380 Idle Control Valve Test
 

My 380SL has been plagued by a lumpy or "hunting" idle since I bought it. The idle will fluctuate between about 500 and 650 RPM, and is worse in Park or Neutral than one of the drive gears. The problem occurs as soon as coolant temp hits 40 at which point the idle speed is supposed to drop from 750 to 500. It does not seem to be a miss, since there are no other running problems.

The fuel distributor, injectors and seals, all air hoses, and all intake gaskets have been replaced. The coolant sensor is working correctly, and all the wiring to the relay is OK.

So it seemed to come down to the idle valve or the idle relay. But how to tell? The valve decreases air flow and hence idle speed as the relay increases voltage to it. The relay was putting out a voltage that fluctuated between 3.8 and 4.1 volts, which is within spec. Was the relay causing the "hunting" or reacting to a problem (like a bad valve) and trying to correct the idle speed? These are big ticket, non-returnable parts, so it was important to find out.

What I ended up doing was buying a low-end model train transformer capable of putting out between 0 and 19 volts. I put my DMM on it and "calibrated" the 0-100 scale to voltage.

So I warmed the engine up. I disconnected the leads to the idle control valve and hooked the transformer up, setting it at 5V. Got a stable but high idle. Cranked up the voltage, and it took approximately 9V to the valve to get the idle down to the 500 RPM range.

So my conclusion is that the valve is bad, requiring more voltage than the relay can put out to hold the idle at spec. Why this would cause the relay voltage to fluctuate is unclear. Maybe the relay would try to lower the RPM, give up, try again, give up, and so on.

As a side note, at 9V, the valve started to buzz and I let it do that for mayber 10 seconds. Since then the idle has been noticeably better. So maybe a little higher voltage is good for these older valves.

Comments welcome.

Whatever you do that reduces idle speed does it by reducing net airflow. The idle air valve is much less likely to be bad than the possibility of an air leak elsewhere defeating its efforts to control. Also, any air leak will bypass the airflow meter deflection plate, which means it will improperly meter fuel injection flow at the fuel distributor - creating a lean mixture.

Look for hardened and cracked, loose-fitting idle air circuit hoses and vacuum hoses.

Steve

Maybe, but I have followed the vacuum road as far as it goes. As I stated, the intake was completely re-sealed, all the air plumbing is new, injector seals are new, and vacuum fittings checked. No change in idle as carb cleaner is sprayed around common leak points.

And air valves do seem to be a popular replacement item, according to the parts guys.

Another possibility would be a bad spot on the airflow meter pot. An erratic signal from this device due to faulty contact would fool the ECU into interpreting a large change in demand - causing enrichment. The idle speed would then increase as the ECU attempts to lean the mixture back out via the idle control valve.

The easy check for this is to see if symptoms disappear by pulling the plug on the pot.

Steve

This is an early CIS (1981). No pot to check. It is pretty much mechanical except for the O2 sensor, lambda, and idle control.

Just to close this out.

I trawled e-Bay for an idle valve. Last week one came up and I was able to get a VDO OEM unit for less than $20. I installed it last night and it fixed the lumpy idle.