March 20, 2004 notes

 

Working engines

 

We finally have a solid engine configuration!

 

The angle bar flameholders were causing combustion instability at high thrust and melting things underneath them, so we focused on using the heavy perforated metal plates as flameholders.  A single plate retaining the cold pack with the spark plug underneath it worked ok, but if we added a second one right below the spark plug, we got much stronger and more even flameholding.  We also started using the longest reach spark plugs I could find – NGK PLFR5A-11, which gives over an inch of reach to get it past the 5.5” to 7” step transition.  This would light reliably if the pack was completely cold, but it still had problems when we tried to light it a second time.

 

We had found in the past that the voodoo ritual of briefly opening the valve very wide when it wasn’t starting would allow it to start again, so we pursued that and found that it did seem to be a reliable way of starting a hot engine.  John Carr, a local combustion chemist that I had been corresponding with, had stopped by on Tuesday to see if he could help, and he was able to offer us a logical explanation for the behavior.  When we had a run that wouldn’t start, you could smell formaldehyde, which is what methanol breaks down to on a catalyst (along with hydrogen).  He also pointed out that at 100C and the molar fraction of methanol in our mixture, there actually isn’t all that much methanol completely vaporized.  Apparently, if the monolith catalyst in the cold pack has gotten much of any heat soaked back into it then it will start attacking the methanol as well as the peroxide, and if much of it gets catalytically burned, then there isn’t an ignitable O/F ratio by the spark plug.  A large slug of propellant will cool the catalyst back down so it operates at the level we want it to.

 

We have now done over a dozen full thrust firings with this procedure, with no problems.  If the engine is hot, it isn’t even worth trying to start it at the normal warmup flow rate, we just start it right off with a big slug of propellant to wash the heat down, then drop it to the normal warmup level in the 12% - 20% throttle range.  We also found that it didn’t require a particularly strong spark, so we went all the way from the MSD10-Pro back to the MSD small engine ignition controller with no problems.  This saves a non-trival amount of mass, power, space, and several thousand dollars.

 

With two thermocouple probes in the engine we have seen a couple anomalies in the hot pack behavior, where sometimes both the upper and lower sections of it stay hot for the entire run, and on other runs the top section collapses down to about 250 C, while the bottom section stays at 1000 C.  It doesn’t seem to effect the engine, but it implies that we have more catalyst than necessary, and I don’t like the distinction.  The existing hot packs have two layers of 500 grams of rings each in a 7” diameter, which is quite a bit more than what we had in the 5.5” engines.  When we originally tried a single layer of 600 grams of rings in a 5.5” engine the rings on the bottom had too much pressure on them, resulting in many of them collapsing and increasing the pressure drop.  With the 7” ID chambers we should be able to have more rings in a single layer both because of the increased area and because the pressure drop is lower due to the greater chamber to throat contraction ratio. 

 

We considered making a 600 gram layer, but many of our catalysts seem to have a minimum critical depth regardless of the mass or flow, so we bumped it up a little bit to 700 grams.  In the interest of avoiding an unwanted flameholder right above the rings and to save some mass, we used a 10 mesh 316 screen on top to hold the rings in place instead of a disk of perforated metal.  This engine warmed up considerably faster than the older engines, which is to be expected due to having less than half the mass in the hot pack (700 g vs 1000 g of catalyst, only one heavy perf plate, and a screen instead of a light perf plate).  The temperature profile was great, actually slightly hotter than normal.  We then did a 15 gallon run with no problems, but the side of the engine showed some uneven heating.  When we did another long run there were combustion oscillations.  We cut the engine apart and found that the 316 screen on top had melted / oxidized  away and clumped onto the rings below it, allowing the rings to get all out of shape and partially block some of the flow.  The fact that the engine ran completely clear even with the rings not evenly distributed implies that we can get by with even less, so our next build will probably only use 600 grams of rings, and we will go back to using a medium weight perf plate as a top retainer for the rings.

 

We also did a test with opening up the main propellant flow a bit earlier, and found that 600 C at the bottom of the engine isn’t enough, but 800 C is.  While we could probably get by with a fixed warmup profile on the vehicle, I decided it is worthwhile to go ahead and add thermocouple amplifiers on the vehicle so we can get chamber temperature as well as pressure.  This will mainly be useful at startup, but it may give us useful data during flight if something goes wrong.  I am going with these amplifiers: http://www.axiomatic.com/thermo.html To avoid having to build new cables for each engine, I am going to start sharing power and ground between the pressure transducer, valve pot feedback, and thermocouple.  Considering the magnitude of the noise we have at the PC104 system, I don’t think this is going to hurt us.

 

We have all the materials to put together a ship set of engines with the new design, so we are shooting for another full vehicle warmup test next weekend.  If that works, we will be aiming for a hover test the following week.

 

While driving home, I think I also put together another piece of the puzzle.  While scavenging parts from some older engines, Matt noticed that there was a blued section right in the middle of some of the screens above a monolith in one of the cold packs.  We always see some heat effected areas around the outer edge of the screens, but we don’t usually notice them in the middle.  I realized that there is something in common between the center and the outer edge – the top spreading plate has the center blanked off and the holes are kept some extra distance from the outer edge.  This design makes sense for a purely catalytic motor, because you want to avoid channeling around the outer rim and taking too much of a high velocity hit right under the inlet, but with our mixed catalytic / combustion engine, the areas with no propellant feed (there are 10 screens between the spreading plate and the monolith, so it spreads out some, but the edge and center are still going to get less) become effectively internal flameholders.  This probably explains why the outside edge of the monoliths gets all pruned up after running for a while – the outer edges aren’t sitting at 100 C like the rest of the catalyst, they have actual burning in them.  These are probably also the local hot spots that cause the startup problems.  If we eliminate these, we may not even need the initial slug of propellant to cool it down.  We aren’t going to change them for this engine, but the next time I make spreading plates I am going to add an explicit radial ring as close to the outside edge as we can get and still braze the plate in place, and just run the grid all the way across the center without a blanking spot.  A little channeled propellant doesn’t hurt us, because it will just be burned in the hot pack.  If that keeps the monoliths in better shape and possibly avoids the hot start issue completely, it will be great.