lean on full load

[Milos_Balunovic]

That diagram is based conventional engines, where the tip of the plug protrudes into the combustion chamber.

This isn't the case with the rotary, so the visible section is the most indicative.

Sent from my SM-G398FN using Tapatalk

[350matt]

So rescaling the MAF has brought the AFR back into line and the car does 'feel' better although this coudl be pure placebo of course

next stpe is to add a standalone wideband , which I do have but haven't used in many years

[350matt]

So update time
I've fitted my wideband kit and its showing about 6-7% richer than the OEM sensor
now to give this some caveat I've mounted the new wideband sensor post cat in the narrow band sensor port and the wideband kit I've got is a bit old and has no gauge so I'm measuring voltage output from the amplifier and applying this curve

wblambda.gif

as I have the NTK sensor this refers to

and I can't remember if the O2 sensor should read leaner or richer post cat - I would have thought leaner post cat? however my colleagues tell me there's little difference in reality

pulling the plugs and they look fine, a nice light milk chocolate brown and using some spark testers shows all 4 coils are lighting up

your thoughts

I was planning on fitting wideband into my Mini to see what it does in that car to confirm the wideband is actually OK

[350matt]

And further update I've tried the wideband on the Mini and checked its lambda with a scan tool so it looks like that curve doesn't quite fit the output of my device..... as I ge 2.37V for lambda 1 on the mini ( which I trust ) this basically means the rexi O2 sensor is reading within a percent or so of truth

so I've tried a different approach to the issue and disconnected the post cat lambda sensor and disabled all the DTC warnings for it , thinking that if that was faulty it coudl be skewing the LTFT - also reverted the MAF scaling back to what it was previously ( less fuel)
Mondays drive in will reveal all....
perhaps

[Milos_Balunovic]

Nice and organised approach. Hats down

[350matt]

Right then that seems to have fixed it
I'll keep an eye on the trims over the next few days and see how it behaves with no rear O2 sensor and I have actually reverted back to my earlier MAF scaling
might be the way forward and saves about 300g in weight...

[Milos_Balunovic]

Mine is mapped with no rear o2, it’s been performing and logging fine so I think you’re right that’s the best way.

[Phil Bate]

Apologies - I'm just catching up with this thread. The rear O2 sensor does influence the feedback calculated with the front sensor though it's mostly just to ensure the cat is happy. Under normal steady state running in closed loop you should see the STFT oscillate above and below a steady value in regular intervals - this is the PCM adjusting the target AFR based on what it can see at the rear sensor to enable 'purifcation' of the cat - in other words to keep it on the knife edge of having too much and not enough free oxygen.

It shouldn't build any wild correction values though, and I suspect your LTFT problems were definitely down to a MAF scaling issue. As MadTaz points out it learns this trim at fixed rpm/load bands so it could take a while before you see it return on WOT runs. What method are you using to scale your MAF table?

[350matt]

Hello Phil
thanks for chiming in
the odd thing is / was the tune was untouched since February and the car is driven almost every day and it was only over the last few weeks I noticed the LTFT getting bigger than normal
so I don't quite see how this could be a MAF scaling issue / error as the LTFT was very stable prior to this ( I run a turbogauge OBD device so I can see the LTFT all the time when driving )

my MAF scale method is take a long data log with mazda edit on the laptop and use the custom log table to capture the data with a filter to remove any events that are less than 5 occurrences in the same rpm / load bin
then cross reference this AFR data with MAF voltage and scale the MAF curve appropriately to achieve the target AFR in the map

with this method I got the LTFT down to less than 1 %

[ChrisHolmes]

That sounds good to me You have to be careful of transient conditions though, the car genuinely needs to apply trim for certain circumstances. One day it might need +5% for some reason and you scale the MAF upwards accordingly, then afterwards that condition changes and it might no longer need +5% or even needs to take fuel out - variations between tanks of fuel for example - this can then leave the car having to correct a lot the other way and could conceivably reach what you are seeing.

I just read back and see you haven't actually changed the intake - is the Cosworth filter similar to the standard paper one? The problem with scaling the MAF like that is you're using the results of a relatively variable process (combustion of petrol in a ridiculously quirky engine) to calibrate air flow. It's a good way to re-establish a base line after changing the intake for example, and you're doing it the right way, but in doing so without such changes you might have simply adjusted it the wrong way just because that's what the car needed at the time.

I'd put the table back to standard and see what happens - if it stays under about 5 either way then I would sleep easily. If it started getting above 7 then I'd start suspecting coils or fuel issues. I used to chase zero trim too until I realised it was something that didn't need fixing

[350matt]

Thing is if I revert to the stock MAF table then it won't hit my target lambda as the stock fuelling is a bit unpleasent as we all know

I've not touched the VE table which may be the better method to go instead of scaling the MAF differently to skew the system response