Alcohol Stove Comparison: Trail Designs (12-10 and Kojin) vs. Zelph (Starlyte and Starlyte Mod)

[David Hosmer]

Companion forum thread to: Alcohol Stove Comparison: Trail Designs (12-10 and Kojin) vs. Zelph (Starlyte and Starlyte Mod)

This review compares the boil time and fuel consumption performance of the Trail Designs 12-10 and Kojin alcohol stoves, and the Zelph Starlyte and Starlyte Mod alcohol stoves, under a variety of wind conditions.

[James Marco]

Hmmm, I didn’t see where you noted the type of fuel used.
Methanol has the lowest heat value at around 9800BTU/lb and Ethanol has the highest at 12800BTU/lb. Roughly a 3000BTU difference. Ethanol has around 25-30% more than methanol (depending on water content.) This could effect boiling points/flame temps and overall heat output rather drastically at the close results you indicate.

All are easily available, but like canister gas, not necessarily available at all camp stores or general stores/pit stops. Yellow HEET is pure methanol. Denatured “fuel” alcohol is usually around 50/50 but can vary a lot with different manufacturers. Ethanol is usually 190proof (95% pure since it forms a eutectic at that temp while distilling) and needs to be picked up at a liquor store.

Methanol boils at ~65C, Ethanol boils at ~78C. This causes Methanol to flash off faster than ethanol causing methanol to burn hotter and faster than ethanol. Good and bad.

The fuels are different enough to warrant a note and could effect the test outcomes. Otherwise a nice set of tests. Thanks!

[DAN-Y]

This video shows a comparison of the Starlyte and 12-10 stove and another one side by side.

(video removed)

 

[Ken Thompson]

In the wind tests were the bottom intake holes on the windward or leeward side?

[Dan @ Durston Gear]

An important variable that wasn’t controlled here is the distance between the stove and the pot. All of these tests used the same pot and cone, but the stoves differ substantially in height (from 0.8″ to 1.5″ tall). This has a large effect on performance. Generally speaking, as you increase this gap you get more power (faster boils) but less efficiency. This is something you want to design a cook system around.

Here is a graph showing how changing the gap between the stove and pot affects boils times and efficiency of the Starlyte (x axis is the stove-pot gap, y-axis of left plot is minutes, y-axis of right pot is grams of fuel).

As you can see, a 0.4″ increase in height (from 1.6″ to 2.0″) reduces boil times from 10 minutes to 8.5 minutes, but increases fuel consumption from 16g to 18g. This is basically the same difference, as you find for the Kojin vs Starlyte in the outdoor test (the Kojin is 0.45″ shorter than the Starlyte). This may explain why the Kojin was observed to have more power but lower efficiency. If you increased the Starlyte gap by 0.45″, it would have boiled a couple minutes faster but used a little more fuel – so potentially similar performance.

Having a short stove like the Kojin is a good thing, as it allows for shorter cones (lighter, more stable). So I like that it is so short. But it’s important to consider “setup variables” like the stove-pot gap separately from “stove differences” because it allows for a better understanding of how things work and thus how you can tune them. If you were to put these two stoves in another cone-pot combo that naturally had a quite a large gap, the taller Starlyte would likely burn well while the shorter Kojin would likely burn even more vigorously and get poor economy.  Or vice versa, in a cone-pot setup with a naturally small gap, the short Kojin would perform well, while the taller Starlyte would take forever.

It would also be good to have some replication in this testing. It appears that each test was only done once, which doesn’t let you identify if an anomaly occurred. Some of the results here look strange, such as the low speed wind test where the Starlyte Mod boiled faster than the Starlyte. This doesn’t make sense, nor does it match with your other results, nor the results of other testers.

[d k]

Thanks for this study.  I’m pleased to read of the virtues of the 12-10, since I’ve been eyeing the modded Starlyte but haven’t yet spring for one.  My solution for the leftover fuel in the 12-10 is to snuff the stove after cooking, with the bottom of a soda can I keep in my kit, then after a little cooling I use a plastic pipette to remove the leftover fuel and squirt it back in my fuel bottle.

[Daniel Collins]

I am disappointed in this test as being unscientific, does not meet criteria for DOE (Design of Experiments).

You cannot start with “approximate” water temperatures and compare to other tests with “approximate” temperatures. We are dealing with tiny amounts of fuel containing tiny amounts of BTU’s.

For that reason any testing is intolerant of variances. (Container/windscreen, water amount and temp., altitude, wind, and  type of fuel such as which brand of denatured alcohol).

On the other hand, I appreciate the time you spent to do this and share it.

I own three of the four stoves tested and am keenly interested in seeing an accurate test, which I have not had the time to do. Starting with my 83 degree faucet water doesn’t help either.

After recently receiving my Kojin I was disappointed in the first couple of tests. It boiled pretty quickly then ran out of fuel pretty quickly afterwards. This is with the 40 CC max fill. This doesn’t seem very efficient to me and seems to trade efficiency for speed.

I wanted to be able to have a longer burn for actual cooking beyond the initial boil. (Or at least lower consumption for the two cups of water boiled).

So I made a couple of different stoves based on the same design. The 3 oz stove required setting same as the 12-10 by lifting the pot higher and resting on the stakes. The second one is a 2 oz and with the wick material set (compressed as in the Kojin) to the same height as the Kojin. With the 2 oz stove, I got the same performance as the Kojin, but with a longer burn time due to the stove holding two ounces.

The 2 oz stove tapered to a simmer after the extended boil, whereas the Kojin simply flamed out.

Since I have not yet tested ( fully scientifically) any of these stoves, I would be happy to send them to someone who would do so.  Whoever does this can keep them of course. I’m not selling these stoves – I just want to see if I can improve on the Kojin for my own personal use.

My initial setup was the 12-10 for boiling and the Zelf Starlyte Mod for simmering if that is needed.

I prefer the screw top stoves, so I will eventually try to make a “simmer” stove.

I encourage you to repeat the experiment but with tighter controls, and list “time to reach boil” and time to fuel exhaustion/flameout. If I want to cook something beyond the initial boil, starting out with more fuel might affect the efficiency, since the unburned fuel may act as a heat sink, thus slowing the flash off/vaporization rate. It might even be more efficient if it is used only for boil, leaving whatever is left over for next time. In other words, an ounce of fuel burned in a stove that holds just an ounce may be less efficiently burned than in a stove that holds two ounces and filled with two ounces. Only time consuming experimentation will tell.

I will post photos of my stove prototypes if requested.

DC

 

[Gary Dunckel]

I am with Dan regarding the stove top-to-pot bottom distance. Although, there is likely a range wherein the efficiency is about the same, say 1.0″ to 1.5″. And, like he mentioned, using different height stoves with the same cone and pot will surely yield varying results. After a lot of testing, I have settled upon a 1.0″ distance, which I understand is rather consistent among many alcohol stove users.

I have had some experience with several commercial stove types (a number of Dan’s and a Stormin Stove from Norm in the U.K.), and I have found that they all have different ‘personalities’. Certainly each one would require a separate dedicated cone for best efficiency. Since I can’t see the waste of titanium foil to make a bunch of cones, I just figure out a way to modify each stove to work with my go-to Titan kettle cone. An example was that I needed to raise the Starlyte 1/2″ higher for its optimal efficiency. Setting it on top of an inverted wedding tin lid works great for that.

Your article was nicely written, and I think you addressed most of the key points regarding those various stoves. There are certainly numerous variables that will affect any alcohol stove’s efficiency. Among those are (1) starting water temperature, (2) ambient air temperature, (3) the presence or absence of any breeze or wind, (4) elevation above sea level, (5) wind screen design, and as Dan  and James alluded to, (6) the specific type or brand of alcohol used.

[Brad Rogers]

Interesting article. I have been using a Caldera Cone and 12-10 stove for ten years and have found that I can pretty consistently boil 12oz of water with 15ml of fuel.

I mostly use Denatured Alcohol but on fly out trips out west, normally by HEET as I can procure it in a smaller bottle and easily at a gas station on the way from the airport to the trail head.

[IVO K]

I mostly agree with Dan Durston’s take, however Dan is misstating facts by saying
“x axis is the stove-pot gap”. It is not.

The X-axis actually represented the distance between the pot’s bottom and the ground (on which ground the Starlyte burner was positioned) when these graphs first appeared here on BPL.com.
These graphs at the time helped me make and tweak my own cones for different pots, with whose performance I’ve been quite happy ever since.

I also agree that ignoring the pot to stove distance, and the starting water temperature, makes this study one of a questionable value. It is regretful that the researcher David Hosmer wasted time, effort, and funds to execute this study, as he clearly had no clue how such tests should be done.

[Drew Smith]

Hmm… I have to say your results with the Starlyte Mod are completely at odds with my experience.

When I first got it, I did a set of similar experiments – 1 pt water at 50F  in a Caldera Cone Sidewinder, in a couple of different Ti pots (0.9L Evernew and 1.3L Toaks). Fuel – which could be a critical difference, was denatured ethanol. Ambient temp 65F, no wind. Elevation 5400 ft.

Times to boil in four runs ranged from 9:20 to 10:10, roughly half of what you observed. Fuel use averaged 17g. I’ve used this setup hundreds of times on trail, and although I’ve never timed it, I think this is pretty representative of performance.

Prime suspects for the different results are fuel type, elevation, and pot type.

[Dan @ Durston Gear]

“I mostly agree with Dan Durston’s take, however Dan is misstating facts by saying
“x axis is the stove-pot gap”. It is not. The X-axis actually represented the distance between the pot’s bottom and the ground (on which ground the Starlyte burner was positioned) when these graphs first appeared here on BPL.com.”

Indeed you are correct. My memory is a little fuzzy after 6 years. Original thread is here in case anyone is interested. Either way though, the main point is that the stove-pot gap has a large effect which should be considered.

Also confusing things slightly is that the height of the Starlyte has changed over the years. It used to be shorter (0.8″) whereas now it’s 1.25″ tall apparently. So with an older Starlyte, I was using a ~2″ pot height to get a ~1.2″ gap, whereas now you’d need a 2.5″ pot height to have that same gap. With sidewinder cones, you’re limited with how high you can hold the pot (unless using stakes), so the old shorter Starlyte often allowed for more powerful configurations.

[Rod Braithwaite]

A gentle reminder of two important acronyms:

YMMV – your mileage may vary, and

IMHO – in my HUMBLE opinion.

The author was not offering a peer reviewed document. His results are valid in context and are perhaps akin to quality forum discussions, where useful experiences are shared, sans rigorously enforced experimental controls.

[Daniel Collins]

I just wanted to clarify my request for future testing: I don’t need “time to flameout”.

The OP shows how much fuel was consumed to reach boil and that is fine by me.

If I were to attempt a test , what water temperature would people want to start with ?

60 degrees might be good if we are replicating mountain stream water.

I am at sea level and I can set my house temperature for the ambient temperature.

I don’t think I would have time to do all of the wind tests.

I wouldn’t be able to supply all of the cool graphs.

My batting is pretty much the same as what’s used in the Kojin – the first prototype  was using ceramic batting from a jet engine, but I would love to find that unidirectional batting used in the Zelph as I think that is the key.

And any testing would list the distance from the batting surface to the pot lower surface.

DC

[Rod Braithwaite]

Daniel C. – for my own stove tests, I put two large “double boiler” pots of water in my fridge overnight. That way, I was able to start each test with two cups of cold water at a consistent and realistic starting temperature – perhap something similar might work for you?

[James Marco]

Well, I was using about 40F based on my tap water in winter, but that varied about 5F over the time snow was on the ground. The most consistent was to pour a full tray of ice cubes in a bowl and flood it with water. After about 3-5minutes, you get 33F plus/minus about a degree rather consistently. I just pour it off into a measuring cup for 16oz at a time just before lighting the stove.

Funny, but I was always getting between 5:30 and 7:00 minute times to boil with about 23g/~3/4oz of methanol or 20gm/~5/8oz of SLX or 19gm/~5/8oz of EverClear95. with a 12/10 stove starting with 40F water. Hmmm….I believe this was with a modded pot, though. Hard to remember, it was about ~9-10 years ago.

I also called a “boil” with a digital probe at 210 degrees, since at my altitude water boils at closer to 211. Boiling temp is NOT consistent. I didn’t want to get involved with all the potential vapor pressure so once it was sealed it stayed that way until boil temp. Then the stove was extinguished and weighed. There was always a bit extra fuel since I always added about an ounce or 28.5gm. Note that you need to WEIGH the fuel because alcohol has a density of about .79. Without allowing for this and just measuring with an “ounce” measuring device you will be about 20% off with fuel calculations. A fluid ounce (volume) will actually weigh about .79oz (mass.) Yes it varies a bit between types. They are only the same with water and some mixes.

Anyway, as an unrefined approach to testing stoves this was a great first attempt. I have fiddled with many different types and found that they don’t differ significantly. Chimney, pressurized, unpressurized, special air mixing, etc. the OP’s results don’t tell me anything new. His results were about the same for all the stoves once the setup was tweaked to optimize it. Alky is partially combusted already and just has low heat values compared with WG, Diesel, Butane, and Propane. Only the technique of using the heat and this will be more or less successful based on any one particular design.

[Dan @ Durston Gear]

Daniel C: I think 40F is a good starting point. A lot of mountain water is cold, so I prefer to optimize for these conditions, and then if I can find 60F water or higher, than a bonus.

A common error to avoid is repeated tests on the same surface. Any surface you cook on will absorb some heat, and then in subsequent test this can warm the stove so it burns more vigorous. This is a big part of why you’ll often observe a faster boil in subsequent replicates, if you repeat the exact test several times. I prefer to do my testing on a cement floor, but move over a foot or two for each test so you’re not using a warm spot.

[DAN-Y]

@dandydan…..I went to visit your link and was reminded what rand said about fuel efficiency…take a look:

[Eric Blumensaadt]

Thanks for a very thorough review. The graphs are the heart of this report and easy to understand.

I have a TD Sidewinder ti cone W/3 cup pot and 12-10 stove that came with it.

But I almost always use ESBIT tabs with a Brian Green style tab burner to get longer burn times. IT’s what I used for my rim-to-rim Grand Canyon backpack last November and several other trips in Utah and Las Vegas’ Spring Mountains. I just prefer ESBIT over alcohol but may try the Kojin alcohol stove IF it works well above 8,000 ft.

[DAN-Y]

Eric, throw away your esbit and get some of these:

[Rex Sanders]

Thermite matches apparently allowed in carry-on by TSA:

You go first :-)

— Rex

[DAN-Y]

I AM “KOJIN”, These I have created in the likeness of “STARLYTE” Queen of Stoves:

[Michael Ray]

I collected a bunch of cat food cans and made some with our Boy Scout troop.

FYI, sadly alcohol stoves have been prohibited in the Guide to Safe Scouting for several years now. One could try to argue these are “commercially manufactured”, but it’s clear what the intent is. I still use one when out with my son on our own trips.

[DAN-Y]

Fancy Feast cat food can stove. Most popular DIY stove.

[David Hosmer]

James Marko: For the testing in the article I used yellow Heet.  Thanks for bringing that up!

Ken T.: The holes were on the leeward side.

Dan Durston: I hear what you’re saying about the pot hight issue and not replicating each test.  The test was how each stove performed in the Caldera Cone and their hight is part of that.  Also, take note that I used the Stanco Grease Pot and that many others may use a different pot with different dimensions and made of a different material.  Because the Caldera Cone is made specifically for each pot, everyone who doesn’t use the same pot will have different heights between the stove and pot and thus different results.

As for replicability: yup, it sure would have been nice to have enough time to do each test multiple times.  Because this article is is just a brief overview and comparison I didn’t worry about it.

d k: Nice workaround!

Daniel Collins: In preparing for the tests I filled a jug with tap water and set it in my garage.  I tested the temperature before each test and did find a variation up to 0.02 in temperature.  Because it wasn’t exactly the same temperature for each test but the difference was minuscule I opted to use the adjective ‘approximately’ rather than exhaustively explain the minute temperature difference.

I’d be interested in seeing any results of your further testing! I did use this setup with the Starlyte to boil water for three scouts who forgot to bring dinner at our last campout.  The ambient temperature was around 65F (much warmer than the testing).  I boiled 2 cups of water for two kids sharing a Mountain House and set a 9-minute timer so that they would know when the rehydrating was done.  I had 2 cups boiling for the next boy’s potatoes before my timer went off. After all that, I still had enough fuel to start my dinner but had to add fuel after about 5 minutes.

Gary Dunckel: As I understand it, the optimal distance is in fact 1 inch.  If I’m not mistaken, when Trail Designs makes their Caldera Cone, they take the pot into account and cut the cone to achieve that height. If anyone is planning on switching up stoves or pots then I hope that they have the good sense to optimize their setup.

Ivo K: Sorry to disappoint you.

Drew Smith: I suspect that the largest contributor was ambient temperature.  With a temperature in the min 60’s I got much faster boil times.

Rod Braithwaite: Solid point, thanks for the acronyms!

James Marco: Sorry that I didn’t specify this in the original article, but the measurements are weights, not fluid ounces. You hit the nail on the head about boil temperatures.  Where I’m at water boils at about 194F.  For cooking, this isn’t very important.  For sterilizing, however, this can be very significant, especially as you continue to gain altitude!

Eric Blumensaadt: My elevation is only about 4.700′ so I can’t say how they do at 8.000′

Dan Y: That’s awesome!

Michael Ray: I’m aware of the prohibited use of alcohol stoves during BSA activities but have made several “homebrew” stoves with them so that they have an inexpensive and easy to use stoves on their own outings.

[Author]

Posts