Design notes

[Thor Johnson]

After our limited experience with Methanol/M85, I thought I'd share some 
of our notes:

  1. Meth+Blue Goop = Trouble
      The "Ultra Blue" just does not hold up under the influence of 
meth.  It doesn't dissolve, but it seems to lose all adhesion, and leave 
a sticky substance behind.  Anybody know a substitute?

  2. .91V Oxy is a little lean
	We still have a few bugs to work out (anyone know how to get air 
out of a VW rabbit FI pump...I can't!), but it seems that we have been 
running our engine a little lean, even though the oxy readings are at .9V 
or so (if I'm reading the darned thing right...).

  3. Meth is not nice!
	Engine/corrosion aside (though we have destroyed 2 O-rings), Meth 
produces formeldehyde when it burns.  I originally thought it was an 
insignificant byproduct, but it is much more bothersome than that!  We 
are running our exhause through a 12 foot, 4" dia tube to keep exhause 
away from the garage.  Wish I could coil up the tube and mount it.  
Sounds very cool.

  4. Stay away from full/complete maps w/o interpolation.
	Software issue.  We werent able to get the engine running until I 
did this; for a strict EE (no shade tree experience), full maps were too 
complicated to tune.  I am getting much better results using 2 MAP points 
(atmospheric & 30"vacuum) + a cylinder/RPM correction factor.

  5. NEVER trust loc-tite on high vibration components.
	A sequence of events behind this one.  We have broken the main 
drive sprocket twice, the first time by breaking the welds (poor welds), 
but the second time was by shearing 3 of the drive bolts off after the 
fourth had vibrated loose.  We used loc-tite 274 (blue 
anti-vibration)+lockwashers, but that wasn't enough.  We have now 
lockwired the entire drivetrain.  Hope this holds!

  6. An audi 4WD box makes a great differential.  Nuff said, we can skid 
all four tires with a single rear brake.

  7. Put the ECU in a waterproof location.
	Just one of theose things- a severe thunderstorm sprang from 
nowhere and soaked the ECU board before we could get the garage closed 
up.

  8. You can't use just plain MAP.
	I guess we might not be sampling/doing something fast enough, but 
if I run plain MAP, I get the equivalent of a carb w/o an accelerations 
circuit- you somp it, it stutters.  Added delta-throttle, and it works 
much better now.

  9. Use a COP/failsafe.
	The 68hc11 has a COP timer that mus be reset every so many 
seconds, otherwise it will reset and run a routine that you specify.  One 
of my interrupt handlers got out of control and I almost didn't notice 
that Injector #4 was stuck on while the computer appeared to be idle.  If 
I hadn't noticed it in time, we would have hydraulicked the engine (not a 
great way to complete the race...)  I now have the COP run a routine that 
shuts off all injecotrs/ other outputs, and prints a diagnostic message 
to the screen and serial port.  BEWARE OF "small" SOFTWARE ERRORS.  This 
was caused by an uninitialized varaible- it would work just after I 
programmed the system, but would not work after the first run was stopped.
Be safe.  Use a COP.  Otherwise you could lose an engine!

I'm sure I'll have more once we get through.  I'll add a few pages to my 
formula SAE project notes on the web once the race-push is over. 

Once I get the working elegantly and my code cleaned up (after the 
race!), I'll post the entire details of the schematics and code to run 
this.  This is a four cylinder EFI unit that is running Meth so we can 
use Toyota fuel injectors, so YMMV.


                Thor Johnson
          johnsont at falcon.cs.mercer.edu
     http://falcon.cs.mercer.edu/~johnsont
                                                    
         Have you seen the WarpMap lately?                    
    http://falcon.cs.mercer.edu/~johnsont/warpmap