The Joule Thief: Yet Another HX Pot

[Tyler R]

I posted this in Jetboil Stash thread and realized I was probably steering it off topic with my post so here it is in its own thread for y’all to ask questions, love, hate, give feedback/suggestions, share ideas, etc.

The Joule Thief:

A heat exchanger pot that weighs 2.9 Oz (with lid) and is 17% more efficient than a Jetboil Stash System (indoor test) when paired with a BRS 3000t.

Inception:

I’ve been lurking on this forum for some years now, but out of all the threads I’m always finding my way back to Roger Caffin’s article and Colin Kruser’s thread on Heat Exchanger Pots. The complexity of the problem and the theoretical ways to attack it have always intrigued me. This past summer after an extended JMT hike, I finally decided this could be a fun project if it means I got a working pot out of it.

The goal going into the project was to create a HX pot that answers the question Roger posed in his article “Is a heat exchanger pot worth the weight?” with a yes. Because IMO, most HX pots are only worth the extra weight in hyper specific circumstances.

 

Efficiency of The Joule Thief:

So efficiency is tricky, and I’m no expert on the subject, but I choose to measure it in terms of Joules absorbed by the pot per gram of fuel burned. I choose this metric because of the following reasons.

• its not dependent on maintaining a constant water start temperature across all test
• its not dependent on maintaining the same elevation across all test
• its not dependent on maintaining the same volume of water across all test
• and it doesn’t require you to go all the way to a boil

Thus, I think it is a finer method for strangers to compare cook systems across the internet vs the traditional grams of fuel burned to reach a boil method most of us have been using for years now.

I made a video on how to use this method here.. Its not without its flaws but I do think its an improvement over the common alternative methods. PLEASE let me know if I have made a mistake in calculations or an error in fundamental understanding. As I said, i’m no pro.

But to the point, the Joule Thief absorbs 33,514 Joules per gram of fuel burned when paired with a BRS 300T stove. That is a 17% increase of efficiency over a similar system, the Jetboil Stash at 33,752 Joules per gram of fuel.

 

Construction:

• The Joule Thief HX is 3D printed Aluminum made via the Selective Laser Sintering process (so no binding agents are used).
• The carbon fiber wall is infused with a high temperature epoxy and cured around a mandrill
• The CF wall is than coated with a top coat and epoxied to the Aluminum HX and Aluminum lip ring using an epoxy that is Title 21 FDA CFR 175.105 complaint for food coatings and also happens to be great at joining to materials with dissimilar coefficient of thermal expansions(its proven wonderful so far!). I also fill this epoxy with microglass spheroids to create an electrically insulative layer between the CF and the aluminum to protect against galvanic corrosion.
• The inside walls of the pot are then coated with an additively cured silicone coating that is FDA compliant for direct food contact. I put the pot on a homemade centrifuge and dispense a premeasured volume of silicone to achieve a .2mm coating. The pot is left on the centrifuge for the silicone to cure under the centrifugal force. This coating leaves no epoxy exposed on the inside of the pot.
• I then bend Titanium handles out of 2.4mm titanium welding rod using a cheap amazon wire bender. Next these handles simply popped into the handle holes, then a Joule Thief is born!

 

Longevity: 

I just finished up thermal cycle testing with two of the pots. 1 cycle consisted of performing a boil, shaking the pot vertical with hot water in it 20 times (trying to simulate handle loads on the joints while epoxy joints were hot), than dumping the water and cooling in a freezer to below freezing. the point of thermal cycle testing was because I wanted to see if there would be any signs of delamination between the carbon fiber and the aluminum HX, or between the carbon fiber and the silicone coating. One pot endured 280 thermal cycles before someone convinced me to sell it to him. My own personal pot just finished up my goal of 450 thermal cycles. I chose 450 because I wanted to see if it could survive the quantity needed for an average PCT thru hike, assuming 150 trail days and 3 boils a day. Both pots made it to those number with no signs of delamination, so i’m really stoked about that!

Here is a video I made of the pot. A few things said in there don’t apply anymore, but the general gist is there.

 

So, is it worth the extra weight?:

To answer this question, I made a fuel consumption calculator that is based on the Joules per gram metric of pot/stove combo efficiency mentioned above. Then I compared it to a sister system without an HX. But in short, I personally think it is worth the weight!

Joule Thief with BRS 300T vs Toaks 550ml with BRS 300T.

This graph shows system weight on day one vs the trip duration. Blue line is Joule Thief, red line is the Toaks 550ml. The conditions for this calculation are 9,000 ft elevation, 51 degree F source water temperature, and 950 ml of water boiled per day.

As you can see, it’s your typical step function you encounter with HX pots. They often can only save you weight in the scenarios where it saves you from having to bring the larger fuel can. With the Joule Thief though, these margins in trip duration cenarios are much wider. And the weights are so close, that the added 14 grams even on trips where its efficiency isn’t saving you from a larger can, may be worth the benefit of faster boils(near ondemand coffee folks!).

What about Custom Filling your own cans?:

This is where the Joule Thief really shines. If you have gotten into the practice of filling your own cans, than the Joule Thief is actually the lighter option over the Toaks 550 in all trip durations due to its efficiency.

 

Of course it’s not always this simple, sometimes you have to factor in weight carried per day as well. A more efficient pot/stove combo burns less fuel, which ends up being more weight your carrying in fuel if your not custom filling your cans. Here are two graphs to show this. Both generated using the same environmental conditions mentioned above. They graph system weight per day of the Joule Thief vs the Toaks 550ml (JF in Blue, Toaks in Red).

For these graph, I purposely chose a trip duration of 5 days because it shows something interesting in both cases. If your buying fuel cans full, the efficiency of the Joule Thief ends up causing you to carry more weight per day over those 5 days because its not burning as much fuel. And so although the start weights of the two systems are similar, they diverge greater each day.

And if your custom filling your cans, we can see the Joule Thief is the lighter complete system on day one, but eventually becomes heavier as it sips its fuel each day and the toaks overtakes it (or undertakes it i guess) and both systems reach a constant weight on day 5 when they run out of fuel.

 

A Note on the Graphs/Calculator:

If you like those graphs, and want to compare your own stove system against others you own using these, I made the calculator spreadsheet that produced those graphs available to mess with. I actually go over it and how to use it in the video link I left above when talking about the Joules per gram metric of efficiency. You can find a link to that calculator spreadsheet in the video description.

 

Health Concerns:

This pot uses carbon fiber, there are health concerns posed by this both because of toxicity of different types of epoxies, and the chance of tiny CF fibers abrading off the material and into your food if the top coat is ever abraded away down to the CF core.

I chose combat both health concerns by coating the carbon fiber and epoxy joints in direct food contact FDA compliant silicone. A widely industry accepted food safe material. As an extra precaution, the epoxy used as the top coat of the CF and joiner to the Aluminum parts is Title 21 FDA CFR 175.105 complaint for food coatings. I decided to use this just in case a part of the silicone coating were to ever delaminate during a trip and I couldn’t recoat it in the field. So far the silicone coating on my pot has been holding up way better than I expected, and I havent had the need to recoat it yet, but I went the extra mile to source the epoxy out of an abundance of caution.

As an extra remptive precaution, I have switched to using a bamboo spoon versus a titanium spoon. The idea here is that it will cause less abrading the on the silicone over time. Giving me longer service life between recoatings.

Another thing I need to look into is getting the HX hard anodized, as untreated aluminum isn’t bad for you per say… but it’s also not recognized by experts as being anywhere near the best either.

 

If you made it this far, wow you read a lot! thanks for your time and I look forward to all aspect of what will be discussed here!

[dirtbag]

I have no idea what you just said! But i think I like it!!! What size pots can you make??? That is my question.. and will you sell them?

[Pithawat V]

This is really cool, and thanks for being open about how you collected the data. Agreed joules / g is a much better metric than boil time. I’m a stove nerd so looking forward to more development in this area!

[John K]

This is quite interesting!

[Jerry Adams]

wow, amazing job.  Are you planning on selling these or is this just a project to amuse yourself?

since you took the time to post this I’m trying to figure it out.  I think I generally get what you’re saying.

did you make the aluminum lip ring?

how are the handle holes constructed?

what is the weight of a joule thief plus stove?  and a Toaks plus stove?  and how many grams of fuel per day for each?  On the plot “initial start weight given a trip duration” (filling canisters) it looks like for 1 day duration they both weigh about 135 grams.  I’m trying to figure out what that 135 grams is derived from.

“But to the point, the Joule Thief absorbs 33,514 Joules per gram of fuel burned when paired with a BRS 300T stove. That is a 17% increase of efficiency over a similar system, the Jetboil Stash at 33,752 Joules per gram of fuel.” – 33,752 is not 17% greater than 33,512.  Is there supposed to be any relationship between those Joules per gram numbers and 17%?

Sorry, I have no “hates” but Joules per gram of fuel is not an intuitive value.  Grams of fuel to heat 500ml of water 80 degree C is more intuitive.  You definitely have to factor in amount of water and the amount of temperature it’s raised.  Maybe I’ll warm up to it : )

Too bad in ‘merica we don’t just bite the bullet and switch to metric.  Rip the band aid off.  It’ll hurt initially but eventually things will be better.

I assume you used 4 ounce canisters?

[Jerry Adams]

how many ml does the Joule Thief hold?

on your initial start weight plots there are these jumps in weight where for a particular duration you have to carry an extra canister.  I think that’s the biggest value of efficiency.  If your stove is more efficient which allows you to not have to carry an extra canister it will save some significant weight.  (if you call the weight of a canister significant : ) )

[Roger Caffin]

I suspect that the 33,514 is a most unfortunate typo, and should be 39,514, according to the end of the comments in the other thread.

Cheers

[Piney]

One pot endured 280 thermal cycles before someone convinced me to sell it to him.

Yeah, that was me :) Just starting to test it, but the Joule Thief is pretty awesome!! I am really enjoying following along with Tyler’s project – he is a very talented individual.

[Roger Caffin]

Well, I am very impressed by the level of care and detail which went into the making of the CF pots. Huge. :) :)

I note the comment that the CF should never go above 110 C in use: it will be limited by the boiling water inside it. Well, very true, although I would be a shade cautious about the very bottom edge of the CF, where it joins onto the aluminium HX. That needs testing over time. Mind you, the aluminium HX base should not go much above boiling point either.

There have been problems with the difference in coefficient of thermal expansion between aluminium and CF. Easton tried to make arrows and tent poles with CF wrapped around an aluminium core. They were not very successful as tent poles as they delaminated when cycled between hot weather and sub-zero conditions. Testing will be needed for this.

The use of a food-grade silicone coating on the inside of the pot is smart. That should resolve a lot of worries. It does mean that the inside of the pot is a shade more delicate than the inside of a metal pot when you go to pack stuff inside it. However, that can be managed. I wrap my stove in a little bit of old tea towel before I put it into my pot. Then I use the tea towel to wipe the pot dry after coking!

Where the idea runs into a (big) speed hump is the cost of manufacture. Laser sintering aluminium powder is fine for aerospace, but I can’t see it being cost effective in the consumer market. You can get Toaks Ti pots on ebay for <$40, and aluminium HX pots for about the same. Very competitive.

Could one do it by stamping? Maybe, but the dies would not be cheap. Much CNC machining and testing.

But a very impressive project!

Cheers

[Pithawat V]

Sorry, I have no “hates” but Joules per gram of fuel is not an intuitive value.  Grams of fuel to heat 500ml of water 80 degree C is more intuitive.

For reference: the energy to heat up 500ml water by 80C = Specific heat capacity of water (4.18 J/gC) * weight (500g) * temp change (80) = 167440 = 167kJ

Converting to grams

Joule Thief + BRS: 39kJ/g -> 167/39 = 4.2g

Stash: 34kJ/g -> 167/34 = 4.9g

[Jon Fong / Flat Cat Gear]

A standard powdered metal injection molding process should be cost effective and much cheaper than SLS.  Similar to injection molding followed by a debinder process and sintering.

[Roger Caffin]

Injection moulding may be cheaper, but by how much?

We need hard dollars costs.

Cheers

[Jon Fong / Flat Cat Gear]

Of course, it is a tricky question as you are talking about and mold which will be Capital Equipment.  The only way to answer the question is to develop a business model that estimates sales rate as well as cost as well as how long you plan to sell the device.  If I recall the article correctly, the SLS aluminum parts were north of $200 so the payback might be quite.

Let’s assume that the Joule Thief is mug nirvana and can sell for $150  And the Total Manufacturing cost for the mug alone is $40.  And assume that the sunk Capital equipment coat for the mold is $50k.  You would have to sell over 500 mugs just to recover the cost of the mold.

[CS]

Jon, do you have an idea of how big the stove market is in total? Or does it just get lumped into outdoor stuff?

[Tyler R]

I suspect that the 33,514 is a most unfortunate typo

Aww darn… it is a typo! Thanks for pointing it out Roger.

[Jon Fong / Flat Cat Gear]

I have no idea how big the market is.  UL stuff is a very small market and the Joule Thief idea is probably in that market.  REI level stuff is huge, unbelievably huge.

[Tyler R]

Hi Jerry! Thanks!

1. The aluminum lip ring is also 3D printed. I don’t have access to many tools or machines, so the SLS 3D printing process has been a godsend in allowing me to iteratively design the pot with little overhead cost.
2. The handle holes are apart of both the HX and the lip ring, they are integral to both bodies and printed as a hole when made.
3. The weight depends on the stove you use, but a BRS 300t goes for about 24 grams. The current iteration of the Joule Thief plus the BRS is 109 grams and the same stove plus a toaks 550ml is 95.15 on my scale.
4. With my elevation and start water temperature of 71 degrees, the JT boils 2 cups of water with 3.82 grams of fuel.
5. Something is a bit off about the 1 day mark weight, I’ll have to look into that. But that weight should be the Joule Thief + BRS 3000t + 19 gram fuel can + fuel. Which should be in the 150 gram range for both of those so something isn’t right. I made the mistake of opening this sheet on the R/ultralight subreddit and spent many hours fixing things people were messing with for some reason. I’ll dive back into the calculator later and see what’s going on. Thanks for pointing that out!
6. Yeah, as Roger pointed out, a typo on my part.
7. I agree! its not an intuitive metric by any means. It tells you next to nothing about your fuel consumption without doing some math. But it provides a consistent means to compare efficiency of dissimilar systems via test performed by different individuals with different elevations and different source water temperature.
8. Totally agree! We need phase into metric! I’ve always hated fractions. haha.

 

[Tyler R]

how many ml does the Joule Thief hold?

wow I cant believe I wrote that entire post and never mentioned it, I sure can babble!

The Joule Thief is 550 ml, though it’s not to difficult to vary the volume by using a shorter or longer carbon fiber sleeve.

[Roger Caffin]

Forgive me for asking, but to whom does the SLS printer belong?
(Some jealousy if it is yours!)

Cheers

[Tyler R]

Thanks Roger! The stuff coming from you and Jon mean alot to me, I’m an admirer of both your contributions to the forums.

delaminated when cycled between hot weather and sub-zero conditions. Testing will be needed for this

This has been one of my biggest concerns too. I just finished up thermal cycle testing my own pot 450 times (I briefly mention the hacked together procedure above), but that is a case study of N=1… so not the embodiment of confidence. But it was my goal so I am very much pleased it has survived that long with no signs of delamination. I don’t think I’ll continue thermal cycle testing till I can find a way to automate the process… I’ve done a ton of boils for the last 2 months and it got old on the 2nd day. Once I figure that out, I’d like to run the thermal cycle testing till destruction.

Where the idea runs into a (big) speed hump is the cost of manufacture. Laser sintering aluminium powder is fine for aerospace, but I can’t see it being cost effective in the consumer market. You can get Toaks Ti pots on ebay for <$40, and aluminium HX pots for about the same. Very competitive.

Yeah this is where i’m currently sitting with this project. I’m a college student with no money for tooling, and way out of my comfort zone when it comes to anything business. But I would like to try and sell these. They cost about $200 to make (it’s been an expensive hobby for this college student), which means were looking at $350-$385 to make a safe margin after running some numbers(or so i’m told by some business friends).

So my current plan is I might start offering to make these for people who are willing to pay the premium and are interested; with the hope there is a large enough demand for this hyper niche to support the cost of me exploring different avenues of production methods. Which would hopefully one day lead to a design that greatly reduces the per unit cost after amortization of the tooling for whatever that different production technique may be. I have to much sunk cost into this hobby as it is, so I don’t foresee myself going any further until I can find a way to at least make is self sustaining as weekend hobby if not profitable.

[Roger Caffin]

$$ Ouch! $$

My GUESS is that stamping the base and the base ring out of sheet Al and then compression-fitting the ring to the base would have to be the only financially viable route. Once you have the dies the cost of stamping the parts is trivial.

You could even do the stamping with a single hydraulic press., one at a time. I have done that once to make a base for one of my stoves. I made the very simple dies out of well-aged Australian hardwood, which may sound a silly idea if you have never met the stuff. Strictly limited production quantity though.

Can you write g-code, or is it all CAD/CAM? Just curious.

Cheers

[Tyler R]

Thanks for the suggestion Jon!

I have actually looked into this, and ignoring the tooling cost, the limiting factor is the thinnest possible walls one could make using this method. Every manufacture I contacted so far couldn’t do anything less than 1.5mm thick for aluminum. The Joule Thief HX is .6 mm thick and represents 60 grams of the 83 grams the entire pot weighs. Even if I changed the geometry of the design to allow for the injection process, they will still need to be machined down to .6mm thick to achieve the same weight. The injection molding + added machining process may actually be more cost prohibitive than the SLS printing process. Do you know of another way to uniformly decrease cross sectional thickness? I was toying with the idea of abrading methods like sand blasting, or chemical etching, but I don’t know much about those processes or if that’s a practical route to take or not.

[Tyler R]

Forgive me for asking, but to whom does the SLS printer belong?

Oh I wish it was mine! lol. No I have been working with an AM broker to help source a manufacturer who has the ability to print the part. I have about 6 failed prints from various shops with an SLS machine, I’ve spent a lot of time sourcing someone who has a machine able to print unsupported .6mm walls with the aspect ratio they have for the monthly budget I allowed myself during the iterative design phase of the project. So I wont give away my secret just yet, but I promise to completely open source the project with documentation (including prefered manufacturers) if I decide bringing these to market is a lost cause.

[Jon Fong / Flat Cat Gear]

Given the thinness of the materials, you might look it making out of sheet aluminum and dip brazing.  I’ve seen many examples of steel parts done this way but not aluminum.  NC punching can be really cheap in qwuantities, I do not know the cost of dip brazing.  Good luck.

[CS]

Regardless of the market viability, I think it would be great if your normalized measurement process was adopted and formalized. That would be a win.

[Author]

Posts