David Luddy wrote:

> Thank you...
>   to all who pointed out this possibility in my new Intermittent Miss
> Behavior "IMB".
>    However there is some conflicting info on what size capacitor, and which
> direction it should work in.
> its either 1.  25volt 10uF + side to pin 4 of AFM connector , other side to
> pin 2
>                                 -or-
>                2.  25volt 22uF  + side to pin 2 of AFM , other side to pin 4
> soooooo...
> Is there a majority vote on this?
>           tanks
>              <bugle>

Check your Bentley.  On the 91, pin 4 (brown and white wire) goes to ground,
negative.  Pin 2 (blue and red wire)  was at about 4.5 or 5 volts positive when
I measured it with a DVM.  Someone in the archives felt that 10 mfd was
insufficient and recommended 20 mfd, so that was what I used.
  More capacitance is more effective in preventing voltage fluctuations, but too
much could keep the voltage from changing quickly enough when you press the
accelerator.  A rough calculation of the RC time constant, seems to indicate
that we could probably use 100 mfd with no problem, although its physical size
might make it more difficult to fit inside the boot of the connector.  Radio
Shack sells 10 mfd tantalums off the rack; I hooked two of them in parallel (+
to +,  - to - ).  Tantalum capacitors are generally smaller physically than
other types and of higher quality in other ways. The voltage rating of the R.S.
ones seemed a little low to me until I saw that the voltage from the ECU was
apparently 5 V or lower.
  Since the intermittent power loss in my 91 only happened after an hour of
driving at speed in cold weather, I am not sure if the fix has been effective
for me.  Installed it a week ago.  It may be my imagination, but the idling
seems smoother and slower when cold now.
Good luck.
Larry Alofs   Chicago