Idle hunt

[Bitsyncmaster]

It bothered me that my ECU showed that problem starting closed loop early and bench testing showed no problems.

Well I did a couple scope shots. One with the scope ground on the signal ground pin and one on the power ground pin. The stock wiring has two signal ground pins (one connected to the shielded sensor input wire). The power ground is where the FV driver conducts current. Well I connected my buffer op-amp to the signal ground and my micro to the power ground. That is a problem as you can see in the attached scope shots. My micro has the A/D ground reference to the power ground.

[Bitsyncmaster]

Here is a scope shot of the stock ECU O2 voltage. I slowed the time base on my O-scope so getting the same slow swinging O2 voltage as my ECU. Don't know why that voltage swings like that. It is not caused by the dwell changing because my ECU shows the same and I slowed the dwell change real slow. As it warms up the voltage is very flat and says flat even when closed loop starts but about one minute later voltage starts swinging like that. That is what is causing the idle hunt because the dwell on the stock ECU is following that voltage change. My ECU I can slow the dwell change so as not to swing dwell that much.

[Helirich]

So read the first post and some of the others. (I haven?t read it all) Is your goal to make an after market part that will not hunt? That would be great. My car hunts sometimes and some times it works perfect. There seems to be no rhyme or reason. It hasn?t been a big problem, so I haven?t really investigated it.

Most of your posts are over my head, but please keep at it. If anyone can do it, you are the one. I await the announcement of your new product. Please take my money. Lol.

[Bitsyncmaster]

So read the first post and some of the others. (I haven?t read it all) Is your goal to make an after market part that will not hunt? That would be great. My car hunts sometimes and some times it works perfect. There seems to be no rhyme or reason. It hasn?t been a big problem, so I haven?t really investigated it.

Most of your posts are over my head, but please keep at it. If anyone can do it, you are the one. I await the announcement of your new product. Please take my money. Lol.

Yes, cure idle hunt caused by the lambda ECU. A lot of owners unplug the O2 sensor to cure their idle hunt but I just had two owners with high duration cams get idle hunt with my idle ECU and not get it with the stock ECU. So I decided to cure the cause. It turns out high duration cams make the O2 sensor show lean at idle RPM (does not affect normal RPM).

I bet that O2 voltage swing increases with those cams. I'm going to look into what causes that O2 swing but it may just be normal for narrow band sensors. My ECU can eliminate the hunt by slow dwell changes so as not to follow that O2 swing but still maintain the 14.7 AFR.

[Bitsyncmaster]

I put a user switch onto my ECU so I could stop the update (holds current dwell). So I found out holding the dwell does stop that O2 voltage from swinging. So the stock lambda ECU is swinging the mixture in that 4.5 second sequence. My ECU will currently do an occasional swing with the present timing. I think running slower would stop that but it does not hunt as is.

I guess there is not really a reason to run fast since the many owners that pull the O2 wire don't get any mixture adjustments. Have not heard of any complaints with the fixed duty cycle at 45 deg which is what you get pulling the O2.

[Bitsyncmaster]

The more I test the more I hate a narrow banc O2 sensor. If you let the car idle from cold the sensor gets into that swinging voltage change (My ECU is slow enough not to hunt with that). If you do a normal RPM after that the sensor seems to clear out after a while. If you punch the throttle voltage jumps and then falls to zero volts. I guess the punch cools the sensor off so it goes high impedance.

Now I'm going to add a circuit to my ECU to detect when the sensor is high impedance. Need that to set when to go closed loop. It also lets the ECU set the fixed 45 deg. dwell when a user pulls the O2 wire. This will also let me know if the user is testing with a short circuit on the O2 input to get minimum dwell.

[dmcman73]

The more I test the more I hate a narrow banc O2 sensor. If you let the car idle from cold the sensor gets into that swinging voltage change (My ECU is slow enough not to hunt with that). If you do a normal RPM after that the sensor seems to clear out after a while. If you punch the throttle voltage jumps and then falls to zero volts. I guess the punch cools the sensor off so it goes high impedance.

Now I'm going to add a circuit to my ECU to detect when the sensor is high impedance. Need that to set when to go closed loop. It also lets the ECU set the fixed 45 deg. dwell when a user pulls the O2 wire. This will also let me know if the user is testing with a short circuit on the O2 input to get minimum dwell.

Why not use a wideband O2? If it works with that, the end user will just need to switch over to a wideband O2, you can even make it where the power comes off your new unit (if you're creating one.

[Bitsyncmaster]

Why not use a wideband O2? If it works with that, the end user will just need to switch over to a wideband O2, you can even make it where the power comes off your new unit (if you're creating one.

The wideband devices take very complex hardware and software which has already been designed. They output an analog output which can drive the stock lambda ECU input. I did have one installed in my car before but I removed it when I was trouble shooting a skip in my engine. Wish I left it in because it never hunted but then I would not have started this R&D. After I finish this design I may get a heated narrow band sensor just to measure if it corrects some of the problems I've seen without the heated sensor.

Now you give me an idea to have a user switch to run my unit without a delay if the owner has installed a wideband.

[Bitsyncmaster]

Well with more testing, it's not as bad as I thought. I keep thinking the O2 voltage is an analog device showing AFR but it is really almost a switch showing above or below 14.7 AFR. So even if the AFR is holding solid at 14.7 +/- .1 AFR your going to get swinging O2 voltage.

That and controlling the AFR with the frequency valve is also not that precise.

So I'm getting close to finishing up the software to do road testing and have a few other owners do testing.

[Timeless]

So I'm getting close to finishing up the software to do road testing and have a few other owners do testing.

I'm happy to test on my DMC stage 2 car if you need.