[AlexE]

Hi,

Im trying to tune of my MX5 1.6 nbfl using LSU4.9 with spartan2 controller and getting weird results.

I made sniffer based on
https://www.wbo2.com/lsu/position.htm

Idle accuracy is bad but I expected problems there, mid load-rpm seems to be ok but problem is on WOT i get LEAN reading (AFR 20). Its not true because adding fuel does not change anything except cas running worse.

So i decided to compare Spartan reading on my MC5 NC (factory WB equipped, myself tuned) and logging both at the same time using Tactrix Openport device. And situation is similar: cruise area ok but WOT showing LEAN on Spartan but rich on original sensor installed in header

I made second test on NC prepared to maintain lambda 1.0 all the time and this time both sensors are behaving more correlated. At least no unusual lean reading on Spartan.

What is going on?

[PSIG]

I don't have time to review your tune and logs right now, but I've had that issue a couple times before. With a short exhaust or tailpipe O2 adapter, it's usually reversion at certain rpm and load ranges. As yours has both low and high load issues, it wouldn't be your fuel flow.

The way I proved it was to wear some gloves, and mostly cap the adapter outlet with my hand to reduce inversion. Note O2 readings. If no change, I use my fingers to mostly block the exhaust pipe outlet to reduce reversion. Note the O2 readings.

If the adapter outlet blocking changes O2 richer, it's likely the adapter pipe outlet end is too short.
If the exhaust tip outlet blocking changes O2 richer, it's likely the adapter inserted pipe is too short.

However, this assumes the exhaust has been tested for either exhaust or air leaks, and a perfectly sealed exhaust. Blocking the exhaust can create more positive pressure and not allow an upstream leak to suck air. The exhaust may be tested cold by capping the intake or throttle body and lightly pressurizing from the exhaust tip, spraying soapy water from a spray bottle to look for bubbles (like a tire leak).

A running test may not show bubbles, as most leaks will both suck and blow, making little to no bubbles. Note some exhaust leaks only appear with metal expansion at higher temperatures. An emissions exhaust sniffer or propane torch detector may locate those with a restricted exhaust tip while running.

I hope that helps, and I will have a look at your files when I can later. Good luck!

[EDIT] Adding WBO2 tailpipe sniffer tweaking info from wbo2.com:

WBO2_Tailpipe_Sniffer.jpg (295.17 KiB) Viewed 5744 times

[AlexE]

I made some tests you described but cannot see any consistent results. Even if I see output signal chnge when made some changes (blocking) on exhaust it does not come back when I remove it, rather stays on the same voltage

It seems exhaust sniffing - at least with my device - is more or less waste of time and fuel. Will try to remove original o2 sensor and install temporarily LSU there

[AlexE]

coming back after some time.

I installed LSU4.9 in place of original o2 sensor ... and it made things worse! I got signal from Spartan only for some time. As long as I had stoich sensor seemed to be working. But any RICH conditions for few seconds gave signal of LEAN. And lean stayed there ragerdless of anything. Only making really LEAN (DFCO) for few seconds made sensor run again ... for a while.

My friend bought brand new set of another WB controller (Stag AFR) with sniffer, so we could make testing. Found LSU sensor faulty

After some investigation I found LSU4.9 sensors are used in lots of cars, including diesels. Difference is for instance cable length. I bought two for about 10 bucks and will be testing. First test on the bench using gas lighter show they are both working