Apr 10, 2001 Meeting Notes

 

In Attendance:

 

John Carmack

Russ Blink

Phil Eaton

Neil Milburn

 

New supplies:

 

Batteries and plugs for an Air Port power pack.

 

On order:

 

PC104 16 channel, 16 bit A/D board

Life jacket for pilot of manned ship

VOX headset for pilot of manned ship

 

To get:

 

Mil-Spec wiring, terminals, crimpers, etc

Several gallons of distilled water

Tide for washing tanks and plumbing

Dilute nitric acid

Neil: short section of hard line for pressure transducer, 1/8” NPT male on one side, ¼” NPT female on the other

Silicone break fluid for pressure transducer

 

 

We haven’t gotten a confirmed delivery date for our peroxide yet, but we are hoping it will be before our next meeting.

 

ShowDataq.exe now takes parameters for sampling a single channel at 240hz, or both channels at 120hz.  Output is in hex, decimal, and volts.

 

Gps.exe now accepts both radio modem input and UDP input, so it should be usable by both rockets this weekend.

 

During testing, I did notice that my consumer Garmin GPS-III was displaying noticeably different positions than the GPS-35 OEM module connected to the flight computer.  I am still curious how much of the difference is due to firmware / hardware issues versus fundamental issues.  At some point, I want to try comparing two absolutely identical GPS units.  If the errors are exactly correlated, an attitude sensing system could be built out of the GPS-35 modules with the double precision floating point output.  It would only be once a second, but the timing pulse signal line could be used to exactly mark the time, and the flight computer could adjust it’s gyro integration based on the values at the pulse, even though the serial data wouldn’t be in for another quarter second or so.  I don’t have high hopes for this approach, but we are still looking for an “obvious winner” in the absolute attitude sensing department.

 

Russ has the sensor board mostly working with the Gyration ASICs now.  The roll axis isn’t working right, but since we aren’t using that yet, we should be ok for a while.  The precision is drastically better than before – it only has LSB noise with a full 12 bit range.  I haven’t checked the integration error yet, but I expect much better results now.  Assuming we don’t wreck our electronics at the HPR launch this weekend, we should be ready for another guided flight test next week.

 

The new Crossbow sensor catalog arrived yesterday, and they have a few new fiber optic laser rate gyro packages.  The prices aren’t listed, but the micromachined ones were in the neighborhood of $3k - $5k, so I wouldn’t be surprised to see the laser ones up around $10k for the 6DOF inertial packages.  Another interesting point was that the laser products all require 15V power.

 

We got the GPS-35 mounted to a PC104 board, and the entire stack put together with metal standoffs.  We added a tie wrap around the PCMCIA card, so it shouldn’t pop out.  Neil still has to do a little more work to get everything mounted correctly, so we are going to have a brief meeting on Friday evening for final integration.

 

I finished the new flight computer software, but for some reason, the new CPU board’s serial ports don’t seem to be working.  I need to track that down in the next few days.

 

We should be launching two different electronics packages this weekend:

 

Neil’s level 3 rocket will be carrying our full electronics package, minus the solid state relay driver board.  We will be getting GPS data, accelerometer data (although it will be pegged at 4G during the initial boost), and rate gyro data.  This will also let us test the range of the IEEE 802.11b wireless network.

 

My level 2 certification rocket is a 4” diameter PML Tethys stretched with a 2’ electronics section.  It has a consumer Garmin GPS-III connected to a Tigertronics raw 1200 baud radio modem, connected to a 200 mW radio.  This will give us another GPS trace, and let us compare the conventional radio modem with 802.11b.

 

 

We did a set of water flow tests tonight to compare the effects of check valve size, solenoid size, and line size.

 

100 psi pressure, 500 msec pulse, three feet of line.

 

-6 line   big solenoid      no check          :           64 ml

-6 line   big solenoid      ¼ check           :           62 ml

-6 line   big solenoid      1/8 check         :           41 ml

-6 line   small solenoid   ¼ check           :           34 ml

-6 line   small solenoid   1/8 check         :           31 ml

 

-4 line   big solenoid      no check          :           58 ml

-4 line   big solenoid      ¼ check           :           56 ml

-4 line   big solenoid      1/8 check         :           40 ml

-4 line   small solenoid   ¼ check           :           33 ml

-4 line   small solenoid   1/8 check         :           24 ml

 

While back to back runs were very repeatable, we had some longer term repeatability problems, where it seemed like the longer the tank sat pressurized, the more volume seemed to flow.  There may be some effect going on with the pressurizing nitrogen dissolving in or out of the water.

 

Our nov16_00 notes had the small solenoid flowing 70 ml/s, which is roughly in line with what we saw today, because that was with a 1000 msec pulse, which should flow slightly more than twice a 500 msec pulse due to momentum.

 

Our nov21_00 notes had us testing the latching bottle valve at 200 ml / s flow at 100 psi, with a longer –4 hose, which is quite a bit more flow than we saw here.

 

We want 400 to 500 ml/s of flow to each engine for the manned craft, which is a lot more than we were getting today.  We could always increase the tank pressure (even the 100 psi tested today is probably optimistic for total pressure drop into the engine), but it would have a big impact on our 10-20 gallon pressure vessel selection for the main tank on the manned craft.

 

It is clear that the small check valve was a major restriction, but the ¼ check is not a significant restriction.  The hose size did not seem to be a restriction, so the likely bet is that even the big solenoid is the restricting point now.

 

The unscientific test of blowing through the valves a while ago seemed to show the big solenoid flowing more freely than the latching bottle valve, so we should retest that combination and see if it still shows the 60% greater flow.  The latching valve (NOS remote bottle valve) is a non-momentary valve, but it doesn’t switch fast enough for pulsing.  In fact, we should test it at several different durations to make sure that it didn’t get the good flow rating by just taking longer to close.

 

There is only one size larger to go in the NOS solenoid valves, and that is a slightly different configuration with a bottom discharge, instead of an in-line in and out.  They only rate it for slightly more max horsepower, so I don’t think it will cut it.  I am getting back with Bob Fortune about getting some custom pulsing solenoids made.

 

The other possibility is that the bottle manifold is the restriction.  The manual valves on the nitrous tanks don’t open up a whole lot, so there may be something to it.  We can test that by running the same tests with one of the tank manifolds that Russ made.  We can use the manifold from Spider, but it only has 1/8” ports on it, and we probably don’t want to drill on it.

 

Russ: when you get a chance, dig out that first manifold you made with the single central 1/8” port, and tap it out to ¼”.

 

It is possible that we won’t be able to get a solenoid that flows enough and can still pulse at 20hz, in which case we will have to investigate actuating rotary valves of some sort, or move to multiple solenoids per engine.  Multiple solenoids would be easy to implement with our current facilities, and they could be brought in incrementally to avoid no-flow during pulsing, but would be obviously pushing a technique past its sweet spot.