Jan 16 meeting notes:

Location: Long Range Systems

In attendance:

John Carmack
Phil Eaton
Russ Blink

Next meeting on Tuesday at LRS for VTVL system checkout with water and more engine runs.

We now have more than sufficient thrust out of the small engines for the VTVL.  Russ is going to start cutting four engines (we should probably keep the current one as a spare test mule) this week.  Russ: you might want to think about leaving space for a pressure port if you can.

We made new catalyst packs out of our batch of newly silver plated foam.  This foam was plated for longer, and was noticeably more brittle.  We tried making the packs this time by just rolling it up instead of cutting out individual disks.  It still catalyzed fine, but we believe that the rolled packs contributed to the more ragged results today.  We will be going back to cut disks.

media.armadilloaerospace.com/misc/jan16.xls

All runs after the first were with 200ml of propellant.

1: new rolled catalyst pack with a couple disks at the top, slightly compressed. 100ml run.  This is more thrust than last time, due to the less compressed pack letting more peroxide in.

2: increased pressure to 450psi, 200ml run.  A little more thrust, holding 5kg, but a bit ragged.

3: new rolled catalyst pack with no disks and no compression, slightly better flowing, increased the peroxide concentration to a bit over 90%.  The run made a bit more power, but was more ragged.  The nozzle did show a little bit of silver deposited on the inside, so this is probably pushing it too much on the concentration.

4: drilled the nozzle throat out from 0.20” to 0.25”.  Averaged 7.5 kg thrust.  Yeah!

5: verification run with the same parameters as five.  Similar good results.

6: Our first attempt at a pulse width modulated hot fire, 20 hz, 50% duty cycle   It was pretty ragged (20% variation), and gave about two thirds the power of the 100% duty cycle.

We didn’t touch the engine between runs four, five, and six, so there was still a little residual tension in the test stand, giving a bit higher starting values.  It was pretty cool to just be able to suck more peroxide in and run the engine over and over.

The later runs with the drilled out nozzle are still making a fair amount of thrust just from the nitrogen blowing through the hot catalyst pack after the peroxide runs out.

The 7.5kg of thrust we are getting is more than sufficient for our needs, so to try and smooth things out a bit, we are going to go back to fairly tightly compressed disk packs like we had last week.  Our pressure drop seems to be very large, but our tank is designed for 1100 psi use, so we aren’t going to sweat it right now.  When we build an optimized system, we will need to investigate more.

The nozzle isn’t very optimal right now, because Russ couldn’t cut a narrow diverging cone at the time, and we don’t have anything resembling a radius on the throat, so I expect we will get another five or ten percent when we make better ones.

We have a few directions to go with developing the PWM.  We will do another run with a compressed disk pack, and we will try some different frequency ranges.  I am also interested in experimenting with a small pressure accumulator between the solenoid and the engine.

One issue that may be significant is that our current load cell meter only reads at 12hz, and we aren’t sure how much averaging it does over that period.  We may have a huge aliasing problem.  We will try a 24 hz cycle time and see if the runs look smoother.  If they do, we are probably aliasing, and the returned values would be in an unknown part of the pressure cycle.  I may need to break down and buy the high frequency sampling meter that I wish I had seen in the first place.

Material test:

We did some clothing peroxide safety tests today so we could see for ourselves exactly what is likely to happen if there is an accident.

media.armadilloaerospace.com/misc/materialtest.mpg

Cotton shirt and denim jeans were basically unaffected in the five minutes or so that we left it on before dunking it in water.  It was only 45 degrees out, so results may be worse in the middle of summer, but this concurs with the advice we had heard before: cotton is good, because it will soak through and give you peroxide itch long before it will combust.

90% Nylon  / 10% lycra and 50% cotton / 50% polyester had minimal effect.

Pure polyester and silk melted into goo, but did not combust.

Leather fizzed a little bit, then no-shit burst into flames quite violently after it had been sitting there for about a minute.

The lessons are simple  wear cotton, and make sure you don’t have any leather shoes, belts, or jackets.

Misc:

Russ finished the new bottle manifold for the VTVL, and it is going to work great.

media.armadilloaerospace.com/misc/vtvlmanifold.jpg

The bottom four ports go to the engines, and we are going to use the top ports for the fill cart quick connect, a pressure gauge, and dual blow off valves.  We can save a fitting chain if we tap one of the 1/8” NPT ports out to 4 AN for the quick connect.  Bob has a tap for that if Russ or Phil want to run by there some evening with the manifold.

I got new, all-metal check valves from McMaster-Carr.  They are just machined pipe fittings, which means that we can replace the male/male pipe union we had between the solenoid and the engine with the check valve and simplify our plumbing.

It turns out that the nice swivel 90 degree fittings I got won’t clear the solenoid mounting bracket, so we are either going to have to add a spacer, or use the non-adjustable 90 degree fittings and wrap with Teflon until they clock to the right angle.

Russ/Phil: when you find an allen wrench, take the screws out of the solenoid mounts and go buy us a box of them.  If we go the spacer plate route, we are going to need longer ones as well.

My wife had her graphic artist make us an Armadillo Aerospace mylar sticker for the dumb rocket airfame.

Phil dug up some candidate battery packs for the VTVL.  We will be tying three 4.4V (is that correct?) packs together for the unregulated supply to drive the solenoids and serve as input to the 5V regulated power supply on the flight computer.  We are apparently going to need to remove the current limiter to get enough power out of them for our needs.  We should build and test a couple packs with appropriate mounting hardware and charging stations soon.

Terry Parks has gotten a quote from Murata for the rate gyros, but they have a minimum order of 100 parts.  Prices are:

ENV-05D 100pcs $29.90ea
ENC-03JA 100pcs $13.00ea
ENC-03JB 100pcs $13.00ea

Russ also found a Delphi auto part rate gyro:

http://www.delphiauto.com/pdf/delcopdfs/IntellekARS.pdf

http://www.delphiauto.com/pdf/delcopdfs/IntellekARLASM.pdf


I just got a Delphi distributor number from Bob, so I am going to check into getting some of these.  If that doesn’t work out, I will probably buy 100 of the Murata gyros and just give them out to any experimenter that wants some.

The magnetometer gives nice clean data, but the inclinometer is useless on a moving platform.  We will probably run the magnetometer just to collect the data, but without another axis, we can’t use it for attitude control.

John Dom sent me a link to a supplier of pure silver foam, which I have requested a quote from:

http://www.porvair.com/fuelcell/metalfoam.htm

I have been working with the single board computer a bit.  I installed linux on the dev system drive, and I have booted dos off of the onboard flash Disk On Chip, but I haven’t configured a linux system on the DOC yet.  I did a little bit of work making sure I know how to talk to our sensors over the serial ports, but I haven’t built a real architecture yet.

I am still fretting about the exact system choices.  The current dev system has a power supply, HDD, floppy, and CD in a nice metal enclosure, and the SBC (EBX form factor) sits on top, with power and two ribbon cables connecting it.

The SBC has video and Ethernet on it, which could have been part of the development system, and might have then let me use a completely PC104 form factor.

I need to get some more experience one way or the other, in any case.  I ordered a couple grab bags of standoffs and connectors for us to use in putting various boards together.

Russ and I discussed possibly putting the solid state relays on PC104 boards, either just for mounting purposes, or actually having it decode an IO port so we don’t need to ribbon cable from the parallel port to that board.

Stuff to get:

Refilled nitrogen tanks (get two)
A new vacuum pump for when this one dies
Light chain for tethering the VTVL.
Female 4 to male 1/8 NPT if we don’t tap the manifold for the quick connect
At least 2 more 1/8 to 3 90 degree fittings if we don’t make spacers for the solenoids
Pressure gauge for the VTVL manifold
Blow off valves for the VTVL manifold
Change the PWM program to allow it to vent the engine for 30 seconds without logging the data.
Long parallel port cable for the tethered VTVL testing
Polycarbonate safety glass
Still need to find hazmat stickers