ignition control....

[Steve=Ravet%Prj=Eng%PCPD=Hou at bangate.compaq.com]

lusky at knuth.mtsu.edu (Jonathan R. Lusky) Wrote:
| 
| 
| Steve=Ravet%Prj=Eng%PCPD=Hou at bangate.compaq.com writes:
| > ObEFI:  What are possible algorithms for electronically 
| controlled ignition?  
| > Ithink a closed loop system using feedback to continuously 
| modify timing is a 
| > better solution than static lookup tables, but what do you use 
| for feedback?  
| > Exhaust temp/composition?
| 
| Exhaust temp and composition vary widely with a bunch of factors besides
| spark timing, so I don't think either are appropriate.  I think the only
| effective way to do closed loop spark is with a pressure sensor, a crank
| angle sensor, and calculating the location (degrees ATDC) of peak
| cylinder pressure.  You'd still want to have a static map beneath 
| this
| I think.

Well, how about this:  it takes a certain amount of time for the fuel to
completely burn.  This should be relatively constant regardless of rpm.
You want most of the burn to happen at max compression.  So the timing
has to be advanced with increasing rpm to give the burn a "head start" with
respect to compression, which happens faster at higher rpms.  This would
imply that once you know how fast the flame front travels, you could
calculate the advance for any rpm.  But this doesn't take into account
the type of fuel, etc.  so there must be something else.  Do higher octane
fuels burn faster/slower?

actually, re-reading your reply above is causing light to dawn.  As the fuel
burns, the pressure in the cylinder increases, right?  for max torque, you
want this pressure peak to occur at a certain angle atdc, right?  so if you
have a pressure sensor in the cylinder, you can measure pressure vs. crank
angle, and vary the timing to keep the peak where you want it.  is this
basically correct?  so are there pressure sensors that can survive in the
cylinder, and from where can they be had?

on another note, incoming mail is addressed to 'diy_efi at whatever'
when i use reply, it gets addressed to 'diy_efi-owner at whatever',
causing john much grief.  if anyone has a suggestion for why
this is happening, i'd like to fix it.  

--steve