You Are Here: Home » Tech » By Vehicle » SU Carb (very) Fine Tuning

SU Carb (very) Fine Tuning

I spent most of my working life as owner and operator of a dyno tune-up and diagnostic shop in New Jersey.  This started as a service for fellow SCCA racers and sports car owners, but after about 10 years I realized that if I ever wanted to retire I’d have to switch to American cars. Happily, I discovered in my retirement years that there was an active British car club here in retirement heaven (the Ozarks). So now it’s back to fun cars.

Let’s consider a real-life problem and potential solutions for one of these fun cars.

The problem

This season, our modified 1275cc A-H Sprite engine in the Austin A40 would very occasionally backfire through the carburetor at highway speeds under light load. It’s a very slight problem, considering the potential for troubles … —the engine has a non-standard aluminum cylinder head, a mild ‘hot rod’ camshaft, and the carburetor, intake and exhaust systems are all upgraded, non-standard items. I know from experience that the symptoms indicate an excessively lean fuel mixture (that means not enough gasoline in the air stream). The question is: how do we prove and correct that condition?

Me and my wife Shirley in the 1959 Austin A40 Farina. Mechanically it was originally similar to a bugeye. This is the first time BMC hired Farina (became Pininfarina later) to style their sedans. After that they all got the Farina touch, for better or worse. Mostly worse. Car was introduced in 1958 as their ‘world car’ to answer the VW Beetle – well, you know how that went! Mine was a rust-free Canadian car with an amazing history that I imported in 2003.

Background

SU carburetors are marvelously simple devices that work wonderfully well, despite everything you have heard or even have experienced on old, worn-out British cars that are decades away from the condition they were in when they left the factory. (Exception: discounting the horrible early era of the US federally mandated emission control devices. You can’t blame the English for that! Our cars were just as bad.) But all that original goodness was created by many hours of testing in engine research labs and field testing by skilled technicians. There are few moving parts in SUs, but throttle size, piston weight, piston spring pressure, needle diameter at 1/8 inch increments, float level, dashpot check valve design (accelerator pump substitute), dashpot oil viscosity, etc., are all critical to faultless performance. When things are worn and the engine is badly tuned (you know – “The carbs are messed up”, spoken without having first done a compression test, replacing the burned points or discovering the vacuum leak at the intake manifold) the carbs are too often blamed first.

When we go mucking about – All of that careful factory calibration goes out the window when we start modifying (sometimes improving!) our engines for more performance. A lot of experimentation is needed to optimize a race engine and even more is needed for a modified street engine, because the latter must perform well under a wider variety of operating conditions – not just mainly full throttle, as on a race track.

How does a SU carb control the fuel mixture to meet the current need? Simplified: The carburetor piston carries a tapered needle, hanging into a jet. The fuel in the jet is at a level controlled by the float. The piston is controlled by vacuum and, carrying the needle, rises and lowers in response to the engine load and, to a lesser extent, the speed of the engine. Fat needle position in the jet = less open area in the jet to flow fuel. Skinny needle position = more fuel flow to compensate for greater air flow.

How to duplicate a problem under scientific conditions – It really is quite simple with about $40,000 or $50,000 dollars worth of chassis dynamometer, oscilloscope and exhaust gas analyzers. Let’s say that our Austin occasionally spits back at 3,200 rpm in fifth gear on a level road, when applying a hair more throttle. It means that the portion of the carburetor needle in the jet at that moment is too fat – causing a borderline leanness. But the car runs smoothly and well at all other speeds and loads. How can we precisely duplicate that speed/load combination to determine the exact part of the needle then in the jet? Easy; we just connect a vacuum hose to the intake manifold, route the hose back into the car where a vacuum gauge can be seen and drive the car on a flat road at 3,200 rpm in fifth gear. The intake manifold vacuum, normally expressed in inches of mercury, is our precise indicator of load under those conditions. When the car is then driven on a chassis dynamometer capable of variable speeds and loads, we repeat that speed/gear combination and then use the dyno to load the engine to the previously noted vacuum gauge reading. The exhaust gas analyzer will confirm our lean mixture suspicions. We don’t need a different carb needle; we need a slight modification to this one. With the air filter off we can measure the exact piston height at that load. It is then easy to determine the exact needle position relative to the jet.

“Are you crazy!?” OK, OK. I know – the equipment needed is not readily available everywhere and a shop with that equipment and skilled technicians will (deservedly) charge a bundle for their time. Nevertheless, that is the best way to go and in the long run it may be the cheapest. After all, we are only talking about a half-hour of dyno time, or their minimum testing fee. You can perform the cure yourself.

Summary so far

At a now-determined distance the tapered needle it is too fat, causing an excessively lean fuel condition. Let’s say that the particular spot on the needle is four eighths of an inch down from the mounting shoulder (larger diameter). We are going to think in terms of eighths of an inch.

Doing it

Remove the needle. Just as in a SU needle catalog, we first need to have accurate measurements of the needle (needles, if a multi-carb engine) before starting any mods. Using a ruler and a fine-point pen, start from the mounting shoulder and make fine lines across the needle at precise 1/8th inch increments. With a micrometer, measure exactly on those marks and record the diameters on a sketch of the needle. Chuck the needle up to allow high speed spinning and carefully remove 1-½ to 2 thousands of an inch diameter four eighths of an inch down, using very fine wet & dry sandpaper. 400-grit works well. Remove material slowly and mic the diameter often. The critical area should be blended into the needle taper 1/8 inch above and below our measured length. Reassemble and try it again!

Is there a cheaper way to do this? Without a dyno, there are a few on-the-road methods that might work well enough to serve. With the air filter off, a homemade scale (ruler) with bold 1/8th inch marks could be affixed near the carburetor mouth. Your video cam recorder – or a rented camera, capable of short focal lengths and fitted with a long release cable – could then be solidly mounted in the engine compartment facing the carburetor mouth. Drive the car to the problem speed/load combination and then record. Hopefully, the camera would reveal the piston height with sufficient accuracy. Alternately, having removed the hood (bonnet), perhaps a large mirror could be solidly mounted and angled to allow a riding mechanic to see the piston height. You watch the road, please! Or, with the dashpot plunger and oil removed, a calibrated, balsa stick could rest in the piston and readings taken above the top of the carb. There is another option, but I can’t recommend having your small child ride in the engine compartment with a pair of calipers.

Conclusions

1. A vacuum gauge is the best way to reproduce an exact load condition for a given engine speed and gear. 2. A chassis dynamometer, of the proper type, is the best way to solve tuning problems or misfiring. 3. A determined individual can cure fine-tuning glitches normally resolved at the factory.

Footnote: Many/most popular dynos used these days have a new twist: they use a quick full throttle acceleration run to determine horsepower. The time needed to accelerate a given mass (weight) by the car’s drive wheels (working the dyno’s roller/s) is digested by computers and a neat horsepower curve is printed out. That’s great if you are using your car for racing, but make sure the dyno shop you select offers ‘steady-state’ and variable load testing. Water brake or eddy current electric dynos are needed.

And remember! There is only one sequence to successful tuning and diagnosis: Mechanical, ignition, and ONLY THEN, carburetion. Don’t be that guy who spends a fortune on wrong parts and speed equipment, and then finds he has low compression.

By Wil Wing
Proud member of British Iron Touring Club of Northwest Arkansas

Comments (4)

  • Mike Simon

    Sound advice and an excellent example of how state-of-the-art equipment can be used to precisely diial in the diagnosis of a common problem. For the rest of us, we compromise. Wil’s fundamental advice: “mechanical, ignition ONLY THEN carburetion” is something I’ve learned the hard way! Nothing like rebuilding your carbs only to find the points were not set correctly (I now run a Pertronix electronic ignition)…

    My idle is too fast now, but before I get out the screwdrivers and carb adjustment tools I’m going to hunt down the likely cause of the fast idle: vacuum leak.
    Modern gas lacks the lead our motors once depended upon for smooth running—and modern gas has alcohol in it which attacks the rubber in the fuel lines and in the carbs. Our cars were not designed for ethonol in the fuel and I find the major source of my carb problems today is rubber that is “perished” from exposure to alcohol in the gas.

    Reply
  • Thomas Abbe

    Very fine article, thanks for sharing. Not having a dynomomter, I have resorted to postulating the critical part of the needle (that part needing modification) can be arrived at by dividing the needle generally into basic areas: idle, off idle, transition, cruising and flat out. These I generally correlate to the first 1/8″, 2nd and 3rd 1/8″, 1/4 of throttle size, 1/2 of throttle size, and 7/8 to full throttle size (all measured from the shoulder of the needle). Whether correcting lean or over rich conditions, tabulate the corrections and look through the needle charts until the best match is found. One other common problem is inability to obtain adequate idle richness without spoiling running at every other speed. Look in available charts for needles with small (0.088 or 0.098″) first station but with similar profile otherwise. Best of Luck.

    Reply
  • Harrison

    I have a 68 b20 in a 59 Volvo pick up, and the carbs are spitting and rocking, but i checked for vaccum leak and got nothing. What should I check? Coukd I rebuild them with a kit and they work better? Any advice would be great!

    Reply
  • John E Wilson

    Spent the winter converting my 71 Mk4 to dual SU Hs4″s. After alot of head scratching and fiddling around, I finally have them balanced, running very smoothly and air leak free. The issue that I’m having envolves the idle speed. It is over 1100rpms with the engine timed at 8 degrees btdc, 10 flats out on the mixture adjusting nuts, the throttle cross rods not touching the throttle shafts, and the idle and choke fast idle screws turned completely out. Could my distributor be out of adjustment via the gear drive by one or two teeth? The carbs really make the car come alive, just need some final adjustment advice. By the way I used to work at Morgan’s Auto Parts around the corner from your Goleta operations, had the priveledge of meeting Art Moss for the 5yrs that I worked there. No hurry on the reply, have lots of other things to do with the Mk4, Thanks

    Reply

Leave a Comment

Scroll to top