Mission Impossible: Ground Offset

[casper980]

I am trying to use an LM-1 to run AutoTune on my Autronic PNP in my Evo. I have everything wired correctly, and have even grounded the LM-1 power ground to the ecu power ground. The LM-1 and Autronic still don't read the same AFR. they are off by a little bit .1 and then after the engine warms they are off more like .4 which says Ground Offset, to me. I have tested the LM-1 Analog out to the o2 input by adjusting the LM-1 to output linear voltages and the Autronic cable is doing its thing. My problem is how do I fix the Offset? Do I need to wire in a direct ground to the engine from the LM-1 power, the LM-1 analog output and the ECU, to ensure they all get the same ground? What do I need to do, to diagnose and solve this? I have wired, searched the forums and called all over the country and even to Australia to talk to Autronics for over a month now to no avail. HELP!! I really don't want to buy an Autronic Exhaust Gas Analyzer.

[jfitzpat]

Is the Autronics input single ended (one wire) or differential (+ and - inputs)?

It actually sounds to me like the units are on different scales (the meaning of volts is different). But if it is a ground offset there are several ways to fix it depending on the input being fed.

-jjf

[cbernhardt]

Originally Posted by casper980

I am trying to use an LM-1 to run AutoTune on my Autronic PNP in my Evo. I have everything wired correctly, and have even grounded the LM-1 power ground to the ecu power ground. The LM-1 and Autronic still don't read the same AFR. they are off by a little bit .1 and then after the engine warms they are off more like .4 which says Ground Offset, to me. I have tested the LM-1 Analog out to the o2 input by adjusting the LM-1 to output linear voltages and the Autronic cable is doing its thing. My problem is how do I fix the Offset? Do I need to wire in a direct ground to the engine from the LM-1 power, the LM-1 analog output and the ECU, to ensure they all get the same ground? What do I need to do, to diagnose and solve this? I have wired, searched the forums and called all over the country and even to Australia to talk to Autronics for over a month now to no avail. HELP!! I really don't want to buy an Autronic Exhaust Gas Analyzer.

I don't have a lot of experience with my LM-1, but based on what little I do have I would say that you are doing well to get the two readings as close as you say. I have the analog (linear 1v-2v) output #2 feeding into an MEFI controller and I consistently see 0.1 to 0.6 difference in AFR between the LM-1 display and the analog output voltage. In my case I would not think that a ground offset would be the cause since the LM-1 is generating both readings.

When I first connected my LM-1 I just plugged it into the cigarette lighter and I was getting a lot of voltage spikes from the analog output. I tried it connected to a separate battery and the voltage spikes were virtually eliminated. I then tried connecting directly to the battery and the spikes returned. I now have an in-line filter that seems to work as well as the separate battery.

Charles

[casper980]

Originally Posted by jfitzpat

Is the Autronics input single ended (one wire) or differential (+ and - inputs)?

It actually sounds to me like the units are on different scales (the meaning of volts is different). But if it is a ground offset there are several ways to fix it depending on the input being fed.

-jjf

The autronic cable has 12v power, o2 sensor input and sensor ground.

[jfitzpat]

Originally Posted by cbernhardt

I don't have a lot of experience with my LM-1, but based on what little I do have I would say that you are doing well to get the two readings as close as you say. I have the analog (linear 1v-2v) output #2 feeding into an MEFI controller and I consistently see 0.1 to 0.6 difference in AFR between the LM-1 display and the analog output voltage. In my case I would not think that a ground offset would be the cause since the LM-1 is generating both readings.

When I first connected my LM-1 I just plugged it into the cigarette lighter and I was getting a lot of voltage spikes from the analog output. I tried it connected to a separate battery and the voltage spikes were virtually eliminated. I then tried connecting directly to the battery and the spikes returned. I now have an in-line filter that seems to work as well as the separate battery.

Charles

The LM-1 has a good DAC in it, but I can definately see how ripple in Vbat could turn into spiking on the analog output. If you don't mind my asking, what are you using as an in-line filter? A simple low pass would seem to do it, but I'm curious what you ended up with.

One thing to keep in mind is that, by default, the analog output is updating every measurement, but the display is updating at a lower rate. When I put the analog output of an LM-1 or LC-1 on an o-scope, I can see individual missfires as spikes, and valve action as ripples. This can make it hard to distinguish electrical noise from actual lambda changes. I generally program the output for flat line, then look at the remaining ripple, to measure noise.

-jjf

[jfitzpat]

Originally Posted by casper980

The autronic cable has 12v power, o2 sensor input and sensor ground.

What I would try is programming the LM-1 so that it is putting out a steady voltage. You can basically set .5 lambda = 3V and 1.523 lambda = 3V. The unit should then be always outputing a steady 3V.

Then look at what reading you are getting on your autronic unit. You could do this with several voltages, then just program the LM-1 output to give the autronic what it expects, that is, cancel out ground offset by 'shifting' the output of the LM-1.

Other considerations are update speed, which I mentioned in the post above, and time delay. The display on the LM-1 and the display on the other unit can be shifted from each other in time. Lambda is always on the move, so the different time lags can be deceiving. I usually use butane from a lighter to take the sensor pig rich, then gently shake it to empty the chamber and take it back to free air. Time delay differences are easier to spot on a nice, relatively smooth ramp from rich to lean.

Good Luck,
-jjf

[casper980]

Originally Posted by jfitzpat

What I would try is programming the LM-1 so that it is putting out a steady voltage. You can basically set .5 lambda = 3V and 1.523 lambda = 3V. The unit should then be always outputing a steady 3V.

Then look at what reading you are getting on your autronic unit. You could do this with several voltages, then just program the LM-1 output to give the autronic what it expects, that is, cancel out ground offset by 'shifting' the output of the LM-1.

Other considerations are update speed, which I mentioned in the post above, and time delay. The display on the LM-1 and the display on the other unit can be shifted from each other in time. Lambda is always on the move, so the different time lags can be deceiving. I usually use butane from a lighter to take the sensor pig rich, then gently shake it to empty the chamber and take it back to free air. Time delay differences are easier to spot on a nice, relatively smooth ramp from rich to lean.


Good Luck,
-jjf

I will try setting it to a linear setting and see what it reads on the other side. I was also considering shifting the output, so I will try that too.
I have the update speed at 1/12 sec but I will try other speeds. I thought it could be a time delay as well, when the car is just started and in warm up thats what it looks like and the AFR is very close between the two. Then when the car warms up the LM-1 will continue to hold with little fluctuation at 14.6-14.7 and the Autronic display's 14.4-15.3 constant fluctuation. Maybe I am getting too much noise in the system when the fan kicks on.


Thank You, I need it

[cbernhardt]

Originally Posted by jfitzpat

The LM-1 has a good DAC in it, but I can definately see how ripple in Vbat could turn into spiking on the analog output. If you don't mind my asking, what are you using as an in-line filter? A simple low pass would seem to do it, but I'm curious what you ended up with.-jjf

I used a Bandit NF-40 CB/ Stereo Noise Filter.
Charles

[jfitzpat]

Originally Posted by cbernhardt

I used a Bandit NF-40 CB/ Stereo Noise Filter.
Charles

Clever! Thanks for the info.
-jjf

[casper980]

Originally Posted by cbernhardt

I used a Bandit NF-40 CB/ Stereo Noise Filter.
Charles

I am going to try that, Thanks

[casper980]

Originally Posted by cbernhardt

I used a Bandit NF-40 CB/ Stereo Noise Filter.
Charles

OK, I got the same one...does it go between the power and ground of the LM-1 or does is go on the analog output?

Thanks
Jon

[cbernhardt]

Originally Posted by casper980

OK, I got the same one...does it go between the power and ground of the LM-1 or does is go on the analog output?
Thanks
Jon

It has positive and negative leads on both the input and output sides and it is wired inline (not between power and ground) on the power supply side.
Charles

[casper980]

Originally Posted by cbernhardt

It has positive and negative leads on both the input and output sides and it is wired inline (not between power and ground) on the power supply side.
Charles

Thank you for the clarification, I'll will let everyone know where I am after tomorrow.

Jon

[casper980]

Mission : Ground Offset - ACCOMPLISHED!!!!!

Ok, after a month of head ache, I have it so that the AFR's between an LM-1 and the Autronic Software are within about .1, ALL the time Here is how it is wired for an 03 Evo 8 with an Autronic SMC 1.99 PNP board (which is actually an SMD) Note: I am not a professional, this wiring is unique to the application and was found by trial and error and LOTS of research.

LM-1
Power - Switched 12V power doesn't matter where from.
Ground - ECU Ground pin A26 Black w/Gold Spots (A has 26pins, B12, C16, D22)
With 20A Noise filter between LM-1 and ECU Not sure if it does anything but its staying in just the same, because its finally working and that was put in the same day I figured it out.

LM-1 Analog Out, Note: I used the Innovate Autronic Cable but I did not use the 10:1 Divider, final wiring was done with standard analog side white and Green Wires. I had wired this up in lots of combinations and settings.

Sensor out White Unmodified - D6 o2 Sensor Input
Sensor Ground Green Unmodified - D22 Sensor Ground, Not Chassis Ground

LM Programmer Settings This is the tricky part

0V = 10 AFR
2.45V= 20 AFR
Delay = 1/6, I think, I will double check

The 2.45V is the tricky part, I had to go back and forth from the LM Programmer and the Autronic software and adjust the voltage until the numbers were close; not the most scientific way, but it seems accurate, so far. The same procedure will prob have to be done from vehical to vehical to be more accurate. Also remeber that I am not trying to Fully tune my car with this, only get the VE some what close in the low to mid load cells under 5000RPM's so that I can break in a new engine.

Autronic Settings

M1> Base Settings > A/F Sensor >

A/F Input Circuit > SMD 0 - 5V

A/F Input Sensor > Linear I/P A/F Meter

I hope this can help others, it could be useful in other applications as well.
I was not able to get the "autotune" feature itself to work but I was able to use the Datalog Autotune. Which is done by the following

1. Run Data Log - F8

2. Drive Car, starting slowly in my case from Idle, at desired load cells

3. Stop Data Log -F8

4. Map Data Log - F10

5. Save Data Log - Alt-E/Edit > Save Logged File, I always just overwrite the first one, by just pressing enter over and over, If needed you can save all of the runs to different names. I am just saving time.

6. While looking at data log, Tab over by hitting L, to Mixture Table Press F5 to Extract Data, tap down and Always hit OK, unless you know something I don't.

7. Select the Cells that you want to change and just press Enter twice. The software changes the Fuel Table for you

8. Press F4 to upload to car

9. Press F2 to save file I do this every time just to be safe

10. Repeat, as you can see, takes a little more time then the standard Autotune feature. I am going to work on that next.

This is in no way a replacement for an Autronic Exhaust Gas Analyzer, but it does its job I need it for. Sorry for the long post.

[jfitzpat]

Thanks for the info (and your perseverence)! Analog connectivity in cars, particular with somethign like a 0-5V signal, is always a pain. The environment is just spectacularly noisy.

-jjf