Jetboil cold weather heat exchanger

[jimmer ultralight]

Bob et al,

What would you say is a safe maximum ambient temp you could safely use that HX rig in without a danger of overheating the cannister?

40F ?

50F ?

[Bob Moulder]

Jimmer, I don't think you could overheat it even at an ambient temperature of 60°F or above, but of course it is not needed then.

This excludes, obviously, doing something stupid like additionally overheating the canister with a windscreen that covers the whole rig, or some other gross error or misuse.

However, I did not test in such temperatures because it is a cold-weather item.

[Stephen Parks]

Fun thread and great results, Bob! To get better transfer between the copper and the canister, instead of thermal grease you might be able to use thermal pad from Bergquist (or 3M, etc.)

With the thin copper, you could scrunch it up along the length between the flame and contact with the canister and then cover that with some Teflon shrink tube to reduce heat loss along its length, but it isn't that long so probably not worth bothering with.

You've demonstrated that this works great for longer burn times, but if you were only doing 2-minute boils, I wonder if you would still quickly lose most of the propane in the canister (not enough time for the hx to start working its magic?). Even so, perhaps just warming the canister up a bit first might still get things started when the canister starts to get low on propane.

[Bob Moulder]

Hi, Stephen:

The velcro strap and silicone insulator (or CF pad as used by Gary D for the insulator) works very well to create excellent contact between strip and canister without further help.

Earlier in the thread (date 2/19) I did a test burn using a 200g canister with 17g of fuel remaining at +7°F with windy conditions and was able to get a boil reasonably fast without a canister cozy. Works better with a cozy when using a larger canister at quite cold temperatures with very little fuel remaining in the canister, but it does indeed work. When canister fuel levels are this low there is very little, if any, propane remaining. I'm convinced that this system would very likely work well using pure isobutane — with no propane at all — but I don't know where/how to get pure isobutane canisters to test.

Even at very cold temperatures (I tested down to -6°F) a surprisingly small flame is all that is necessary to get the feedback loop going, so even leaving the stove/canister out all night around 0°F or slightly lower temperatures it nearly always started without an external heat boost. Two times I had to help it a bit with external heat, but even then it was very little.

Give this a whirl if you're interested in summer-like performance in the winter! :^)

Finally, a reason to give up the XGKs and Whisperlites that I thought could never be replaced for seriously cold weather.

[Stephen Parks]

"Give this a whirl if you're interested in summer-like performance in the winter! :^)"

Bob, I live in Houston (sad, I know) and don't go anywhere cold in the winter so I'm not your target audience, but I think it is great to have his tested and shared with the community. Great contribution.

[Gary Dunckel]

In the interest in furthering Bob Moulder's (and my own) data base for the purpose of improving/tweaking the copper heat exchanger concept, I thought I would show/tell what I've been doing since last April in this regard. As you might recall from my above post, here is my first version. The copper strip is a 1.125" wide strip of .021" copper, and it is attached to the canister with /.625" Velcro. Afraid of melting the Velcro by it being in direct contact with the hot copper, I placed a small piece of carbon felt between them. Notice that I try to adapt the copper to have as much contact with the canister as possible. In May while cruising the aisles at Lowe's, I found a roll of what I think is called copper plumbers tape. It was cheap, so I thought, "Why not"? As you might imagine, it was a miserable failure. The holes let too much heat escape, and it hardly warmed the canister at all. In June, I decided that there was too much of the copper strip being exposed to the cold air. That couldn't be good. So I decided that most of the copper strip should be insulated. I also thought that a 1.125" wide strip might be overkill, so I had my copper guy cut the strips .75" wide to see if I am right about that. Here was what I was thinking should be done: But Scotch tape won't work, especially in sub-freezing temperatures. I tried some duct tape, employing a unique design which the following photo shows. First problem–duct tape trashed the carbon felt when I tried to remove it. Solution–place heavy duty aluminum foil where the duct tape would be to protect the felt. As an aside, I decided to see just what would happen when the Velcro was in direct contact with the hot copper strip. After 5-6 minutes in tight contact with a very hot copper strip, the Velcro showed no hint of melting. What's this stuff made from, anyway? So I thought, why even use Velcro if duct tape will do the job? I can lighten the kit up a wee bit, maybe. I don't quite know how Bob does it, but I am a bit concerned about possibly overheating the JB stove's regulator, so I countoured the copper strip to give it some breathing room. There's just one place to put the copper strip, which is right next to the regulator. Sorry about the out-of-focus photo, but you can see what I've done. Not liking the flimsiness of the aluminum foil, I decided to accept the minor weight penalty and use titanium foil to protect the carbon felt instead. You'll see two different pieces of ti foil, which meet where the canister curves upward. The reason there are two sections is so that I can easily remove the copper strip for ease of storage when I pack things up. Just open one side of the top duct tape strip and slip the copper thingy out. The bottom line is that this version weighs 0.7 grams more than the one that employed the Velcro strap, but this is primarily due to the much longer strip of carbon felt. The duct tape and titanium pieces weigh less than the Velcro. The final weight of this last prototype (using .75" wide, .021 thick copper) is .58 oz. for a 110 g. canister, and .74 oz. for a 220 g. canister. I want to compare the performances of a short felt strip vs. the long version, and also with a Velcro only/copper strip. Also, I will record the time it takes for the stove to reach peak efficiency, as well as the time to bring a fixed volume of cold water to a boil, and also the amount of fuel to reach that boil. Note that there is a bit of a fiddle factor involved with this approach. Most of that would happen at home before the trip–placing the copper strip and felt/duct tape in the precise location for whatever stove one will be using. Once in the field everything is easy, assuming that the duct tape holds up properly. Now I just have to wait for that first 0* F cold spell to test these setups. I figure that there will be several things to watch for: 1. Will the duct tape hold up under those conditions? Maybe if I wrap the entire lower portion all the way around the canister? 2. Is there an inherent difference between a 110 g and a 220 g canister as far as performance with this heat exchanger concept (greater surface area of the 220 g)? 3. Is there any possibility that a well insulated copper strip employed this way can actually overheat a canister? 4. Did it turn out that I have totally wasted my time doing this, and that I should instead have been doing other, much more productive things? Feedback, anyone?

[David Thomas]

Quickly on one point cause I have to run: >"I found a roll of what I think is called copper plumbers tape. It was cheap" That is copper-plated something. Probably steel. Because steel is cheap. But steel's thermal conductivity is 10-20x less than copper (it varies a lot with the type of steel due to grain boundaries). Yes, the holes hurt. Imagine heat flow like water flow – you want a continuous cross section, the thicker the better, but thermal conductivity trumps cross-sectional area.

[Gary Dunckel]

Thanks, David. That and the holes explain everything about why it failed. Now, what do I do with it–maybe make some cheap jewelry to sell at the Christmas craft fairs…?

[Jerry Adams]

thermal conductivity per weight is about half for aluminum copper has a better conductivity per cross sectional area, but aluminum weighs less I just take a loop of #18 galvanized steel wire that's a bit bigger than the canister, but friction fits the copper (or aluminum) strip between it and the canister. Simpler. Doesn't melt. weighs very little.

[Bob Moulder]

Hi, Gary: Once you get to experience the operation of the unit in some really cold weather, you will see that it is amazingly effective without any embellishment, except for the cozy for the larger (220g) canister when the temperature drops below about 15°F or so. I noticed last winter that you had bent the copper strip to match the curvature of the top of the canister to create more contact area. I did that myself early on but found that the approximately 2 square inches of contact on the side of the canister were totally adequate for operation down to -6°F without any risk of overheating. I am fairly certain that with a cozy this would remain true down to -40°F, but -6°F was the coldest it got here last winter. Today I assembled some strip/strap/cozy sets for field evaluation by others (as mentioned in the other, more recent thread). Just for the heck of it — but mainly to verify the safety if it is misused in warmer weather — I ran the stoves (JB MiniMo and BRS-3000) with HX strips and canister cozies at an ambient temperature of 75°F, which was the temperature on my deck. No problem… passed the 'touch test' easily. Yes, these runs were with pots in place with water, using the JB MiniMo pot and the Primus 1l Eta with the BRS-3000. As to the proximity of the HX strip to the MiniMo regulator, I have never had a problem with it after a combined several hours of operation both in testing and in the field.

[David Thomas]

For more moderate temps, aluminum is great. Light. Cheap. It is used in all sorts of heat sinks for electronics. But copper handles a much higher temperature. I wouldn't expect aluminum directly in a pre-mixed blue flame to last. I'd expect it to melt because I've seen that happen frequently. Different point: if you're ever really pushing the limit of this approach (-20F? -40?), put a thin film of something between the copper strip and the canister. Heat sink compound would work great but is really sticky and messy. Toothpaste? Sunscreen? Ear wax? Any solid will be better than the air gap which is under 90% of the apparent contact area.

[Bob Moulder]

Yes, Aluminum melts quickly under a blue flame… been there, abandoned that. >>Any solid will be better than the air gap which is under 90% of the apparent contact area. David, I agree with your estimate that a large percentage (maybe even greater than 90%) of the apparent contact area probably isn't touching the canister directly, although I bend and shape it to try to match the radius somewhat. But it is interesting how well it works even with those gaps, and its performance is unaffected with temperatures down to below 0°F. During those long winter nights in the tent last January and February I gave this some thought. My conclusion is that the imperfect contact doesn't matter because the silicone insulator on the strap traps enough of the heat from the strip and holds it there long enough that it is transmitted to the canister efficiently enough by convection and radiation. What makes me think this is true — besides extensive direct experience! :^) — is noting how well water baths can work, despite the fact that they're imparting far less energy to the canister than this copper strip. I have also considered the possibility that the gaps might actually be desired, perhaps being the key reason that this system simply doesn't overheat.

[Jerry Adams]

I've only done aluminum a little. Maybe if I used it more it would eventually melt? If it's aluminum foil, yeah it melts, but if it's thicker, like 1/32 inch, and only the tip is in the flame, then it conducts heat away fast enough it doesn't get hot enough to melt. Even if the very tip melted a little, it wouldn't matter. Doesn't really matter that much, aluminum is only a little lighter. More a matter of convenience.

[Jerry Adams]

Aluminum – 1.75 x 7 x 1/32 inch – 0.5 oz Steel wire – 0.1 oz I haven't tried this particular model, but I've done similar versions. Probably pushes the lower limit temperature down 10 F or so. Now waiting for cold weather : )

[David Thomas]

Jerry: okay, I stand corrected. I guess the aluminum is thick enough, a big enough "pipe" to conduct heat away. After I posted, I saw RogerC has also used Al, also pretty thick. When I've melted Al in a flame, it was much long and not connected to a heat sink (the canister). And ambient temps were room temperature, not sub-freezing. Bob: you mentioned radiant transfer (from the metal strip to the canister). Painting the Cu will greatly increase radiant transfer compared to shiny metal. Just darkening it with a Sharpie would help. If it is "too effective" maybe in mild conditions, put a leaf between most of the contact area. Then trim it down in size when you get home (save weight!).

[Jerry Adams]

If it's too effective, just don't use it. If air temperature is greater than 25 F. When I've experimented, the problem is it doesn't heat the canister up very much. Never a problem over heating. I sort of gave up on it and switched to water bath method, but that's too fiddly. I'm going to try the strip again. If I could by using the strip, lower the minimum temperature from 25 F to 15 F, that would be all I'd ever need, because I don't like camping below that, everything gets too difficult and it doesn't get that cold around here.

[Bob Moulder]

Because this thread is something of a repository for info about my tests with the HX strip, I am linking to a post in another thread with the preliminary results of cold-weather operation with pure n-butane.

Bottom line: With the HX strip, a BRS3000-T operated just fine at 13°F (-10.5°C) using a canister containing only pure n-butane… NO propane and NO isobutane.

[Stuart R]

I’m impressed – good work!

What dimensions of copper strip did you finally settle on?

[Roger Caffin]

I will make a wild guess:

16 gauge (1.6 mm) SOFT aluminium 1″ (25 mm) wide could be almost overkill.
On my winter stove I use a heat shunt of 1.0 mm aluminium 15 mm wide. But I found that thinner (eg 0.5 mm) is too thin to conduct enough heat.

As others have pointed out: aluminium may not be as thermally conductive as copper, but it has a much lower density. On a scale of ‘energy conducted per gram of weight’, I find aluminium beats copper easily.

Cheers

 

[Bob Moulder]

Thanks, Stuart!

The copper strip I’m using now is 1 inch wide and .020″ (20mil) thick.

Well if you like that, it gets even better. :^)

Yesterday morning the temperature was considerably cooler (1°F, -17.2C) and I did some tests with other stoves. It was a lot less windy for this round of burns than for the first round with the BRS3000-T. This time there was 30.2g of fuel remaining in the canister (100g capacity), which had remained out on the deck all night.

Still using pure n-butane — no propane or isobutane.

In each test the canister was in the cozy and warmed up for 1-2 minutes by breathing warm air over the top of the canister with hands (gloved!) cupped around it.

In each case the stoves started, but it took a little more loving care and coddling to nurse the thermal feedback loop to life, which was done by shielding the burner with hands curved around to block out wisps of wind. In every case the best technique for setting the fuel valve is to turn it wide open.

The additional stoves tested were the Snow Peak GigaPower, Primus Classic Trail (a big-burner, 8oz behemoth that no self-respecting ULer will admit to owning), and, fittingly for this thread, my beloved JetBoil MiniMo.

First, the bad… the one that got away… but I’m not crying too much because it was the heavy Primus Classic Trail. The burner head is just a giant heat sink and the spread-around flame pattern does not impart nearly enough energy to the HX strip to get/keep the feedback loop going.

SP GigaPower, no problem, very similar to the BRS3000-T.

JB MiniMo, leaps of joy because my cold-weather darling delivered in spades! It helps that the pot support structure and the fins and shroud of the MiniMo HX pot helped in this regard, because they both blocked the wind somewhat, and they helped keep some convective (and maybe a little radiated) heat near the top of the canister.

Durn tootin’ it was pretty cold, lol

Primus Trail Classic. Don’t worry, it dies soon enough

SP Gigapower, performed very well, similar to BRS3000-T

And still the champ and my fave, JB MiniMo (with that stylish FourDog Stoves topper!)

[Jerry Adams]

1 inch wide, 0.02 inch thick copper

I used 1.75 inch wide, 0.02 inch thick aluminum, which is 1/2 the thermal conductivity, but since it has bigger area to radiate heat to the air, it’s probably worse.  I should try 1 inch wide and 0.04 inch thick

copper is 3.3 denser than aluminum, so it should weigh 1.65 more for the same thermal conductivity

[Bob Moulder]

Jerry, what does your aluminum strip weigh?

My HX strip weighs 16.9g, a skosh more or less depending upon what burner is being used. This for the 100g canisters. The strips for the larger canisters weigh a little more, maybe about 20-21g.

[Anonymous]

Whether you’re using Cu or Al, wouldn’t it make sense to have the strip thicker and narrower at the top, and then gradually shape to thinner, but wider on the part that is going over the canister?

[Bob Moulder]

That’s a good question, Justin, and something I’ve thought about.

Plenty of scientists here to answer it, but my intuitive answer is that heat might behave like electrical flow and lose some efficiency of transfer if the ‘channel’ is narrowed.

I stand by, bent over, awaiting my intellectual @$$ kicking. :^)

Whatever the verdict, I’m getting very good stove performance with n-butane at 1°F with a simple, 1oz setup, which speaks for itself. I like the fact that anybody can make one with a simple instructional.

If someone can make it work better, I’m all for it. I don’t think it’s going to make a difference. However, Gary Dunckel has been testing with wider strips, and I have a feeling that those might be better for super-cold applications. And I also like Gary’s idea of using carbon fiber felt for the cozy because when I did the “idiot test” to evaluate the safety of the system when abused with too-warm ambient temperatures I did get some deformation (melting) of the foam cozy.

The bottom line is probably going to turn out to be practicality. It’s easy to buy standard-width strips, but it would (for the average person) be extremely expensive to produce custom-spec’d pieces. I have some 2″ wide stuff to experiment with and might very well give your idea a shot at some point. Copper is very easy to work with, and I do have a good digital micrometer, a metal snipper and a hammer. ;^) Gimme your spec’s, bro!

[Anonymous]

Ha, i don’t have any specs to offer Bob.  I was asking a very tentative question and hoping that the science folks would have some insight into whether or not it would make sense.  I don’t have any certainty about it and since i don’t own a single upright canister stove–i won’t be testing this any time soon myself.

Besides lack of training/education/experience, the thing that most separates me from the average scientist type, is that i don’t have much filter and tend to talk/wonder out loud, whereas the average scientist type tends to be a lot more cautious and keep things closer to the chest until they are more certain about it.  As to common or average psychological reasons why, seems to be a combo of a desire/need for accuracy, and also a great, deep social and personal fear of ever being wrong or inaccurate.  But that’s the philosopher in me talking ; )

[Author]

Posts