Adventures with a horizontal feed

[thierrylah]

Hi satamax
yes I am French, I live in the mountains next to Rome in Italy

thierry

[satamax]

Bello, io abito vincino della frontiera, nelle alpi. Ma sul lato francese

[thierrylah]

C'est ici que les grands esprit ce rencontre ;D , je vie dans les abruzze a 70 km de ROME l'hiver est froid et je vais expérimenter ce rocket cet hiver . Ensuite j'ai deux autre projet une cuisine a bois et une chaudiere

Tu cherche a te faire une forge sur base rocket ?

thierry

[satamax]

Salut Thierry. Bon, j'vais quand même continuer en Anglais. Par égard pour nos hotes.

Yep, i want to make a rocket forge. I need to find a way increase the draft. and to manage to spread that draft through either wood embers or coal. Feels like a grate is in order. But grates can't seem to wistant the heat of a rocket. Thought, cast iron resists for a while with coal fire. I've gone to the forge.fr for a bit more "comprehension" of the forge.

If it's your own house in the Abruzes, make yourself an horizontal feed, described here, perfect way to have a nice looking rocket with an open and seeable fire.

[peterberg]

@max,

The temperature of the glowing embers is 1152 centigrade (2110 F.) maximum. Not for long, the highest sustainable temperature wouldn't be over 950 centigrade (1740 F.).

[peterberg]

Since the Testo analyzer is operational again, results seem to be getting worse. That's to say, I've done one testrun before which wasn't up to expectations. But now, 6 runs in two days came out rather bad.

I've had the riser replaced by a round stainless steel duct complete with insulation around it. This riser consisted of one brick on its side above the tunnel and the duct on top of that. Presumably, this created a restriction in the riser where the flames had to pass through. Not a good thing: this disadvantage may be alleviated by placing the round duct higher up, so the restriction wouldn't be in the flame path but above it instead. I didn't try that and replaced the round riser by the bricks arranged in a rectangle. The bad results persisted, so there ought to be some other difference compared to, say, three weeks ago.

High res picture:
Averages: 15,2 % O2, 76,4 % eff., CO 2189 ppm, temp 192 C.

Delving into the results of the past month, there were some differences, though. I wanted the best graphics to be the standard, which isn't a good thing to expect. And another, very real difference: I'm used to close the stove by means of dry stacked fire bricks, leaving an opening at floor level. Those bricks leave little cracks and crevices so this wasn't tight at all. The window I've been using since is sealed with superwool, so this won't admit any air inlet higher up in the stove entrance. To mimic that effect, I've opened up the superwool strip above the glass pane somewhat. The next graphic do show some improvement, although not as much as I've been used to.

High res picture:
A little bit better, but the mountain range of CO in the first half is still there. Averages: O2 15 %, eff. 79,1 %, CO 1429 ppm, temp 181 C.
There had to be some other deviation, and the same day I've found this: I've been cramming the thing as full as possible too enthusiastically, up to the top and against the wall with the tunnel opening in it. Maybe this could be the reason of the bad results. So directly after the burn I've filled the thing again, behind and on top of a small bed of glowing embers, leaving room above and behind the fuel. This burn was developing very slowly, maybe I've been too cautious. The Testo was switched on after the stove picked up, so this developing phase isn't in the graphic.

High res picture:
Averages: O2 14,9 %, eff. 75,8 %, CO 1092 ppm, temp 194 C.

Obviously, the dangerous territory is in the warming up of the thing, leading to the top of the burn. When this is passed safely, there should be no troubles ahead anymore. To check this, I've done another run 6 hours later to check this presumption, with the stove mostly cooled down. All the causes named before I'd try to avoid this time. Conclusion is now: this stove is a sanguineous one. But there's a remarkable attribute: it's very capable of burning hydrocarbons, the filters of the analyzer come out colored from very light grey to dingy, whatever the results are. But at burning CO it's a bit fussy, surprisingly. Not necessarily a bad thing and a perfect fit to the American market because PM is regulated here, instead of CO in Europe.

High res picture:
This last run, I would love to have it like that at all times.
Averages: O2 15,1 %, eff. 78,7 %, CO 676 ppm, temp 183 C.
This is a very nice burn, challenging the best woodstoves around.

Edit:
This above graphic do show an oxygen level which is too high to my taste. That would be a matter of tuning the air inlet openings, nothing too drastic.

[peterberg]

Yesterday, I've tried to eliminate some of the enhancements described in my post from July 14th. In short, I've stacked the fuel up to the ceiling of the batch box in the first run. No discernible difference.
Second, I've blocked the top front air intake, widening the bottom front intake at the same time. No discernible difference again.
Third, I've stacked the fuel against the far wall around the tunnel opening. The result: a very dirty burn and unpredictable behavior like CO spikes. So, leaving space between fuel and wall is required in order to obtain a healthy burn.

By the way, I've rebuilt the thing partly. The riser is turned around a quarter, so the wider edge of the rectangle footprint is at the side of the batch box. This can be done without repercussions to the quality of the burn. Or at least, it looks like that.

Today I've tried to minimize the air intake in order to obtain a higher efficiency. And I've stumbled over another phenomenon, that of the ratio between inlet and tunnel. The original one had the inlet fixed to a quarter system size, tunnel nearly the same as riser size. At 1000 C., the incoming air will expand at about a factor 4. So in theory, the velocity of the gas stream will be the same at inlet and tunnel. This air speed is very important to the mixture of air and combustible gases.

I've made the tunnel c.s.a. smaller, 72% of riser size in my case. To maintain the same air speed, the inlet could be smaller without repercussions to the quality of the combustion. By minimizing the air intake even further, the air speed in the tunnel would drop. The roar of the flames will go down, turbulence will be lower resulting in poorer mixing. The testrun supported that theory, in effect the quality of the burn went down quite dramatically. Not that the entire run would be worse, far from that. But a stable burn would not look like this one. This testrun has been re-filled twice and at one point the CO level has been very low.

High-res picture:
Conclusion: the tunnel should be 4 times as large as the total air intake, or smaller to maintain air speed?

[peterberg]

For some weeks, I've been busy to rebuild the batch-box rocket stove about twice a week. And testing the darn thing, of course. Maybe it's a bit strange to use such quantities of firewood during summer. ;D

I'm still in the department of variable results. Sometimes with top and bottom front air, sometimes with bottom front only. A cast refractory piece with a longer tunnel formed as a roman arch, equipped with a trip wire didn't do the trick either. Longer or shorter tunnels won't produce better or worse results, there's one combination which is not good though. I've had the riser turned around a quarter so the wider side was facing the fire box. Combined with a short tunnel this is not a happy solution, presumably the flames need to have some length horizontally.

High res:
Here's one of those diagrams, on the top of the burn everything is quite good. But around that valley there's nothing but the Black Mountains.

This week I've tried something different: I've thrown the complicated tunnels out of the window and built a short tunnel of 9" high. The width of this opening is a little bit more than 2", because I'd stick to the 70% of the riser c.s.a.. The riser behind this letterbox on its side is back to it's old position, the shortest side to the fire box.

Above this tunnel, I've mounted the old p-channel plate of 4 inch wide. This stubbornly wouldn't work as hoped for: only with fiddling and prodding, changing the inlets top and bottom to keep the results in a sort of acceptable range. Somehow it dawned on me that the p-channel couldn't deliver enough air at the right place. The mistake being the assumption that the whole wide of the channel would draw air. But when that wasn't the case, the small under-pressure should be present in the tunnel entrance only. So the active channel c.s.a. would only a little bit more than half the 5% heat riser c.s.a..

To make a long story short, I've made another channel out of a 4x4" duct but larger this time. In such a way the tunnel was used as the width so the gap rose to 0.67". This time the results were not much better, mainly because I was still prodding and fiddling with the inlets.

High res:
About 20 minutes into the run it looked like the thing was behaving itself, but shortly after that the glowing phase started and the CO went up again. Out of cheer frustration I've loaded the thing again on the glowing bed and everything couldn't be better. I like to have more of these diagrams, please.

High res:
Averages are very nice: O2 12.5 %, eff. 77.7 %, CO 540 ppm. end temp 241 C.

Next: I think it over to make a new p-channel with a larger c.s.a. Total air supply was not too much, I let it ran with fixed air entrance and all seemed quite good. The stove produced a real roar during the entire run, only impaired when a piece of fuel fell against the inlet opening.

[Donkey]

Peter,
Do you have an image of the current configuration?
I think that I can follow along with your descriptions..
On the other hand, one (or two) images might remove doubt.

[peterberg]

Right, I can do that. Here they are, the first is how the bricks looks like at the outside.



The second is the inside, the bricks are clearly visible and the p-channel is the same width as the tunnel. Overhang is about 0.8 inches, the same as the gap. Explanation in the next post.

[peterberg]

Yesterday, I'd produce another p-channel. The width of the tunnel, together with a larger gap. The size is now 6.8 % of the riser c.s.a. Because the channel is wider now, I've closed the front air a bit to maintain the same total air intake opening. Total opening has been 5.25 sq. inch for some time, giving the best results overall.

I've started the next testrun with a small fire, warming up the tunnel and lower riser end seems a good idea.


High res:

Not particularly good, but the averages were not bad: O2 12,5 %, eff. 85,2 %, CO 1276 ppm, Tr 177 C. The burn was very short, but the oxygen level quite low. During the entire duration of this run a quick puffing could be heard, like a small steam engine. Variable in speed and sometimes only the usual roar.



After that I've loaded the fuel behind and on top of the glowing bed with the Testo already running.


High res:

The loading took about 3 minutes, after that the stove was closed. This is clearly visible in all the lines of the graphic. After that, everything went very well except for two thin high peaks. This was caused by a burning piece of wood which fell down in the air intake opening. As I've said before, the total inlet size has been maintained , but the oxygen went as far down as 4.7 % this time, while still running impeccably. About every good stove can stand a level as low as 8%, anything less and the CO will go through the roof.

Quite extreme, together with another surprising effect. The burn lasted for 44 minutes, while usually it's over in 30 minutes. The paint on the 55 gallon barrel began to stink again, so I've had to open the doors front and rear. Averages for this run: O2 11,5 %, eff. 81,1 %, CO 715 ppm, Tr 228 C. The filter of the Testo had been used for both runs and came out very light grey.

It looks like I'm making progress now, the violence of the burn do scare me now and then. I'm wondering how this thrown together mock-up can stand this explosion of heat. The weight of one load of fuel is about 8 lbs, small bits and pieces, very dry, even smaller than kindling. When anything would produce an overload of hydrocarbons, this will.

Amusing, isn't it?

[peterberg]

Today, I've tried an even larger p-channel, close to 9 % of riser size. This run didn't go well, the fire had a tendency to burn back to the inlet and the CO went up and down like a intoxicated submarine.


High res:

At one spot there's a large dip in the red line where the stove has been open after the small start fire. There's also a moment, a couple of minutes, when everything is going well but the rest is really bad. Probably I've overdone it with this larger p-channel so I've waited for 5 hours.

After that, I've changed the channel back to the one of 6.8 % of riser size. The stove was pretty warm, I've had to use gloves to lift the top plate. Starting the stove again proved its potential of being a real good appliance. There's been a moment where I could see over the shrinking pile of burning wood into the tunnel. A blue-purple flame of about 8 inch long roared through the tunnel and into the riser.


High res:

The thermometer on the barrel did creep to the 570 F. mark, the bricks close the bottom of the riser went to 455 F. Again, I'm surprised about the sheer violence which is taking place inside this humble stove. One of the fire bricks is cracked, all those are used before but still...

[matthewwalker]

This is really, really interesting stuff Peter. A question; are the bricks you are using "light fire brick" or regular old heavy fire bricks? I wonder if you could improve your performance if the firebox was highly insulated right against the burn.

I too am heavily prototyping, as we speak. I'm less concerned with efficiency and more with construction techniques and alternate applications of this type of stove, but your research is giving me all sorts of ideas. Thank you very much for sharing what you are doing.

[peterberg]

Aug 10, 2012 12:03:00 GMT -8 matthewwalker said:

A question; are the bricks you are using "light fire brick" or regular old heavy fire bricks? I wonder if you could improve your performance if the firebox was highly insulated right against the burn.

The bricks are not the insulating kind, so these are heavy bricks. I know, it's best to insulate the hottest parts. In fact I planned to start with the lowest half of the riser this weekend. This stove isn't built with insulation in mind, but we'll see how far I can take this.

[peterberg]

Saturday I've been busy to insulate the lower part of the heat riser. With just enough 1/2 and 3/4 fire bricks and some sand/clay I've built a thin wall about 2" from the riser. The space between is now filled with vermiculite, after two hours of burning there's hardly any heat coming through.

www.ecologieforum.eu/download/file.php?id=3986&mode=view

Some sand/clay was left over, so the firebox walls were increased with some half bricks lying around. Clearly visible in the picture. The p-channel of last run, the 6.8 % part, was used in the testrun. Not a success, the CO line was going up and down again like crazy.


High res:

I wasn't satisfied by this run, the highly appreciated low and straight portion of the CO wasn't there. I'd need a day brooding over it, and I think I do see the solution now.

The air coming into the p-channel has to be sucked down by the under pressure just in front of the tunnel. When the metal is heated up, the air will get hot and is resisting to go down. There's a relationship between the draw produced by the riser and this resistance. The piece of metal that's inside the firebox I would like to call the active part of the channel. So, there's an active channel part versus the riser length.

At first, the riser above the tunnel has been 6 times the active channel length. The channel being 5 % of riser c.s.a., the system was running happily. In the past month or two I've been busy to change the tunnel height several times together with increasing the height of the batch box. The last testrun before this one I'd use a wider channel of 6.8 %, the active channel part being about 5" long. The riser length in that configuration measured 5 times the channel.

By raising the walls of the batch box again, I've created an imbalance. That could be alleviated by using a wider channel once more. This widening isn't a real solution, the only effect is a creation of a laminar air stream. The friction of air over air is lower than the same air over a steel surface.

Tomorrow, I like to run a test using a channel of 8.8 %, with an active part of 7.48 inches and a riser which is 3.5 times this hot end.