mx5cartalk.com

Seeing as we are in the mood for tech write ups (nice work Uncle Arthur!) I thought I'd share my plans.

I am running my GReddy Emanage in closed loop while in boost using a fancy little device called an Autotune. It adds fuel based on the voltage output of the stock O2 sensor (just like the stock sensor does when the car is in closed-loop). The O2 sensor sends a lean signal when it is kaput (buggered, faulty) or below operating temp, which means my car is bogging like a bi-atch on boost before the car gets really hot, and then again if I have been sitting at a red light for too long. The O2 sensor says \"it's pretty lean in here\", when actually, it's nicely rich. The emanage then says, \"Lean? That's not good, I better dump a metric sh!tload of fuel in to fix that...\" \"Hang on, that metric sh!tload didn't fix it, it's still lean! I better keep dumping fuel...\"

So, going to a heated O2 sensor will do 4 things: get to temp quicker, stay there for the entire time the car is on, save me fuel and stop the car from bogging and driving me crazy.

I did some searching, and of course, someone else has done it before me.

http://mymiata.paladinmicro.com/Miata4-WireO2.htm

I bought the sensor today from Repco - around $60-70 will get you a universal 4 wire.

This mod will work for anyone who has an old sensor (they don't usually last more than 10 years or 150,000kms) or anyone who has installed aftermarket extractors (which move the O2 sensor bung further away from the head, and therefore keep it cooler for longer).

I hope this becomes helpful for someone! I will keep everyone posted on my success.

Boags

Don't you want a 5 wire wide band sensor for autotune?
My understanding is that a narrow band sensor can only tell if the engine is rich or lean, but not by how much... so I'd be careful using readings from it on boost.

This graph gives you some indication of how both types of sensors respond:
http://www.megamanual.com/v22manual/wbvolts.gif

There is still a finite value for the a/f ratio I am trying to hit (12.5:1) so it works. NB O2 sensors have a range of about 11.1:1 to 17.0:1 according to the write up in the link I posted.

I will be using a wideband to monitor the a/f ratio, but not to run the autotune. The stock computer still needs a NB sensor too for closed-loop operation.

Geoff

I agree with your comment about the range, but I don't think the voltage for a particular A/F ratio is deterministic, not at least in my experience. My new, heated NB sensor barely sees 0.05v movement between 11:1 and 13:1 (tested against an exhaust probe with the car on a dyno), and the specific voltage varies once the car goes through a few runs (due to increased temperature in the exhaust).

I would at least verify the tune on a dyno with equipment that is known to be accurate

I'm afraid I don't understand the question...

Well, it worked! Took a lot of work getting the bugger in, but now it's in and the massive bogging is gone. The senspr gets up to temp MUCH quicker - the end of the road instead of 5-10 minutes driving!

A worthwhile mod IMO.

Boags

Good to see it worked out and was benificial Geoff.

Just for the record the sensor he used was a four wire for a VS-- commonwhore, it was the same sensor as the universal one but worked out to be cheaper

Tim

rhysk wrote:I would at least verify the tune on a dyno with equipment that is known to be accurate

Absolutely! I will be on the dyno in the next week or two. I will also be monitoring via my own wideband sensor. It doesn't have a 1V output though, so I can't use it to run the stock ECU.

Thanks to zoomzoom for finding the part in a very vague catalogue.

It drives *almost* like stock now, though I still get the tiniest of little coughs if I really stomp it suddenly. Should be able to remedy this fairly easily.

Boags