Sequential fuel Tuning using EGTs

[daxtojeiro]

Hi all,
anyone done any EGT tuning?

I've put the standard plenum back on, ordered a large inlet from V8D to replace the siemese one which I flogged on EBay.

I've flow tested the injectors and found them around 2-3% different max, so not too bad at all.

I've just been playing with an infa red temp sensor on the exhausts. I know that these are far from acurate and I will be fitting EGTs over the next few days, but here are my findings.

AFR leaner on Even bank by around 1.0 AFR

Temp readings (infa red) showed cy 2,4,6 significatly hotter (leaner) than the rest (80ish degC)

Increased fueling to cys 2,4,6 to get roughly the same temps and now AFRs even across banks.

Anyone had this or know why? Im aware this may change when I do it using EGT, but I would guess its close enough to prove that they were lean.

Found injector angle of 40deg before inlet valve opens(-300deg) gives an increadably smooth idle.

Phil

[conrod]

Anyone had this or know why? Im aware this may change when I do it using EGT, but I would guess its close enough to prove that they were lean.

Not had this on a V8, but had it on a 6 cyl BMW engine many years ago.

Inlet manifold was a "log" shape with throttle body at the middle of the "log" With 1 lambda downstream in the exhaust it said all was fine. BUT, with a lambda in each primary pipe (as well as EGT's) we found the centre two cylinders (3&4) were running dangerously lean, and the outer two (1 & 6) were very rich with 2 & 3 about right.

Quite simply this was because of the uneven air distribution from the inlet manifold, the centre two cylinders getting a direct shot from the throttle body, making them run lean, while the end two cylinders got less air as it had to take two turns and travel down a fairly narrow shaped plenum.

The only way to fix this is to either redesign the inlet maniold to give even flow to all cylinders, or to "trim" each cylinder individually so that the AFR is corrected. We picked up 6hp on a 2L 200hp six cylinder engine by trimming the injector outputs. Unfortunately it is usually only high end ECU's that have this facility.

It is probably important to point out that you need to take the readings over the full rpm range, and not just at idle. But it sounds like you may have figured this all out anyway!

Conrad

[daxtojeiro]

Cheers Conrad,
I will be fitting the EGTs very soon. Doesn't look like this has been done by many people, everyone I speak to has no idea about it
Phil

[Eliot]

what about 8 widebands

[conrod]

what about 8 widebands

we tuned the BMW with six! EGT's should give a similar result though.

Cheers Conrad,
I will be fitting the EGTs very soon. Doesn't look like this has been done by many people, everyone I speak to has no idea about it
Phil

You obviously have a good understanding of why you want to do this. It really is a super accurate way to tune an engine, especially one where the inlet plenum shape can give rise to differences in air distribution between cylinders. Where there is a significant difference (and in your case it sounds like there is) there can be some good gains to be had, it is almost "free" hp. Look at any serious race engine builder, and when they dyno an engine they will be taking readings (either lambda or EGT) from each cylinder. One lambda in an exhaust pipe (or two in a V8 ) is really just giving an average reading from the cylinders that it is reading from.

[DaveEFI]

what about 8 widebands

we tuned the BMW with six! EGT's should give a similar result though.

Cheers Conrad,
I will be fitting the EGTs very soon. Doesn't look like this has been done by many people, everyone I speak to has no idea about it
Phil

You obviously have a good understanding of why you want to do this. It really is a super accurate way to tune an engine, especially one where the inlet plenum shape can give rise to differences in air distribution between cylinders. Where there is a significant difference (and in your case it sounds like there is) there can be some good gains to be had, it is almost "free" hp. Look at any serious race engine builder, and when they dyno an engine they will be taking readings (either lambda or EGT) from each cylinder. One lambda in an exhaust pipe (or two in a V8 ) is really just giving an average reading from the cylinders that it is reading from.

When I fitted my wideband, there were warnings about it being too close to the exhaust port and overheating. So it had to go further downstream than a narrow band type. How do you get round this?

[conrod]

maybe it was a problem with that particular sensor?

I have owned 2 x Escort Cosworths, and they had the factory lambda fitted in the turbine housing just behind the turbine wheel- it doesn't get much hotter than that! Only problem I have had with lambdas is mounting them too far away from the manifold, and they won't read properly or at all without enough heat in them.

[Eliot]

maybe it was a problem with that particular sensor?

I have owned 2 x Escort Cosworths, and they had the factory lambda fitted in the turbine housing just behind the turbine wheel- it doesn't get much hotter than that! Only problem I have had with lambdas is mounting them too far away from the manifold, and they won't read properly or at all without enough heat in them.

I've got T34's on the dakar and indeed there is a 12mm hole on the turbine - which suggests titania or some such perhaps.

[conrod]

maybe it was a problem with that particular sensor?

I have owned 2 x Escort Cosworths, and they had the factory lambda fitted in the turbine housing just behind the turbine wheel- it doesn't get much hotter than that! Only problem I have had with lambdas is mounting them too far away from the manifold, and they won't read properly or at all without enough heat in them.

I've got T34's on the dakar and indeed there is a 12mm hole on the turbine - which suggests titania or some such perhaps.

yes you are right, they were 12mm, a lot smaller than the usual 18x1.5mm.

[stevieturbo]

Plenty of people at the top end do this...but in most cases, is it worth the expense or extra work for more normal builds ??

Likely not.

You also need the ability to tune/adjust both AFR's, and ignition timing on an individual basis to make the most of the effort.

Inlet manifold, exhaust manifold, and even condition of camshaft/lifters may all play a small part. Are all the injectors 100% flow matched too ?
3% flow difference at various duties is very common...so could easily account for it. Some standard injectors could be much worse.
Simply testing an injector at 100% but means nothing, and far too many come flow rated at 100%
At 30% they might flow a lot different, and same again around 60% for example.


Injection timing is another debate. Logic might say to inject only when the intake valve is open. But that is a small window, and likely little point injecting any fuel as the valve is closing. Im sure I read that Motec say that at times, injecting all fuel against a closed valve is preferable.

So really....actual testing is the only way to optimise for each and every engine.

For injection timing, tune AFR's via timing until they read the richest ( making NO changes to the actual fuel map itself whatsoever )
This will indicate that more of the fuel being injected is actually getting burnt.

But again, obviously this will require a full mapping session at various rpm and loads if possible.


As for fitting EGT's though. Innovate's TC-4 is a relatively cheap way of achieving it. Each unit can read four thermocouples, so daisychain two of em for an 8cyl.

8 Widebands would be nice....but that's getting into a lot of work and cost !!

[stevieturbo]

maybe it was a problem with that particular sensor?

I have owned 2 x Escort Cosworths, and they had the factory lambda fitted in the turbine housing just behind the turbine wheel- it doesn't get much hotter than that! Only problem I have had with lambdas is mounting them too far away from the manifold, and they won't read properly or at all without enough heat in them.

I've got T34's on the dakar and indeed there is a 12mm hole on the turbine - which suggests titania or some such perhaps.

yes you are right, they were 12mm, a lot smaller than the usual 18x1.5mm.

Likely for EGT as opposed to lambda

[DaveEFI]

maybe it was a problem with that particular sensor?

I have owned 2 x Escort Cosworths, and they had the factory lambda fitted in the turbine housing just behind the turbine wheel- it doesn't get much hotter than that! Only problem I have had with lambdas is mounting them too far away from the manifold, and they won't read properly or at all without enough heat in them.

I'd guess a factory fit one is narrowband. Modern cars don't vary much from stoic. Wideband are very different.

[daxtojeiro]

Plenty of people at the top end do this...but in most cases, is it worth the expense or extra work for more normal builds ??

Likely not.

You also need the ability to tune/adjust both AFR's, and ignition timing on an individual basis to make the most of the effort.

Inlet manifold, exhaust manifold, and even condition of camshaft/lifters may all play a small part. Are all the injectors 100% flow matched too ?
3% flow difference at various duties is very common...so could easily account for it. Some standard injectors could be much worse.
Simply testing an injector at 100% but means nothing, and far too many come flow rated at 100%
At 30% they might flow a lot different, and same again around 60% for example.


Injection timing is another debate. Logic might say to inject only when the intake valve is open. But that is a small window, and likely little point injecting any fuel as the valve is closing. Im sure I read that Motec say that at times, injecting all fuel against a closed valve is preferable.

So really....actual testing is the only way to optimise for each and every engine.

For injection timing, tune AFR's via timing until they read the richest ( making NO changes to the actual fuel map itself whatsoever )
This will indicate that more of the fuel being injected is actually getting burnt.

But again, obviously this will require a full mapping session at various rpm and loads if possible.


As for fitting EGT's though. Innovate's TC-4 is a relatively cheap way of achieving it. Each unit can read four thermocouples, so daisychain two of em for an 8cyl.

8 Widebands would be nice....but that's getting into a lot of work and cost !!

Hi Steve,
the MS can tweak each cylinders fueling individually as well as the timing (although timing would be harder to get right than fuel)

I tried tuning the fuel angle for the richest AFR, but it didnt seem to make any difference, well I couldnt see it . In theory its supposed to finish squirting just as the valve opens, depending on air speed. I simply tweeked it untill the idle was at its smoothest, which wasnt far off the theory in the end.

I have the Innovates TC4 you mention, thats whats going on it over the weekend, will be very interesting to see what I get.

Ive gone back to the original plenum for now, sold the siemese. I have blended trumpet base, so this could be an issue, I will try a normal trumpet base on it when I get the chance to see if that makes any difference.

As for the injectors, I tested them using the MS, firing them 2000 times at various PWs in order to find the dead time for each one and the flow rates. Altough my measuring jug wasnt the best thing in the world, Im happy to think they were within 2-3% of each other.

One thing I found out, I was sure I had 320cc injectors, turns out they were 250cc, I now remember selling the 320s and fitting 250s Must be getting old

Phil

[stevieturbo]

Topic on Motec about it

http://www.motec.com.au/forum/viewtopic.php?f=39&t=769

[daxtojeiro]

OK, findings so far, using EGTs, backup the lazer infa red detector, just this makes it more accurate:

Cy1 needs 8%
Cy2 needs 7%
Cy3 needs -3%
Cy4 needs 12.7%
Cy5 needs 12%
Cy6 needs 9%
Cy7 needs 7%
Cy8 needs -1%



Now all EGTs are around 800C at idle and within 10-20C of each other.

And it is sooo smooth now its untrue, I could balance a pound coin on the plenum, with a fair amount of bluetack

Can anyone explain the differences?

Have yet to do any EGTs at higher revs, but these all stayed close to eachother with the same trim in the table up to 2500RPM from Idle

Phil[/url]