here's a mod that offers a small potential mileage improvement under engine warm-up conditions: spoofing the signal from the water temp sensor (WTS) / engine coolant temp (ECT) sensor.
The coolant sensor is often called the “master” sensor because the powertrain control module (PCM) uses the sensor’s input along with that from the oxygen sensor to go into the “closed loop” mode of controlling the fuel mixture.
The coolant sensor also controls a number of other systems, including:
- Start-up fuel enrichment on fuel-injected engines.
- Spark advance. Spark advance is often limited until the engine reaches normal operating temperature.
- automatic transmission torque converter clutch lockup
the theory behind this mod is that the manufacturer has played it conservative in programming the PCM's response to engine temperature during warm-up, i.e. erring on the side of drivability.
the goal of the mod is to alter the WTS signal to "hurry up" the warm-up process as far as the PCM is concerned, while still maintaining satisfactory engine performance.
the mod requires a basic understanding of electrical resistors (the WTS is a heat variable resistor after all - a thermistor): the PCM feeds the WTS a fixed reference voltage in one side, and reads the return voltage from the other wire.
most coolant sensors' resistance decreases as they warm up.
an elegant approach to this mod is described at teamswift.net: one metro owner attached a second WTS - externally - beside the original one. it's in direct contact with the housing where the factory one is installed, so they are heated together (though obviously not to the same extent).
he then wired them in parallel, effectively cutting the resistance in half - the PCM sees his engine warming up apparently twice as fast as it actually is.
he claims no drivability problems. (he actually says the amount of soot around his tailpipe has decreased since adding this mod.) he lives in a cold climate (germany) and his commute is short (15 km), so he says he has seen a positive mileage improvement - though I suspect the improvement would be quite small).
to prevent the ECM from reading a false overheating situation once the engine actually reaches normal operating temps, he wired an appropriate value resistor in series with the additional WTS. this retains enough resistance in the modified circuit once warm and the temp gauge reads "normal".
the elegance of the above approach is it's fully automatic, and relatively self-limiting.
the down-sides are that it's (a) non-adjustable on the fly, and (b) the "warm" engine temp isn't 100% accurate (though if you're just going by the temp gauge, he says it's in the usual position).
my idea was to do this mod using a variable resistor - potentiometer - in place of the second WTS/resistor combo.
that way i can fully dial in as much or as little "heat" as i want the PCM to see, and then switch it off once actually warmed up, to revert to the "un-modified" coolant temp signal.
but i'm no electronics expert. (i never had any electrical classes at school.) if you have a look at the thread on the teamswift site, you'll see me asking some pretty basic questions, as i learned about how this mod works. in particular, i needed to understand the difference between resistors wired in parallel vs. series. (one halves, the other doubles the sum of the values of the resistors, assuming equal value resistors).
i have a decent grasp now, but could use some help with the potentiometer idea. i've already scrounged one to test, plus a switch to take it in/out of the circuit, but could use some help if there are any electronics gurus here.
i'm looking to clarify the wiring before i try it out.
i believe the pot i have is a 20k. though i have a BAAAASIC question about using the DMM to verify this.
on the "20k" ohms DMM setting, the figure I see on the readout is 19.9 when i'm dialed almost fully on the pot. when i turn the pot knob a tiny bit more to the end, the display goes to '1' (which i assume to be "over range" for the 20k setting). does this mean it's a 20k pot?
oddly, the pot itself doesnt have a rating anywhere on it, as far as i can see.
I don't mean to be a party-pooper, but tricking sensors has always concerned me. The torque converter requires a certain temperature to engage, or else damage could occur to the mechanism over time with the cold, stiff fluid. Secondly, start-up emissions would definitely be increased, since you're running at closed-loop conditions, even though the coolant temp is way below recommended levels. I know, I know, it's "Big-Brother" under our hoods, but cold-start emissions are a big deal -- hence the whole Winter gas crap. That's why I'm an advocate of "actual" temperature mods -- Engine Block Heater, hotter thermostat, and warm-air intake. It's real heat instead of perceived heat. I even feel guilty thinking about modding my 02 sensor because the CAT's bad -- it's about emissions too, not just mileage. I dunno, feel free to correct me if I'm wrong -- I won't take it personally.
Wow that's a big range of resistance change on the sensor - looks like a 20k pot is a good choice. The 1 on the meter is the over range indicator so you need to kick it up a range to read the 20.0 k ohms - measure the outer contacts not the wiper to get the resistance - your measurement accuracy is close enough. Having the pot is a good idea in case you run into a problem with fuel mixture when cold.
The torque converter requires a certain temperature to engage, or else damage could occur to the mechanism over time with the cold, stiff fluid.
i don't think lock-up delay has anything to do with protecting the tranny; it's delayed to speed up engine warming. the lock-up mechanism is just a clutch after all - earlier lockup won't hurt the transmission. (and think about how happy you were that it was locking up sooner from block heater use; yet the block heater has no effect on heating the transaxle fluid.)
Secondly, start-up emissions would definitely be increased, since you're running at closed-loop conditions
closed loop won't activate independently of 02 sensor feedback, so spoofing the coolant sensor won't result in closed loop at start-up. i doubt the car would even start - from a cold soak - in closed loop mode.
my understanding is that high cold-start emission levels are predominantly the result of low catalytic converter temps. the catalyst doesn't work until it's heated up by the exhaust.
based on that, i can see how spoofing the WTS signal may marginally increase emissions: if i have the car running slightly less rich than it otherwise might be, the cat may not warm up as fast.
and let's not forget you can't aim for a *radical* change in the WTS signal; it would create cold driveability problems. my goal is just to see if i can get a bit ahead of the pre-programmed curve.
it's largely academic anyway. the fuel savings will be small - more than likely negligible.
part of the appeal of this mod is just to see if i can do it. i like to hack stuff. i've already scrounged and altered a small circuit board, potentiometer and switch from a dead 12v cooler, and soldered a few jumper wires to make it do what i want.
i just want to confirm the wiring with jangeo before i hook it up. i suspect the greatest measure of success on this trial will be that i'll be able to make the temperature gauge & readout go up and down by twisting a dial inside the car
truth be told, i'm basically goofing around waiting for the weather to warm up so i can carry on with aero mods. that's where real gains are to be had.