October 25, 2003 notes

 

Engine Improvements

 

On Tuesday, we did several things to test improvements to our highly successful combination of last week.

 

We used the short (2”) chamber extension, and packed it with one 1” thick foil monolith catalyst and covered that with 110 grams of catalyst bale.  This extension is 6” ID, and the monoliths are only 5.5” OD, so we packed the bale catalyst around the sides as much as possible.  It turns out that the Champion diesel glow plugs we are using have the same thread diameter and pitch as 1/8” NPT, but with a straight thread.  This allows us to weld a 1/8 F-F pipe union on the outside of the chamber, and thread the glow plug into it.  With some anti-seize on it and a bit of torque, it seems to seal even though the taper is only on the female threads.  The glow plug tip does protrude slightly past the end of the union into the chamber, but we didn’t expect quenching to be a problem.

 

I made new three-pass catalyst plates.  Because the upper catalyst doesn’t get much over 250F, the monolith has plenty of structural strength to support itself without the cross bars, so I removed those for better flow.  The upper plate is also 0.050” thicker to give a bit more turn-around volume.  Russ welded on some spacers at the top of this plate to guarantee that it wouldn’t bow up and close off any of the top flow.  Instead of standing the perforated flame holder away from the bottom of the lower plate, we welded it directly to the plate, which gives a 0.25” gap between the bottom of the catalyst and the perforated plate.

 

Somewhat to our surprise, the short chamber did hold plenty of catalyst to function properly.  We made several good runs, proving the following:

 

We don’t need all the volume and catalyst we had in the long chamber.

 

The glow plug doesn’t need to protrude into the chamber to light the preheat vapors.

 

A single glow plug works fine.

 

The glow plug is only necessary for the preheating, it doesn’t play a part in the long run functioning.  We made some fairly long runs without the glow plug powered after opening the throttle.

 

The new three-pass-pack-plates with the forced standoff at the top did provide even flow all around the engine.

 

The perforated plate worked fine welded directly to the plate, the standoff wasn’t important.  It looks like the perforated plate, in combination with the 3-pass pack, was the magic ingredient, inducing lots of little turbulent zones instead of letting the output of the top pack rush straight down.

 

Now, a couple things that were only partially successful:

 

We added a small solenoid in parallel with the ball valve so we could do preheats with more control.  I had hoped that it would be possible to just run a small amount of propellant through until it completely preheated, but it definitely needs pauses to get the bottom pack glowing red hot.  A constant low flow will get the pack hotter than simple peroxide decomposition levels, but not to the point where the exiting water vapor is superheated to the point of clarity, which usually coincides with a red-hot pack.  My theory is that the output of the 3-pass pack has free oxygen, methanol vapor, and quite a bit of liquid, moderate concentration peroxide.  Because a lot of the oxidizer is still in liquid form, this gives a very rich, relatively cool burning, that isn’t hot enough to force the rest of the peroxide to decompose.  If you let a slug of propellant into the engine and close the valve, it will sit in the first two channels of the 3-pass pack, giving lots of time for the peroxide to more fully decompose.  This fills the engine with pure oxygen, so when the next slug of propellant comes in, it has a lot more oxygen to burn with.

 

We tried a couple runs with food grade peroxide.  The first run took much longer to preheat, and started out cloudy, but did clear up and run clean.  The second try ran cloudy the entire time, probably because we weren’t patient enough with the preheat.  After these runs, we did another run with unstabilized peroxide, which behaved normally, so the catalyst didn’t seem to be poisoned.  We can probably make food grade work, but we are going to try unstabilized semiconductor grade next, if we can’t get FMC to sell us any more of the 50% propulsion grade.

 

On Saturday, we replaced the foil monolith catalyst in the extension chamber with more of the catalyst bale.  We like the monoliths a lot more for repeatability, but we are somewhat motivated by the fact that we have exactly enough monoliths to possibly build a ship-set of four engines if we use bale catalyst on the bottom.  We have more catalyst on order (we are going to 600 pores-per-square-inch, from 400, for all of them), but it is 4-5 weeks out.

 

Phil built the catalyst with 680 grams of bale, and pressed it to 1000 psi indicated on our press gauge.  This was a very tight press.  The engine worked fine, preheating seemingly better than before, but we weren’t logging data at the time.

 

We then swapped the small nozzle for the 2” throat nozzles, and made a couple test runs.  The thrust was surprisingly about the same as the much smaller nozzle.  It turns out that the catalyst was packed tightly enough to create a huge pressure drop.

 

When we pulled the engine apart, we had one surprise – the perforated plate on the bottom of the new 3-pass pack had burned through in a spot.  We believe this was due to it being pressed directly onto the lower catalyst without an air-gap.

 

We rebuilt the catalyst with 330 grams of fresh catalyst bale, and only pressed it enough to get it seated with a half inch or so gap under the top pack, which barely registered on the press gauge.

 

Unfortunately, we decided to change two things on the next test, the catalyst density and the mixture ratio.  We went from our normal 5:1 by volume O:F ratio, which is lean-burn, to 2.5:1 by volume, which is rich burn.  They both, in theory, give about the same Isp, but the rich burn is 170 F cooler in operation.  We were not able to get this self-preheated very well, and it didn’t burn clean.  There was lots of pops and flame coming out of the nozzle during the preheat, but the insides only got partially red hot.  We tried again with the mixture ratio at 3:1 by volume, but we couldn’t get that to work either.  At this point, we wanted to go back and try our normal mixture ratio, but we were completely out of methanol, because my last order didn’t arrive on time.

 

Just for something to do, we tried an ethanol mixture at 8:1 by volume, which is lean burn at about the ratio we normally run the methanol, but it couldn’t be preheated either.  Assuming that our normal mixtures functions properly on Tuesday, this does show that using methanol for its catalytic properties is actually important.

 

I have started machining the 3-pass plates for the 12” diameter engine, and we are having an extension chamber for that fabricated.  Firing is probably two or three weeks away.  If we hit our expected numbers with that, we will start the fabrication work for the complete X-Prize set of four.

 

 

Mixture Ratios

 

Here are the results for various 50% peroxide / alcohol mixtures using the Isp code available at http://www.dunnspace.com/.  Surprisingly, I had to add data for the various alcohols to the propellant library.  Mass and volume are O:F ratios.  Isp is with a 300 psia chamber pressure with a nozzle sized for expanding to 14 psia at sea level, shifting equilibrium calculation.

 

Good online chemical data book: http://webbook.nist.gov

 

Heats of formation:

 C1H4O Methanol       –57.1 kC/mole             –239 kJ/mole

C2H6O Ethanol           -66.4 kC/mole              -278 kJ/mole

C3H8O Isopropanol    -76.0 kC/mole              –318 kJ/mole

 

                                                            methanol                       ethanol                         isopropanol

Mass    Volume            Density             Isp       Temp (K)         Isp       Temp (K)         Isp       Temp (K)

----      ----                  ----                  ----      ----                  ----      ----                  ----      ----

12        9.6                   1.138               133      923                  151      1174                160      1299

10        6.7                   1.130               144      1060                163      1341                170      1454

8.6       5.7                   1.123               153      1186                170      1448                167      1384

7.5       5.0                   1.116               161      1304                168      1388                165      1317

6.7       4.5                   1.109               168      1414                165      1330                162      1252

6          4.0                   1.103               170      1434                163      1275                160      1190

5.5       3.6                   1.097               168      1393                161      1222                158      1131

5          3.3                   1.091               166      1354                159      1170                156      1077

4.6       3.1                   1.085               165      1315                157      1122                155      1036

4.3       2.9                   1.080               163      1278                156      1077                154      1009

4          2.7                   1.075               162      1243

3.8       2.5                   1.070               161      1208

3.5       2.4                   1.065               160      1175

 

For comparison, 90% peroxide monoprop gives density 1.381 / Isp 134 / temp 1030 K, and 98% peroxide monoprop gives density 1.430 / Isp 146 / temp 1226 K.

 

Obviously, there isn’t a heck of a lot of difference in the performance of the various alcohols.  In theory, isopropanol offers a couple percent better bulk density for a given Isp, but that is way down in the just-doesn’t-matter range.  Methanol reacts better on platinum catalysts, which is a far more important factor for us.  We are 15% or so off from the theoretical Isp values, so there may be some room for improvement, but we aren’t going to do any engine revs specifically targeting performance, because what we are getting is good enough.