This page is about how to control an evaporator canister - sometimes called as charcoal purge canister

The evaporator canister is ment as a temporary storage for gases from the fuel tank, instead of releasing them to atmosphere.

A simple valve controlled by GenBoard releases the gases to the intake given some conditions are met.

Maybe with WBO2 closed loop control the condition becomes very simple.

Since we haven't seen much research/feedback on this (anyone want to contribute links or experience?) ...

I plan to install a solenoid to suck out the evaporator canister from time to time, using the same condition as for the variable intake with MembersPage/MarcellGal/EngineSwap/AirFeed : 1500..3000 RPM, 80..101kPa MAP. Actually planning to use same ECM output (a FET), but different solenoid (the 2 solenoid connected in parallel electronically).

google words: ...

pages: ...

Fero (or someone else), can you collect info about valve control that I believe was collected before but misplaced?

Hooked up the scan tool to an 01 civic, here is what I get.

I can do more in depth testing if needed.


Thanx, very good info. A bit more complex than I thought. Apparently, it tries to avoid sucking the hell out of the canister (or adding extra air at low-load, whatever we call). Also, it changes slowly, so the EGO-correction can follow.

Also looking at this here: MembersPage/DavidBlades/FanTest

Carbon can operation

Based on the system in my TVR.



Carbon can is always connected to atmosphere.

Carbon can is always connected to fuel tank vent.

Inlet is connected to carbon can ONLY during cycle 2.


During cycle 1 the tank vent emissions pass into the carbon can where they are absorbed by the carbon. Non-hydrocarbon vapours are allowed to pass to atmosphere. (The tank is vented to atmosphere so the tank doesn't collapse as fuel is used.) There are no non-return valves in the system. The inlet port is closed.

During cycle 2, the inlet is connected to the carbon can. This allows the engine to consume the hydrocarbons stored in the tank. The tank is still connected to the carbon can at this stage, so it would be possible to collapse the tank due to excessive vacuum.

On turbo engines, carbon canister valve is only operational up to 100pka and not on boost. Therefore on full throttle the valve goes to 0% duty cycle and the valve should be openened on all occasions where the lamda is actually controlling the mixture,in which case it is a narrow band lamda,therefore on closed loop conditions.

Unfortunatelly VEMS does not have any misc outputs that can have more options for setting such a valve,which is why im actually using one of the misc settings to open the valve (on/off type unfortunatelly) on deceleration where there are no implications regarding the driveability of the car. I tried on cruising and small TPS values to see if it works,but unfortunatelly the car does not run right.

It would be a good idea if this was implemented in future vems firmwares to have: 1) engine temperature 2) duty cycle control of the valve, along with the rest of the misc output settings that are already there.

My car is an Audi RS2 which i have tried it on.

See also