servion

I've got a techedge unit (v1.0), and its reaction time is really getting SLOW. For example, when I let off the gas from, say, 4k rpms, it drops to extreme lean really quickly (as expected, because the injectors are shutting off), but can sometimes take ~5-7 seconds of idling after that to return to reading a reasonable value. Any thoughts? I"m afraid my sensor may be bad.... if so, me = HOSED

BigMoose

That is the "failure mode" I have seen with the sensors. They go slow.... then slower.......... Might be time for another 300 buck NTK sensor. By the way the calibration is drifting also, but without a synthetic gas cal bench to reference it you wouldn't know that it is.

servion

The thing is: it's only seen probably ~15 hours of exhaust gases..... do they normally die this quick?

BigMoose

No. Unless you ran highly leaded fuel or ran so rich that the porosity that controls gas diffusion within the sensor plugged with carbon.

The heater must be on anytime the sensor is in an exhaust flow. If you perhaps ran the engine with the heater unpowered that would explain it, or ran so rich you were throwing carbon at/into the sensor.

Any of this apply?

servion

There was a few times where I accidantelly ran the car while the turbo timer was counting down, and when the turbo timer is counting down, the wideband controller is unpowered.... but the timer is set for usually 1 minute, and it has happened less than 10 times since I've owned the wideband.

I usually tune my boost tables to approx. 12:1. There have been some times where its been richer for a few passes (10:1) until I've tuned it down, but nothing more than that... just regular tuning. I always wipe some carbon off the sensor with a soft cloth after using it for a while. Does any of this sound like enough to cause the problem?

Is it possible fro me use use some brake parts cleaner or something similar to clean/renew/help the sensor?

BigMoose

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by servion &raquo;</TD></TR><TR><TD CLASS="quote">but can sometimes take ~5-7 seconds of idling after that to return to reading a reasonable value. Any thoughts? </TD></TR></TABLE>

OK, expand the description of the above in the timeline. Can you give specific a/F versus time?

I am assuming your wideband is installed before the Cat? Idle is sometimes a funky region. If you back off from 4k to idle it responds very lean....OK. If you jump on it will it immediately go to your 12/1? Or is it slow there also?

The safest first way to "clean" it is to leave it powered and hammer it hard as long as you can.

If your turbo'd is the wideband around 18 to 24 inches downstream from the turbo?
If N/A where do you have it?

Sorry for the questions but lets try to trace the problem down.............

servion

I'll try... here goes:
Say I'm curising at 5k RPM at 14:1 (doesn't seem to matter how long I'm doing this). If I let off the gas completely, the a/f reading in my romeditor drops to 17:1 (The value would read leaner than this, but this is the most lean voltage-&gt;lambda mapping I have in my conversion table. The actual voltage is approx 3.7V off the top of my head)

The motor behaves normally, (i.e. the rpms will drop to the idling range and the motor will begin to idle), but the A/F reading will not rise from 17:1 for another 5-7 seconds. (Keep in mind, the actual reading is much leaner than this... I know because of the voltage reading from the controller... 3.7x Volts... it only displays this because of my mapping table) After these 5-7 seconds have passed, it will begin to read what I believe to be accurate for the idle reading of this motor (approx 14:1). The whole time, however, the motor is idling consitently.. there is not change at all after these 5-7 seconds have passed in the sound or any other actions of the idling motor.

My setup is a b18c turbocharged with eagles and CP pistons. The wideband is before the cat, in the stock O2 location. (Previously I had ran it 2 inches off the turbo because my full-race.com DP has a bung there, but Peter (h-t member from techedge) warned against this).

I don't think I tried gassing it again while its still reading lean from a previous rpm drop... I think I always waited until it started reading sensible values again before continuing. I can try this out tonight if needed. Very strange problem... any suggestions BigMoose?

BigMoose

Try the 5K drop to idle then goose it. Let us know what the time response is.

I have built a tech edge simulator (both in hardware and spice- software) and the integration coefficients (R and C values) in their design when under a wide a/f excersion like slamming from lean limit to 12/1 can take up to 7 seconds to reach 90ish% of its final steady state output voltage value.

If that is what you are seeing it is normal (in my experience base) with the tech edge. Perhaps Peter of tech edge will comment? These sensors are relatively slow under rail to rail slams no matter what controller is used. ... note this is not a slam against tech edge.........

I get (with my controller designs) around 100 to 150 milliSecond response for a more normal A/F swing.... like 11.5 to 13 to 1...... as does the tech edge.

Hope this helps some..................