1227747 Diacom output?

[thergen at svn.net]

How about using something that queries the ecm over the aldl and sets it's
output based on whether the block learn value is over or under a certain
number (maybe 128)?  The idea could be tested on a bench with a PC and
using a printer port output (couple passive components to limit voltage
range).  It's not a universal solution for every ecm, but I think it would
work for a number of them.  This approach does close the feedback loop.

Using the block learn from the ecm should be alot easier than trying to
drive the O2 sensor input based on injector pulse width, tps, map, iat,
rpm, maf, and/or ve :) 

I'd try it, but I don't have the time or resources at the moment. 

Tom


On Mon, 19 Apr 1999 jgwynne at mrcday.com wrote:

> 
> 
> | Ya, for me the brain dead solution would be to record a 'run' with diacom
> | (all the inputs to the 7747, including o2), and just feed them back to the
> | benched ecm.  With a little software, it seems that you could 'lean out' the
> | o2 sensor for the next run, 'richen it up' for another software simulated
> | run, etc.  Hence the simulator.  I figured Diacom is an easy way to record
> | this run.
> 
> Unless the O2 input back into the ecm is made a function of injector
> pulse width, the feedback loop will not be stable. I recall someone
> posted having found code in a gm ecm that a PID (proportion integral
> differential) controller was used to control mixture and keep the O2
> sensor cycling (because of hysteresis). It shouldn't be that hard to
> make something simulate the feedback.
> 
> john
>