Idle hunt

[Bitsyncmaster]

Got my chip and hooked it to the car. Well 200 Kohm pull up is way to low resistance. I increased it to 3.6 Mohm and it could get down to the 0.5 volt threshold. But I need to run a test of what the sensor puts out without the lambda duty changing. I was getting hunting (more so then stock) and the O2 sensor was jumping all over.

So needs much more work. I think I can modify the warmup software by waiting for the O2 to drop to a lower voltage and maybe adding a short time delay after that before going closed loop.

[Bitsyncmaster]

I changed my O2 input to load it to ground with a 3.6 Mohm. I monitored the O2 voltage and it takes about 2 minuets warming up to hit 0.2 volts and at 3 minuets it is up to 0.50 volts. So for now I wrote my software to wait till it goes above 0.50 volts and then go closed loop. Results are the same as the stock setup. About the same amount of hunt with my setup or stock.

I'm going to play with filtering in software the O2 reading and see if that helps hunting. If that does not work then I will resort to my micro producing the dwell output. That way I can find out if faster or slower dwell changes help or hurt hunting.

[Bitsyncmaster]

Check out this info when using high duration camshafts. Only seems to affect idle as far as the lambda system.

https://www.enginelabs.com/engine-te...s-limitations/

[Bitsyncmaster]

I think I know the problem with the stock lambda ECU. It very slow responding to changes with the O2 voltage which I think is why I get hunting.

Only thing left to do is use my micro to eliminate the stock lambda ECU.

[Bitsyncmaster]

I'm writing the software to produce the duty cycle with my micro. Here is what I see happens to cause the mixture (duty cycle) to get into an over adjustment result.

1). The frequency of the frequency valve is 67 Hz. So a change can only take place 67 times a second.
2). If your reading the O2 input faster then a change of the FV then your reading the still unchanged mixture again.
3). It takes time (delayed) for the mixture to change from a change in the FV
4). It takes time (delayed) for the O2 sensor to respond to the mixture change at the exhaust.

So I'm making my software wait till the dwell completes one cycle and then read the O2 to output a duty cycle change to the FV. But even that does not account for 3 and 4 causes. I would have to wait (delay) making another change to the FV.

I can set the frequency of the FV higher to get faster response but not sure how fast the FV can work at. How fast does the mixture change? It would not be a problem with fixed RPM but when you punch the pedal you will go lean but the lambda system can't respond that fast.

[Bitsyncmaster]

First test in the car works like a champ. No hunting.

I set a one second wait delay before the next PWM change which makes the new reading of the O2 sensor accurate. Now I will reduce that wait time but even with one second it looks like the dwell does not need to change much when I go to 2500 RPM.

My ECU circuit could be put on a one square inch circuit board but the stock circuit board is 5.4" x 4.6". I may just get a few prototype full size boards to get a few working units for testing. With the full size I would just replace the original one with my board and everything would be plug and play.

[WHO1DMC]

I was wondering when you would be working on something like this. I'm curious if circuitry could be added that could indicate failures in the system. Say for example that the FV, sensor switch and probe wasn't functioning as far as a current failure?



I'm definitely interested when you bring them to market.
Dave B.

[Bitsyncmaster]

I was wondering when you would be working on something like this. I'm curious if circuitry could be added that could indicate failures in the system. Say for example that the FV, sensor switch and probe wasn't functioning as far as a current failure?



I'm definitely interested when you bring them to market.
Dave B.

I can try to add extra test functions. Only problem is how to show the user. An LED would not be seen in the closed up unit. There are a few spare pins on the connector but no mating pins in the harness.

[WHO1DMC]

I can try to add extra test functions. Only problem is how to show the user. An LED would not be seen in the closed up unit. There are a few spare pins on the connector but no mating pins in the harness.

Myself personally I would have no problem making a window in the side of the case. Not trying to be difficult it was just a thought. I have the utmost respect for you and your products. If it wasn't for you I don't think anyone else would be willing to undertake such reverse engineering for obsolete electronics. I certainly don't have the skills!



Dave B.

[Bitsyncmaster]

Myself personally I would have no problem making a window in the side of the case. Not trying to be difficult it was just a thought. I have the utmost respect for you and your products. If it wasn't for you I don't think anyone else would be willing to undertake such reverse engineering for obsolete electronics. I certainly don't have the skills!



Dave B.

I can put an LED on the board to use for signaling the user. I have all the space to put down extras.