Which Wideband?
 

I'm hoping to have my car running MSPnP by the end of next month, but I'm still on the fence about which WBO2 kit I want to go with. DIYAutotune suggests the Innovate LC-1, but from what I understand [any?] other systems will work with the ECU as well, like the AEM UEGO. Anyone have valuable input on this? Should I bite the bullet and just pick up the LC-1 or would it be worth it to save a few bucks and get the AEM or another unmentioned kit?

Also on a side note, could something like the JAW (<-- linked) work with the MS input? For a DIYer looking to save a few bucks it could be a viable option.

The MS can be made to work with any wideband that has a linear output in the 0-5v range. Innovate, AEM, JAW, Autometer, etc.

IMO, Innovate makes the best one. I've not used the JAW, but I did have an AEM for a couple of years and didn't care for it. The JAW does support free-air calibration, so it might be OK. They're cheap enough that there's little harm in trying one.

Why is it that no one seems to like the PLX? No one ever even lists it as an option...

Maybe because no one has ever heard of them?

I know what they are I wouldn't use them anyway...... But here's the thing ask your self this? How much time do you wanna spend on tunning your car and how nice do you want it to run? I have tried my luck with the LC-1 and well It died (not knocking it though) Got A Uego and I won't look back now. The thing is the LC-1 Will be more precise no matter what, It's just a pain trying to get it to work. There are soo many "Help Me with my LC-1" Threads, And that's mainly due to incompetence, or in my case faulty equipment. If you want something to work and be fairly accurate get the AEM if you want the most sophisticated piece of tech, then get the LC-1, Also you can get the AEM for around $200 new so it's a little savings, Over all It's really up too you.

The LC-1 rarely works properly because people can't seem to read the instructions. I've installed and fixed many that work flawlessly every time.

Its nice to know that, when properly installed and calibrated, my WBO2 is more accurate than what's on most dynos.

LC-1

Installed mine (correctly) without a hitch, and it's served me well for over 6months. I love it.

Yeah, also hitchless here. My LC-1 replaced an AEM that had never-ending calibration drift on the analog output.

I honestly don't see how people can have problems with this unit. The instructions are clearly written, the wires are all color coded, they really couldn't make it any easier.

I have a PLX, and it is preferred by my local dyno shop. It self calibrates. It was dead on with the dyno's wb. The manual is a little bit confusing, read the documentation a couple of times and look at all the wiring before hooking it up wrong.

Installed the LC-1 in my 95 back in Jan. It has worked flawlessly except for the fact that it is .1 to .2 of a point off when compared to what megatune sees. After talking to some however, this is not too far off and seems to be somewhat common. I have had no problems with the actual unit itself, and I, like some before me have already said, believe that you will not have problems with the LC-1 if it is installed correctly.

Sounds like the LC-1 it is then! I'd much rather spend a little extra to get a known reliable, accurate unit than take the chance with my tuning on a cheaper wideband. Thanks guys!

Dont update to the latest firmware on the LC1 and you wont have any problems.

Oh, meant to tell you... got mine for $185/free shipping on ebay with a best offer auction... Your mileage may vary...

There's a few of us that like PLX, but unfortunately it's like most things in the Miata community.......

A bit fruity?

question for those of you with an lc-1. when you use your lc-1 to simulate a narrowband o2 sensor does it just scale down voltages so a factory computer can understand it or is there other things going on in there? and when hooked up this way on an otherwise stock car, does the car still sweep back and forth like a narrowband sensor or is it more stable?

reason for asking is for another car with a realtime rom tuning board inside the ecu, i'd like to run it with this system and when i log it, i can covert o2 values and to represent afr's and have them accurate.

Of course it "sweeps".

The cool thing about the LC1 is that via a graphical application, you can independantly scale the calibration of both outputs to whatever you want. By default, the narrowband calibration more-or-less faithfully reproduces the non-linear characteristics of a Nernst sensor.

thanks, that's what i was meaning to ask. now will a factory car run well if the voltage range was setup to 0-1v but linear?

Yes, but how well I honestly do not know. Really depends on what you mean by "linear".

The ECU is expecting to see a hugely non-linear swing right at the 1 lambda (14.7:1) point. The output of a stock narrowband sensor looks like this:

http://mysite.verizon.net/vzerdcib/s...tion_graph.jpg

Because the output varies both with temperature and age of sensor, there is a bit of a deadband in the ECU's interpretation of the signal, however in actual practice this does not matter because once 1λ is crossed, the voltage swings radically one way or the other.


The LC-1 does not precisely emulate the curve of the signal, because it doesn't have to. Here's what they do:

http://img37.imagefra.me/img/img37/2...ym_facf1df.gif

They can get away with this, because the ECU is really only looking at two states on the sensor input: Richer than 1λ (say, >600mv maybe), and leaner than 1λ (maybe <400mv or so). It's essentially a two-state machine, and it can get away with that since the sensor itself, under normal operation, passes through the deadband between these ranges almost instantaneously.


If you were to give the ECU a signal that was totally linear, my guess would be that the oscillation cycle would slow down, and it would tend to overshoot both positive and negative by a considerable degree each time.

Let's reference the first chart above, and arbitrarily say that for every step of 1% of fuel trim, AFR increases or decreases by 0.0025λ. We start at exactly 1λ, and are in the negative part of the cycle. After only 2 steps, mixture is at 1.005λ, the voltage swings down, and the ECU registers that the engine has gone lean and switches to positive trim mode.

By contrast, if we are scaled linearly where (to pull a scaling out of my ass) 0v = 1.5λ and 1v = 0.5λ, it would take a lot more steps before the voltage goes low enough for the ECU to register a change.

In my experience, it will run... but far from "well".

FWIW, I had to get my car smog-checked recently. I installed the MAF, stock injectors, EGR, and stock ECU (usually run stand alone MS). During the conversion to smog-readyness, the one thing I did NOT want to deal with was installing the NBo2. So I configured one of my analog LC1 outs to 0-1v. The car ran very lean in cruise (16:1). Then I sacked up and installed the NBo2 and all was fine. So unless I was doing something wrong, I can say that the LC1 analog out won’t mimic a Nbo2 by simply configuring it to 0-1v. I read something about a “knee” somewhere but don’t remember where.