Re: Armadillo May 2, 2004

>
>How much of it is really that vanes are easier and how much is
>differences in your own experiences/knowledge and time working out the
>wrinkles? Or to put it another way - if you had started out with the
>vanes and then switched to differential throttling would you have gone
>through much of the same process? In reading your updates it seems like
>you have worked out a lot of bugs that would affect both vehicles.
>
>The other more technical thing - I'm not recalling if you were still
>using the solenoids and rapid on/off for throttling the differential
>setup. The vanes are continuously variable (even if the controller
>isn't), did you try a continuously variable valve setup on the
>differential throttling vehicle?

The last two vehicles have been continuously variable differential 
steering.  The difficulty with that over solenoid operation is that 
solenoids give you instant response, where you can go from full on to full 
of in 10 msec at the actuator, while the ball valve driven big engines take 
800 msec to go full open to full close.  Even though you have the same 
amount of control authority with each engine supporting 1/4 of the vehicle, 
the responsiveness is much lower.

The decision is a no-brainer if you have more complex engines, like our 
mixed monoprops or a biprop.  For 90% peroxide it is less clear cut, but I 
still think there are compelling advantages.

Our very first vehicle that just used four solenoid driven 90% motors might 
have been easier to put together and get working than the jet vane vehicle, 
but every vehicle after that would have been better off with vanes.  The 
solenoid design can't scale up, so it would have been better to start with 
something that could.

The jet vanes don't have any cross axis coupling.  Our differentially 
throttled vehicles have four axis (pitch, roll, yaw, lift) coupled onto 
four engines.  The computer deals with it ok, but having to balance and 
correct for it reduces the responsiveness of the system.

The response to vane tilt is pretty much linear, unlike engine 
throttling.  Again, the computer dealt with it ok, but it is nice to have 
it go away.

The jet vanes provide instant torque on the vehicle, without needing to 
accelerate liquid or pressurize the engines.

Making four or more engines run exactly (or very close to) the same is 
often a bit of a challenge.  We have done it seven or eight times during 
our development, and we usually have a "problem engine" of some kind, and 
it sometimes changes after they have been run in.  We know a lot more about 
how to do it reliably today, but it still isn't without difficulty.

Vanes are much easier to manually check out -- tilt the vehicle and watch 
which way they turn.  Building a jet vane system is rather like building an 
RC plane.

Much less plumbing.  Our test vehicle still has some hoses because of the 
dual tank design, but with a single tank, the engine will just be a 
straight shot through the cutoff and throttle valve, which will probably 
get us at least 10% better engine performance.

Scaling is lots easier, you just build one bigger engine instead of four.

There are some manufacturing and handling benefits to having multiple 
smaller engines, but the variability that we currently get overwhelms it 
for us.

John Carmack