Nov 7 meeting notes:

Location: Long Range Systems

In attendance:

John Carmack

Phil Eaton

Russ Blink

Darin Smith

Next meeting at Long Range again.

Tested side by side, the silver foam from the catalyst packs used unsuccessfully last week did show lower reactivity than fresh foam, so we cut a completely new foam catalyst pack for the motor tonight.

The peroxide used last week may have had some boil over from the unpurified peroxide, which could have poisoned the pack.  The peroxide was denser this time (84%), which gives some support to that.  We can test some of the disks from today's tests and compare.

We need to get some nitric acid to experiment with cleaning the catalyst packs.

The silver screen order finally got processed, so we will be able to test screen versus foam catalysts soon.  A benefit of the screens is that they are pure silver, not just plated.

After our runs tonight, there were parts of the foam where the silver had been completely stripped away.  The metal plating shop had suggested sintering the silver to the nickel foam, which we should consider if the problems are still there after we get the distribution worked out.  I expect that it was mostly "washed off" by 450 psi flow, rather than melted off.

We have the cart to assemble all the fill system onto, but it needs some repair work before we can tie everything down.

We finally got a DC vacuum pump wired into our fill system, so the entire loading process is now pushbutton operated from the fill console.  We also have a vacuum gauge inline now, which is helpful.  The only remaining thing to add to the loading system is a pressure outlet on the vacuum catch can as a safety feature.  Even just adding another glass tube out of the stopper and putting a vacuum cap on it would help, although that probably wouldn't solve the case of accidentally opening a pressurized tank into the vacuum catch tank instead of the peroxide dump.

I left eight sold state relays with Phil for the VTVL platform driver board.

We added some bright LED in parallel with the solenoid, so we could watch the delay between the warm up pulse and the exhaust plume.  This was a good thing.  We should plan on having LED on both the driver board, and at each solenoid.

Even though we weren't testing partial duty cycles today, we used the PWM program's script driven mode to precisely time our warm up pulses, which worked great.

Our test stand isn't very well suited to these small motor firings, having a 100lb load cell and a bit of friction in the sliding bed.  I will order a 25lb load cell for us, but we are still waiting on our larger load cells to come in.

We really need to pick up a big crescent wrench for the regulator and engine disassembly.  I have to hit Home Depot tonight, so I'll grab one.

X-L Space Systems is looking like a better peroxide supplier for our needs than FMC.

We need to get a Teflon siphon pump for transferring from larger containers.

We should rescale the load cell meter from kg to newtons.

I have a few programming things to change to make based on today's work:

Single pulse command line option for PWM didn't turn the solenoid off correctly.

The load cell logging program should be modified to strip extra cr/lf pairs.

The PWM program should integrate load cell logging and add another column of output next to the load cell reading for the current duty cycle.  This will simplify testing, completely eliminate manual cropping of the data, and allow us to see the exact delays between solenoid command and thrust.  I was switching between two consoles today to run LOADCELL and PWM.

I will modify PWM to optionally control the launch valve by using separate drive channels for on and off.

We had our act together pretty well today.  We coordinated load cell reading, PWM control, and video capture without any hitches.

Test 1

-------

100 ml peroxide

Large (45mm?) metering jet

250 psi

Three 200 ms warm up pulses separated by two seconds each, then full open.

The first warm up pulse had a long delay before catalyzing, but the second one was quite active, so we decided to only use a single warm up pulse on the following tests.

The main firing started off perfectly clear, but clouded up after a half second.  We decided we were still pumping too much peroxide through, so we moved to a smaller metering jet for the next run.

Test 2

-------

100 ml peroxide

Small (23mm?) metering jet

250 psi

200 ms on, 2000 ms off, full open.

The motor was obviously still warm, as the single warm up pulse was fairly vigorous.

The first quarter second was cloudy, but it cleared for rest of the run.

Peak thrust decreased, but it held together for the entire run.  It did demonstrate an odd thrust bump just as the last of the peroxide was spraying into the engine.

We decided to increase the pressure, while keeping the small jet.

Test 3

-------

100 ml peroxide

Small metering jet

450 psi

200 ms on, 2000 ms off, full open.

The first quarter second was cloudy again, but it cleared up briefly, then clouded up for the remainder of the run.  The spike before blow down at the end was more pronounced.

The thrust when the engine was fully catalyzing was about 35% greater than the 250 psi run.

Test 4

-------

Remaining peroxide, about 300 ml

Small metering jet

450 psi

200 ms on, 2000 ms off, full open.

The behavior was similar to test 3, but the longer running time showed that once it went cloudy it never does clear back up.  There was a longer period of sputtering higher thrust at the end.

Observations

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

Even though the runs were close enough together that the motor did not have time to cool all the way down, the warm up pulses were generating les thrust on each succeeding run.  The catalyst pack may have been degrading with each hot run.  When we took it apart, the core had some areas completely stripped of silver.  On the other hand, we may just be aliasing in the 12hz sampling rate of the load cell meter.

Our current theory for the increase in thrust just as the peroxide runs out is that the gas flows through around the remaining liquid in the catalyst pack instead of forcing it out, allowing the last remains to completely catalyze, unlike the steady state, which was forcing it out through just the center of the pack before it had time to decompose.

We were aiming for about 5kg of thrust from the small motor at 100% duty cycle with 450psi.  Test 1 had a period of apparently proper functioning at 250psi, but only gave about 1.6kg of thrust.  Scaled up, that is still less than half what we are aiming for.  Once we get things working properly on the existing jets, we will need to try larger ones.  We don't know yet where the solenoid becomes the limiting factor, but even the large jet was a very noticeable restriction over no jet at all when we were water testing.

Upcoming Tests

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

Once we get a larger bulk supply of peroxide, we will test Juan's motor.  While working with our distillation unit, the small motor lets us get a lot more results each week.

The large jet at 250psi should be our test case for improving the catalyst pack.  If we get a clean run all the way through, we have made an improvement.

We are going to design several different spreading plates for the forward end of the catalyst pack.  Darin has a friend at a laser cutting shop, so we are going to get a batch of different ideas tested.

Our current plate has fairly sizable holes in it, which may be allowing the entire flow to channel primarily down only a few holes in the center.  The holes should probably be 20mm or smaller for this size motor.

If you pour water into our foam packs, it naturally tends to flows out the center even if you spread the water over the entire top surface, so it is probably best to have injector holes only near the edge, and let it flow towards the center by itself.  There may be more danger of channeling around the edges, but we are currently burning out the center of the pack, so I don't think that is a primary problem.

We want to try significantly compressing one of the foam catalyst packs to increase its pressure drop.

Cutting out all the little circles of foam is a pain, and cutting silver screen is going to be worse, so we want to try making a pack by just rolling the foam up into a cylinder, then crunching it into the motor.  We will probably have to shim it with a few disks to keep the injector plate at the correct height.

We had about 600ml of peroxide this week, so if we have a similar amount next week I think we should make two firings in each of three new configurations:

New plate over a hand packed catalyst.

New plate over a compressed catalyst pack.

New plate over a rolled and compressed catalyst pack.

If the first run is clear at 250psi, we will try the second at 450.  Otherwise, we will just see if the behavior improves on the second test when the engine is warmer.

Phil: we should get some more of the foam plated, and get it sintered on this time.  Better flow distribution will probably help avoid stripping the silver on the runs we made today, but we are going to need to flow over twice as much in the final form, so it will probably remain an issue.  If we have enough of the existing plated foam for three separate catalyst packs, we can probably live with that at the low pressure / flow rates.

Russ: don't forget to refill the nitrogen bottle and get more peroxide feedstock.

We still need a 1000+ psi regulator for the nitrogen.  We should plan on a full test matrix of pressure and jet size changes once we nail the catalyst pack.

Eventually we will want a couple pressure taps and multiple channels of data collection.  It may also be interesting to allow the flight computer to watch chamber pressure, so it can tell if one of the engines is flooding.  If catalyst stripping or poisoning remains an issue, this would be a valuable thing to have.