can anyone recommend a decent air/fuel meter gauge(kit) which is accurate enough for me to rely on to use for setting up the fuelling on my car.

I dont mind having to add another sensor in the exhaust or whatever it takes. I just want a really reliable gauge which i can use as an accurate tuning tool.

Cheers

No. EGT is much more useful.

yeah, and A/F or stoichometer (sp?) is basically a light show.

get an EGT guage. Greddy sells a nice unit.

ok chaps so could you point me to a site which explains the use of an egt aa a tuning tool

cheers in advance

You're looking at one... Some people get a little confused about C* vs. F*, but the information is here. Above a certain temp, you're generally too lean, below a certain temp, you're generally too rich. It's all been posted here before.

You may also want to hit the DSM sites for more information, as they seem to be just as good about documentation as we (SR20 folk) are, if not better at times. :)

But does anyone know where I can get a wideband O2 sensor so that I could use the a/f gauge accurately?

There's a group called DIY EFI or something like that out on the internet. Funny group of guys, always tinkering. They made some sort of an O2 meter using a Honda five-wire sensor, but there are a myriad of problems with using one of those sensors, not to mention the fact that you'd have to make your own gauge, too (can't use one from Autometer or VDO or somebody like that).

In Aviation, all they have is an EGT, and it's marked off in A:F ratio. I'd consider EGT more than enough information, since that's really the more important measure. You can have a perfectly stoich mix and still be too cold to make power or too hot to last another five seconds, and not know the difference. If you know the temperature, though, you always know where in or how far from the safe zone you are for component life.

Why is everybody so obcessed with this? It's worse than noisy BOVs...

Yea, there are actually several different wideband o2 sensors available at relativley good prices. Building a readout system for them is pretty easy too, poke around the plans are out there. The thing to keep in mind is these sensors tend to be pretty temperature specific, so you really need to run an egt with it just to make sure what its telling you is actually reality. I've got the plans for a few different variations if yer intrested.

I would be very interested to see these plans for a wideband o2 sensor could you post a link or send them via email.

Cheers

EGTs are good for High speed runs, where as A/F ratios are better for low end driving. like 1st 2nd gear. I have a Greddy EGT probably one of the best on the market today. Greddy also makes an A/F guage with it's own O2 sensor. It reads in volts and works extremely well. You can tune with an EGT only, but an A/F guage aids you a long. You will notice if you ever make a dyno run, it reads A/F ratios from a wideband not EGT numbers. FYI
www.roadraceengineering.com
Originally posted by Surry
can anyone recommend a decent air/fuel meter gauge(kit) which is accurate enough for me to rely on to use for setting up the fuelling on my car.

I dont mind having to add another sensor in the exhaust or whatever it takes. I just want a really reliable gauge which i can use as an accurate tuning tool.

Cheers

Originally posted by Surry
I would be very interested to see these plans for a wideband o2 sensor could you post a link or send them via email.

Cheers

No problem. Here (http://diy-efi.org/diy_efi/projects/diy_wb/) is the page for the previously mentioned version using the honda sensor. If your not much at building neato toys like this, check out this (http://www.techedge.com.au/vehicle/wbo2/default.htm) company in australia that builds everything for you at affordable prices. They ahip overseas but there was (I havn't looked into this stuff in over a year) a wait for there stuff.

The other sensor that was popular in the build-it your self market was Bosch's LSM-11 sensor. The electronics needed for that one are a bit simpler, but the sensor will run your $250-300 bucks. Just enter LSM-11 on google or whatever and you should run across some good links.

Good luck, and keep me posted if you end up with one of these setups. Its been on my list of things to build for quite a while, but I've havn't had the time. :)

You know, I never figured out why somebody doesn't just make a stepper box to hook a standard O2 sensor gauge up to this Honda sensor. I know it's more complicated than a simple transformer circuit, but still. Seems like a much better idea to me than using a bulky square project box with LEDs on it...

Originally posted by CowboyDren
You know, I never figured out why somebody doesn't just make a stepper box to hook a standard O2 sensor gauge up to this Honda sensor. I know it's more complicated than a simple transformer circuit, but still. Seems like a much better idea to me than using a bulky square project box with LEDs on it...

Its because the output of the honda o2 sensor isn't the same as that of a normal o2 sensor. A normal o2 sensor reads from 0 to 1 volt, the honda sensor is from 0 to 5 volts (actually I think its like from 1-4 realistic range, don't quote me on that). Also, the output of the o2 sensor (either one) is not linear. So you need a little electronics (well, one micro should do the job) to calibrate exactly what voltage matches with what A/F mixture. You could just slap a voltmeter on it and just correlate the voltage by hand, but that'd be ghetto man. :) Oh, and the honda one is heated (5 wire if I recall) so you do need something to drive that.

I know that it's not the same signal. :rolleyes: If you're going to all the trouble of making a box to decode it and recode it into digital information for display, surely an analog repeater circuit would be childs' play.

Originally posted by CowboyDren
I know that it's not the same signal. :rolleyes: If you're going to all the trouble of making a box to decode it and recode it into digital information for display, surely an analog repeater circuit would be childs' play.

No really, the problem is in the signal. Think of it this way (ready for the bad analogies?): the wideband sensor speaks Japanese, your good old flip-flopper o2 sensor speaks english, and your pretty autometer A/F meter has at best a rough grasp on the english language. "Repeat" the japanese to your old A/F gauge and you'll end up with a thoroughly confused gauge. It won't manage to learn Japanese, trust me. Garbage in, garbage out.

What you really is something to translate Japanese. Take a look at the two signals. Heres what the Honda is talking:

http://gemini.tntech.edu/~cjj6558/table.jpg

And heres what the output of the traditional o2 sensor looks like:

http://www.gadgetseller.com/gauges/pictures/voltage_vs_af.gif

Notice the difference? Note how the honda output is almost linear but the other is drastically not. Also note that the they're flipped around; ones reading a large voltage as lean and the other rich and vice-versa. Translator needed. If I was going to do it, I'd use (basically) a micro (not that this is the only way, but its just plain easy). It'd be pretty simple, just a matter of implementing a look up table. Almost any micro you pick will do A/D and D/A conversions for you, but heres the dealeo: why bother converting it back to analog when you can simply write the result to an LCD screen? And heres the other can of worms: how well do you really think that autometer A/F gauge is calibrated? Its just not made for this application.

You're preaching to the chior, Calum, I know the curves look and act different. ;) A translator (as you correctly put it) circuit is exactly what I want. Here's my motive: I hate the LED display on the black box, and I don't have the other stuff to display it on a big LCD screen like an $1800 Kenwood DVD display. Let's say that I hypothetically have an AutoMeter gauge in my garage...

What I'm saying, is that instead of making that complicated digital circuit that takes the processed (linear) data from the sensor side of the circuit and displays it, you make an analog circuit that ouputs a 0-1v range, just like the stock O2 sensor puts out. The difference in my idea is that the black box won't have an integrated DISPLAY, it will have a 0-1v OUTPUT. Instead of putting a bar graph inside the black box, put an analog voltage output in the box, and use a standard gauge in a standard pod somewhere in your cabin.

Of course you're going to have to keep all of the circuitry inside the box to massage the information; that's a given. The stock O2 sensor has hideous output performance, but the gauges we use to read them are linear in nature. IOW, an AutoMeter gauge is nearly identical to the bar-graph side of that AU-based kit, but in a narrow notch of the range. The guage does not display the full 0-1v range, though, it only displays from about 0.6v+ IIRC. The gauge isn't really calibrated much at all, it just runs. A:FR meters are not worthless because of the gauge, they're worthless because of the sensor. There would be calibration necessary in my design, but I think that's perfectly doable in analog circuitry, and it'd be mostly set-and-forget, once you get the ranges right.

Originally posted by CowboyDren
What I'm saying, is that instead of making that complicated digital circuit that takes the processed (linear) data from the sensor side of the circuit and displays it, you make an analog circuit that ouputs a 0-1v range, just like the stock O2 sensor puts out. The difference in my idea is that the black box won't have an integrated DISPLAY, it will have a 0-1v OUTPUT. Instead of putting a bar graph inside the black box, put an analog voltage output in the box, and use a standard gauge in a standard pod somewhere in your cabin.

This is a non-trivial analog problem. Its cake to do with a microcontroller. Either way, you'll end up with a project box. Pick the right micro and you should be able to get away with a very small box. But I'd still be leery about using an autometer gauge as your display. Do you know where 12:1 is on an autometer gauge? Me either. Its used to a ham-fisted input, its just not designed for the kind of resolution that a real wideband sensor can provide. Buuuut, what you could do, and I think this would be neat, is lets say you figure out where your car should be running. Lets say you think 12:1 is best for you. Calibrate your look-up table in your micro so that 12:1 centers around a .5 volt output (or whatever your gauge corresponds to as being in the middle). Then your gauge would be great for daily driving as a warning light for to lean or to rich conditions. And when you really wanted to do some tuning, just plug back in that ugly led screen.

Now we're on the same page. :)