Well, there's quite a bit to report after last weekend. To begin, I was
a little low on coolant, but not as much as I'd feared. Bleeding the
system as directed here did help me, and after that step the car would
occasionally rise *to* the N but not get over it, typically at a stoplight
for a long rest; once underway again it cooled down quickly.
So on Saturday morning I motored over to the Dime Bank Garage (aka Chris
Kantarjiev's shop) and we set about tuning the car before the threatening
rainclouds decided to open up. (Since the top to my car is still in my
garage, and Chris' place is 10 miles from my house, this proved wise.)
First we hooked up Chris' new, shiny, unbelievably cool Fluke automotive
DMM to the M.G. -- an act that reminded me of "Forbidden Planet," where
the humans with their puny ape brains (the M.G.'s ancient electrical
system) try to comprehend the intellectual capacity of the Krel machines
(the Fluke). (So Chris, has your DMM started ripping people apart yet?)
Anyway, this device makes a dandy dwell meter. Chris got a reading of
49 degrees, which he thought was too much; we fiddled with it for fifteen
minutes or so, succeeding only in getting the dwell to about 52 degrees,
when it finally dawned on me that Charlie Rockwell had achieved the best
results on the dyno (same motor, different chassis) with the dwell set to
about 52-53 degrees. (Not enough coffee.)
We then tackled the timing, which was seriously retarded. After learning
that the later Lucas 45D4 distributors use 12mm bolts on the clamping
plate, we set the timing to the setting that the Abingdon factory Special
Tuning manual suggested for the factory's engine in roughly comparable tune
to mine -- about Stage Four (compression, cam, porting). That setting, if
you're making your own shopping list, is supposed to range between 6 and 8
degrees BTDC -- on no account more advanced than 8 degrees. We set it at
about 5, opting for a little more conservatism in respect to 92-octane gas.
The car already ran better; in fact, adjusting the timing raised the idle
speed and smoothed it out. So I took off the air cleaners and set about
fixing the SU carbs. Chris had a Colortune and a Unisyn. I now view
both of these devices as absolute necessities -- as critical to the ability
to tune an engine as a spark plug wrench or gasoline. More later.
As with my earlier efforts, I began by the book and synchronized
the airflow of the two carbs with the Unisyn, a simple process but one
that told an interesting tale. Putting the Unisyn on the carb throat
restricts the air flow, and normally causes the idle to drop. When I
held the Unisyn up to the carb, the idle speed roared to 3000 RPM due to
the increased richness. We had some work ahead of us. I got the carbs to
produce identical gauge readings and Chris installed the Colortune while
I reset the mixture nuts to zero, then lowered them 14 flats each. (A
"flat," for this operation, means one of the six flat faces on the side
of a hex nut, and represents 1/6 of a full turn. This is the usual LSB
of SU tuning operations.)
The Colortune, if you've never used or even heard of one before, is a
spark plug with a quartz-crystal window in it that lets you see into
the combustion chamber and actually see the living, breathing heart of
the engine at work (Kim says that's too poetic, but she hasn't gone for
a ride in this car yet). The Colortune manual says that yellow is too
rich, whitish-blue is too lean, and "bunsen blue" is spot-on.
The #1 plug was flickering dimly, with occasional flashes of white and
otherwise little fire at all. It was a classic case of terrible lean-burn
misfire, the kind that wrecks pistons if left unchecked. I twiddled the
adjusting nut for all I was worth. After what seemed like 12 *more* flats,
the light from within the chamber was blue, just barely tinged with orange.
(The A Series engine, my reading told me, runs best at air:fuel ratios of
between 12.5:1 and 13.5:1, with 14:1 being stoichiometric, and the B Series
engine is basically a bigger version of the A Series, so I wanted to
be a flat or two richer than dead even.)
Chris replaced the spark plug and installed the Colortune in the #3 plug
hole and I did the same thing: another mad session fiddling the nut while
the plug appeared to do nothing, then a gradual progression from white
flashes of misfire down toward blue happiness tinged with peach. Lovely
color, really.
I finished up by getting the idle setting where I wanted it, about 1100 RPM
(hey, it's a slightly lumpy cam), reinstalled the air cleaners, and
consulted with Chris for a few minutes on the best PCV installation.
"We need to make an instrumented test run," Chris said, and brought out
his Vericom computer. I like this almost as much as the Colortune. It
is dead simple to use, entirely internally driven, with sensors to read
acceleration inputs against time and calculate various components of the
car's performance based on the parameters it measures. It's said to be
very accurate. I hope so, because I'm very excited by what it said to
us.
Chris hooked the computer up by sticking it to the inside of the windscreen,
punched in a few figures, and gave me the driving instructions. We were
using a simple 10 to 40 mph run to eliminate problems with wheelspin or
bogging at the start, and to give us a chance to shift gears once (at about
35 mph in the B) without being too fast for the light-industrial street
near the Dime Bank. With the test parameters keyed in, I revved the
daylights out of the motor, dropped the clutch, and laid a nice streak
of Yokohama's finest A008RS rubber down Olive avenue. "Okay, shut it off
and let's read the data," Chris said. We pulled into the driveway and
dumped the readout.
10 to 40 mph took 3.8 seconds, but what was really impressive were the
g-loading and calculated bhp figures. Maximum linear (not lateral) g in
this car was .87, at 11.3 mph, which I reached in 1 foot (from 10 mph, I
presume, arrived at after wheelspin stopped). Then the Vericom displayed
the calculated bhp per 1000 lb: 57.
57 bhp? I thought. No, that's 57 times (weight / 1000).
"So how much does your car weigh?" Chris asked.
"About 2000 pounds," I said.
"Then your motor makes 114 bhp," he said. For comparison, the stock
MGB engine was rated at 92 to 95 bhp (depending on the source) up to
1967; power dropped from then to 68 bhp by 1975's single-carb, log
manifold, and catalyst equipped version.
This was very exciting news, for several reasons. For starters, in the
short dyno session I'd had previously, it had only measured 87 bhp at the
wheels, which the conversion factor said works out to 110 bhp at the
crank -- a good figure, a little conservative, but about what I was
shooting for with my cam, head, and compression values. (That's 1 bhp/
cubic inch, for those who remember the fuel-injected '57 Chevy V8.)
I explained this wheels-to-crankshaft relationship to Chris, who looked
skeptically at me.
"But the Vericom is supposed to measure power AT the wheels," he said.
This means several very interesting things, particularly after looking up
the actual curb weight of the MGB (and factoring in the presence of myself
and Chris during the instrumented testing).
First of all, it means that I really *have* been right all along in
saying that the factory M.G. exhaust is a highly effective unit. The
lengths of the tubes are all decent, the shape is good, and I have
proven to my satisfaction that it works well with even fairly highly
tuned motors. (It's worth pointing out that even the Stage 6 engine
in the factory Special Tuning manual uses the stock exhaust manifold
setup.) I'd suspected in fact that the smaller tubes and better
spacing (the LCB design, q.v.) of the primaries and secondaries with
respect to the Huffaker design I'd used in the Black Car would help
this engine breathe better. That appears to be the case. In short,
headers are not necessary for high performance MGB motors. (And yes,
I'd still like to try some quality LCB designs on the same engine,
but I'm quite satisfied with the performance of the stock system.)
More thrilling for the numbers geeks among us, however, is the estimate
of 2400 pounds as the actual test weight. (The car might weigh even
more than that; it's 2100 and some pounds dry, and we tested it with two
adults and most of a full tank of gas. And I'm, well, more than 10% of
that total...) If you don't feel like pulling up the calculator, that
puts the Vericom's calculation for total overall horsepower at 136 bhp.
One hundred and thirty-six ponies.
Not bad for a motor built by two technical writers on their lunch hour
using a hair dryer and a block of wood. For the record, the factory's
own Le Mans cars were said to produce 130 bhp, at what the Special Tuning
guide said was Stage 6 and included a Weber.
Anyway, effusive thanks go to Andy Banta for originally helping me put
this motor together (and for the "two technical writers" comment when
Jeff Zurschmeide and I were puzzling over which way to install the
offset connecting rods), and to other Britcars list members for helping
me install it during most of 1993: Daren Stone, Robert Keller, Tim
Pettenati (absent but not forgotten), and Chris Kantarjiev for physical
help, and Pat Vilbrandt, Roger Garnett, Mike Burdick, Randy Taylor, and
no doubt many others who provided assistance electronically.
--Scott Fisher
|