The Evolution of a Winter Stove – Part 1

[Roger Caffin]

Hi Mike

A couple of things. First, I have followed the literature on flame chemistry and combustion engineering for some time now, and I do understand the basic chemistry and thermodynamics. Second, I have also followed the literature on CO deaths in small enclosures, including tents and huts for some time.

Can one kill oneself by using a stove in a moderately (or even partly) sealed enclosure such as a tent? Oh yes, absolutely, and people have done just that on a number of occasions, around the world. We have had that happen here in Oz in a snow cave, with heavy snow fall blocking vents, wet snow stopping air flow, and a well-known white gas stove. The four bodies were found in and out of sleeping bags, with the stove set up, valve open and tank empty.

> They had mis-assembled their lantern
User error. Unfortunate of course, but the responsibility rests with the user. A bit like driving a car at 100 mph on a wet road.

> a plume that registered 500-1000ppm CO. That's a monstrous amount of CO
True, I agree. MSR seem happy to sell the Reactor stove which emits (by my measurements) around 2,000 ppm. User responsibility to take the necessary precautions.

On to techie details.

> Flame quenching is a second order effect.
I am not sure what you mean by 'second order' here. I don't think the term means anything. What I am sure of is that flame quenching does happen in practice, as I have run enough enough experiments under controlled conditiosn to verify this. That is, I have put cold steel and titanium in and out of a flame and monitored the effect on the CO levels. It happens.

> A recirculation explanation seems to fit your data quite well.
We will have to disagree on this. It does not fit the data at all. See next.

> but it's hardly a failsafe design.
Nothing in this world is failsafe. However, experimental data from actual measurement shows that one can have a stove emitting under 10 ppm of CO in the exhaust stream, sometimes as low as 2-3 ppm. If the hazard from recirculation was that severe, you would not get that result. the recirculation theory fails the experimental data. So I don't think it is all that preposterous.

> the kind of attitude that the buying public wants to see in their would-be stove designers
Me, I go for experimental results over theory every time. Sure, the theory is valuable in helping you design something and understand what may be going on, but only real measured data tells you whether you got the theory right! I have the data.

Cheers

[Nick Gatel]

Less than 16:1 produces CO. Cars use catalytic converts to reduce the CO because the A/F ratio is under 16:1. At 16:1 the combustion temperature is too high, which creates too much NOx. Lower combustion temperatures are critical for NOx reduction. By recycling exhaust gases into the combustion chamber we can reduce combustion temperatures, which even at 14.7:1 would be excessive.

Also at 16:1 too much HC is produced.

So cars run at 14.7:1 because we are concerned about HC and NOx too. Too much NOx is very bad for our atmosphere. Here is what my 200,000 mile SUV currently produces out the tailpipe:

CO2 = 15%
O2 = 0%
HC = 1 ppm
CO = .01%
NOx = 5 ppm

ppm = parts per million

Allowable emissions in Calif for my year vehicle are:
HC = 98 ppm
CO = .55%
NOx = 978 ppm

There is no standard for CO2.

Edit: this is response to Jerry's question.

[Michael Gillenwater]

Roger,

I'm excited to see your series on this coming out. As I have posted before, I am eager to see someone better optimize a design for a winter canister stove. Keep it coming and I want to be an early customer.

[Tanner M]

Are you suggesting that the base of a pot could create a circulation pattern that could lead exhaust to the intake ?

I wonder how much exhaust would have to enter the mix chamber for appreciable amounts of CO to be created.

There was an article a little while back that explained how the bottom of a pot could work together with the burner to create a roiling effect. This particular stove created very little CO and is actually designed to create this effect. The description just seems like a good illustration of what I believe is your point.

You might have to jump to 'Fire Maple FMS-300T' at: http://www.backpackinglight.com/cgi-bin/backpackinglight/2013_developments_canister_stoves.html#.UdZKUsUXJv8

I don't get the idea you exclude flame quenching as a source. I guess that is a secondary effect because it happens after combustion and exhaust gas recirculation is a primary effect because it happens to/during combustion. I don't think you call it 'second-order' in an attempt to minimize its contribution.

If I put cold metal in the flame and that cold metal is a pot, I could contribute CO generation to quenching but that doesn't mean some exhaust gas didn't recirculate. I could isolate the intake to remove the possibility and test one variable at a time. I suppose if I just put cold bar stock into the flame there wouldn't be much likelihood of exhaust gas being redirected… I guess it would still be a good practice to isolate the intake.

Mike Cecot-Scherer
> I know recirculation can cause CO generation because I have made CO with a camp stove or two in just this way. And I've found that wind screens are a terrific way to cause these recirculation currents. Put them on: significant CO generated. Take them off: low to no CO. Isolate the intake from the exhaust and then put on a windscreen: again, low to no CO (all done with a pot of water on the burner). Seems pretty convincing.

[Stuart R]

> carbon monoxide results from incomplete combustion CAUSED by the fuel-to-oxygen ratios being shifted to too rich by low oxygen exhaust mixing into a stove's intake.

In a stove the fuel-to-air ratio of the pre-mixed gas is rich in any case, without any exhaust recirculation. Complete combustion relies on secondary air mixing with the flame.

> But I must point out that you have no actual data as to where CO is generated and destroyed in the flame.
The light blue part of the flame is where all the CO is generated and the dark blue upper part of the flame is where this residual CO burns in the secondary air.

> I know recirculation can cause CO generation because I have made CO with a camp stove or two in just this way. And I've found that wind screens are a terrific way to cause these recirculation currents. Put them on: significant CO generated. Take them off: low to no CO.

Yes, if you restrict the availability of secondary air (or quench the flame) then you will get CO. Wrapping a windscreen tightly around a stove, or using it in a sealed up tent are both good ways to restrict the availability of secondary air.

[Jerry Adams]

Is CO only an issue if you run stove in closed tent, room, snow cave, or whatever?

If it's raining, I run stove under one side of tent right next to door which I leave open. And I just boil water so I don't run it long. I don't think I have to worry.

If you're creating CO does that mean you're not running the stove efficiently? Or does making stove more efficient tend to result in CO? Or are the two independent?

[Damien Tougas]

Very cool, I am looking forward to the rest of the articles!

Having the needle valve after the pre-heat loop, is there any danger of excessive pressure building up in the fuel line due to the high volume of expansion that happens when the fuel vaporizes?

[Mike Cecot-Scherer]

Hi Tanner,

Exactly! a circulation pattern (not necessarily symmetrical or uniform) that brings exhaust back to the intake.

My understanding is that combustion engineers will always shoot for a slightly lean mix (including secondary air entrainment) as a safety feature even though it decreases the efficiency due to the heating of extra gas. Lean flame becomes surprisingly inefficient.

I definitely don't want to be seen as trying to exclude flame quenching as a source of CO. I refer to recirculation as primary because it messes with the most basic part of the combustion reaction – the fuel/air mix, because it's the dominant effect in many accidents, and it's able to create incredibly high CO concentrations for long periods of time.

That roiling pot/stove combo is very interesting. Seems like they made the circulation strong enough that it entrained fresh air and isolated the intake – very cool! A nice example of the fractal nature of reality – loosely speaking, a bad thing that becomes good after a certain point.

After thinking about it I think I can offer a couple of characteristics that should be not-unusual for recirculation caused CO in some stoves:
— uneven CO distribution around a centered pot that has maximums in-line with the intake ports.
— a burner that fires more upward than outward might show high CO because of what I'm going to call "flame bounce" off the pot; essentially driving exhaust downward. Stoves with more outward directed burners should be more resistant to recirculation. A Pocket Rocket might fit this.

Best,
Mike

[Mike Cecot-Scherer]

Hi jerry,
Yes, exactly. But the reduced oxygen I'm referring to is the percentage of oxygen, not the absolute up or down variation from altitude. Altitude is another thing. Maybe Roger can explain some of the ins and outs in his next article.
Best,
Mike

[Nick Gatel]

Roger's testing under controlled conditions show that some stoves emit more CO than others. If I remember correctly, MSR Pocket Rocket among the worst and Snow Peak GigaPower among the best. With proper ventilation you can safely cook in a vestibule. The problem with CO poisoning is that it is a silent silent killer. You are dead before you know there is a problem.

[Jerry Adams]

Yeah, Roger hates Pocket Rockets : )

[Mike Cecot-Scherer]

Hi Jerry,

The vast majority of stove users have found no issues with CO in well ventilated areas. Your description of how you use your stove sounds pretty safe. It seems to match the protocol of the Rainier Guides. You could always check things with a little CO monitor…

If the burning device causes concentrations of CO over 500ppm it could literally knock a person out so my personal opinion is that there is no safe way to use such a product.

(BTW I should probably mention here, given what's been written here in BPL, I have a production MRS Reactor, love it, and mine produces no CO.)

That said, I believe that hikers and climbers that are struggling to acclimate to higher altitude are probably routinely being impaired by low levels of CO and, in the case of sealed-in stove users, by lowered oxygen in their tent. My 2¢.

Best,
Mike

[David Thomas]

Jerry: Being in an enclosed area has several effects.

For the stove: if you're in such a tightly sealed environment (unventilated tent, snow cave without vent holes or in which vent holes have been covered by snowfall) that the percentage of O2 falls and CO2 rises, then the stove will see less oxygen in the pre-mix air and hence the flame will be richer in fuel (i.e. prone to make more CO) – the effect Mike is discussing, although he is also arguing for smaller-scale recirculation within a windscreen, for instance.

For you: not only may the stove be making more CO per minute, but all the CO (and CO2) being generated is being retained in your breathing space and increasing in concentration the longer you run your stove.

Usage pattern makes a difference. If, in a very-tight house, someone bakes a batch of cookies, they end up with a batch of cookies. If, however, they attempt to heat their very-tight house by running an unvented stove continuously, they can end up dead. Similarly, your brief usage of a stove, near the entrance, just to boil water is well within my personal safety limits. While running a stove in an attempt to heat up a sealed tent or a snow cave is a bad idea. If the area is sealed enough to retain heat, it is inadequately ventilated to run combustion equipment.

All stoves make CO. Some make a little, some make a lot. If you use it in a way that you are venting that poisonous gas away, fine. If you are in an environment in which the stove's heat or water vapor effluent (two things you can detect) are being retained, you need more ventilation.

[Nick Gatel]

Jerry,

We have a furnace in our camper with a vent to the outside, so we can keep the camper closed up when using it. But most often we use a catalytic heater because it uses about 75% less propane and no electricity. When using the catalytic heater, the camper must be vented otherwise the heater will deplete the oxygen, which then cause more CO. Some catalytic heaters have oxygen depletion sensors that will turn the heater off. Most of these heaters with the depletion sensors don't work above 7,000 feet.

That is my blue-collar explanation since I am not a scientist. Sort of my translation of what David Thomas said :)

[Roger Caffin]

Let's have some measured facts here.

Measurements show many stoves emit quite low levels of CO (<30 ppm). This is incompatible with the recirculation theory.

Measurements show that a stove emitting a very low level of CO can be made to emit a whole lot more by inserting cold steel into the upper part of the flame and quenching the combustion process. Removing the steel from the flame drops the CO level back.

Measurements show that a simple windscreen placed 3/4 of the way around a stove does not increase the CO emitted. A totally enclosing windscreen might bve very different – don't do that.

Measurements show negligable amounts of CO in a tent when using a good stove in a ventilated vestibule. (Actually, very often the readings are down around 2 – 3 ppm, in the 'noise'.)

Basic physics/chemistry also says that an incomplete combustion process gives off less heat than complete combustion.

Cheers

[Roger Caffin]

Hi Damien

> Having the needle valve after the pre-heat loop, is there any danger of excessive
> pressure building up in the fuel line due to the high volume of expansion that
> happens when the fuel vaporizes?

Nope. If the pressure in the line near the needle valve gets higher than the pressure in the canister, it will simply stop the fuel coming from the canister. Actually, that happens all the time in any remote canister stove: fuel oscillates back and forth.

If you take the fuel in the region of the needle valve and heat it to 1,000 C, you might think that could raise the pressure. Nope: it will still push back into the canister and recompress fuel vapour back into liquid.

On the other hand … if you heat the canister to 100 C, you can expect instant transmogrification. See http://www.backpackinglight.com/cgi-bin/backpackinglight/exploding_gas_canisters_the_hazard_of_overheating.html

Cheers

[Roger Caffin]

Hi Nick

> MSR Pocket Rocket among the worst and Snow Peak GigaPower among the best.
Well, the PR was not good. The full data is at http://www.backpackinglight.com/cgi-bin/backpackinglight/stoves_tents_carbon_monoxide_pt_3.html but a few stoves extracted are:

Optimus Crux: 300/260 (low/high, ppm)
MSR Pocket Rocket: 240/140
Coleman Xtreme: 5/5
Snow Peak GST-100: 5/21
FMS-116T: 10-50/50-100 (The FMS stoves are from later testing)
FMS-300T: 20/3

The variable results with the FMS-116T are due to some solid Titanium in the flame: while that metal is 'cool' the CO emission is high, but it drops as the metal heats up. (No recirculation was harmed during the testing.)

As far as the Reactor goes – we tested a preproduction model and found >1,000 ppm emission at low power. Boosting the air inflow (with a compressed air line!) reduced the emissions enormously. We sent these results in to MSR and they withdrew the Reactor from the market weeks before its release as a result of our report.

Several months later MSR reissued the Reactor with modifications to improve the air flow. It still tested very high (350 ppm at low power), but obviously not as bad. We reported on this at http://www.backpackinglight.com/cgi-bin/backpackinglight/co_series_supplement_1_msr_reactor.html Yes, we still have all the correspondance with MSR over this.

If you only use a Reactor at full power the CO emissions are lower. It was not designed for low power or cooking, but may be useful outdoors for melting snow.

Cheers

[David Thomas]

>"since I am not a scientist."

I'm not a scientist either. Scientists do things like entangle more photons than any human ever had before or find an exception to the Heisenberg Uncertainty Principle. Both of which my BIL has done. Engineers clean up toxic waste sites, build snowball cannons and backyard zip lines for their kids, and bring portable hot tubs on backpacking trips. All of which I've done.

Engineers (generally) don't publish in Science or Nature.

Scientists (generally) don't use pipe wrenches and voltmeters. Nor are they satisfied with one significant figure.

Scientists discover things no one ever knew before. Engineers get paid better.

[Roger Caffin]

> Scientists discover things no one ever knew before. Engineers get paid better.
Sulk.
Snark.

Sigh
PS: I own about 6+ voltmeters…

[peter vacco]

" I go for experimental results over theory every time. "
yes. quite correct. i have read that the laws of thermodynamics owe more to James Watt's engine, than his engine will ever owe to the laws of thermodynamics. (it was a horrible thing efficiency wise).
also somewhere along those lines, is, "if it looks good on paper, but doesn't work in real life .. you've got it written down wrong."
—
anybody hanker'n to produce something in quantity might be well advised to read/study/come-to-Believe, Rajan Suri's principles of Quick Response Manufacturing.
they are a godsend of common sense, that is a good leap forward from what is openly obvious.
(and to boot, you'll make a better product with less scrap)
—
if ever the little voice in your head shuts off, and you are still alive, there is a fine chance you have CO poisoned yourself.
(tuck that tidbit of information away. you may need it someday.)
—
great article. we can all see how as a product evolves, several generations of equipment may be purchased and quickly discarded as knowledge matures during a manic development cycle.
—
peter gives not a squat of snot about CO coming off a pot.

[David Thomas]

> Sulk.
> Snark.

In my major, we'd say, "Chemical Engineers just want to be chemists. Chemists just want to be physicists. Physicists just want to be mathematicians. Mathematicians just want to be philosophers. Philosophers just want to be God. Unfortunately (for the rest of them), the pay scale is the other way around."

[Roger Caffin]

Hi David

Where'd you put the programmers?

Cheers

[Jason Brinkman]

What do programmers have to do with the price of tea in China?! :)

For that matter, what do exhaust recirculation patterns, carbon monoxide pools, snow caves, or faulty lanterns have to do with Roger's new stove?

If someone were to try hard enough, they could probably off themselves with their titanium spork, but the prudent user has predictably few problems. I'll give you that canister stoves (or mis-assembled lanterns) are slightly more risky than sporks, but with rare (albeit notable) exception, the average user survives more times that not.

Oh, and I am definitely intrigued by the cliffhanger, Roger… Let's see that baby!

[Jason Brinkman]

BTW, I've learned more than a couple things about prototyping and manufacturing from the responses here – good stuff! I like the notion of molded-in threaded inserts the most!

[David Thomas]

>"Where'd you put the programmers?"

Roger: that's what the physicists and mathematicians end up doing to pay the rent, IME.

[Author]

Posts