Note: Allpar does not take responsibility for the veracity of any information or opinions here, does not claim expertise, and is not responsible for any consequences. Please proceed at your own risk.

Cars by name
Trucks and Jeeps

Engines / Trans
Repairs / Fixes
Tests and Reviews

Driving the 840-hp Demon

Durango SRT at the track

Widebody Challenger drive

Troubleshooting Cars with Electronic Carburetors

This is a description of the procedure I've used to diagnose air/fuel mixture problems in computer controlled carbureted engines; the outlined procedure can also be used to set the idle air/fuel mixture without resorting to infrared exhaust gas analysis. The procedure is based on the General Motors "System Performance Test" which was developed for diagnosing problems in GM's carbureted engines. This procedure will not work with carburetors which use a stepping motor to control the a/f mixture.


The a/f mixture is controlled by a MIXTURE CONTROL SOLENOID (MC solenoid). This is a valve which operates at a fixed frequency (typically, 10 Hz) and whose duty cycle (valve's ON time divided by period) is varied. That is, the valve is pulse width modulated. When the valve is turned on, the incoming a/f mixture is fully leaned; when off, fully enrichened. The former is called a "lean command" whereas the latter is called a "rich command." By varying the duty cycle of the MC solenoid, the AVERAGE a/f mixture can be varied. In GM products, this valve directly varies the incoming fuel and air flow. In Fords, only the incoming air is directly varied. In Chryslers, only the incoming fuel flow is directly varied.

 The valve has a two wires electrical connector. On wire is connected to switched battery voltage whereas the other is connected to a power transistor in the computer and is a source of switched ground.

 During closed-loop operation the following will occur (assume the oxygen sensor is sensing a lean condition -- its voltage will be low):

  1. The computer gradually decreases the MC solenoid's duty cycle.
  2. The exhaust eventually becomes rich enough that the oxygen sensor's output will swing high (about 1 volt).
  3. The computer gradually increases the MC solenoid's duty cycle.
  4. The exhaust eventually becomes lean enough that the oxygen sensor's output will swing low (about 0 volt).

The cycle now repeats. A device for monitoring the solenoid's duty cycle (such as a dwell meter) will show a constantly varying duty cycle. The frequency of the oscillations will depend on the how fast the computer varies the duty cycle and the engine's RPM. An AVERAGE duty cycle of 50% corresponds to, on the average, NO average a/f correction. Stated differently, everything is operating correctly. An average duty cycle of LESS THAN 50% corresponds to, on the average, a rich command (the computer is compensating for a lean condition). An average duty cycle GREATER THAN 50% corresponds to, on the average, a lean command.


Monitoring the MC solenoid's average duty requires (for most people) the use of high impedance dwell meter. A low impedance dwell meter may be used unless it affects engine operation; my recommendation is to not use a low impedance dwell meter (that is, stay away from self-powered dwell meters). Following the GM procedure, set the dwell meter to the six cylinders scale REGARDLESS of the number of cylinders in the engine. At this setting, 30 degrees will correspond to a 50% duty cycle, 60 to a 100% duty cycle, and 0 to a 0% duty cycle. Run the engine until closed loop operation is present; this will be indicated by a varying dwell (see footnote 1 for deviations from this procedure). Once the engine is hot, not the average dwell -- the reading should vary equally above 30 degrees and equally below 30 degrees. The following is a brief trouble listing:

  1. DWELL NOT VARYING: system is operating in open loop.
  2. DWELL STUCK AT 10 DEGREES OR LOWER: full rich command is present; the computer is compensating for WHAT APPEARS TO BE a massive fuel flow reduction (check for dirt in carburetor, air injection system stuck in upstream position, vacuum leaks, improper a/f mixture setting...).
  3. DWELL STUCK AT 50 DEGREES OR HIGHER: full lean command is present (check for float stuck low, valve seat damage, oxygen sensor's sense lead shorted to battery voltage, etc.)
  4. DWELL OSCILLATING, BUT AVERAGE READING IS BELOW 30 DEGREES: average rich command is present (check for vacuum leaks, dirt in carburetor's jets, improperly set a/f mixture...)
  5. DWELL OSCILLATING, BUT AVERAGE READING IS ABOVE 30 DEGREES: average lean command is present. Check for incorrectly set a/f mixture, float stuck low, valve seat damage, clogged air filter, etc...).

Based on the above descriptions, it should be fairly clear on how to set the idle a/f mixture: merely set the mixture so that the average dwell is 30 degrees. Now, suppose the system's dwell is not varying, but the sensors are working properly, the upper radiator hose is hot...

 Several cars with small engines have the oxygen sensor mounted fairly far away from the engines. Indeed, during idle conditions, the sensor may cool off to the point that it will not operate (I had this experience in a 1986 Mustang with 2.3 liters engine and EEC-IV system). My recommendation is that all electrical accessories be turned off (so as to provide a minimal load on the engine) and use the idle stop screw on the carburetor to gradually increase the idle rpm until the sensor begins oscillating. Ensuring a negligible load on the engine guarantees that the carburetor will be operating mostly on its idle circuit. Now, set the a/f mixture so that the average dwell is 30 degrees. On the Mustang, this was done at about 1500 rpm.

 Note that the a/f mixture setting procedure assumes that NO fuel delivery problems (vacuum leaks, clogged carburetor, etc.) are present.  


[1] In some engines (e.g., GM cars with the "min-T" system -- Chevette) the a/f mixture is varied REGARDLESS of whether the engine is in closed loop operation or not. Consider setting the a/f mixture or diagnosing at a slightly increased rpm.

Additional note

Ted Mittlestaedt wrote:

I have a 1984 Chevy Celebrity with a Rochester Varajet II computer controlled carburetor on it. These carbs have a problem where many of them (perhaps most) came from the factory with a mixture control solenoid in them which had a rubber (viton) tip in the needle. The tips can break off and become lodged inside the M/C solenoid. It is not easy to see the broken off tip as the needle is inside the M/C solenoid. You can shine a very strong light into the bottom of the valve and look sideways and down though the holes in the side of the solenoid and see the end of the needle, if it is flat, the tip has come off. If you are lucky (as I was) the tip will be retained inside the M/C solenoid and you will not have to take the carb off and tear it down, you can just replace the solenoid.

New solenoids from NAPA are about $160. You can find them online for about $100. Supposedly the replacements come with steel tips. You can try wrecking yards but the ones in the yards are either the rubber tipped ones and are worse off than yours, or someone has already bought them.

Thexton did make a needle tip repair kit for these solenoids for a number of years, part #568. I don't think they still make it. The procedure for tip repair is pretty hairy.

The textbook symptom of this problem is surging and backfiring at part throttle. Full throttle is fine as is idling. Sometimes as the tip bounces around inside the M/C solenoid the tip will stop obstructing the valve for a period of time and the surging and backfiring will go away. Mine started surging and backfiring and did this for 2 months then quit for a month then started doing it again. Unfortunately when it quit happened to correspond to when I replaced the ignition wires, which of course made me positive the problem wasn't the carburetor!

You do not want to rush into a rebuild on this carburetor if you can in any way avoid doing it. The carbs are simple enough to rebuild, no different than most other carburetors but the mechanical adjustments to the carb are a nightmare. Half of them are sealed because they were made at the factory and aren't supposed to be altered again. Steel plugs have to be drilled out to get to them and it is easy to go past the plug and wreck the adjustment screw head. There are also some plastic parts that will dissolve in carburetor cleaner, and need to be removed first.

You can buy rebuilt Varajets but as of 2004, street price for them is about $350-$400 from the parts places. They may of course also not have your exact carb number in stock. GM seemed to take great delight in modifying this carb at the drop of a hat. I even have 2 that have identical carb numbers but one has an extra vacuum port.

The Fix Page

Driven: Demon, Widebody Hellcat

New merger-talk record set?

Former-Chrysler: low inventories

More Mopar Car
and Truck News