Took the ISV apart this week to find out that it wasn't just my imagination that it wasn't operating properly. There was sign of wear from condensation on the commutator and the contact brushes. But my worries about the unit itself being bad, leading to the transistor/DME damage came true. More inside.

First thing I noticed upon opening up the ISV was the surface rust that collected on the unit. Both the contact brushes inside the cap and on the commutator were gunked up and corroded. After a quick cleaning, they looked like this!

But even the cleaning didn't show proper connectivity across the circuits. There should be closed circuits, with resistance, across the three terminals of the commutator (and also the terminals on the port as a result). Intermittantly I could get it to do so, but once the cap was put back on, the circuits would open up and I'd get a one-way valve, instead of a two-way operation.

What is hard to see on the pictures is that the little black plastic cap at the top of the motor is burned and distorted. We believe this acts as a guide for the brushes to contact the commutator when the cap is reinstalled. With it all bent, the brushes get stuck and never quite sit correctly onto the contacts. Also, one of the commutator blocks looked to had been broken and wanted to fall out. Water damage? High temperatures under the intake? Who knows. It hadn't worked since the first winter I'd been driving the car. That said, it was broken for one reason or another.

We explored the DME as well. The Darlington Transistors (ON588) which drive the current to the ISV were likely shot if the ISV took a dump. The design of the circuits are such that if the ISV sees a short, the end result is smoke. From this BMW article which uses a similar ISV setup:

If the exterior body or the axle inside rotates a little bit, brushes will short circuit the commutator segments. Now instead of 20ohm load, current will flow directly from battery 12V to transistors. And since there is no current limiting for the transistors, smoke is the end result.When ignition is now turned on, T560/T561 will start to conduct and 12V goes directly to ground instead of DC motor coils, destroying the transistors or the foils in the ECU.

What happens is the current has no fuse in-line and the transistors burn out. This was confirmed by one of the tests we did on one of the ON588 transistors that we pulled from the DME. It had a melted clip tab so it was easy to spot. One had a short between the collector and emitter ends and as a result, the ISV was continuously being driven in one direction - full air flow. This resulted in a leaner mixture than necessary to run the car at idle. As a result, in the cold, the engine would hover its RPMs at ~400, far less than optimal. A good functioning ISV will close the air valve, richening up the air/fuel mixture which keeps the car idling much higher, ~1000rpm.

We swapped out the ON588s for some Radio Shack Darlington transistors of similar design and am awaiting a new ISV to arrive via eBay. I snagged one for $9.99 on an auction which was a pretty good steal, as new ones are $160 or so, even if they're labelled Volvo or BMW. No Porsche Markup on this one.