Date:         Fri, 14 Apr 2000 13:13:13 EDT
Reply-To:     FrankGRUN@AOL.COM
Sender:       Vanagon Mailing List <vanagon@gerry.vanagon.com>
From:         Frank Grunthaner <FrankGRUN@AOL.COM>
Subject:      The Waveform is Different. Was: Extra Grounds on 1.9l
Comments: To: stuart@cobaltgroup.com
Content-Type: text/plain; charset="US-ASCII"

Interested Parties.

Surprising result. While I can't directly comment on the 1.9 L Waterboxer as
I run a 1.8L 8V GTi Digifant in line 4, I recently discovered something I had
not anticipated. I had always assumed that the key ground issue was contact
to met. As in not a contact through an insulating substance (oxides, crud,
crude oil, kerrogens, plastic, the ether, etc.). Wrong. For the electronic
fuel injection systems, the details of the resistive path are very important.
Stereo (big thumper) system installers are well versed in a series of
superstitious problems related to ground loops. I say superstitious because
they are not analytically resolved but rather empirically treated with
warnings and a recipe for the correct way. The aluminum block/head systems we
work with have many surface elastomers (lubricants, anti-seize, etc.) and
thin film insulators (oxides, see above) in the series path to the battery or
computer. Also, the engine components are actually cast of mechanical grade
alloys (not electrical grade conductors). The actual resistance across the
block from connector to connector (or to the ground strap) can range from a
few tenths of an ohm to 3 ohms or even more.

The surprising result is the major effect on the fuel injector trigger pulse
waveform. I haven't quantified this yet, but I have wired in taps to follow
the trigger (on/off) pulse waveform, the ignition coil pulse waveform (think
rpm and tach), the spark pulse and the O2 sensor response. Without a data
logger, I can't follow all four signals simultaneously, but I can do two at a
time. I'm setting up for real time road test the weekend after I do the taxes
to get true load waveforms (no dyno). And I data set is modest (related to
the fact that I can only rev the engine for extended periods if
she-who-must-be-obeyed is away from her post. Something about neighbors who
can't appreciate the sweet sound of a well-tuned engine).

Anyway, when I did my conversion (many years ago, diesel to gas). I placed
the ground wire from the Digifant system that is supposed to go to the head
to the body instead. Ground is ground I said! Well, triggered by recent posts
by Ken Lewis over my starter saga, I rigged up a switched system to go back
and forth (ignition off!) between the body ground and the head ground. I saw
on the oscilloscope (250 MHz bandwidth) an amazing change in the fuel
injector waveform, and in the O2 sensor waveform. The fuel injector pulses
were shorter and sharper (faster rise and fall times) and the same was
observed for the O2 signal. The integrated or mean voltage of the O2 signal
was also different by 0.072 volts indicating a leaner running condition.

It appears that even though the currents are modest the issue is at least an
RC circuit filter modifying the input to the computer. I also find that the
computer outputs are considerably cleaner (less high frequency noise) which
suggest a possible ground loop problem with which the Digifant computer is
trying to cope.

Summary, more ground wire is good. Heavy ground wire is better. My friendly
local electronic goodie store (FLEGS?) sells AWG 1, 2, and 4 insulated wire
with solderable connectors for minimal amounts of negotiable securities. I
now have the Digifant head ground wire in the Bentley mandated location. I
have run an AWG #2 wire (GM side terminal 42 inch ground strap from the
accessible (not clutch slave) starter mounting bolt to the battery, an AWG #2
wire from the AC mount (traditional I-4 ground point) to the battery, same
wire gauge from the coil/head mount, and from the transmission to the body.
May soon have to adjust the rear tire pressures to compensate for excess
copper.

I will try to post waveform images after I clean everything up if there is a
image-web-site-poster-hoster available.

Sorry for the length, but I think this has big MPG and Digifant/Digijet
gremlin implications for stock systems as well as engine swaps.

Hope this helps someone,

Frank Grunthaner