Here we describe our second run of preliminary small-scale Steam Balloon experiments. The conclusions are at the end...

We sewed up another miniature test envelope, using the same pattern as the first run of of experiments - thus, again, the envelope form was the classic ball-and-cone balloon shape made from twelve gores, and the radius of the ball was 50 cm and the bottom cone angle was 60 degrees, with the area being 3.5 m2 and the volume 0.6 m3, so that again the lift when filled with steam was 3.76 newtons - about 383 gm. The details of the material used for this second small envelope can be found here. Again, we left the bottom of the envelope open and led an umbilical tube into its side.

We then faced the problem of sealing the seams. This was not as easy as it appeared, because we could not find any adhesive that stuck to the fabric; no adhesive had any affinity at all with the siliconized coating material. Enquiries from 3M and two specialist adhesive companies drew blanks. Finally we tried using the obvious: bathroom silicone rubber sealant. This sticks beautifully - a sandwich of multiple layers of fabric and sealant, over time, turns into a sort of strong leathery mass, and boiling only seems to make it stronger. We discovered that the proper stuff - Dow Corning 785 - was far stickier and more glutinous than cheap no-name brands, and we used it to glue strips of the fabric itself over the seams. The final result was excellent.

Our test rig was only slightly modified from the one used in the first run of experiments, in that three boilers were connected rather than one. Here is a rather crude diagram:

Again, the steam sources were 2 KW boilers from do-it-yourself steam wallpaper strippers. The total steaming capacity with all boilers running was about 9.5, kg of steam per hour. As a preliminary test to assess the amount of water carried over as droplets in the steam, two of the boilers were set to blow off directly into a collection bucket (with no envelope being involved). In the first 30 minutes 340 gm of water was collected, and in the second 30 minutes 300 gm. In the final 15 minutes 165 gm of water accumulated. Therefore we can say with some confidence that each boiler carries over about 160 gm of non-vaporized water in its steam in 30 minutes.


The first test of the envelope was done naked without any insulation. As before the steam was fed into the side of the envelope through the umbilicus, and a clip was put on the bottom to restrict the escape of the steam, so as to build up pressure to fill the envelope. With two boilers operating (total 4 KW, i.e. 6+ kg of steam per hour) the envelope inflated slowly but not satisfactorily. However when we connected the third boiler as well (total 6 KW, i.e. 9.5 kg of steam per hour), the envelope inflated and tightened in a most impressive manner, and we were obliged to turn one of the boilers off again.

Here is a picture showing this third time that, as far as I know, a naked steam balloon has ever been inflated:

You can just see the vapor from the excess steam puffing out of the bottom! Meanwhile the bucket is collecting the condensed water which drips down.

This next photo shows the three boilers:

This photo shows the umbilical steam supply tubing:

And this photo shows the setup as a whole:

The detailed results of naked testing were as follows.

First 30 minutes, 3000 gm condensate collected;

Second 30 minutes, 3000 gm condensate collected. Exactly the same!

Therefore, deducting 3 x 2 x 160 gm/hour for carry-over water, i.e. roughly 1000 gm/hour, and dividing by 3.5 square meters of area, we get the basic result:

Naked condensation rate - 1.43 kg/m2.hour



We next made an insulating jacket of bubble wrap ("buffer" for our Japanese readers). This was "AirCap" barrier-sealed bubble wrap, nominal bubble diameter 30 mm, approximate thickness 2 cm, weighing about 52 gm/m2, marketed by Armstrong in London.

We made this jacket out of gores like those for the envelope, sticking them together with tape (with the bubbles facing towards the inside), and using the envelope (inflated now with air) as a former. Building the jacket and making it fit was much more laborious than we had anticipated. Developing the three-dimensional shape with the gores is quite hard work, which can be tolerated for making the envelope; but one grudges the labor when struggling with bubble wrap....Here is a picture halfway through:

And here is a picture of the completed jacket upon the envelope:

We then filled the envelope with steam, finding this time that the use of two boilers provided quite sufficient steam, and collected the condensate every half hour. The results were:

First 30 minutes, 1750 gm condensate collected;

Second 30 minutes, 1750 gm condensate collected. Exactly the same!

We also realized now that the area of the umbilicus should not be ignored any longer, since it was becoming comparatively more important for condensation because it wasn't insulated. We estimated the area of the two umbilical branches in use at 0.1 m2, which according to the previous result should condense 70 gm of steam in 30 minutes. Therefore, deducting 140 gm/hour for the umbilicus and 2 x 2 x 160 = 640 gm/hour for carry-over water (total deductions 780 gm/hour with two boilers), and dividing by 3.5 square meters of area, we get the result:

Condensation rate - 780 gm/m2.hour

In other words, about half of the naked value.



We next added, on top of this bubble wrap jacket, a further jacket made of "Mylar" type polyester film aluminized upon both sides. This was of high quality, "Tonzon" trademark, nominal thickness 19 microns, weight about 31 gm/m2.

As before, we made this jacket out of gores and stuck them together with tape, using the envelope and bubble wrap jacket as a former, this time while keeping them full of steam (rather than air). Again, building this jacket was rather laborious. Here is a picture halfway through:

Looking closely at the above photo, one can see one of the problems with using bubble wrap as insulation material in the Steam Balloon context. Inevitably some steam does leak through the fabric of the envelope, since it is not absolutely impermeable (and in fact is not required to be), and this steam is accumulating upon the inside of the bubble wrap; you can see the minute condensed water droplets. In fact, the steam is actually penetrating into the insides of the bubble wrap bubbles and condensing on their inner surfaces! From the insulation point of view this condensed water may not make a great deal of difference, but if this were an actual Steam Balloon and flying, the weight of condensed water would be entirely parasitic, and would accumulate steadily. There would be no possibility of dumping it during flight, and even between flights it would be very difficult to get rid of.

Here is a picture of the finished article:

We filled the envelope with steam and collected the condensate every half hour. The results were:

First 30 minutes, 1100 gm condensate collected;

Second 30 minutes, 1250 gm condensate collected.

Therefore, deducting 780 gm/hour and dividing by 3.5 square meters of area as before, we get the result:

Condensation rate - 450 gm/m2.hour

We see that, as the insulation performance increases, the deductions become more important.



We next ran a test using this "Mylar" type aluminized film jacket only, i.e. with no bubble wrap layer. The results were:

First 30 minutes, 1550 gm condensate collected;

Second 30 minutes, 1500 gm condensate collected.

Third 30 minutes, 1600 gm condensate collected.

Therefore, deducting 780 gm/hour and dividing by 3.5 square meters of area as before, we get the result:

Condensation rate - 660 gm/m2.hour

The aluminized film by itself is surprisingly effective! actually better than the bubble wrap by itself. However, as was the case with the first run of of experiments, we think that this result would not scale up, because it seemed that a great proportion of the insulating effect was due to a layer of air between the envelope and the metallized film, which stood away from the envelope in many places due to its stiffness. This would not be the case with a metallized film jacket upon a full-size balloon envelope, unfortunately; the weight of the jacket would inevitably pull it down tight upon the envelope, at least over the upper two-thirds or so.



In the next test, we reversed the order of the jackets in Test Three, and put the "Mylar" type jacket next to the envelope, with the bubble wrap layer over it, again with the bubbles facing towards the inside. The results were:

First 30 minutes, 1000 gm condensate collected;

Second 30 minutes, 1100 gm condensate collected.

Therefore, deducting 780 gm/hour and dividing by 3.5 square meters of area as before, we get the result:

Condensation rate - 380 gm/m2.hour

We found it surprising that this order of the two jackets, with the Mylar on the inside, proved actually more effective that the opposite order employed in Test Three.



In the final test, we added another "Mylar" type jacket over the two jackets in Test Five, so that now we had a Mylar jacket next to the envelope, then a bubble wrap layer over it with the bubbles facing towards the inside, and finally another Mylar jacket on the outside. The results were:

First 30 minutes, 780 gm condensate collected;

Second 30 minutes, 950 gm condensate collected.
(This last result seems to us slightly anomalous, but there it is)

Therefore, deducting 780 gm/hour and dividing by 3.5 square meters of area as before, we get the result:

Condensation rate - 270 gm/m2.hour



First, it is encouraging to have obtained a definite value for the condensation rate of a naked envelope, but we should be cautious about extrapolating this value to the case of a full-size Steam Balloon. Heat is removed from the surface of the envelope, very largely, by the air near it becoming warmed up (by means which, in my opinion, are slightly mysterious), and rising due to its decreased density, so breaking away from the surface in swirls; this is popularly termed convection, but that is a simple name for a very complicated process. I can't imagine that the heat loss value per square meter for a large Steam Balloon would be greater than for a small one - probably the reverse because the swirls have to move further to get properly away from the envelope - but only full-scale testing will give us the real picture.

Second, even with these crude types of insulation, we are approaching a tolerable value for heat loss. In the detailed analysis of Steam Balloon operation here, a steam condensation rate of 200 gm/m2.hour is postulated with a jacket weighing 200 gm/m2, and we see that in Test Six we got within shouting distance of that figure with three layers of insulation whose total weight per square meter was only about 115 grams. So we are within shouting distance!

Third, this general type of insulation is wrong, because it is impermeable. We need insulation that can breathe; in other words, it should allow air from the outside to percolate through it little by little, to carry away the wisps of steam that inevitably must escape from the envelope.

However, we have clearly reached the limits of our experimental apparatus with Test Six... That is, the amount of blow-over steam to be deducted, etc., has become comparable to the amount of condensation to be measured, and if the insulation becomes any more efficient we will be subtracting one large number from another large number to get a small number, so that the inaccuracies will come to dominate our results. In other words, we need a much larger test rig, preferably with smaller initial inaccuracies. This was the motivation for building the setup for the mid-scale experiments.

But before that we did another run of small-scale experiments, the third series, with a somewhat refined procedure.

Back to the small experiments page...

Back to the experiments page...

Site map

Back to the top page...