Older Classics, Modern Technology

In exchange for loaning a car to a friend, I took advantage of his extensive knowledge of carburetors and fuel injection systems by making him help diagnose the SU on my MG Midget. The Midget, aka the Flaming Cockroach, has been annoying fellow motorists for several years now with its roarty exhaust and plenty of supercharger boost.

Since it lacks an accelerator pump, the SU struggles to provide enough fuel at tip-in. This is easily countered by rolling into the throttle with a bit of moderation, but I’m not here to gently “roll on” the throttle. The Cockroach knows only instant on, like a Labrador presented with unguarded food.

I am inclined to approach problems in a haphazard fashion, which generally results in little forward progress. After much cajoling from my car-borrowing friend, and with the help of the good folks at Innovate Motorsports, we’ve turned a corner and are now taking a more scientific tack. We hooked up Innovate’s LC-1 wide-band O2 sensor controller and air/fuel ratio gauge (Moss part No. 982-100).

The wide-band O2 sensor gives instant visual feedback of the air/fuel ratio. When used in conjunction with an accurate AFR gauge, one can see not only when things happen, but also to what degree. In other words, before using the sensor, I knew I was lean on tip-in, but not exactly how lean, and for exactly how long after stomping the loud pedal.

Installing the O2 sensor requires welding a bung into the exhaust system, but the results are well worth it. Once the Innovate was installed, we took the car out for some tests. First up was a simple drive around to see how the AFR looked. Then it was time to see if we could “move the needle,” meaning we wanted to see how much change we could cause with any given adjustment.

We tested with no dashpot oil, and then with 90 weight. We removed the piston return spring, then installed one which had been stretched to a greater length. We removed the air bleed bushings from the bottom of the piston, and tried various combinations of the above. The theory here is to not creep up on a solution; by making massive changes, we can develop a sense of the effect and scale of changes in a given piece of the system.

The interesting and unexpected conclusion of this experimenting was the realization that there are only two ways to significantly affect how the SU runs: by either changing needle profiles, or by resetting the idle mixture adjustment with a given needle. Mucking around with the dashpot piston can destabilize air velocity across the bridge, but it won’t dramatically change the inherent tune.

There is one exception to the above statement. As we here at Moss, and as some of our customers have seen, if the air bleed holes in the dashpot piston are too small, the piston may rise too slowly or not at all. This effect could well be diagnosed as a lean tip-in condition, when in fact the problem is that no air is being allowed past the piston. The Innovate gives us a clear indication of whether any bogging is caused by lean mixture or a too-slow piston rise.

A day spent learning has given us a clear indication of where we need to go. Throughout the last few weeks I’ve been testing and refining my own version of SU carburetor “secret sauce.” No, you can’t see it yet. First, I want to know if the project will end in success. That way I’ll have time to craft my next article. It will be about the slick new product I’ve designed, or a funny story about Rube Goldman’s latest abject failure.

By Robert Goldman