O2 Signal error
 

Ever since I installed my megasquirt, my Zeitronix display has always read lower than the Megasquirt. Last night I decided to track the issue down.

At first, I was pulling power and ground from the O2 heater. With this, the reading was WAY off. So I moved the ground to the megasquirt's ground. This got it close, but when it went to the lean end of the range, it got farther and farther off. Then I looked at the wiring and realized there was a sensor ground on the Zeitronix that I hadn't hooked up, so I hooked it up to the megasquirt, and moved the main ground back to the heated O2 circuit. This made no difference. Next, I moved both power and ground and sensor ground to the wires going into the MS. Again, no difference.

Next I opened the Zeitronix sensor include file, and compared it against Zeitronix specs. Other than extending the curve below 9.7 and above 21, the values were all right on.

So I opened MTConfig and set my wideband to "Unknown" so the gauge would display volts. Sure enough, when the wideband outputs 2.5v for stoich during sensor warmup, the megasquirt was reading 2.55v. The error got worse as the sensor warmed up and voltage increased.

This morning, I hooked up my multi-meter across the WB Signal and ground wires, and it read 2.50v, when the megasquirt was reading 2.55v.

So, it looks like there's something internal in the megasquirt that's skewing the O2 signal. Any ideas what it could be?

-Mike
(PS- This is unrelated to my issue in my other thread about AFRs getting leaner with higher intake temps)

hmm i have the same issue on my MS i am consistanly 2 points leaner than my AC vent mounted gauge. I hope you find the problem it might help me as well. Where are your grounds at? Have you calibrated the sensor? Also if you do real time display it will show 02 sensor volts, as well as afr.

whats is the output of the zietronix? can you program it? If so, program it for 0-5v to equal 10-20:1 AFR, then set MT configurator to innovate 0-5V

I investigated this issue more today while I had the MS on the bench for coolant fan and PWM idle mods.

I hooked up an old computer power supply to power the MS. I hooked two grounds (one from the 5v and one from the 12v) up to pins 7 and 8. I hooked the 12v up to pin 28 (or whatever the input for that is supposed to be). Then I used a potentiometer across 0-5v, with the wiper going into pin 23 (the O2 input pin). Then I fired up megatune. I verified that when the display said 2.55, there was only 2.50 volts across 23 and 7. I then took the next step, and measured voltage at each step on the circuit... DB37 end of R10, other end of R10, pin AD5 on the CPU (don't remember the number, it's 8 down starting from the pin diagonally opposite pin 1). I even tried measuring from pin AD5 to pin 2 (ground), and AD5 to the other ground pin... All 2.50v. So it appears to actually be internal to the CPU.

So... I guess using my bench setup, I can figure out all the error at all the voltages, and just make a new O2 sensor map. It's almost exactly right upt about 1v, then it starts drifting, and is almost a full .1v off by 4v. Still seems pretty strange, though.

-Mike

I think I finally have this figured out.

My Voltage regulator (U5) has 4.98 volts across pins 2 and 3 (2= ground, 3= vout). I only have 4.93-4.94 volts across the ground and +5v pin pairs on the ECU. I'm expecting the voltage drop is normal due to the number of things drawing current off that regulator. The value I see displayed for O2 voltage in MegaTune is approximately equal to Vo2 * 5.0 / 4.94. For some Vo2 values it displays a hair higher, and for some a hair lower, and I think that's because of the resolution of whatever is turning the voltage into the number.

-Mike