Right now I have an S5 TPS hooked up using the full-range cam. If i set the TPSdot Threshold to ~0.4 v/s, the car responds much more nicely just off-throttle and when coming back from overrun fuel cut. However, when cruising steady-state on the highway (and I assume other times too), the car will randomly go 1-2 AFR points richer (accompanied by a slight change in exhaust note and a little bit more power). Then, if i set the TPSdot Threshold to ~0.6 v/s, I dont notice any signal noise or random enrichments but the car has the tendancy to stumble a bit more in the situations I described above (and pretty much anywhere when you'd roll onto the throttle slowly).

What would be the best way to hook up a capacitor (or would an inductor be a better choice since I think the concern is more with high frequency noise) to try to get rid of the noise on the line? Any recommendations on capacitor size? Thanks a lot!

 

a low pass filter?

you need a circuit with two resisters and a cap between them feeding noise to ground




this is a diag of the infamous corvette tach filter
( its a low pass filter ranged in the 6 - 12 V range for noises outside a normal tachos signal ranges )
[ to kill the bell ring echo from doubling up the tacho signal that HEI coils are known for ]

( there is some math involved that i am reluctant to chase knowing the real world limit is actually the ranges that caps are actually available as )
recommend cut offs using .047 mf to .1mf caps
will range thing close enough on the guestiguage

if using polarized one way type caps get them around right !!
( i use unpolarized polyester film type )


going lower than .047 mf for the cap is going to raise the cut off
-and allow lots of noise through
going more than .1mf and to higher resistances may feed most of the signal away !

PS
noises may not just be coming down the 5 v signal from TPS
- but instead " soaking " the boards power input wire
a low pass filter and some flyback protection may help on the main power and earth feeds
( some types of common ECU are very bad for this )

 

Another thing to consider is the cabling you are using for the TPS line. Optimum is twisted pair shiekded wire with the shield connected to ground at the MS end only. I use microphone cable and have no problems with noise on any of the inputs. YMMV...

 

Thanks for the responses!

I'm just using the stock wiring for the TPS right now (and all other sensors for that matter), and the TPS is the only one i've seen any problems with. I will try to grab some shielded cable for the TPS and run it in the harness the next time I have the UIM off to see if it helps a bit.

I'll try the low-pass filter, but wouldnt the addition of those extra resistors increase the closed throttle ADC count, effectively reducing the sensitivity of the TPS? I can see how it'd work for a pulsing digital PWM signal like a tach signal though.

 

yes the Resistances added will knock down the voltage as seen at the ECU
- but going through the normal calibration technique should automatically counter the small loses
( in effect its just narrowed the available range as seen by the ECU, and added the filter to feed HV cycles away )

the cap needs SOME resisters to work as it should
( mainly to stop the cap feeding flyback on shut down )
though you are welcome to drop back the resistances as posted
( since they are for a 12V tacho, not 5V TPS )


again, i will point out that the problem is likely not the 5 V input to ECU picking up radiant noises as such from the air
BUT instead ignition noises being fed into the ECUs 12V power feed
( and finding there way to earth along the 5V bus to the TPS and back again to earth )

so, the low pass filter and flyback stuff may actually be more effective on the 12V feed rather than the 5 V as such

PS
there is much more complex clipping circuits that do what you want without the drastic tenuation that these simple low pass filters do
you need to talk directly to the electronics whizzes for them,
as my experience with anything more complex than these simple filters is low