It was a pretty scary piece of improvisation, really. All the assorted
circuit boards and wiring that made up the prototype ECU were shovelled
into a cardboard box, which was cable-tied onto the GR’s luggage rack.
The fuel pump was wedged in between a couple of frame rails without being
tied down to anything and was wired directly to the battery, without any
kind of switch whatsoever. We did, however, deliberately leave the wiring
hanging out in the breeze so that in the likely event of something going
wrong, the crash-test idiot could pull the plug. Gotta have a token nod
to safety, even on a rolling death-trap piloted by a lunatic.
Fuel supply on the bike was a bit more of an issue, because the GR’s
vacuum fuel tap was in a pretty bad way. Like the starter motor and the
speedo worm gear, it must have been designed by Suzuki’s work experience
student while the rest of the design team enjoyed an extended liquid lunch
at the nearest karaoke bar. Either that, or it was designed specifically
as a horrible practical joke on everyone silly enough to buy a Suzuki. Bet
they're still laughing about that one. Anyway, the bypass mechanism on the
fuel tap was dead, so it wouldn’t flow fuel without a vacuum supply. We didn’t
have a spare vacuum feed on the manifold, so we needed another way.
Fortunately, Richard had some spare plastic tubing lying around that
he usually uses for bleeding brakes. We slid one end over the vacuum connection,
opened the tap by sucking on the other end (Note: don’t try this at home,
kiddies. Brake fluid tastes very bad, stunts your growth and impedes brain
function. Perfect example: Neil) and made sure the fuel kept flowing by
pinching the pipe with a common bulldog clip. It took a couple of goes before
we managed to apply the clip to the hose without getting bits of Neil’s face
in there as well, but we got there eventually. Amazingly enough, the small
amount of suck we trapped in the tube was enough to keep the tap open. No
jokes about golf balls and garden hoses, please. Oh, all right then. If you
must. Actually, it was a cricket ball, but… moving on…
That only sorted out the intake side of the fuel supply problem, though.
What could we do about the fuel return from the regulator? The only option
was to remove the fuel cap and sticky-tape the return line to the top
of the tank. Not the neatest or safest solution, but we’re Team Feral!
We don’t care about neat or safe! Not in the prototype stage, anyway.
If we were injecting the GR as a guinea pig because it was expendable,
then it made sense to use an expendable crash-test idiot to ride it. Of
course, Steve was the most obvious choice. If something unfortunate happened,
nobody would miss him and there would be plenty of spare bikes to share
around. Actually, that was almost a good enough reason to get some sabotage
happening, but we decided against it because the awful lash-up we gave him
to ride was almost certain to blow up anyway.
We strapped Steve onto the bike and pushed him down the hill (carefully,
in case he exploded) but Project Shitbox only popped a few times and then
died. A backfire through the manifold had popped the pressure sensor off,
so the ECU cut the fuel. Bummer. We pushed the bike back up the hill and
then lay around wheezing for five minutes. Too many brewskis and doughnuts
were never regretted so much. No question, the GR was waaay too fat and
lardy (like us, really) to be pushing around too much. If we didn’t get
it right next time, we’d need a defibrillator or three coffins.
Fortunately, the second attempt was much more successfulererer (er).
No need for five thousand volts across the nipples (much to the disappointment
of Richard. He’s kinky like that), the world’s first fuel injected GR650
coughed into life on both cylinders about halfway down the hill.
Right away, Steve had his hands full. The bodgy manifold
didn’t have a throttle stop, so he had to be very careful not to let the
revs drop too much or the bike would stall. Not a huge problem normally,
but the throttle was incredibly touchy and the engine so responsive without
load that it only took a millisecond of inattention and it was pushstart
time again. A better throttle linkage would have made the bike much easier
to ride. A small example to illustrate exactly how touchy the GRi's throttle
was: After riding the bike for probably 15 minutes, Steve had pretty much
got the hang of blipping the extra-touchy throttle for downshifts. Once
the carbies were bolted back on, he blipped the throttle as he had been
earlier in the afternoon and the carburetted motor didn't even respond.
The throttle butterflies opened and closed again so quickly that the engine
didn't have time to pick up even a single RPM.
The fuel map proved to need a bit more development as well.
It had been very carefully “interpolated” (fudged) from some “initial” (wildly
inaccurate) calculations using “empirical” (guessed) inputs gleaned from
“self evident” (fabricated) data.
Translation for those unused to academic-speak: Richard pulled some
random numbers outta his bum, scrawled some bodgy calculations on the back
of an envelope whilst drunk and got most of it wrong. Then he got his envelopes
mixed up and typed in his shopping list instead. The whole fuel map was
basically sourced from where the sun shineth not. Even so, the bike ran
just fine. It wasn’t well tuned, but it was just barely rideable. That was
fine, though, ‘coz Steve can only just barely ride. On a good day.
Wearing carburettors, the GR engine is very torquey and flexible, especially
at low rpm. With injectors pumping in a very arbitrary amount of fuel,
all the torque had disappeared and the bike was completely gutless. However,
once it coughed past about 3,000rpm there was a sudden change. The engine
roared like an enraged bull with an electric cattle prod jammed firmly
where the fuel map came from and took off in much the same fashion. It
was like someone had hooked up another two cylinders, and it really made
the carburetted motor feel and perform like the boat anchor that it is.
Unfortunately, the grins didn’t last for very long. Above 4,000rpm, the
old girl coughed and spluttered like an asthmatic chain-smoker waking up
in a gas chamber and all the lovely power vanished as quickly as it had
Steve made a couple of circuits around the test track and returned
to report on how it felt. After some head scratching, we decided to add
more fuel over the entire map and see how it changed things. Three mouse
clicks on the crapslop and we were in business with more juice everywhere.
Push start number three was a success as well, and then Steve was slipping
the clutch back up the hill. The clutch abuse was necessary because the
bike was now even worse below 3,000rpm and sluggish for the next thousand
revs. The 1,000rpm-wide power band was unchanged, but it had moved north
to between 4,000rpm and 5,000rpm. So now we knew we were too rich down
low, too lean up top and about right in the middle. I guess you’d call
it tuning by the Goldilocks method. If so, we must be the three bears:
Bearly Competent, Bearly Thinking and Bearly Sober.
By this stage we were running out of weekend, and Steve still had to
ride back to Gumpville on the GR. We had successfully demonstrated that
we could inject a bike without blowing it up, fiddle with the mapping
and improve (or at least alter) the state of tune, and we knew what we
had to improve (everything!) to make progress. We ripped out the bodgy
additional wiring (as opposed to the bodgy standard GR wiring), bolted
the carbies back on, and started making plans to begin work on the CBR.