open. In order to avoid smoke in my workshop i started the thing with a small fire in front of the port.



The Testo was running at the time, the situation above was about the point that I loaded some more fuel. Later in the burn, at 13 minutes, I added three more pieces.



As can be seen, there's lots of space between the different pieces so air is able to run through. The CO reacted to this by a slight bump in the level and corrected itself again rather quickly. At 23 minutes the fuel pile collapsed, visable in the O² and efficiency lines. And also ever so slightly in the CO line.



Not bad at all, given the fuel wasn't particular large and very, very dry, 9% moist to be precise. The fuel raced through, making a lot of noise and a short burn time, just 42 minutes. The numbers are obviously not as good as the former testrun. O² 13.08%, eff 81.74%, CO 359 ppm, TR 194.5 ºC. The higher levels of O² and TR both had a knock-on effect on overall efficiency but that's to be expected.

Next plan: test how the open system would behave with larger fuel and top lit.