belize1334
Well-known member
This is a copy of a topic that I stared on the DSMap board in their section for development ideas. I wanted to throw it out here also and see what people think.
My VR4 has a bizarre behavior that not many people mention but which is, I believe, fundamental to the design of the DSM throttle body. Under normal driving conditions or when at WOT the ECU opens the ISC all the way. This is done so that if you abruptly close the throttle the car won't buck as hard. The problem is that if you go from WOT to closed throttle for a hard shift, the ISC hangs open for about a second which lets a fair amount of fairly dense air past the throttle plate and causes the rpms to spike. At first I thought it was driver error but I've look at my logs and the spike occurs AFTER the closed throttle switch is triggered.
As an example, if I let off hard and engage the clutch at 3,000rpm and 10psi then my rpms jump to 4,100rpm. The higher the boost at let-off the more the spike, the higher the rpm at let off the lower the spike (which makes sense since higher rpm would consume more air naturally so the amount let through the ISC is effectively less significant).
As an experiment, I blocked off my ISC (note that the FIAV was already blocked) and this almost completely eliminated the behavior. I say almost because I still get about 300rpm spike at 3k and 10psi but by 5k it's basically gone. I attribute the fact that it remains to a certain degree to the BISS which still provides a pathway around the throttle plate under all conditions.
I've also done some looking around and this behavior seems to exist in all cars with forced induction that have a BISS and/or ISC which are inline with the compressor. Two examples are Evos and WRXs, both of which have an idle control which draws air from inside the charge path and around the throttle plate. I've looked on their respective forums and both cars have this same reported symptom. As soon as you start looking at cars which are operating in SD and draw their idle air from a separate source (which is never above ambient pressure) then the effect is gone.
I've considered two possible solutions.
1) Modify the throttle-body so that idle air is drawn from either ambient air in the engine bay (works if your in SD) or from a separate path that bypasses the compressor (metered air provided by, say, a line tapped from your intake pipe through a check-valve to a tapped fitting on the ISC housing). The problem here is that it seems hard to accomplish physically, especially for the BISS which is part of the TB and harder to cut/tap/weld than the ISC housing.
2) This is teh haxors come in ... Modify the ECU code to change the ISC behavior. I'm not sure what the best way to affect this would be or whether it's even possible to mess with the code but I'm imagining a scheme like ... an upper limit on ISC steps that falls off with rpm so that once you get to about 2500 the ISC is always closed ... or ... an overwrite that forces the ISC to clamp shut any time the TPS signal goes abruptly to zero above a certain rpm range.
So...that's all I've got on this, except to say that for combustion you need both air and fuel so perhaps another solution would be to make the injectors shut down faster as I've noticed in my logs that my IPW hangs at like 1ms for about a second after the throttle plate closes before they go to zero for coast-down.
Thoughts?
My VR4 has a bizarre behavior that not many people mention but which is, I believe, fundamental to the design of the DSM throttle body. Under normal driving conditions or when at WOT the ECU opens the ISC all the way. This is done so that if you abruptly close the throttle the car won't buck as hard. The problem is that if you go from WOT to closed throttle for a hard shift, the ISC hangs open for about a second which lets a fair amount of fairly dense air past the throttle plate and causes the rpms to spike. At first I thought it was driver error but I've look at my logs and the spike occurs AFTER the closed throttle switch is triggered.
As an example, if I let off hard and engage the clutch at 3,000rpm and 10psi then my rpms jump to 4,100rpm. The higher the boost at let-off the more the spike, the higher the rpm at let off the lower the spike (which makes sense since higher rpm would consume more air naturally so the amount let through the ISC is effectively less significant).
As an experiment, I blocked off my ISC (note that the FIAV was already blocked) and this almost completely eliminated the behavior. I say almost because I still get about 300rpm spike at 3k and 10psi but by 5k it's basically gone. I attribute the fact that it remains to a certain degree to the BISS which still provides a pathway around the throttle plate under all conditions.
I've also done some looking around and this behavior seems to exist in all cars with forced induction that have a BISS and/or ISC which are inline with the compressor. Two examples are Evos and WRXs, both of which have an idle control which draws air from inside the charge path and around the throttle plate. I've looked on their respective forums and both cars have this same reported symptom. As soon as you start looking at cars which are operating in SD and draw their idle air from a separate source (which is never above ambient pressure) then the effect is gone.
I've considered two possible solutions.
1) Modify the throttle-body so that idle air is drawn from either ambient air in the engine bay (works if your in SD) or from a separate path that bypasses the compressor (metered air provided by, say, a line tapped from your intake pipe through a check-valve to a tapped fitting on the ISC housing). The problem here is that it seems hard to accomplish physically, especially for the BISS which is part of the TB and harder to cut/tap/weld than the ISC housing.
2) This is teh haxors come in ... Modify the ECU code to change the ISC behavior. I'm not sure what the best way to affect this would be or whether it's even possible to mess with the code but I'm imagining a scheme like ... an upper limit on ISC steps that falls off with rpm so that once you get to about 2500 the ISC is always closed ... or ... an overwrite that forces the ISC to clamp shut any time the TPS signal goes abruptly to zero above a certain rpm range.
So...that's all I've got on this, except to say that for combustion you need both air and fuel so perhaps another solution would be to make the injectors shut down faster as I've noticed in my logs that my IPW hangs at like 1ms for about a second after the throttle plate closes before they go to zero for coast-down.
Thoughts?