Jetboil cold weather heat exchanger

[Jerry Adams]

my aluminium strip is 1.75 x 7 x 0.02 ” and weighs 0.46 ounces = 13 grams

if the strip has more cross sectional area – that is, is wider or thicker, there will be less thermal resistance, so more heat will get to the canister

if you look at how tip of strip gets hot from flame, loses heat to the air along it’s length, and how the heat is conducted to the canister it’s rather complicated.

 

 

[David Thomas]

Really, Jerry!  Yours is 7 inches long and 1-3/4 inches wide?

TMI.

[David Thomas]

Okay, now for a response that won’t get deleted as soon as the Australians wake up.

My sense is that any bit of metal sticking into the flame will get to about the same temperature.  And a “rounder” or “squarer” conductor below the flame won’t lose as much heat as the thin stock that you are using.

You don’t need much surface area in the flame and prefer minimum cross-sectional area between the flame and the bottom of the canister.  So I’m thinking of an inverted “T” in which the inverted stem is round or square and then transitions to a square inch or so of aluminum / copper for reasonable HX to the canister.

I recently replaced an igniter for a gas oven and the wire leads had a super-high-temperature woven insulation slipped over them.  I’ve seen similar loose-fitting insulation over wires in toasters and hair dryers.  Something like that over the vertical stem would reduce heat losses and let you use less total metal for the same effect.

[Jerry Adams]

ha, ha, ha,…

that makes sense the tip of the strip gets the same temperature from the flame

I wonder how much heat is lost to the air, along the strip, vs how much heat goes to the canister?

The thermal resistance between strip and canister is much less than strip to air, but the surface area is small, compared to the entire strip

[Bob Moulder]

I guess I should have included that…

Mine is 5-7/8″ x 1″. Seems so inadequate compared to Jerry’s. But it’s made of copper!

Serioiusly though, the distance from burner to canister is a smidge over 2 inches for most of the upright stoves, so, while the added insulation wouldn’t hurt, it probably wouldn’t help that much, either.

 

[Bob Moulder]

I got my Primus “Summer” gas in the mail today from exxpozed.com, a German online retail outfit.

I ordered 5 canisters and there were only 4 on the box. Great. And of course the extra shipping charge for “hazardous materials”, but they just threw them in the box with some of those plastic balloon thingies as if they were shipping a pair of socks, and the loose canisters clanged and banged together all the way from Germany, with each of them sustaining multiple dents.

What a joke. Be forewarned, don’t do business with exxpozed.com! I’m going to email with a complaint, but that is as far as I’m going to take it. Not worth the hassle, and it’s not as if I have a screaming need for “summer gas” anyway.

[Stuart R]

I should imagine that the shape of the conductive strip has only a minor effect. The thing that amazes me is that there is enough contact between the strip and the canister to transfer a useful amount of heat. It would be worth trying a thermal pad, of the type used between processors and heatsinks, to see if that improved the heat transfer.

BTW, n-butane in canisters is not “pure”. There will be up to 5% (or more) impurities (butenes, propane, propene) but not enough to make much of a difference.

[Jerry Adams]

If the strip loses significant heat to the air along it’s length, then a wire would be better – smaller surface area

#6 wire is 0.16 inch diameter which is the same cross sectional area as 0.02″ x 1″

leave the insulation on it and it will provide some insulation (except the tip where it gets too hot)

Flatten the wire where it contacts the canister for better conduction of heat to canister

or #3 wire of aluminum would weigh a little less

I think this is common practice, for example, Bob Gross mentioned it a couple winters ago

[Bob Moulder]

I should imagine that the shape of the conductive strip has only a minor effect. The thing that amazes me is that there is enough contact between the strip and the canister to transfer a useful amount of heat. It would be worth trying a thermal pad, of the type used between processors and heatsinks, to see if that improved the heat transfer.

I’ll check into getting a thermal pad and test with it. But since I don’t have a proper lab and can’t replicate the ambient conditions for temperature and wind, my impressions are somewhat subjective, but I am developing a pretty good “feel” for things with all the testing. Although sometimes the difference is so obvious it jumps up and grabs you by the ears. :^)

BTW, n-butane in canisters is not “pure”. There will be up to 5% (or more) impurities (butenes, propane, propene) but not enough to make much of a difference.

Point well taken! I just meant it wasn’t a blend with isobutane and/or propane. As cheap as this stuff is, there’s gotta be some junk in there.

[Bob Moulder]

OK, I just ordered THIS on eBay. Hopefully it’ll arrive in time for the next cold snap.

[David Thomas]

It not that it is necessarily “junk”, but that when fuels are distilled, the refinery only aims to meet certain metrics of vapor pressure, heat value, etc; not purity for purity’s sake.  There are some refineries that have additional separation equipment to produce fairly pure streams of hydrocarbons to be used as chemical precursors, especially for plastics production.  An example would be the separation of benzene for producing the intermediary ethylbenzene for the production of styrene.

Building the taller distillation columns and operating them with higher heat inputs and higher reflux rates costs more for that better separation and is kept to a minimum when producing fuels.  If you want a fuel with a higher vapor pressure, you don’t add less pure n-butane to more pure iso-butane and more pure propane, you simply take the distillate off the column a little higher up.  You’ll get more higher vapor pressure compounds in the mix with a smaller fraction of less volatile ones.

[Bob Moulder]

Thanks, David.

I’m guessing this stuff is pretty low on the column, and probably toward the higher end of Stuart’s estimate for impurities!

Re thermal pad testing… although I don’t have a proper lab, I do have 2 of the BRS3000-Ts and can run them side-by-side to compare with/without thermal pad.

[Jim Sweeney]

What about soldering some insulated # 6 wire to a patch of much thinner copper?  That way there would be no hammering needed (copper work-hardens quickly, which might makes getting a smooth contact surface with the canister difficult in any hammering scheme ), there would be insulation of the wire except near the flame and at the solder joint, and the thin copper (say ~1″ wide, or even wider) could run sideways to wrap further around the canister.  I’m pretty sure that a solder joint is a good thermal joint — someone else no doubt has more experience here — and that the solder joint would easily withstand the temps involved, as it would be in the environment of the canister, not the flame.

[Jerry Adams]

Great minds think alike

I was trying that a bit, but I have a soldering iron for electronics so it doesn’t get it hot enough.

Except I used #10 copper wire which is as big as I have, flattened it with hammer, then took two more pieces that I also flattened and put them in parallel, so there are three widths of flattened wire touching the canister.  Solder the three together.

I have to get out the propane torch.

[Roger Caffin]

Okay, now for a response that won’t get deleted as soon as the Australians wake up.

Yawn. I haven’t woken up fully yet. I’ll think about that when I do.

Cheers

 

[Jerry Adams]

I used propane torch for plumbing soldering

I put the 3 flattened copper #10 wires on an empty canister just to position them and soldered them together

Predictably, the copper wires are now soldered to the canister – it must be steel not aluminum

Not what I intended, but if I now refilled from another source of butane that would work pretty good – very good thermal contact between copper wire and canister.  Wouldn’t need three pieces of wire though.

 

[Jerry Adams]

A couple back of envelope calculations

Steel and copper glow red when hot, not aluminum, based on color charts for color vs temp, maybe the tip of the strip is 800 C.  Maybe it’s 50 C next to the canister at the other end – I touched it and it was warm but not hot.

I burn 1/4 ounce of butane in 5 minutes.  That’s 7 grams in 300 seconds.  Heat of vaporization is 386 Ws/g.  I need 9 Watts of power to vaporize my fuel.

I use canisters that are 8 ounces of butane and 5.4 ounces of steel.  Maybe in the 5 minutes it heats this up 10 C – it starts slow and speeds up after a few minutes of heating.  Specific heat of butane is 1675 J/kgK.  Steel – 500.  It’ll take 4589 J to heat up the canister 10 C.  If it does it over 300 sec, that would be 15 W.

If the strip is 5 inches long and 1 inch wide, that would be 10 in2 area (both sides).  Maybe the average temperature of the strip is 300 C.  There’s an air layer of about 1/2 inch around the strip providing insulation.  The thermal conduction of air is 0.024 W/mK.  We are losing 3.6 W due to conduction to air.

So, there’s a total of 27.6 W flowing down the strip.

The cross sectional area of the strip is 0.02 in2.  It’s 5 inches from tip to canister = 0.075 m.  The conduction of copper is 401 W/mK.  If 27.6 W is flowing, that would produce a 697 C temperature drop from the tip to the canister.  That’s pretty consistent with my guess that the tip is 800 C at tip and 50 C at canister.

According to this calculation thermal conductivity from strip to air is a small effect compared to heating and vaporizing butane.

My aluminum strip was 1.75 x 7 inches – about 5 inches from tip to canister.  1.75 times bigger than Bob, but conduction is half, so it’s a little less.  So I would expect a little more temperature drop.  Bob’s seems to work a little better than mine so maybe this explains it.

Radiative heat loss is emissivity * sigma * area + (T strip ^ 4 – T air ^ 4).  Maybe emissivity is 0.5.  sigma is 5.67e-8 W/m2K4.  Maybe approximate strip as a 800 C area of 1 inch^2 = 0.0006 m2.  Air is 0 C.  Then radiative heat loss = 22 W.   If it was 600 C it would radiate 9 W.  400 C – 3 W. So, only the hottest part of the strip will have large radiative heat loss.  So, most of the radiative heat loss is right at the tip where it’s hot.  This is consistent with observation – it’s very hot at the tip but quickly cools down further away from tip.

Copper that’s been heated and heavily oxidized has an emissivity of 0.8.  Heavily oxidized aluminum is 0.25.  So maybe aluminum would be better?  But maybe this isn’t important, because this is only important at the tip where it’s really hot.

One other question is how much heat is radiated from canister to the air.  Let’s say canister is 0 C and air is -10 C (14 F).  My 8 ounce canister is 4.25 inch diameter and 3 inches high so total area is 68 in2 = 0.04 m2.  The boundary layer of air around it is 0.5 inch = 0.0125 m.  Conduction of air is 0.024 W/mK.  So it loses 0.8 W.  This is insignificant compared to vaporizing butane and heating up canister.  Don’t worry about insulating your canister.

 

 

 

[Bob Moulder]

That must be a big envelope. :^)

I’m definitely more of an empirical type.

[Jim Sweeney]

Thanks Jerry.  That’s a great result, not having to insulate the canister.

[Roger Caffin]

Hi Jerry

Well, interesting calculations, but in the case of aluminium definitely slightly wrong. The melting point for aluminium is 660 C, and it goes quite soft above about 550 C. 800 C is out of the question.
Such a heat exchanger strip does not melt or even go squooshy soft in a stove flame on my Winter Stoves. With adequate thickness, the heat is carried down the strip fast enough. See Part 5 of my
Winter Stove articles, where John Linn measured the temperature of the HX strip with an IR thermometer.

Cheers

 

[Bob Moulder]

Thanks Jerry.  That’s a great result, not having to insulate the canister.

I would recommend some actual testing on the patio and in the field before committing to this for a big trip. Larger (8 oz) canisters with low fuel levels need something that provides more heat retention, in my experience.

I have not had a chance yet to test the larger canisters with n-butane only, however. I didn’t have an empty 8 oz canister available for refilling, and my lab warmed up. ;^)

[Jerry Adams]

“… not having to insulate the canister…I would recommend some actual testing on the patio…”

I agree, just calculation with assumptions.  Easy enough to test.  Maybe the weight and bulk of a canister cozy is small enough it doesn’t matter.

“Well, interesting calculations, but in the case of aluminium definitely slightly wrong. The melting point for aluminium is 660 C, and it goes quite soft above about 550 C. 800 C is out of the question.”

Okay.  I was just basing it on how hot steel gets in the flame.  Probably not that important because it quickly gets down to colder temps (100s of C) because of radiative heat loss.

Your heat shunt has to provide the same heat of vaporizing fuel as the Moulder strip.  Doesn’t have to heat up the canister and butane though, although that’s just a short delay in heating up.

Your strip is 0.047 inch thick.  That should provide about the same conduction as 0.02 inch copper.

How long and wide is your strip?

[David Thomas]

Nice calcs, Jerry.  Nicely presented.

To add another heat source to the discussion, for scale:

A mini Bic weighs 11 grams full.  So I’m estimating 4 to 5 grams of butane.  0.01 pounds.  x 19,944 BTU/pound.  = 200 BTUs.  Mini Bic duration = 33 minutes.  So 360 BTU/hour = 105 watts.

A flame delivers about 30% of its heat to a pot (maybe better for a cold butane canister due to condensation of water vapor from the flame) so 105 watts x 30% = 32 watts (or more) delivered to the canister.  So about two minutes to deliver the 4589 Joules to rise the canister 10C / 18F.

I should go cool off a few canisters and give it a try.  To all those who fret about directing a flame onto a fuel canister – jeez!, grow a pair.

[Bob Moulder]

 To all those who fret about directing a flame onto a fuel canister – jeez!, grow a pair.

LOL! Well, I now do this and it’s not such a daring trick, after all.

When it was 1°F the other morning, I’m pretty sure I could’ve held the flame on the bottom of that canister for 5 minutes with no problem. And this was the larger Bic, not the mini. Easy enough to feel if the canister gets warm. I actually held it there for about 10 seconds and with the n-butane it was barely enough to get a flicker out of the burner when I lit the stove.

[Jerry Adams]

video please David

and Bob

that is, the part where it explodes : )

[Author]

Posts