The concept for the Primus Gravity MF stove is for a really flexible, powerful and stable multi-fuel stove handling canisters, white gas and kerosene, to be useful in both summer and winter, and at a fairly low weight. The Primus Gravity MF stove is stable and very low. Using a single connector to mate with both canisters and pumped tanks is innovative and works well. The Primus Gravity MF runs well on canister fuel but it does have a few problems with the other fuels.
- Handles all fuels (except alcohol)
- Light weight, especially for a liquid fuel stove
- Low profile and stable design
What’s Not So Good
- Tends to sputter in some cases
- Priming difficulties, sometimes leading to fireballs
- Flame lift-off possible
- Valve orientation gets in the way for an inverted canister
|2006 Gravity MF|
|Butane/propane mix, white gas and variants, kerosene|
|10,500 BTU/hr (3 kW)|
|8.57 oz (243 g), hard sheet steel, brass, braided fuel line|
Pump and Tank Weight
|3.63 oz (103 g) for the aluminium and plastic pump,|
3.53 oz (100 g) for the aluminium tank with capacity 1 pt (0.6 L)
Minimum Field Weight
|8.57 oz (243 g) + any canister support legs for canister version,|
15.7 oz (446 g) for ‘liquid’ fuel version
Stove Leg Radius
|4 in (100 mm)|
Stove Pot Support Radius
|2.7 in (70 mm)|
Stove Pot Support Height
|2.5 in (65 mm)|
|Claimed: 4 x 4 x 1.6 in (10 x 10 x 4 cm)|
|Canister – not quoted, depends on canister size,|
white gas approximately 1.5 hr, kerosene approximately 2 hr
|Tank cap, multi-tool, extra jets, silicone grease, windshield, base plate|
Lindal-style valve inside pump fitting.
The Primus Gravity MF stove is among the first of a new-generation of multi-fuel stoves which are able to burn all of the common powerful fuels – white gas, kerosene, butane/propane canister fuel, and variants thereof. The previous generation of ‘multi-fuel’ stoves could not handle canisters. As such it is very welcome.
Primus has achieved this versatility very neatly. There is a standard screw-thread fitting on most canisters: they have used this format. Primus puts a standard screw-thread adapter and valve on the end of the fuel line so it can attach to a canister. They also put a canister screw-thread fitting on the pump: this is the brass bit sticking up from the black plastic housing, visible in the picture here. Inside the screw thread fitting on every canister there is what is called a Lindal valve – it keeps the gas inside the canister. It is opened by the stove being screwed onto the canister. There is the same sort of valve inside the brass fitting on the Primus Gravity MF pump, in the hole pointed to by the blue arrow. This is only opened when the fuel line is connected. Having it there means you can pressurise the tank without the stove being attached. What I particularly like here is that Primus has used an existing (and very reliable) industry standard for the connection rather than develop a new one.
One practical caution is in order. The Primus Gravity stove originally came in two forms: the ‘Gravity’ which used butane/propane canisters, and the ‘Gravity MF’ which used a tank or fuel bottle and burnt either white gas or kerosene, although the MF kit did not include the fuel bottle for some reason. Primus has since melded the two together as the ‘MF.’ Some smaller shops may still be carrying the older versions. The current ‘Gravity MF’ handles all fuels and has the fuel bottle tacked onto the outside of the earlier MF box. Close inspection shows that the stove, hose and valve come from the older canister-specific Gravity stove. The new fitting on the pump consists of a screw-thread Lindal valve mated to the bayonet fitting which the pump originally used. The bayonet fitting used to be retained by a spring clip. That clip is still there, but is screwed down to lock the bayonet in place. (It’s a very small screw: don’t lose it!) This neat little adapter allows the older parts and production facilities to be used for the new multi-fuel version, but it does mean the valve is designed to control gas flow, not liquid flow. This will be seen to be significant.
The burner shell is the outer surface of the burner with all the little holes. The fuel/air mixture swirls around inside the shell, mixing, before it comes out the holes. The holes have to be fairly small to prevent the flames from getting inside the burner, and there is a layer of fine mesh inside the visible shell which helps reduce the apparent size of the holes. A large burner shell like this one is often thought to allow a higher peak power, but comparing actual performance figures between different stoves does not really support this idea.
A large burner like this can spread the flames out so your pot does not get so many hot spots which can burn your dinner. However, a large burner puts the flames nearer the edge of a small pot, so that the flames have a greater chance of coming up the side. This seriously reduces efficiency at higher power, and can be a little hazardous as well if the flames come very far up the side.
Three different jets are provided for the stove: 0.45 millimetre for butane/propane, 0.40 for white gas, and 0.35 for kerosene. Changing the jet over is fairly simple. You remove the screw in the middle of the burner (tip of green line in the picture below), remove the legs (slightly fiddly) and then extract the jet from the hole in the fuel/air mixing tube. You may have to prise the reheat tube up a little to get the jet out of the hole, but then it is easy. The multi-tool provided has a straight blade screwdriver on one end which can be used to undo the Phillips head screw (uncouth, and it does not always work). The spanner part of the multitool fits the jet. The jets are labelled with their sizes in fine engraved writing on the faces. The multi-tool (with pricker), a cap for the tank, some silicone grease, a heavy aluminium shim windshield and ground-plate are all included in the kit.
Top view of Gravity MF stove, showing key parts.
The legs fold under at the hinges (blue lines) to further reduce the height of the stove, down to about 1.7 inches (43 millimetres). This makes packing simpler. However, you should be very careful to not bend the flexible fuel line too much, especially where it joins with the rigid fuel preheat tube (red line), and this limits the packed diameter to about 6.7 inches (170 millimetres). Primus claims that you can reduce the folded bundle to about 4 inches (100 millimetres) square, but that is far smaller than I am willing to bend the hose. Primus assures me that the fuel line is lined with Teflon, is very tough and that the company has had no problems from kinks, but I recommend you take no chances. You should also keep an eye on the connection between the flexible and rigid sections to make sure there are no leaks there. The connection has to be quite tight. You should rarely need to undo this connection anyhow.
This is the lowest stove I have seen for these fuels. The fold-up legs are not heavy, but they are quite strong enough. The result is a very stable stove indeed – yet quite compact. The only problem with having four legs rather than three is that sometimes the fourth leg doesn’t touch the ground without a little packing under it, but this is a minor matter, easily fixed in the field. Primus has achieved this low height by placing the normally vertical fuel/air mixing tube horizontally inside the burner. There is obviously some clever baffling inside the burner as well, to distribute the vaporised fuel/air mix evenly around the burner. This works fairly well as the flame pattern is very good, with only a slight tendency for some extra flame to appear opposite the jet. However, as detailed below, the burner design does present problems with priming.
The Primus Gravity MF stove is claimed to have a peak power output of about 3 kilowatts, which puts it up there with the other mainstream stoves of all sorts. You pay no penalty in peak performance for the versatility itself, but note my comment above about large burners. However, I never run my stoves flat out as that is very inefficient. As always, using the supplied windscreen or a lighter one is advisable for reliable and efficient operation.
Field Testing – Canister fuel
The Primus Gravity MF by a small creek in the summer running off a canister.
Operation of the stove with the canister upright is very straight forward, and the stove runs just fine like this. This is the normal summer method of operation for any upright canister stove, where the gas evaporates inside the canister. The operation is very flexible, ranging from a very low flame suitable for a simmer to a quite aggressive flame for a very fast boil. Control over this range is smooth and convenient. With the flames just reaching the edge of my 5.9 inch (150 millimetre) diameter Trangia kettle, I can bring two cups of water (500 mL) to the boil from ‘room temperature’ in 4.5 minutes. Sure, I can make it boil water considerably faster, but I choose to be economical with the fuel I carry, and that’s fast enough for two cups of coffee. When cooking dinner I usually have the stove turned down to an even slower rate anyhow. Turning the stove off is simple and obvious – you just turn the valve off. It’s a very nice stove in this configuration.
But that’s not all. The valve on the fuel line can swivel around, so you can invert the canister to get a liquid feed, as described in Selecting a Canister Stove for Cold Weather Backpacking Part II: Commercially Available Canister Stove Systems. One has to wonder why it has taken stove companies so long to reach this design.
However, operation of the Primus Gravity MF stove with the canister upside down for winter use is not without some problems. I tested this at temperatures between freezing and ‘room temperature’ and at altitudes between sea level and 1000 metres (3,300 feet), and found that there are distinct sputtering sounds from somewhere in the fuel line or in the burner, and when this happens the flame height is slightly erratic. I suspect the sputtering may be flash boiling of liquid fuel when it hits very hot metal in the preheat tube. I discuss this below.
Condensation on the valve when running the canister inverted.
I contacted Primus in Sweden about this, and the response implied that this was a known phenomenon. I was advised to try switching from the 0.45 millimetre jet to the 0.40 jet as the former might be a bit oversize, but I am not sure Primus was really confident that this was the problem. I tried changing the jet, but it didn’t seem to make much difference, and a check on the larger jet size showed it was fairly close to the specified size anyhow. The stove does run fairly steadily like this and I have used the stove in the inverted configuration under various conditions, but the sound is a bit of a worry.
In warmer weather when the liquified butane/propane gas coming out of the valve can evaporate at the valve (rather than at the preheat tube), you will see the valve getting cold and collecting condensation. This is illustrated in the picture here. This doesn’t seem to affect the operation at all, and won’t damage the valve as it is brass.
The picture does show a small problem with the design of the Primus valve for inverted operation. It extends out from the canister a long way, unlike some other valves which stick out sideways (eg MSR WindPro). This means the canister has to be held much higher in the air with the Primus Gravity MF stove than with other stoves. With other stoves I can use a very simple stand of some sort, but for this stove I ended up leaning the canister up against something and letting it rest on the black valve knob. I have to say it works just as well this way.
Flash Boiling – Examination
The flash boiling seems to produce a surge in pressure in the line, which pushes the liquid fuel back a bit and the flame height increases briefly as well, due to the higher pressure. Then as the vaporized fuel burns away the pressure in the line drops, the flame height drops, and the liquid fuel pushes forward again to do some more flash boiling. This sputtering can happen with all fuels when the driving pressure is high.
One very possible explanation for the flash boiling is that the liquid fuel comes along the chilled Teflon-lined fuel hose and hits the hot metal preheat tube suddenly, and flash boils when it makes contact. I checked the temperature of the connecting nut between the flexible and rigid sections and found it was well over 100 C (212 F): this is certainly hot enough to vaporise canister fuel and white gas, if not kerosene. If this is correct it suggests a design problem with the abrupt transition from the insulated Teflon lining to hot brass. However, I find it hard to believe there can be an abrupt transition from a brass line at a temperature well over 100 C (212 F) to a Teflon line at a temperature under 0 C (32 F).
Listening very carefully to the sputtering sound did not really pinpoint it at the connection. It could equally well have come from within the burner. For this to be correct it would mean that liquid fuel is getting down the entire length of the preheat tube and vaporising at the jet. This would require that small drops of fuel are being carried along by vapour. I find this a little hard to believe because the preheat tube gets extremely hot and it has a fat flexible stainless steel cleaning wire up its length, as do most other liquid fuel stoves. There simply does not seem to be enough room for blobs of liquid fuel to float up the tube. On the other hand, the fuel line on this stove is a lot fatter than other fuel lines which are not associated with sputtering.
It is possible to run a canister stove with the canister inverted without the sputtering. Neither the Coleman Xtreme nor the MSR WindPro do this, but I note the Xtreme does have a brass wire running some distance down the solid preheat tube, while the WindPro has a length of steel wire there. Perhaps this is an internal heat exchanger helping to vaporize the incoming liquid fuel? I tried to copy this by running a bundle of copper wires from the preheat tube down into the flexible tube, but it did not seem to make any difference. So at present this sputtering problem is apparently known to Primus, the cause is not known, but the stove seems to work reasonably well on a feed of liquid butane/propane gas despite it.
Field Testing – White Gas and Kerosene
The stove burns both white gas and kerosene, using a fairly conventional liquid fuel tank and pump. The pump is quite solid and works very easily – much more easily than some plastic pumps which seem to drag at the shaft. The valve at the pump output connection allows you to pump up the pressure without the stove connected. Unlike some other pumps, this pump is tilted when inserted into the tank, so the bottom end of the pump almost touches the inside of the tank. This is visible in the middle of the first picture. The brass fuel inlet line (on the right side of the pump) probably does touch the inside of the fuel tank.
I found that using a too-large size of jet was bad news: long yellow flames went all over the place. This did not happen with the jet sizes nominated by Primus. It is possible to run with smaller jets than specified: you just get reduced peak power. This has never worried me as I never try to run the stove at peak power.
A phenomenon known as ‘flame lift-off’ is a serious problem when it happens. The term means the flame doesn’t start until some distance above the face of the burner: the flame has lifted off the face. I have seen flame lift-off over small areas of the burner with all fuels, and especially when the driving pressure in the pumped fuel tank is too high. It means the fuel/air mix is coming out of the holes in the burner shell too fast – faster than the flame propagation speed. The risk here is that a gust of wind could actually blow the flame out. That said, the very broad area of the Primus Gravity MF burner makes it fairly unlikely that this could happen right across the burner at the same time.
Flame lift-off can happen when you open the control valve too far and the fuel pressure is high. Under these conditions the jet of fuel is traveling very fast and can drag in a lot of air, such that the gas flow out of the burner holes is traveling too fast. This can happen when the canister is too warm or the fuel bottle has been pumped too hard. As explained below, Primus recommends you do not pump the fuel bottle too much.
Flame lift-off does not seem to happen when the jet is too large. I believe this is because the larger jet means the gas velocity is lower, and this means it is not able to drag enough air with it. That is, the air inlets beside the jet are not large enough for the flow of fuel. The result is that the fuel/air mix coming out of the burner shell is far too rich, and this leads to the long yellow flames. Anyhow, opening the control valve too far is something you should not do.
Turning the fuel from the tank off can be done the same way as for a canister, but Primus has added a simple but very nice feature which lets you drain the fuel line before stopping. The black plastic header on the pump has the words ‘ON’ and ‘OFF’ embossed on the rim. When the word ON is facing upwards, the fuel intake is at the bottom of the tank and the stove runs normally. But if you flip the tank over, so the word OFF is uppermost, the fuel intake is at the top, above the fuel, in the air-space. Let the stove run like this for a short while and the remaining liquid fuel in the line will be exhausted, and the stove will go out. The connection at the pump rotates to permit this flipping easily.
If you don’t do this, you can find quite a lot of potentially flammable fuel dribbling out of the hose and on the ground around you, and this can be very dangerous! It also generates a lot of fumes which stink badly. Finally, it will probably also leak a bit of residual fuel inside your pack.
In practice, I found that it takes about a minute to drain the fuel line, at fairly high power. For someone counting grams of fuel, this extra burn-time can add up. You need to anticipate some time in advance when you will be finished cooking. I note that other stoves mostly seem to have thinner preheat tubes which hold less fuel anyhow, and this seems a good idea.
The instructions which come with the stove specify 15-20 pump strokes for a ‘full’ tank of fuel. Read the Instructions! I can run the stove satisfactorily with only 12 pump strokes and a full tank, although the peak power is not high. That suits me fine. With 20 strokes the stove is going strongly on either liquid fuel, with only a little yellow at the very tips of the flames. I believe this is just getting to the point where the air inlets are starting to limit the mixture. With 40 strokes, which is considerably more than specified, the flames are dancing around the place and very often quite yellow. High pressure also leads to the sputtering noises. The message here is quite clear: do not over-pressurise the tank.
It is worth noting that throttling a liquid fuel is quite different from throttling gaseous fuel. A far smaller flow of liquid is required for the same amount of vapour. The valve on the Gravity MF seems to be the one designed for controlling butane/propane gas from an upright canister, so that much finer adjustments are needed to control a liquid fuel. If the tank pressure is too high even finer control is needed, and I found this to be very difficult. But the manual does warn against using a high tank pressure.
Primus specifies the 0.40 millimetre size jet for white gas. It will also run with the 0.35 jet, just at lower peak power. The normal method for lighting such a stove with white gas is to let a little fuel out into a priming cup and to light that, to warm up the preheat tube. However, the fuel line usually starts off empty, so you have to flush the air out of it first. To do this safely you really need to be able to see the jet, to detect when the fuel starts coming out. However, with the design of this stove you cannot see the jet, and cannot see when fuel starts coming out the jet. This means you cannot see how much fuel has come out, and releasing too much is easy. The results of my first few attempts to light the stove with white gas were therefore somewhat in the nature of a fireball. The first attempt was quite a shock in fact.
Practice does improve on this, and there is a faint hiss when the liquid hits the jet which helps, but it is hard to avoid a bit of a flare-up. In addition, the fuel is initially retained inside the burner, out of sight. This can lead you to thinking there is no fuel inside the burner, and releasing more. This means it can take several minutes to get the stove running on white gas – while the fireball dies down. Primus suggests that the priming fuel will be burnt out in about 45 seconds for white gas, but my experience is that the time is usually longer – and variable.
In the end, I found it best to just barely fill the preheat tube with fuel and then to prime the stove with a small amount of alcohol. This warms the metal-work up enough that some white gas vapour starts coming out of the holes in the shell, and then the stove slowly gets going. Certainly, using a little alcohol means far less of a fireball every time, although I never managed to gauge the amount of fuel in the preheat tube perfectly.
Once the stove is running, the flame can be adjusted from low to high settings. At full throttle with a low tank pressure there may be a little bit of flame lift-off in places around the burner, but it is not enough to cause a worry. Fairly fine adjustment of the valve is needed to get good control, and the higher the tank pressure the more difficult this is. Low tank pressure is required for good operation. The stove runs quite well on white gas, although control is delicate and lighting it seems to me to be distinctly hazardous. I certainly would not light it inside or even anywhere near my tent, and I would be very careful inside a building.
A very sooty stove after priming and running on kerosene.
This fuel is popular with experienced cold-weather travelers who dislike the fireballs associated with white gas. It is popular in the Arctic and Antarctic, and in remote countries like Nepal. I have used it a lot as well, before switching to canisters. The Gravity MF can burn this fuel, but my experiences here were very poor. Primus specifies the 0.35 millimetre jet for kerosene. I did try using the 0.40 jet once, by mistake – it was a bad idea with long flames everywhere. Don’t do it.
One can light a kerosene stove using the same kerosene as the priming fuel, but this is very haphazard and sooty, and tends to make the stove very dirty. The Primus instructions suggest doing it this way, opening the valve ‘for a short time.’ Experienced users always use alcohol or a priming paste to prime a kerosene stove – to get the preheat tube hot. To do this safely you do need to start with a cold stove, otherwise there is alcohol vapour everywhere.
The normal method for lighting a kerosene stove with alcohol is to open the valve briefly to let the fuel line just fill up to the jet with kerosene, then add a teaspoon or so of alcohol in the priming region. The alcohol flame heats up the preheat tube and the jet. As the kerosene heats up it starts to expand and vaporize, come out the jet, mix with the alcohol and vaporize on any hot metal, and slowly take over the priming. However, in this case the jet and the end of the pre-heat tube are concealed inside the burner, and are not pre-heated all that well. I am unable to see when the kerosene reaches the jet, so I don’t know how much liquid kerosene is inside the burner. The kerosene near the jet stays fairly cool as well. The end result is usually quite a smoky flame for some time before the vapour from further up the line reaches the jet. This makes the stove rather sooty.
Sooty base to pot from kerosene flames.
I find that a fair bit of priming is needed with kerosene, and some of the alcohol always flows across the flat base plate to one of the legs, and then runs down the leg to the ground. Sometimes the alcohol on the ground catches alight, other times it doesn’t. This is a bit risky: perhaps paste would be better. If I don’t use enough alcohol for the priming the stove doesn’t light, but the metal is then too hot to safely add more alcohol for a while. So starting this stove with kerosene takes a fair bit of practice. Frankly, I found it erratic.
I would not try to light this stove on kerosene anywhere near my tent, and I would be very careful about starting it inside a building. If there isn’t a large smoky flare, there is a strong smell of kerosene. As with white gas, the priming process can take a while as the fireball has to die down before cooking can start. The Primus instructions suggest 90 seconds for the priming, but I found it was very variable.
The stove is equally sensitive to over-pressure in the tank when running on kerosene. Too much pressure and the flames are often quite yellow at the tips. The yellow flames imply incomplete combustion, making lots of soot. I did check whether the highly lethal carbon monoxide was being generated, but the meter did not detect any. The yellow flames from over-pressure also make a distinct mess on any pot, and the sputtering sound is there too. In addition, at high pressure it is very hard to control the flow of fuel for a low flame. Running the stove on kerosene with the tank at very low pressure does produce a fairly clean blue flame at most settings. Giving the tank a few more pumps does increase the power output, and this is probably the best way of using the stove on kerosene: start at a very low pressure and increase carefully, but not too far.
Changing the Jet
In principle, changing the jet is simple: undo one screw (at the tip of the green arrow in the third picture) and the stove comes apart. However, this was not so simple all the time. When I went to change over from white gas to kerosene I found that this screw was seriously jammed. No screwdriver was able to undo the screw the full way. This problem could not be handled in the field.
I tried WD40 lubricant and all sorts of machine-shop persuasions, to no joy. Eventually I had to machine away the sides of the head of the screw and use a pair of vice-grips to extract it. I found serious damage at the bottom end of the steel screw where it had protruded below the end of the hardened steel nut at the top of the burner, but no damage to the nut itself. I do not know what caused this. I replaced the screw provided with a short stainless steel cup-head: the bottom end of this one does not protrude under the nut and (hopefully) should not suffer the same fate.
Boiling Times for White Gas and Kerosene
Primus does not specify boiling times for this stove on the box or in the instructions. I found I could raise the power output using a canister to the point where there were significant flames coming well up the side of my test kettle, but quoting a boil time with the flames this high is meaningless. Tests of the boiling times for white gas and kerosene gave rather variable results, once again depending on how high I wanted the flames.
In general I found that the peak power output with white gas and kerosene was slightly lower than for butane/propane mix, especially when using a low tank pressure. Owing to the wide range of results obtained on the test bench, I am unwilling to try to be more specific. Anyhow, I don’t subscribe to the machismo of boiling races – I prefer to gently simmer my dinner to perfection.