How I tested the stock engines of Squad's Kerbal Space Program.

[Dispatcher]

Now with photos.

I recently updated my engine comparison chart file at the SpacePort, to add the RAPIER engine. Its the same one which is also available to view or grab using a web browser (see my sig line). Apparently some people prefer to download it, or are not encountering it via browser. A user at the SpacePort commented:

“... My only suggestion would be to add (a) small text file to go along with it that goes into more detail about the testing parameters. Also, maybe try separating the engine types (jets, rockets, radials, etc.) more clearly with more space between them; although I also realize space is at a premium ...â€Â

Since I wanted to minimize the quantity of documents in this download which a person would need to read, I have no plans to add such a text file, although some people might find the information interesting. The commenter ends by mentioning the very reason why I have not expanded the chart in order to separate the engines by category. However, I think that the KSP forum is the perfect place to provide details which were suggested in the comment, since I can link from the chart's SpacePort comment page to this post and anyone who reads this forum post will have a better idea of what was involved in testing for my chart.

Two tests were performed: one for efficiency and one for, in practical terms, “strengthâ€Â. These are specific to the moment of launch, in terms of mass; also of course limited to local Kerbin conditions.

What this thread is not. It is not a discussion of what one's favorite engine is. Nor is it a discussion on how or why strength = thrust and efficiency = specific impulse. For the latter, it is obvious that such relationships generally exist. It is also not a discussion as to whether or not any testing needed to be done at all (hint: I simply wanted to do so, and I enjoyed the process).

As mentioned, I will give details on how I tested KSP's engines. I'll start with the efficiency test. In order to be relatively consistent, I did certain things; some of which apply to both tests. Of course, there were the engines. I used one of each engine tested, launched vertically, with SAS enabled. Since some of the engines are jets (or have a jet mode), I used the circular air intake. In order to equalize mass and drag, I tested all engines by including the intake; with liquid rockets, SRBs, etc., where feasible. This in part offsets the fact that jets do not use the stored oxidizer mass. Since a control pod is required, I used the small RC-001S RGU. I launched each engine from the launch pad surface and allowed it to fly without any piloting. Any engine which did not retain a stable upward flight was tested with 3 AV-R8 winglet control surfaces attached, since there is a trade off with mass between vectoring and non vectoring (gimbaling) engines. This can be readily understood by comparing the LVT 30 and LVT 45 engines. In some cases, the winglets acted as supporting tripods for small engines.

Specific to the efficiency test, I chose to use the small Oscar-B fuel tank for each engine. Obviously this limited the amount of fuel/ oxidizer available to the engine. The test results were of the maximum altitude each engine attained (a combination of thrust and inertia or momentum). Factors affecting each flight were the mass and drag of the engine against its thrust.

There were of course some exceptions and considerations. Jet engines on one hand could not use the oxidizer mass, but on the other hand used oxidizer from the air; of course the atmosphere was a self limiting factor for these. For the ion engine, the PBX-150 xenon container was substituted for the fuel tank (while their mass differs, the size is consistent). It also required electricity, which I did not require adding to the other engines tested. The small FL-R25 monopropellant tank was used by the RCS units. It turns out that the small SRB (Sepratron 1) mass is nearly the same as the mass of the small fuel tank. Since the efficiency test involves limiting the quantity of fuel available, each size of SRB is considered as if it were the small SRB in these figures.

This leads to another consideration: the use of radially attached engines implies that there is more than one engine used (at least for a stable flight). I tested radials using 3 engines (the least number of engines for a more stable rocket flight). So the results for radial engines is for the use of 3 engines. However, the engine thrust listed is still for a single engine. Its actual use thrust is 3 times that in this case. The chart also has figures for 2 radial engines, for use with planes. These were simply calculated based on the 3-engine results. Three engines tested yielded no vertical flights; the RCS7 thruster, ion engine and a Kerbal with EVA pack. Horizontal results were substituted for these.

Now for the “strength†test. The prior criteria remain for this test. In this case, however, the fuel mass was not minimized; rather, it was maximized. I used such combinations of stock fuel tanks that allowed me to do this to a suitable degree of accuracy; sometimes a very small amount of fuel made the difference between an engine reaching a defined apoapsis and it reaching escape velocity. The test results reflect the mass an engine (including its own mass) was able to lift, without crashing or leaving an upwardly stable flight path.

Another difference is that I used the launch stability enhancers (3 where possible), setting the altitude to the maximum start of 150 meters (which is about 80 meters above the launchpad). This allowed some engines to drop and recover prior to attaining their vertical upward flight. Poor Kerbalnaut! For him, I had to provide a pod and a platform from which he “volunteered†to step off. Others which failed to achieve a vertical flight were the RCS thrusters and the ion engine (as expected, really, but I tried anyway). An important difference between the SRBs and the other engines is that most of the mass lifted by the SRBs in this test was dead mass, rather than fuel which could be converted into thrust.

When sorting the engines separately for maximum mass lifted (“strengthâ€Â) and maximum altitude for the efficiency test, some of these were at the top half of the list for both sorts. My chart color codes them as being “strong and efficientâ€Â. It also addresses some other data or questions not discussed here. All engines are sorted in this chart for the maximum mass at launch. While this mixes some types of engines, it can be readily determined which is which by experienced players. I recommend that new players check out the Wiki page on stock parts.

While most of the thrust to weight ratio figures listed come from the Wiki, my figures for the RAPIER (air mode) differ. Since my actual maximum mass results differ significantly between the 2 modes, it seems to me that the TWR between modes should differ as well. In any case, most RAPIER launches will occur from Kerbin's surface, so the air mode TWR would normally prevail in practical terms. It is also true that while in the atmosphere, RAPIER engined craft can take advantage of lifting surfaces, which were not considered in the two tests discussed here. I'm open to suggestions for improving the chart or for correcting any errors in data or methodology.

The chart includes other data not (yet) listed in the Wiki, such as the electricity generated by engines which do so.

Edit: version 0.23.5 has improvements. I changed column Q from useless row sums into an added test: drop test results similar to column O altitudes (with maxed out mass), but with about 50% of the mass being non fuel mass (including unused payload fuel tanks). This is a reasonable test which resides halfway between the maxed fuel test above and what would be a minimal fuel test (which would yield a hover or no flight at all), in terms of flight results. Jets are so efficient that I minimized the fuel mass to 7 or 8% for those two engines, and still had fuel remaining as a portion of the payload. I also found and fixed a few minor data errors and cleaned up the layout a little, while clarifying some of the explanations. I think this version is much more useful than earlier versions.

Thanks for reading!

Edited January 19, 2014 by Dispatcher
Added example pics.

[Dispatcher]

Bumped this because a picture is worth another thousand words.