Stove Efficiency

[atraildreamer]

Here is the article and spreadsheet that I, (with the assistance of Jason Klass), have been working on for the last couple of months. It is an attempt to establish a standard for testing and determining the efficiency of alcohol-fueled backpacking stoves, as well as generating numbers for fuel usage of these stove systems.

I am offering it to all of the major websites for
review and posting.

I would appreciate any comments and suggestions
regarding the article and accompanying spreadsheet,
especially on whether to posting as an article, or
thread.

Regards,
atraildreamer

[Skidsteer]

Thanks for the hard work, Atraildreamer and Jason Klass.

Excellent idea!

[atraildreamer]

If you find the article and spreadsheet useful, please post your test results here. We are interested in the performance of all types of alcohol stoves. :welcome

[highway]

A typo, perhaps in the below quote, page 1:

"As an example, let’s start with 16 ounces (2 cups), (1.25pounds) of water as a standard volume, 2 cups being a typical amount of water used by backpackers when cooking."

If one fluid ounce of water weighs 1.043 ounces by weight, then would not 16 ounces weigh 16.688 ounces or 1.043 pounds?

[rafe]

My comment would be that the main contributor to stove inefficiency -- wind -- isn't dealt with. IMO, that's what makes studies like this not so useful. Also, simple things like... the shape of the pot, the use of a pot cover (or not) can skew the results wildly. When wind is figured in, all bets are off, and the overall efficiency will depend almost entirely on your skill in setting up an effective wind screen.

[Skidsteer]

My comment would be that the main contributor to stove inefficiency -- wind -- isn't dealt with. IMO, that's what makes studies like this not so useful. Also, simple things like... the shape of the pot, the use of a pot cover (or not) can skew the results wildly. When wind is figured in, all bets are off, and the overall efficiency will depend almost entirely on your skill in setting up an effective wind screen.

So set up your stove tests using the windscreen and a box fan to simulate wind if you like. If all other factors are kept equal it should be a useful tool for comparison.

[rafe]

So set up your stove tests using the windscreen and a box fan to simulate wind if you like. If all other factors are kept equal it should be a useful tool for comparison.

For starters, it's not my test :) -- just my comments on someone else's. And yes, it's a cheap shot to criticise. In other regards, it's a nice piece of work.

There are simply too many variables for the "wind" test to be meaningful. I've done my share of sailing (and hiking) so there's one thing I know about wind: it's never constant. Pitching a windscreen can be just as tricky as pitching a tent.

[saimyoji]

To follow on terrapin's comments: I agree that in a real wind, setting up a static windscreen is pretty much meaningless. It must be dynamic, manipulated by the operator to compensate for shifts in wind direction. I can't tell you how many times the wind has pushed my windscreen out of place....

Any engineer will agree that there is an operation RANGE, same with stoves. Read up and study on the stoves' ORs, make a choice based on what kind of conditions you will likely encounter. There are rave reviews on alcohol stoves for the AT, but how many who climb Mt. Everest have them?

A place for everything and everything in its place.

[atraildreamer]

A typo, perhaps in the below quote, page 1:

"As an example, let’s start with 16 ounces (2 cups), (1.25pounds) of water as a standard volume, 2 cups being a typical amount of water used by backpackers when cooking."

If one fluid ounce of water weighs 1.043 ounces by weight, then would not 16 ounces weigh 16.688 ounces or 1.043 pounds?

Thanks for catching this error. :eek:

I rechecked at:

http://www.fourmilab.ch/hackdiet/www/subsection1_4_2_0_7.html

and found that one gallon (128 fluid ounces) of water weighs 8.345404lbs/gallon, the 1 fluid ounce would weigh 0.0651984 lbs. (8.345404/128=0.0651984)

Then:16 fluid ounces weighs 0.0651984 lbs/fluid ounce x 16 fluid ounces=1.0431755 lbs.

I'll be correcting the article and the spreadsheet, and have it reposted.

[ARambler]

It seems that a lot of good points are being raised without the comments being very useful. If someone proposed a real wind test I would want to know how the alcohol stoves operate with no wind.

1) A windscreen that is wrapped around the stove will be more effective than dancing your Nalgene upwind of the stove.

2) Based on the large effect of the relatively small variables we have seen so far, we must define all variable very carefully. I have heard that pot diameter, heat reflection from the wind-screen, and stirring the water can be significant. For home made stoves, complete assembly instructions are needed for others to reproduce the results.

3) I worry about efficiency calculated on 12 ml of alcohol. I think the instantaneous efficiency may vary as the fuel heats up, during the full burn, and as the flame dies down. Since 12 ml of alcohol will not boil 16 oz of water, 20 to 24 ml of fuel may be a better test. Maybe increasing the water to to 20 to 24 oz would be a better representation of bringing 16 oz totally to a boil.
Rambler

[atraildreamer]

So set up your stove tests using the windscreen and a box fan to simulate wind if you like. If all other factors are kept equal it should be a useful tool for comparison.

My intent was not to account for all conditions, just to get a baseline set of comparisons for a particular stove system. Then the system could be "tweaked" for maximum performance by changing the variables: pot size, fuel load, amount of water, etc. Even my limited testing gave some surprising results, and a lot to think about in stove system design. :-?


For starters, it's not my test :) -- just my comments on someone else's. And yes, it's a cheap shot to criticise. In other regards, it's a nice piece of work.

There are simply too many variables for the "wind" test to be meaningful. I've done my share of sailing (and hiking) so there's one thing I know about wind: it's never constant. Pitching a windscreen can be just as tricky as pitching a tent.

Thanks for the compliment! :D

It is impossible to account for all the variables (eg: wind) that can change the performance of a stove system. That is why I put the disclaimer to take plenty of fuel along in the field to account for possible performance changes.

Don't worry about the cheaphots...I've given out my share of them! ;)

[Skidsteer]

For starters, it's not my test :) -- just my comments on someone else's. And yes, it's a cheap shot to criticise. In other regards, it's a nice piece of work.

There are simply too many variables for the "wind" test to be meaningful. I've done my share of sailing (and hiking) so there's one thing I know about wind: it's never constant. Pitching a windscreen can be just as tricky as pitching a tent.


The spreadsheet is set up to compare one stove against another with all things being equal so much as that is possible.

If you want to test the wind resistance of stove A vs. stove B, then add a box fan to the test for both stoves and test. It's a reasonable assumption that if stove A outperforms stove B in the box fan test with all other factors being equal, then stove A would be more wind resistant on the trail as well. It doesn't mean that stove A will boil water in a hurricane and stove B will not. That's not what the test is about.

Atraildreamer's spreadsheet as configured will give a basis of comparison between stoves so long as variables are consistent for each stove.

Simple.

[atraildreamer]

It seems that a lot of good points are being raised without the comments being very useful. If someone proposed a real wind test I would want to know how the alcohol stoves operate with no wind.

1) A windscreen that is wrapped around the stove will be more effective than dancing your Nalgene upwind of the stove.

2) Based on the large effect of the relatively small variables we have seen so far, we must define all variable very carefully. I have heard that pot diameter, heat reflection from the wind-screen, and stirring the water can be significant. For home made stoves, complete assembly instructions are needed for others to reproduce the results.

3) I worry about efficiency calculated on 12 ml of alcohol. I think the instantaneous efficiency may vary as the fuel heats up, during the full burn, and as the flame dies down. Since 12 ml of alcohol will not boil 16 oz of water, 20 to 24 ml of fuel may be a better test. Maybe increasing the water to to 20 to 24 oz would be a better representation of bringing 16 oz totally to a boil.
Rambler

1) The Goya design actually uses the Goya can as the windscreen, but actual performance would probably improve with an extenal windscreen.

2) Pot diameter does affect the efficiency of the system. In the second test burn, with more alcohol, when the efficiency of the Goya system dropped, a larger pot would have trapped more of the heat that was blowing out of the vent holes.

3) The initial test burn of the Goya system was with 11.9 ml of Heet (yellow), the 2 subsequent tests burns got the amount of fuel needed to boil the water up to 26.4 ml. I do not recommend figuring fuel amount based only on one test burn. Fuel amounts can decrease the efficiency of the stove system.

Remember: My intention was to provide a basis for comparison of various stove designs, and subsequent modifications.

Since I don't want to construct every possible type of homemade alcohol stove (My wife is about ready to kill me over the number of stoves I have now! :D) I invite all interested ASA (Alcohol Stove Anonymous) members to run their own tests and post the results. :welcome

[zelph]

Thanks A Million for all the work put into the spread sheet, I'll be sure to use it. I like it just the way it is!!!!!!!

[highway]

A suggestion:

Eliminate wind in tests

Wind is such an erratic, non-defined variable that its most likely impossible for any two testers to arrange the same effects of it to be able to accurately compare (or duplicate) results for means of comparison. Therefore data are flawed for comparison because of it. Even in the field, the cooking conditions may change from minute to minute, given wind shifts, eddys, gusts, updrafts, downdrafts, etc., and the 16 oz water that needed just 6.5 minutes using .41 avdp oz (15 ml) on first pot may require more(or less) on the second. In fact, it usually does, as you most likely have already discovered. The wind effect upon the cooking process may not be exactly the same for even a second heating of water from the one you heated just minutes earlier.

So, since you cannot duplicate it, eliminate it as a variable. Develop data in the absence of it instead, or in a more controlled environment like your garage or room where the conditions would be closer from test to test. There you can pay more attention to using constant water and ambient air temperature for means of comparison.

Just a thought

[rafe]

Eliminate wind in tests


I think the OP essentially did that by doing the tests indoors.

That makes the test more accurate but not terribly useful. Simply because the problem is complex doesn't mean it's not interesting.

Like Dr. Einstein said, "Simplify as much as possible... but no more."

[highway]

I think the OP essentially did that by doing the tests indoors.

That makes the test more accurate but not terribly useful. Simply because the problem is complex doesn't mean it's not interesting.

Like Dr. Einstein said, "Simplify as much as possible... but no more."

Accurate is not useful?

[rafe]

Accurate is not useful?


Not by itself. :) Some complex problems can't really be made simple. These numbers very accurately describe stove performance... in very unrealistic conditions.

[highway]

Not by itself. :) Some complex problems can't really be made simple. These numbers very accurately describe stove performance... in very unrealistic conditions.

In design, one often eliminates some uncontrolled variables (wind?) so as to concentrate on the design which achieves optimum results (efficiency?). Then, once the design/optimum efficiency is complete, subject the design to the variable. Otherwise the variable conflicts far to much in test of the design to achieve the most efficient one.

Some complex problems not only can be made simple but oftentimes must be made so-in order to arrive at the problem's solution. Its like breaking down a complex math problem into a series of simple steps.

[Skidsteer]

I finally got around to using this spreadsheet to actually test a stove and I must say it is a very, very interesting and valuable tool. Thank you Ataildreamer and Jason Klass.

I chose a Turbo Tea-Lite (http://www.whiteblaze.net/forum/showthread.php?t=22524) for my test stove and a Heineken pot/stand to round out the system*.

I've attached a copy of the full test for you analytical types but here's the condensed version:

Initial test burn-59.6%

Second test burn-53.3%

Third test burn-53.6%

From the results of these tests, the program concluded that the Turbo Tea-Lite 'system' would be optimal using 13.9 ML of denatured alcohol(my test fuel), so I tried it.

Final results:

Two cups(16 fluid ounces) tested using 13.9 ML of denatured alcohol achieved rolling boil in 10:56 minutes and continued boiling an additional 3:00 minutes until flameout at 13:56 minutes.

*Total weight of this pot/stand/stove/windscreen system is 2.25 ounces.:)

[Nightwalker]

A typo, perhaps in the below quote, page 1:

"As an example, let’s start with 16 ounces (2 cups), (1.25pounds) of water as a standard volume, 2 cups being a typical amount of water used by backpackers when cooking."

If one fluid ounce of water weighs 1.043 ounces by weight, then would not 16 ounces weigh 16.688 ounces or 1.043 pounds?

Extremely close. One gallon of fresh tap water weighs 8.337 pounds, therefore one pint weighs 1.042125 pounds or 16.674 avoirdupois ounces. I doubt that the number of angels on the head of a pin needs to be worried about to this extent, but it's a number that I accidentally memorized some years ago.

Non-memorized numbers that might help are as follows: one pint of water weighs 472.69994858625 grams (round as you wish). One fluid ounce of water weighs 29.543746786640625 grams. One ounce (avoirdupois) = 28.349523125 grams. Again, round as much as you care to. I do all of the calculations and then round at the very end, thus assuring the accuracy required, whatever that might be. Since my scale only goes to tenths of an ounce and whole grams, the end number gets rounded quite a bit. :rolleyes:

Before I broke my brain, I used to be a toolmaker and numbers were a great big fat deal. I still enjoy them quite a bit, even though they're no longer as big a part of my life.

The one ounce = 28.349523125 grams number came from Wikipedia, and is probably close enough. The temperature where it is most accurate is at four degrees Celsius.

I was much more sure of these numbers before I started researching and writing this post.

If there is anyone that wants to laugh, I think here is where you come in. :)

[Nightwalker]

Thanks for catching this error. :eek:

I rechecked at:

http://www.fourmilab.ch/hackdiet/www/subsection1_4_2_0_7.html

and found that one gallon (128 fluid ounces) of water weighs 8.345404lbs/gallon, the 1 fluid ounce would weigh 0.0651984 lbs. (8.345404/128=0.0651984)

Then:16 fluid ounces weighs 0.0651984 lbs/fluid ounce x 16 fluid ounces=1.0431755 lbs.

I'll be correcting the article and the spreadsheet, and have it reposted.

Maybe your number is more correct than mine. I got it from the 22nd Edition Machinery's Handbook, also known as the toolmaker's bible. My gallon number only goes to three places, and was memorized 15-20 years ago, so maybe I'm remembering it wrong anyway!