Selecting a Canister Stove for Cold Weather BackpackingPart I: Stove and Fuel Fundamentals

[cushing hamlen]

Brian – I am forced to agree with Roger Chaffin’s comments in reply to you: we are talking about a vapor-liquid equilibrium here: the size of the headspace above the liquid should have no effect on the pressure determined by the liquid phase. I do agree that the relative ratios of propane and isobutane remaining in the cannister will have an influence: unstated in my earlier post is the assumption that to get to the “nearly empty” cannister, you get there by slowly….VERY slowly venting the gas in the cannister, so that the cannister does not cool below ambient. This does not cure the issue that the ratio of gasses will change somewhat but will at least mitigate the effect somewhat. Knowing what we have we can compensate somewhat to the composition of the full cannister. The bottom line is, I will not advocate, recommend, and certainly not perform myself, the proposed experiment on a full cannister. Doing so is far, far too dangerous to consider.

[cushing hamlen]

Sigh…..looks like I need to go find a thermocouple and some thermally conductive epoxy :-) Roger: can I reserve the right to contact you to discuss what we think is an appropriate minimum remaining fuel level to preserve a vapor-liquid equilibrium under heating?

[Brian James]

Oops; sorry!

[STRIKE]It seemed to me that 90 g of Butane at 100 degrees C. would be at a lower pressure in a 10 L container than in a 100 mL container — because the larger container has fewer moles of Butane per unit volume. Thus, I felt that the volume/mass ratio would affect the pressure at a given temperature. That said, the last time I was in chemistry was in 1999! Sorry for the misinformation.[/STRIKE]

As for the failure test, I still feel that empty canisters should be connected to high-pressure nitrogen or other inert gas, and then placed underwater and pressurized to failure. This is because the burst characteristics of these devices are unknown, and there is the potential to send shrapnel out farther and faster than many might reasonably estimate. (I’ve seen failed pressure vessels fail at 500 PSIG *underwater* due to bad testing procedure and send pieces of themselves all over the shop and water 80′ vertical onto the ceiling.) A person probably does not want the complication of an explosive propellant added to that equation. That said, it could be that consumer-use devices are designed to fail in a predictable and less-damaging way; maybe the valve blows out or maybe there’s an intentional weak seam that allows them to rupture and bleed rather than disintegrate spectacularly.

I also feel that many canisters must be tested — a test of only one or two canisters could give misleadingly high burst pressures. After all, what if the canisters were from a stronger lot, or of a stronger design? The only way to know safely is to do a series of canisters from different lots and from different manufacturers.

Once a lowest-possible burst pressure is established, calculating the temperature at which this would occur would be just a question of math I believe. (But do post cool pictures of the blown canisters!)

Cheers!

Brian

EDIT: after re-reading the responses, I understand your contradiction more clearly. As long as our volume is *low* enough to keep our fuel in a mixed-phase state, the pressure/temperature relationship can be calculated without needing to know the volume. The fact is that almost all of the fuel is in a liquid phase, and it will be the liquid-phase fuel that generates our pressure. Thus, heating the canister to burst temperature will tell us burst pressure, regardless of volume. Thanks for clarifying, all.

[cushing hamlen]

Brian – that is my understanding of the physics we would expect in this system.

It also occured to me that a safer way to go about this would be use a cannister alone inverted on a calibrated hot-plate with a remote on/off switch….and an inverted container filled with sand surrounding the can. This way we have more control over the temperature reached, and the sand will reduce (hopefully eliminate) the danger from flying shards……

[Coin Page]

My first wind screen for my Snow Peak was wrapped nearly flush around the canister and pot. The heat efficiency was terrific, but way too hot for safety. The little plastic valve in the canister was damaged, (probably partially melted) and after the whole thing cooled down, that canister never worked again, despite being nearly full. I was lucky not to blow myself up. The canister, especially the junction with the stove was way to hot to touch.
Now I use a windscreen with about an inch space all the way around the canister, which allows heat to reach to the ground, and about 1/4 inch around the pot. Works well for me, but I keep an eye (finger) on the canister especially for prolonged cooking.
Remembering my first experiment I can see why Snow Peak doesn’t recommend wind screens.

[Brian James]

Interestingly, I spoke with an engineer at the end of work tonight and he told me that the volume of the system is a contributing factor. The way he explained it to me is that the volume of the system will affect the temperature/pressure *curve*, and in such a small system (i.e. mostly liquid) that change could be significant to the equation.

He actually offered to use SolidWorks/CosmosWorks to do a finite element analisys on a canister. This is a computerized stress+failure analysis that is very easy to conduct when you know the design and materials used in all the the components of a system and the stresses placed on them. It shows you the deflection due to stress, the strongest and weakest points, and predicts the failure point and the magnitude of the stress that will cause the failure. Boeing uses the same software package to design aircraft, and then simulate them under the stress of wind or crash situations. Our company evaluates expansion joint and industrial damper applications with it.

The difficulty is that you have to know the material of the part being evaluated and its’ precise design. This is usually easy as you design a product first, evaluate it, and then build it. In our case we are doing the reverse! I told him that before we try to do that we should post it to a forum and see if any other engineering types have already started. :)

Anyone?

[cushing hamlen]

Brian, at the risk of beating this dog to death, I would not agree with your engineer friend that volume in the cannister does matter. This is a basic vapor-liquid equilibrium issue, and would refer your friend to the Clausius-Clapeyron equation from Thermodynamics: this relates pressure to temperature, and has no term at all to include volume…..

Regarding the proposed finite element analysis, our experience is that you can not trust the analysis unless you have checked it against suitable experiment (too many unknowns: material type, thickness of material, strength of the welds or solder joints, mechanism of failure, etc, etc). Since we would have to do the experiment anyways, why bother with the calculations???

After Coin Page’s post I am wondering whether we are all asking the wrong question: Coin mentioned that he found the plastic parts on his stove damaged, but the cannister did not explode. Perhaps what this is telling us is that the cannisters are strong enough take the heat (pun intended), but our REAL issue is whether we will get the setup hot enough to damage the plastic and rubber components of the stove….??????

[Brian James]

Perhaps I was not clear when I explained the question to him. Hmm.

I agree that an FEA would have to be verified by bursting a series of canisters, just like cars are crashed in simulations first and then in the real world for verification. Interestingly, if I had the fittings for it I am sitting 50 feet from the facilities needed for a 2000 PSI nitrogen burst test. Do we know where a male canister fitting could be sourced, aside from buying a replacement hose for a WindPro and cutting the fitting off? Welded or crimped fittings would be equally ok. Does anyone at least know where to find the spec for this particular valve, so I could ask a fitting supply company?

Regarding the true point of failure in an overheated stove, perhaps that’s a test in and of itself. A poster here described a can “convexing” its’ own bottom when being abused with a Bakepacker oven I believe. Curiously, though, there seems to be a relative absence of actual breached canister reports. For all of the millions of these canisters used in the field by people who may or may not know what they’re doing, shouldn’t there be some reports or at least ubiquitous legends from the community? Dead or terribly injured users? Or are we watching too many movies? We hear about people heating these things to the melting poing of plastic, and even to the point of serious canister deformation, but no one says much about actual breaches.

I’ve heard legends of “pans embedded in the ceilings of ski huts” by canister abusers, but I don’t believe that a canister could possibly fail at the threads and become a kind of vertical cannon, so I give no credence to such reports. In fact I believe that the valve and threads are likely one of the strongest points — when a burner is attached.

A theoretical test, (but not someone anyone should try,) would be a simpler, and perhaps more valuable, test: time. Since most backpackers are unlikely to have thermometers on the surfaces of their canisters when cooking anyway, I would like to know how *long* it takes to cause failure under the *most* extreme conditions. That would tell me how long I could use the thing before worrying about overheating at all: even if the max “safe zone” time was only 3 minutes, it would be a very useful metric.

I’d also like to know what that failure is: melted gaskets, self-sealing canisters, melted stove knobs, popped-out bottoms, or fireball with shrapnel?

Given a common, hot burner (maybe a crux/pocket rocket, or the 16 000 BTU Coleman F1 Ultralight) and a common but extremely hot setup (a 1L pot with 0.5L ambient-temperature water and a tight-fitted gapless windscreen plus reflector plate, stove set to Nuclear,) and warm, windless conditions (85 F), what’s a minimum “safe” burn time? Can I run that stove on max for 3 minutes without experiencing a failure under those conditions? 5? 7? And what is the actual failure?

If I knew (based on extensive testing) that a “max-hot” setup as described, on a hot and windless day, would “pop” *after* 10 minutes, (IN CASE YOU ARE SKIMMING, THIS IS AN EXAMPLE! I DO NOT ACTUALLY KNOW THIS!) that would make me think that I could safely run *any* setup on *any* day for 5 or maybe 7 minutes with a reasonable margin of safety. (IF A TEST GAVE ME THOSE RESULTS — AGAIN THIS IS HYPOTHETICAL.)

Testing would be simpler (and more spectacular) and I do not advocate that anyone even consider doing it: construct the referenced setup, drive 1 mile or more away, and listen for the sound of a poor little canister stove committing suicide. (Video feed would be better .) Wait 60 minutes in case the canister goes to some slow-burn mode, then go see what’s happened. Take many photos, post to BPL as a dire warning about monitoring canister temperatures, and throw the whole assembly (including burner) into the garbage. Lather, rinse, repeat, and we’d have some reliable data about time and failure point.

Ironically, the manufacturers have almost certainly done exactly this dozens or hundreds of times. But if they shared the information with us, they’d be massively liable — so instead we are in the dark and guessing or else conducting dangerous experiments in abandoned quarries. Weird how that works!

Brian

PS I don’t know a lot about the U.S., but I would have to guess that almost anything that blows up almost anywhere in that country would (literally) show up on a screen somewhere. I reiterate that this is utterly unadvisable from a safety and legal standpoint!!

[Douglas Frick]

>I’ve heard legends of “pans embedded in the ceilings of ski huts” by canister abusers, but I don’t believe that a canister could possibly fail at the threads and become a kind of vertical cannon, so I give no credence to such reports.

But that would be the likely result of a canister “convexing” suddenly.

>Curiously, though, there seems to be a relative absence of actual breached canister reports.

Considering that most stove users sit relatively close to their stove while cooking, and the extreme violence of an explosion (RJ: “I tossed one in a fire once. You really have no idea.”), I’m sure somebody would have heard of such a thing. Nothing turned up on Google related to accidental explosion in a camping environment. This suggests that “breach” isn’t a common failure mode, and other ‘hard stop’ modes (meltdown of the stove, convexing, etc.) are more likely than explosion.

[Brian James]

I noticed in the article that the canister valve seems to (strangely) have a plastic component inside the canister. I wonder if this component is meant to become malleable and collapse, thereby sealing the canister in the presence of too much heat?

One would think that ubiquitous Legal Departments wouldn’t allow huge companies to sell on-tank burners unless the tanks were equipped with one or several overheating fail-safes? Otherwise they’d be effectively dispensing millions of small and easily-detonated landmines into the hands of the masses and betting against the collective stupidity of the average group of campers. One would think…

[Robert Miller]

I’m thinking along the same line. I just opened up an empty MSR IsoPro canister and my oven is preheating. I’m thinking start at 200F. Any thoughts on time?

Robert

[Brian James]

I wonder if it would have to be “at pressure” with a fluid flowing through it in order to collapse/seal properly?

You could leave it at 200 and poke it once a minute to see when it goes soft? After that, report on what it does?

MSR, Brunton, SnowPeak, Kovea, Primus, Coleman, are you listening? Could you use the anonymous post feature to enlighten us or at least steer us in the right direction?

[Robert Miller]

I postponed the bake test. I’m looking for input on a test procedure. I’m also taking detailed pics of the valve. From the exterior it looks very similar to the one in the article. Check back in in a few minutes and I’ll get some pics posted.

[cushing hamlen]

Now, now….this is getting really interesting. I just “happen” to have some level of access (please do not ask any detailed questions……) to some pretty sophisticated polymer testing equipment. It might be worth disassembling an empty cannister, taking a close look at the valve assembly, and thermally testing parts that would appear to be most succeptable to thermal damage/degredation. No promises….as this would have to be done in a way that does not interfere with the normal workload on the equipment…….but it just might be very easily done…..

Robert – could you manage to get some picts that make it really clear how the parts are laid out inside the valve?

[Robert Miller]

Detail of blue plastic valve. Approx 1/4″ dia x 3/8″ long. Its function appears to be a guide for a spring. It is crimped into the sheet metal.

Detail of inside of blue plastic valve. At the bottom is a red plastic piece with a slot approx 0.090″ deep across it. The red color is the same as the pin in the screw on portion on the exterior of the canister. I think it is reasonable to assume this is the same piece of plastic but cannot confirm absolutely at this point. The deformation of the blue plastic was from my attempt to dislodge the blue plastic from the canister.

[Brian James]

Seeing the metal liner in that plastic button makes me wonder if the canister could self-seal in the fashion that I proposed.

Do you see any other possible seal-off mechanism? Perhaps the needle in the needle valve is made of a fast-expanding material such that it expands to seal itself when the temperature is too high?

Now I’m definitely interested in an oven test. I’d love to know what happens to each component of the valve all the way up the temperature scale. I also wish I owned canister equipment so that I could deplete a canister and cut it apart! (And that I wasn’t at work…)

Cushing, I’d be interested to know the makeup of the polymers in various brands of canisters; do post!

Brian

[cushing hamlen]

Unfortunately (because this is a more expensive proposition), we also should ask about the plastic components in the stoves themselves….unless we can obtain them via rebuild/maintenance kits.

[Robert Miller]

I’m estimating the red plastic piece is 1/4″ tall. There isn’t much room for more pieces. The slot in the red plastic I believe is revealing. It allows the gas to bipass while keeping the majority of its circumference in close proximity to the steel to facilitate heat transfer.

[Rick Dreher]

This is a good point. I’ve heat-damaged a piezo igniter module from windscreen use, and am a little curious about those o-rings.

Better canister stoves seem to have mostly metal parts, other than the o-rings.

[cushing hamlen]

So Rick – this begs the question as to which are the better stoves…….?

[Robert Miller]

At this point I’m thinking of extracting the red plastic piece, measureing it and throwing it in the oven. Any input would be appreciated. I’ve got some paperwork to do and will check back in a hour.

[Brian James]

Yes please :)

But do two side-by-side tests: one all by itself and one still-intact valve assembly as shown in the picture.

I look forward to the results!

[Robert Miller]

I don’t think I can do two tests. I think it will be beneficial to cut a pie section out of the cartridge and lindal valve so you can see exactly where the parts are in relation to each other. This will make it impossible to do the intact test. I only have the one cartridge.

Besides Brians original idea that the red plastic piece would expand to shut off the gas flow I see another point that may be a better alternative. There is a steel probe in the bottom of the stove that inserts through the lindal valve and presses on a corresponding probe on the red plastic piece to allow the gas to flow. This narrow plastic probe with spring pressure from the blue sleeve could easily deform if the metal probe were to become too hot.

Both ideas can be tested in one bake.

Has any one seen my Dremel?

[cushing hamlen]

Robert – does that blue plastic piece appear to actually be strong enough to compress the spring? If so, it could both be holding the spring in place, AND acting as a heat-sensitive element that when hot enough will fail to hold the spring back, and cause the flow of gas to halt……

[Robert Miller]

The blue plastic piece is both guideing the spring and transfering the tension from the lower end of the spring back to the canister. It appears to be only in contact with plastic parts that I doubt would tranfer any appreciable heat to enable its failure.

UPDATE – plus if the spring loses tension the valve goes full open. Stand Back!

[Author]

Posts