Hi all,

Got an Autometer AFM and going to do it this weekend. After reading a monster thread on this by Paul Wadams and John Banks, I have decided to go for:

Sensor - ECU
Ground - ECU
Live - ? Dont know! :D

Anyone like to advise? Can I use the ECU for this too? And which wire is it?

Many thanks,

MB

Yep, live ECU should be fine, if the wiring loom for the gauge means this will be the easiest spot to take power from, this is where I powered my S-AFC from with no probs. You don't have to get it from here, it is only the sensor and the earth that effect what you are seeing, the power is just to power the chip and LEDs/meter.
On the MY01, there are two ECU wires that can be used to feed 12V, they are D2&D3, Yellow/Blue, and larger than most of the other wires.
But because my AFM is up here:
http://users.bigpond.net.au/tony_h/s...ar_mixture.JPG
I got power, earth, and light dimming from the clock. (I built light dimming into the circuit, and made the accuracy 0.05 instead of 0.1 up in the rich zone as well).
However, I found that turning on the lights would cause the LEDs to move down by up to 3 LEDs.. not good!
I moved the earth to the ECU earth suggested in the S-AFC documentation, and found the LEDs would dance around like crazy (I assume that the injectors or coils firing circuitry cause a voltage change on this particular earth).
On the MY01, and I am sure the previous models, there is a "Sensor Earth" on the ECU where all the sensors earth, this is a great spot to earth! No more jumping LEDs (apart from when the mixture changes!) and no more accessories affecting the reading. This earth is B19 Red/Green, using the Ravensblade (MY02 in US) pinouts as the guide. I also earthed the S-AFC sensor earth here, rather than the suggested ECU earth in the S-AFC documentation due to seeing the TPS% jumping all over the place (and I would have the throttle partially down and see 0% when the engine was running sometimes!). It worked OK where Apexi said to put it, but moving it to B19 I was able to truly see what my readouts were meant to be (and trust that it was sending the right MAF voltages to my ECU too :) )

R's
Tony

[Edited by tonymy01 - 9/14/2002 4:56:45 AM]

Woo hoo! Someone replied :D

Thanks for the info. I have wired it now and it all goes into the ECU.

I let the car warm for 10 mins or so and it seems to work fine. At idle it will sit at rich, then drop to lean. This happens repeatedly and as far as I know this is correct behaviour as the ECU runs in either closed or open (cant remember which) When I blip or hold the throttle it goes rich, and as soon as the foot comes off it goes v lean, I guess as there is very little going through the injectors relative to air flow.

Does this sound ok? If not I might try the sensor earth you used...

MB

That sounds like how it is meant to be. When mine was going silly, it was not a "smooth" transition between rich/lean. Did you check that putting on your headlights doesn't change the reading? If it doesn't, and doesn't go silly, it means you found a suitable place for earthing.
You will find that when revving, then coming to idle, the closed loop control takes about 10 to 15secs to get the mixture steady, you will see it go full lean (when the throttle is backed off, with no LEDs being displayed, due to not having any fuel with throttle back off), rich, then cycling getting closer and closer to stoich each time.
WOT (wide open throttle) should have the last rich LED lit (unless you have a better calibrated unit) without wavering like it does under closed loop.
Good to see you don't have any probs.
R's
Tony


[Edited by tonymy01 - 9/14/2002 10:58:14 AM]

Thanks for the re-assurance Tony.

Your setup looks very neat!

As I had a bit of time today, I rewired the earth to the battery just to check it. Identical results. I have left it there although I would prefer the ECU. My thumbs are burned from the solder so I will leave it til next week! :(

MB :)

I have just asked the same question in general. But I understand that the earth is very critical and should be wired direct to the battery according to the instructions on the dawesdevices.com net.Where is the ecu located on a MY02.
Thanks for the advice so far, very helpful!
Bob

Yep, there was a thread about that ages ago, that was about a mile long!
You won't go wrong if you earth it at the ECU sensor earth (this earth is where ALL the vehicle sensors are referenced). Any of the other ECU earths are no good (well, I have only tried one other, and it was quite noisy on this earth).
There was an argument that if the O2 sensor earth failed, then this would not work earthed here, but I think if the O2 sensor earth failed, then you will have more probs than just not having your LED meter working.
Yes, the battery will be reasonable, but you *MAY* be introducing a small voltage drop earthing it so far away from where the sensor is referenced, i.e. your ECU.
The ECU in the MY01 is on the passenger floor under a kick panel (4 bolts hold the kick panel down), I am assuming it is in the same spot on the MY02? (MY01 & MY02 are the same shape WRX here down under.)
Cheers

Tony / Active

I tried both earth points. Went for ECU first and though it might be iffy, so went for the battery. There is very little difference in behaviour, but I feel it was more acurate on the ECU earth. I only got single lights, but with the battery I can get more than one on.

Once you get used to the way it works its fine.

Just been reading that the lambda is not used at WOT, which is the only time you get a sensible reading from the AFR. The rest of the time it just bounces up and down in open closed loop.

Beginning to wonder if its worth it!

MB

Tony / MB if you look at www.dawesdevices.com and go to instructions you will see an explanation (in A/F meter Instructions) and look under (using the meter) there is an explanation as to why you might get the lights flashing and if the meter is between ranges it is possible to get more than one light illuminated. Also at idle or cruising conditions it is normal for the display lights to flash as the computer adjusts the mixture apparently. According to Dawes, Air/fuel meters can only be used under wide open throttle conditions. Interesting reading!.
Bob

I have read that site a while ago, this (along with John B here) is what gave me the idea to hack mine to have 0.05V resolution instead.
The chip used will at most show 2 LEDs at once, as the transition goes exactly between the voltage for each LED (the chip has a tiny amount of hysteresis so that there won't be a situation where there are NO LEDs lit, apart from when the voltage drops to below the bottom LED.. i.e. super lean).
The reason he says "only used at WOT" is because this is when the ECU goes open loop, and just uses a lookup table for fueling, and so isn't tweaking the mixture trying to make it hit a sweetspot with ever decreasing cycles, this is also the best time to be monitoring the mixture too, as the ECU won't know if the car leans out at this point (apart from feedback from the knock sensor).
Also at WOT, you have a fairly conistant air and fuel stream hammering thru the system, so the mixture is not changing a whole lot.
& at WOT, after the initial WOT accel enrichment, it is fairly easy to see what the ECU is doing with the fuel that isn't going to be influenced by you cycling the throttle (which massively changes the mixtures).
The jumping LED thing I was talking about earlier (with the ECU *power* ground, rather than the ECU *sensor* ground) is where the LEDs would just about light up randomly, rather than having a smooth transition between them (except for throttle back off, where the LEDs instantly extinguish, and full throttle, where the top LED instantly lights up). This was due to voltages induced on the ground, which cause the fairly accurate measurement of only 50mV to obviously be affected.
Hope this helps.
Cheers
Tony

Great! it certainly does, my god mate you know your stuff!
Thanks foe the valuable info. It`s nice to dig a bit deeper and gain a better understanding of how things work.
Tks agn Bob