Subscribe Contribute Advertise Facebook Twitter Instagram Forums Newsletter

Explorations into Candle Stoves

Wax has a reasonably high energy density and unlike other fuels it does not spill (when cold): can you make an adequate DIY stove with it?

Print Jump to Reader Comments

by Mark Hurd and Roger Caffin | 2008-01-02 01:00:00-07

Explorations into Candle Stoves


Wax is a hydrocarbon fuel with an energy density (by weight) comparable to other fuels like canister gas and white gas. It is therefore very similar in some ways to other fuels used in our little stoves, but it has the advantage that it is a normally solid: it won't spill. Small tea-tray candles have been used for some time to keep food warm, but they are a bit too small to bring water to the boil in a reasonable time. A problem with just scaling a tea candle up is the soot generated. We explore here what can be done to improve the heating capacity and reduce the soot from a 'candle stove'.

All the experimental work detailed here was done by Forum Reader Mark Hurd. Backpacking Light Senior Editor Roger Caffin offered some advice and wrote the article text, and accepts responsibility for any errors.

What's Good

  • Robust, won't spill
  • Can be refuelled while running
  • Very cheap to make
  • Stable on the ground

What's Not So Good

  • Low in power
  • Generates considerable soot
  • Generates acrid fumes while burning some waxes
  • Generates lots of fumes after being extinguished
  • Hard to light

Background Discussion

First of all, it is obvious that candle wax burns and makes heat. Small flat candles have long been sold and used to heat 'chafing dishes', and are often known as 'tea candles'. But candle wax has one advantage over most other fuels except for those in the Hexamine class: cold wax is a solid and does not spill or leak in your pack. Compared to Hexamine-class fuels it appears to have extremely low toxicity as well. So for many situations it looks as though it might have some potential.

Just to reinforce this, we list here some properties of some the more common fuels we use, including two forms of wax - beeswax and paraffin. Be assured - there are many others.

Fuel Formula Kcal/g BTU/lb
Propane C311.0 21660
Beeswax C26-4611.0 21650
Butane C410.8 21160
Gasoline C4-12~10.4 ~18500
Kerosene C10-18~10.3 ~18300
Paraffin C19-3610.0 18000
Hexamine  7.3 12930
Ethanol CH3CH2OH6.4 12760
Methanol CH3OH4.7 9750
Dry wood  ~3.9 ~7000

Candle Stove
A candle stove with a cross-wick and a mesh pot-support

Surprise: beeswax and paraffin wax are up there with the best! But in fact that is quite reasonable because they are just higher hydrocarbons, in the same general series as propane, butane, gasoline and kerosene. Incidentally, you should treat many of the values given here with some caution as the composition of many of the fuels can vary. (Just what is in gasoline anyhow?)

But it is also known that candles do have some problems: typical candles have very small flames and make soot. Can these problems be overcome and a more powerful candle stove created?

Candle Stove
A coiled-up wick made of corrugated cardboard - it would be very sooty IF you could get it burning

Flame Design

You might think that this section should be headed 'wick design', but that would be to miss the point. What we are concerned about is the flame, not the wick. We want a flame large enough to heat water at an acceptable rate, while not emitting too much soot. Sadly, at this stage it would seem that some soot is going to be inevitable, so we will try to minimise it.

One way of increasing the flame size is to make a bigger wick. In the photo to the right we show a 'coil wick' (picture courtesy This is a big coil of corrugated carboard embedded in the wax. In fact the coil almost fills the tin, even though it can be hard to see it all under the wax. This design is not going to work very well at all. For a start, getting it lit may prove almost impossible unless you use a small blow-torch to melt enough of the wax that the wick can light. Even if we assume a slight modification so the wick can be lit, it may make a large flame but it will also generate a huge amount of soot and lots of fumes too. This is because the centre region of the wick will be releasing lots of wax vapour but this vapour won't get to burn very well: the outer wall of flame will effectively prevent any oxygen from reaching it. No oxygen means no combustion, but lots of soot (and maybe a bit of carbon monoxide as well).

Candle Stove
A newly cast stove with a cardboard Cross-Wick

The flame needs to have lots of sidewall where oxygen can mix in to support combustion. And both sides of the flame need to be open to the air: a simple ring will still leave the core without oxygen. This makes for poorer combustion and probably more fumes and carbon monoxide. So instead of a ring we look at a cross arrangement. The centre region of a cross is likely to be oxygen poor so we eliminate it. The result looks like the stove shown to the left.

Here we have a new candle stove cooling down from the casting process: the central pool of wax is still hot and clear. There are effectively four wide flat wicks here. The flat design of the wick means that the flame will be flat and will get as much air as possible. Having the four wicks out at the edge means there is minimal oxygen starvation in the centre. The four-wicks design is simple and possibly adequate in a can of this size: a larger can could have more wicks - perhaps.

How big a gap should be left in the centre of the wick? That is hard to say, and has to be judged from the flame pattern. The flames from this example may be seen in the first photo. They are perhaps a little high and smokey, but that seems inevitable with candle stoves.

Other designs are also possible: a tubular wick as shown below to the left has been tried and is called a Circle Wick. It is a simple circle of wick, as opposed to the coil wick shown before. The wick was made quite high to get good power, but there are costs to this as explained below. A Spiral Wick consisting of round 'wick' cord wound around the side of a light tube has also been tried, as shown to the right. This was meant to be similar to some alcohol stoves, but the results were not so good.

Candle Stove
A high Circle Wick stove, burning (needs mesh pot support)

Candle Stove
Spiral Wick stove, with integral pot support

Testing the Efficiency

The testing discussed here and in also the Carbon Monoxide series of articles, plus the basic chemistry of combustion, indicate that having enough air supply to the flames is vital to getting the maximum heat from the fuel. So the obvious question is how can the air supply be maximised? In addition, note that maximising the air supply to the flame should minimise the amount of soot and fumes produced. Reducing the amount of soot is clearly desirable just from the convenience point of view: consider the deposit on the bottom of this pot after boiling half a litre (one pint) of water on the stove arrangement shown above.

Candle Stove
Soot on the underside of a pot

But remember that the soot deposited on the pot is actually unburnt carbon from the fuel: it is wasted fuel and wasted energy. In fact it is the carbon in the fuel which supplies most of the energy (as opposed to the hydrogen), so the loss of power is quite significant. If all that carbon had been fully burnt in the flame instead landing on the pot as shown the stove would have given out more heat per gram of fuel. So soot is very bad for a number of reasons.

It is likely that fuel consumption will be a good guide to the efficiency of the air supply, so the three wick designs illustrated above were tested a number of times in various configurations. For all the tests results shown here the fuel used was beeswax. The reason for using beeswax is given later. The test conditions were the same for all three designs, thus:

Air Temperature75 F / 24 C
Starting water temperature60 F / 15.5 C
Finish temperature (boiling)212 F / 100 C
Volume of water16 fl oz / 455 ml
Pot1.3 liter Evernew Ti

The averaged results from several trials of each design using beeswax are as follows.

Design Wax used Boil Times Best Clearance
Cross-Wick6 g16:00 min Not measured
Circle-Wick10.5 g6:10 1.5 cm
Spiral-Wick11 g8:30 1.0 cm

Clearly the increased air supply to the Cross-Wick design means that a lot more energy was extracted from the fuel, compared to the two very inefficient round wicks where the air supply to the inside of the flame varied from very poor to non-existent. However, the Cross-Wick design was a lot slower to bring the water to the boil compared with the other two. This happened despite the probable loss of some flame energy up the sides of the pot with the Circle-Wick and the Spiral-Wick designs. We believe that this is an inevitable result of wanting to get the best possible air supply, given a finite stove area. That is, the amount of flame surface area was smaller for the cross-wick design compared to the other two.

The Circle-Wick design allowed the height of the wick to be varied easily - with a pair of scissors. It was found that wicks of 1.5 cm and 2 cm length above the rim of the container gave about the same performance, with average boil times 6:18 min and 6:10 min respectively. A wick height of 2 cm did give a boil time of 5:58 min on one occasion. Reducing the height of the wick to 1 cm above the rim gave an 8+ min boil time. The wick to pot distance was kept constant at 1.5 cm for the rest of these tests.

The performance of each stove is fairly sensitive to the height of the pot above the top of the wick. Reducing the height from 1.5 cm to 1 cm seriously extended the time to boil - in some cases to beyond the patience of the Tester. This is explained by the reduced availability of air for combustion and the much poorer flame which resulted.

Operational Considerations

All of these stoves produce a lot of wax fumes. The fumes are very unpleasant and permeate everything around. They are of course a by-product of the boiling wax. An attempt was made to create a design which would allow the fumes to be burnt, but no suitable design was found during these tests. This may leave opportunities for further experiments by someone else.

The paraffin wax put out acrid eye and throat burning fumes. This does not happen with an ordinary candle, but then, an ordinary candle does not have a flame of the size seen here. The beeswax fumes are somewhat acrid, but mostly cloying. Beeswax candles have always been favoured by the more affluent in times past, and this is also why beeswax rather than paraffin wax was used for the tests reported above. However, beeswax apparently has a reputation of attracting bears: residual organics from the honey vaporising from the wax is one reason. Just the smell of the beeswax stove in your pack might also be enough to attract bears too: they have a very sensitive sense of smell.

Candle Stove
A snuffed stove emitting lots of fumes

As mentioned above, in general the stoves made a sooty mess of the pots, and if the soot included some wax condensate it was very hard to clean the pots afterwards. And of course the stuff makes black marks on anything it touches: hands, clothing, gear ...

When making these stoves it is important to make sure the wick is saturated with wax. A dry wick is obviously not going to catch alight at all. But a fully impregnated wick has a fair thermal mass, which means that the stoves are all hard to light, especially the Spiral-Wick design. That one can takes several minutes of flame from a lighter to catch. This is because the flame from the lighter has to melt and then vaporise some of the wax before it can catch alight. The problem can be reduced if a thread or two from the wick is left sticking out from the main wick - provided it too has been impregnated with wax as well. If you do this you have something very similar to an ordinary candle wick.

Extinguishing one of these candle stoves is hard. The Cross-Wick design can be blown out - just, but the other two cannot be blown out. This means they have to be snuffed out. But while this extinguishes the flame (by removing the air supply), it does not stop the emission of vapour/fumes from the hot wax: that still goes on for several minutes. Now there is no flame the amount of fumes is of course even worse than before. The result can be a rather thick fog as shown here, and the fumes get into everything around the stove. In addition, the stove stays very hot for up to 10+ minutes, during which time the liquid wax can be spilt.

Material Sources

Waxes can be bought from most any craft shop and from the web, often by the pound. Sources are not given here.

The first wicks in the experiments were simply corrugated cardboard: as cheap and as simple as you can get. In fact they worked quite well, but they do have the problem that as the wax gets low the wick starts to burn. This makes a bit of extra smoke, which is not so nice. More sophisticated wicks use non-burning synthetics, although one could use cheap cotton tape as well (but it chars like the paper). Sources for synthetic wicks include: -- Lots of different kind of wicks including fiberglass. They sell it by the 100+ yard spool, but more importantly they also sell 10 yard hanks for US$7 each.
Spiral-Wick stoves were made with 1/8 inch and 1/4 inch round braided fiberglass wick, by wrapping the round wick material around a small cat food can, placing the wrapped can in the center of a large cat food can and pouring molten wax in the space between. (suggested by Forum Reader Jason Klass) -- This site caters to "fire-oriented performance art", such as dancing while juggling flaming objects (er - wow!). They sell flat Kevlar/Fiberglass tape wicks in various thicknesses and widths up to 4 inches. (As an aside: these wicks also make good insulators for pots and pot handles, which is what Jason was using it for.)
Two sizes of the K1 Tape Wick, 1/8 inch by 2.5 inches and 1/16 inch by 3 inches, were used for the Circle-Wick stoves. They are stiff enough to stand up by themselves. A length of the tape wick long enough to go around the inside circumference of a small cat food can was placed in position and molten wax was then poured into the can. The wick may be cut to the desired height above the rim, but should be soaked in wax to the top. Cutting the top edge of the wick leaves a raw edge on top of the wick which should be easier to light.


"Explorations into Candle Stoves," by Mark Hurd and Roger Caffin. (ISSN 1537-0364)., 2008-01-02 01:00:00-07.


Reader Comments

You must login to post comments.

New Visitors: Create a new account
Remember my login info.

Explorations into Candle Stoves
Display Avatars
Sort By:
Benjamin Smith
(bugbomb) - F - M

Locale: South Texas
Explorations into Candle Stoves on 01/01/2008 20:59:38 MST Print View

Companion forum thread to:

Explorations into Candle Stoves

Adam Kilpatrick
(oysters) - MLife

Locale: South Australia
Candle Stove fuel efficiency on 01/02/2008 01:15:38 MST Print View

Is it just me or is 6g to boil 455ml sounding remarkably efficient? If you can put up with the crap boil!

I think Mark might be really onto something here.

david plantenga
(dplant) - F
Maybe ... on 01/02/2008 07:12:43 MST Print View

Hey All,

How's about firing up our Caldera Cone with a multipule flames wax burner?

I was wondering about a candle burner the first time I used my Caldera Cone.

I tried a single candle warmer and it didn't work. But, I never thought of multipule flames like "Explorations into Candle Stoves" has shown.

Back to Backpacking Light engineering ...

AT Wann'a Bee

Joshua Mitchell
(jdmitch) - F

Locale: Kansas
Re: Candle Stove fuel efficiency on 01/02/2008 07:18:39 MST Print View

Yeah, it is... that's the beautiful potential of wax... FYI, I was planning on doing it myself, but have found with a young son and another child on the way my time is limited...

I believe the caldera cone system may lend certain advantages to a wax burning setup... the higher internal temps of the cone should result in a more complete combustion (read more heat and less soot from a smaller wick size). Also, the nature of the cone eeks more heat capture from a smaller flame size as well.

Actually, on a SP600 stove I got tap water to tea-temp / hot chocolate temp with JUST a tea candle... not quite boiling, but didn't hardly use any fuel either...

Note, Soy Wax should hold a lot of promise... not nearly the same kinds of organic contaminants that beeswax would hold, and not nearly the inorganic contaminants that makes paraffin nasty to deal with. Basically, it should be purer than either.

HAH! I was still typing whilst David was posting...

Edited by jdmitch on 01/02/2008 07:19:14 MST.

Jason Klass
(jasonklass) - F

Locale: Parker, CO
Interesting idea on 01/02/2008 17:39:50 MST Print View

Hmmmm...kind of interesting. I think I could put up with the low heat output but that soot would drive me crazy. Would you guys say it's more or less than the amount you would get with a typical wood burning stove?

Roger Caffin
(rcaffin) - BPL Staff - MLife

Locale: Wollemi & Kosciusko NPs, Europe
Re: Interesting idea on 01/02/2008 19:49:44 MST Print View

Hi Jason

> Would you guys say it's more or less than the amount you would get with a typical wood burning stove?
Mark needs to answer as well, but from my experience I think the answer really depends on how you run your wood fire. If you are using a good gasifier and wait until the initial flames die down a bit, then I think a clean-running gasifier would be cleaner. I THINK.

Mark Hurd
(markhurd) - M

Locale: South Texas
Re: Interesting idea on 01/02/2008 20:39:22 MST Print View


I think the soot is about equivalent to cooking over a regular wood fire. I haven't used a gasifier type stove, but they would appear to be more efficient and less sooty, so I would agree with Roger.

The major problem I found was the not so much the soot, but the fumes especially after extinguishing the stove. The wax stays very hot long after the stove is out and as the photo in the article shows there is a noxious cloud released if you remove the snuffer too soon. Try blowing out a candle and watch the plume that rises from an eighth inch wick. Now multiply by 30 or 40 to get a small idea of the smoke and fumes you get from a small stove.

Now if one could figure a way to feed a small amount of wax in as needed this would help. Also, if someone is clever enough to "gasify" the wax with a preheat tube or some such arrangement like liquid fuel stoves have, then one would have a potentially great stove.

Also, as far as the 6 grams of wax to heat water to boil- I have found that it takes about 6 gm of canister fuel to do the same thing which is consistent with the estimated heat capacity of the two fuels.

And I have not tried the cone with a candle, but one potential problem I see is that the cone is a "low oxygen" area (as the TrailDesigns folks note) and this could be problematic. But it is probably worth a try.


Edited by markhurd on 01/02/2008 20:41:09 MST.

Joshua Mitchell
(jdmitch) - F

Locale: Kansas
Candle in Cone on 01/03/2008 09:24:18 MST Print View

Yeah, the amount of soot I got from my trial was slightly less than I would have on a basic wood stove (something similar to the J Falk or Nimblewood)

"And I have not tried the cone with a candle, but one potential problem I see is that the cone is a "low oxygen" area (as the TrailDesigns folks note) and this could be problematic. But it is probably worth a try."

Low-Oxygen would only be an issue if you're vaoprizing fuel at a rate faster than it can oxidize (aka combust / burn).

The key is to keep the fuel-vaporization rate lower than the oxygen intake rate. If this is maintained the higher combustion area temps and longer heat retention time would make it advantageous.

"Also, if someone is clever enough to "gasify" the wax with a preheat tube or some such arrangement like liquid fuel stoves have, then one would have a potentially great stove."

Hmm... I wonder what would happen if one paired wax with a BB... probably wouldn't be true gasification but could get interested and be able to burn at a higher rate than the cone...

Ryan Potterton
( - M

Locale: East Cascades
more hot air on 01/03/2008 11:01:35 MST Print View

A doughnut shaped pot, raised off the ground, could provide air to the inside of a circle wick.

Mark Hurd
(markhurd) - M

Locale: South Texas
Re: more hot air on 01/03/2008 13:58:41 MST Print View


Interesting you should mention..., I recently built a donut prototype, but haven't had a chance to test it. I'll post the result when I do but it is likely to be a while before I get to it.donut stove


Adam Kilpatrick
(oysters) - MLife

Locale: South Australia
Re: Re: more hot air on 01/03/2008 17:40:55 MST Print View

Mark and Ryan,

I've been thinking about making an alcohol stove with a donut shape ever since I made my first one (a Pepsi-G).

Mark, what did you use for the weld? Please don't say JB-weld (can't find that in Aus).

Michael Skwarczek
(uberkatzen) - F

Locale: Sudamerica
definitely JB weld on 01/03/2008 19:44:30 MST Print View

Adam, if you're willing to pay for shipping, I'll box up as many as you like and send them out to you. The two part system comes in at about 70g for shipping calculations. The package seems to cost about $8-10. Maybe less if some power shopping techniques were employed.

Lemme know.


Mark Hurd
(markhurd) - M

Locale: South Texas
Re: Re: Re: more hot air on 01/04/2008 19:45:52 MST Print View

Sorry, Adam,- I used JB-Weld to make the join. JB-Weld is just a heat resistant epoxy. ( up to about 350 degrees C, I think,) There must be an equivalent in Aus. I'd check automotive parts shops. They use it to mend engine parts.

The pictured stove has been used as an alcohol stove a couple time, but tends to burn too fast and has too large a flame pattern for the Heineken or Foster's can that I was using for the test. I think a larger pot might work well.

Anyway, my plan was to check it out as an alky stove and then fill it with wax and flat wick to make a circle-wick stove out of it for testing.


Edited by markhurd on 01/04/2008 20:32:44 MST.

Tony Beasley
(tbeasley) - MLife

Locale: Pigeon House Mt from the Castle
Re:Re: more hot air on 01/04/2008 19:52:35 MST Print View

Hi Mark and Roger,

Thanks for a great article I enjoyed reading it, I found it interesting that the waxes have higher heating value than propane/butane fuels and can be as efficient. I have done a lot of testing on canister stoves and with a non flux ring pot I use about 6g of canister gas to boil 0.5l of water. One thing that I have noticed is that with both canister stoves and alcohol stoves it is not less efficient to take longer to boil water but it can be much less efficient to boil too quickly.

I think you might find this interesting. I recently ran some tests on an side burner alcohol stove that I made, with a std aluminium pot and the same stove with a 1.5l JetBoil pot the std pot used 17.18g of denatured alcohol to raise 0.5l water 80C and the JetBoil pot only used 12.4g of denatured alcohol to do the same, but this dramatic increase in efficiency was not the same with other alcohol stoves.

I have done similar work on a Donut stove ( I called mine a central flue stove) I made it out of a piece of bar stock aluminium ( see picture below), it works ok but needs some tuning, the flame is a bit stronger than I would have liked because I think it has too much thermal mass.


Central Flue stove

Mark Hurd
(markhurd) - M

Locale: South Texas
Re: Re:Re: more hot air on 01/04/2008 20:31:51 MST Print View


Nice stove!

You know I have really been thinking that I should give up on all of this stove experimentation and just build a better pot. I seem to recall that you were working on something along that line yourself a while ago, like cutting down a Jetboil pot I think. Anyway thanks for the info on the flux ring.


Franco Darioli
(Franco) - M

Locale: @Tarptent
JB weld on 01/04/2008 22:10:50 MST Print View

JB weld is available in Australia. I have bought some at Bunnings or K Mart.
That coincided with me losing interest in making stoves...

Tony Beasley
(tbeasley) - MLife

Locale: Pigeon House Mt from the Castle
Re:Re: Re:Re: more hot air on 01/05/2008 03:23:58 MST Print View

Hi Mark,

"You know I have really been thinking that I should give up on all of this stove experimentation and just build a better pot. I seem to recall that you were working on something along that line yourself a while ago, like cutting down a Jetboil pot I think."

Yes I removed the mounting ring from a JetBoil 1l pot and I have plans to cut it down to make it a 550 ml pot but have not done it yet as I have been distracted by many other stove projects and family life. I did try the 1l pot on a Gram weenie alcohol stove but it did not work that well, I guessed that the small diameter flux ring inhibited the air flow causing poor combustion, I am thinking about a making a gram weenie with the holes lower down to see if that helps, I will do some more work on the problem soon, I have had more success with alcohol stoves with the 1.5l pot is much larger which allows more air flow.

I have some alcohol mixing questions that I would like to ask you could I PM you.



twig .
(bretthartwig) - MLife

Locale: Australia
JB in SA on 01/05/2008 04:57:46 MST Print View

JB Weld is definitely available in SA at KMart, any problems let me know and I will try and get you some

Jason Klass
(jasonklass) - F

Locale: Parker, CO
Candle Stove on 01/05/2008 15:05:49 MST Print View

Hey Mark & Rodger,
Thanks for the feedback. Well, I'm still an alcohol-stove-guy at heart and if I have to put up with soot, I think I'd rather carry a wood stove. But, as a person who believes that every avenue must be explored, I enjoyed the investigation into candle stoves.

Keith Selbo
(herman666) - F - M

Locale: Northern Virginia
Re: Re:Re: Re:Re: more hot air on 01/05/2008 20:28:51 MST Print View

Have you considered trying the Primus ETA pot? It appears to be wider than the Jet Boil and has a ring under the fins to protect them. I've heard reports that it halves the boil time with an alcohol stove.