Been fiddling with tuning the Digifant on my Boston Engine rebuilt 2.1 (650
miles on it and the ENGINE, at least, runs GREAT) and have run into an odd
behavior. I'm using an Innovate LM-1 wideband oxygen sensor with an output
to the ECU set to emulate a narrowband. At all RPMs, with ECU running in
either open loop or with the lambda sensor wire connected, the LM-1 lambda
reading will periodically "spike". It spikes so high, in fact, that the LM-1
has to switch briefly to %O2 mode instead of AFR. It goes all the way up to
some 20% O2, which is pretty much an atmospheric O2 reading. It does the
same thing with the stock narrowband sensor in place, as evidenced by the
0-1V LCD voltmeter on my dash-- the meter "needle" slams to 0, then the ECU
briefly overcompensates rich (.75V) then dances around the middle (.4-.6V)
for maybe a couple seconds before doing it again. This sounds like a failure
to inject fuel to me; an electrical failure rather than a clog, I'm
thinking, since any fuel in the air charge would bring the O2% down from
atmospheric, right? I guess the next step is to clean all the injector
contacts, "jar test" the injectors for function, and maybe rig the LM-1 to
monitor the injector pulse.

Anyone dealt with anything similar?

--
John Bange
'90 Vanagon
"We'd tell a monkey how to peel a banana, if he said he was peeling one in a
Vanagon."