SQLGuy

From your post, from the manual: "4. Connect the RED wire to a switched 12V source in your car. A switched 12V source
goes on as soon as the ignition on the car is on. Make sure the connection is fused with a minimum fuse size of 5A."

Where do you see something saying that this can't be the 12V that used to supply the stock NB?

Are you maybe under the impression that the supply voltage to the stock EGO is something other than plain old 12V battery power? It is not. Check the schematic. The EGO supply is exactly the same 12V from relay XX and fuse 42 that feeds the fuel pump.

whall

Blown, thanks for your advice - appreciate it. BTW, I used the brown wire to the guage, the yellow is hooked to nothing.

Malcolm, thanks for saying Hi, we need to do a little run again this year.

SQLGuy

This seems contradictory. If the stock EGO ground is not suitable for such a signal, how was the LH getting a good signal to start with. If it is suitable, then aren't you asking for trouble with the LH by changing it? By the way, my AFR gauge is grounded to this same wire as well.

I am seeing slight changes in both idle and idle AFR by changing the stock curve. There is also the question of response time and heater warm up behavior that the LH would like to see. Do you have any suggestions on what these should be?

Although the narrow band signal is far from linear, it is not quite digital either. This can be seen in the 928 factory manual 1A, page 24-120. However, the main thing is the transition voltage, as you stated... it's not quite .5V, though, and small variations can make a big difference if the LH's target AFR. Changing the .958 Lambda voltage from 1.102 (stock) to .750 changes the 1.0 Lamba point from .45V to .33V, which will enrich AFR noticeably. Looking for where 1.0 Lambda intersects the graph from the shop manual, the LH should be expecting more of a .4V transition.

blown 87

I am not going to argue with you, if it works for you great, I just want a clean circuit coming off the fuse block or the battery and a really good ground.

jcorenman

There are multiple issues here. First, the signal ground for the NBO2 is through the exhaust pipe to the engine block. The ground wire in the NBO2 connector is the heater ground, not the signal ground. And because it is a current-carrying ground connection, it is not reliable as a ground-reference for small voltage signals.

The LH also has its signal-ground connected to the engine block, ground point on the right-rear corner of the block (under the heater valve). So for the NBO2 sensor, and the temp-2 sensor, it has the same ground reference as the sensors themselves, and all is happy-- from the LH's point of view.

Now we've got to look at the priorities. The OP was questioning very goofy WBO2 readings, while your concern is the NBO2 signal.

You are absolutely correct, that using a clean chassis ground for the LC-1 is different from what the LH is using as a signal reference, and that will change what it measures for the NBO2 signal.

But for engine tuning, what the LH reads for the NBO2 signal is not nearly as important as getting a reliable/accurate WBO2 signal. If you ground your WBO2 gauge to the LH/engine-block ground then-- in theory-- everybody is using the same reference. The problem with this is that the WBO2 ground itself carries heater current, which will corrupt whatever ground wire you connect it to.

Which is the reason that I suggested connecting the two LC-1 ground wires directly to a clean, separate chassis ground. One wire carries the heater current, the second is the signal ground, and if there is any ground resistance then there will be a voltage on the signal ground. But a clean bolt directly to the chassis has nil resistance and can carry the heater current without messing up the signal ground.

If you then take the LC-1's NBO2 signal back to the LH you have indeed changed what the LH reads for the NBO2 signal. My half-a-volt was ex-rectum, 400mV is probably more correct. But it is indeed a digital signal-- the LH doesn't care a bit about what the actual voltage is, only whether it is a above, or below, the threshold.

For a physical NBO2 sensor, the curve is very steep-- there is only a tenth of an AFR unit (approx) from 100mV to 700mV. Whether that voltage is a tenth of a volt higher or lower, has a very small effect on the final AFR. The LH is not doing anything with the NBO2 signal, other than checking whether it is above, or below the threshold.

This makes sense, and if you move the LC-1 Sim-NBO2 curve left and right across the Llambda/AFR scale then you should see the LH follow that.

During warmup, you want the Sim-NBO2 signal to be open-circuit (high-impedance), that's how a physical NBO2 behaves-- which causes the LH to ignore the (bogus) NBO2 signal while the WBO2 is warming up. WBO2 controllers which set the Sim-NBO2 signal to zero volts during warmup will drive the LH to full-rich (+20% more fuel) until the WBO2 warms up. (This is only an issue on a warm-engine restart, the LH ignores the NBO2 sensor altogether until the engine is warmed up). And the same with an error condition, select high-impedance. (I believe that is the LC-1 default). Response time should be fast, instant or 1/12 sec is fine.

Your numbers are more correct, my half-a-volt was very approximate. My point was that, for a physical NBO2 sensor, the curve is quite steep (steeper than Innovate's default curve) and it didn't matter. But if you change one point, you ought to change the other so that the 0.4v threshold (or whatever it really is) doesn't change. Personally I would make the curve much steeper, 0.99 Lambda at 1.0v and 1.01 at 0v-- just to minimize any uncertainty.

Actually, what I do personally is leave the NBO2 sensor in place, in addition to the WBO2, and use it as a calibration check. If my WBO2 sensor indicates an AFR that varies either side of 14.7 equally, when the LH is running closed-loop, then the NBO2 and WBO2 are in agreement and the WBO2 calibration is good. (They could also both be funky, but the NBO2 is pretty simple and doesn't drift much-- while the WBO2 needs a periodic air-cal-- a couple of times a year, in my experience).

Cheers, Jim

SQLGuy

Hi Jim,

Thanks for the long writeup. I think I see what you're getting at. Here's what I found:

1. The resistance, all the way from the 3-pin ground pin back to the LH pin 17 is about 0.11 Ohms.
2. The LC-1 draws 2A when heating and 1A once warmed up.
3. This translates to, at most, 0.11V operating offset of the NB signal. In reality it's probably more like .07V, since I was including the wiring from chassis MP VII to the LH in the resistance measurement.

By subtracting .07V from the transition point, I'd probably swamp any remaining offset below the noise floor, and that's probably what I'll do for now, but longer term is does sound like a good idea to provide a separate ground for the Blue LC-1 wire (heater ground).

I don't know what algorithm the LH is using for Lambda, but if it's a PID loop, then I think the slope of the transition would probably have an effect on how much overshoot there was on the I part of the equation.

BTW, I had thought the same as you that the LH should be running open loop until coolant temp was high enough to indicate that the engine was warmed up, but this does not appear to be the case. A cold engine, with the programmer hooked up to the LC-1 will stall within 1 minute - this is because programming mode locks the LC-1 output at 0 volts; so it does seem that LH is running closed loop right from cold start.

The fact that more concerns me is that it seems that nobody has actually been able to use the narrow band emulation capability of the LC-1 in place of the stock sensor. I, as you can see, am pretty stubborn. I will continue hammering on it and let you know what I find.

Cheers,
Paul

whall

Okay, today spent a couple hours resetting ground to chassis. I grinded to raw metal and attached.

I then hooked up my laptop and started the car. The results according to the logworks monitor were max afr of 22 with bouncing down when I gave it some gas - min was 12.

The whole time my AFR dash guage read 7.9

Any advice or suggestions would be appreciated. I will post on Innovate forum as well.

Thanks again.

SQLGuy

By default the Brown wire is setup for the gauge and the Yellow wire is set up to emulate narrow band for connection to the ECU.

Whichever output you are using for the gauge needs to be programmed with the LC-1 programmer to use Lambda and to have 0V at .5 Lambda and 5V at 1.523 Lambda.

whall

Hmmm.... Never saw anything about programming the guage but will read the manual again. Still have a problem that most of the time the monitor to the laptop is showing max lean.

jcorenman

OK, let's separate the problems.

Logworks connects to the LC-1's serial bus, so no ground issues, no setup. If it says AFR is 22, then the sensor is telling it 22.

Which I think means either it doesn't have a good air-cal, or the sensor has gone bad somehow. Try the air-cal first, either with the button (if you hooked it up) or the setup program. Also check the LED, should be steady with no blinking error codes.

I would also check the setup for the analog output channels, with the setup program- "LM Programmer", should be on the Innovate CD or download it from their website. It's a bit goofy, but works.

Cheers,

SQLGuy

Well, he said max was 22, min was 12, under load. 12 is a little lean for full load, but believable, and 22 under overrun (fuel cut) makes perfect sense. I don't think there's any problem with the sensor, just with the gauge connection.

whall

It was bouncing off 22 so real was way past 22. I think maybe I should run recalibrations. As for the guage reading I am lost. The whole system is really pissing me off. Major PIA to get working - has to be an easier way to get a accurate AFR measurement.

jcorenman

OK, maybe-- I read that as mostly 22-ish.

Warren, can you clarify? With the car warmed up and idling, what does Logworks show for AFR? If everything is working properly, it should be moving around 14.7.

And I agree, there is a problem with the analog gauge connection- either with the wiring, or the analog-2 output (brown wire) is not programmed per the default setup. The LM-Setup program will answer that question.

whall

To clarify, the 22 was not the high. It was the top of the measured range, the reading was going higher than that. The only time that the reading went down was with a bump of the pedal. I will run the to calibration tests again.