Jan 30 meeting notes:

Location: Long Range Systems

In attendance:

John Carmack
Phil Eaton
Russ Blink
Darin Smith

Next meeting: probably Tuesday at LRS, but I am out of town on Monday, so I might not be back in time.  Is Wednesday or Thursday a better fallback?

We hopped the VTVL tonight.  We had a lot of things wrong, but it was still pretty cool…

media.armadilloaerospace.com/misc/vtvlFly1.mpg

Russ was just finishing up drilling the last engine retaining plates when I got in.

After much complaining about cutting all the discs for new catalyst packs, we decided to roll the new ones.  This was one of our mistakes  we had three different generations of catalyst packs on the platform:  the rolled one from week before last, the discs from last week, and two new rolled ones.

I brought new screws and spacers so we could move the mounting plates far enough away from the solenoids to clear the swivel 90 degree fittings.  We made a second mistake by leaving one of the solenoids with a fixed 90, while all the others had swivel 90s.  The fixed 90 seems to flow noticeably better.  I need to get more washers so we can offset all of them for the swivel 90s.  The only way to get the fixed 90 clocked in the right direction is to add more and more Teflon tape until it is tight in the right direction, which is a pain.

We put some nitrogen into the tank and tightened up all the leaking areas, then we loaded it up with water to system test.  One of the swivel 90 fittings was leaking at the swivel point, but after replacing it with a new one (it’s nice to have lots of spares!), we sealed well.

We didn’t have the cinder blocks for chaining the platform down, so we used some 10 lb exercise weights, which definitely turned out to be a mistake.  While we didn’t have enough thrust to pick them up, it dragged them around enough that they didn’t provide good control over the platform.

The platform is controlled by a joystick with a throttle lever.  We don’t have an attitude sensor on the platform yet, so the control is completely manual, which we know from simulation to be hopeless for actually flying.  It is set up so that engines only fire when the trigger is held down, so letting go of the joystick immediately stops all engines.  If the joystick is centered, equal pulse widths will be sent to each motor.  Offsetting the joystick modifies the distribution.

I tested each engine by leaving the throttle at zero, but pushing the joystick all the way to each side, then pulling the trigger.  That gave a brief pulse to just one engine at a time.  It was clear from the pulses that the engines have plenty of power.

When I throttled all of them up, one of the engines basically didn’t fire, so it jumped sideways.  After we let it use up the rest of the peroxide at light throttle, we saw that one of the LED on the driver board was very dim.  It turns out that the behavior you get when you don’t have enough battery power is that some of the solenoids will just fail to operate, while others still do.  I was sort of expecting them to all stop functioning at the same time, so this was good to determine.

Looking back, we should have noted the odd behavior at the end of the water test run, where the flow seemed to stop and start as it was venting the tank.  I thought it might have been icing at the time, but it was almost certainly overdrawing the batteries.

Phil made up the battery packs we were using out of spare cells from Long Range Systems products, which are pretty small.  Russ still swears that they should be fine if we gave them enough charging time, but we are going to play it safe and move up to some big batteries from remote control cars.

We did another run connected to a car battery, but one of the engines was leaking badly at the top.  We managed to tip it over trying to manually compensate for the weak engine, but nothing was hurt.  On later inspection, several of them were rather loose.  We usually Teflon tape the upper closure on the engines for the test stand, but we skipped that today.  Another mistake.

Next week we are going to be patient and create four identical catalyst packs and run each of the engines on the test stand first.

We are also going to water flow test the entire platform plumbing to graduated cylinders out of each corner. To make sure all of the hoses and fittings are flowing the same.  We have one questionable check valve that looks damaged, although it seems to still work.

We have been just wire wrapping the solenoid leads onto the terminals of the driver board, which needs to stop.  We should move to either screw terminals or automotive connectors.  All of the solenoids have different length leads on them, which should be evened out.

Other stuff:

X-L Space Systems has decided to stop selling small quantities, so I had to place an order for an entire 500 pound drum of peroxide.  Ouch.  We don’t want to keep that much peroxide at LRS, so I am looking for a little plot of land to drop a storage shed on.  In the future, I want us to dilute the peroxide in the five gallon containers instead of before each run, because we have a tendency to slightly change the concentration each time.

We are probably going to run out of peroxide week after next, and the new order probably won’t be ready for a couple weeks after that, so we might be stalled for a little while.  We can use the time to integrate the embedded system onto the VTVL.  We definitely want to make a nice, waterproof enclosure, so we can just hose the entire thing down after each run.  We should think about that for the solenoid connectors.

Terry Parks is going to build the first attitude sensor board out of the MicroGyros, so hopefully that will be ready to go and integrated when our big supply of peroxide gets in.

I have been working with different configurations of the embedded system.  It looks like I am going to be going with a 96mb compact flash for the main storage, but I haven’t found the perfect system to install on it yet.  I was running the LEM embedded linux distribution, but it is missing too many things for my taste.  I can’t make RedHat remove enough packages to install in 96mb, so I am going to have to build up from something smaller instead of tearing down from a full distribution.

I haven’t installed it yet, but RTLinux looks like it is going to nicely do what we need for real time flight control.


Stuff to do / get:

Make catalyst cutter for drill press so cutting discs isn’t such a pain.
Modify engines so we can support shorter catalyst packs.
Finish the hose end protective plate for the manifold.
Tap the manifold for 4 so we can get rid of all the fill adapters.
We might want to exchange the nitrogen tank, even though we have a spare.
Cinder blocks.  We should probably get eight.
Convert another compressor to a vacuum pump.
Piano wire to replace the safety wire between the VTVL legs.
Ratchet set for LRS.
More spacer washers and some extra screws for the solenoid plates.
Extra check valves.
Get a large selection of viton O-rings
Level and plumb bob.
Two sets of two RC car batteries.

Russ/Phil: I might stop by on Thursday or Friday to drop off the AN taps and some other stuff.