Tuning Rb25S For Eflex And E85 .

[discopotato03]

This is breaking away from the thread I started on E85 RB25 vs RB30 .

For those that didn't see it I changed my GTS25T over to Caltex Eflex recently and it was straightforward because I was sort of set up for it .

Basically I have the typical PFC Z32 MAF Datalogit 740s slightly bigger pump and a Tech Edge wide band already .

Previously my PFC had very basic tuning when I ran BP98 and the car did very few miles . When the 740s went it the injector trims were set ta 48% and whatever the lag time was - don't remember off hand . SK and I did a rough road tune to get by on which was sort of Ok for a time .

As mentioned elsewhere the fuel fuel change needed 7 feet of 3/8 EFI hose and pumping the 98 out into fuel containers . The first "start tune" consisted of raising the fuel trims to 62.4% which is 30% more than the 48 setting used with these 740s for PULP . It started up and ran well enough to drive around for a day .

I started altering the PFC fuel and ignition tables to see what gave the most throttle responsive drivability and was surprised to find the around idle bit had been a tad lean . I made some alterations to the cold start coolant temp and cranking enrichment values and I'm not sure if it starts better from stone cold - its not that hard to start cold as it was .

Before changing fuels Gary noticed my 02 probe feedback was not selected so it was ticked - remember this bit .

Tonight I got around to setting up my Tech Edge wide band to see what the mixtures looked like . I use the TE LCD readout which lets you select AFR or Lambda and I use Lambda now because I think it makes more sense .

Anyhow at hot idle the result was around 0.94 to 0.96 which I though may be a smidge rich but not too bad . Driving around the reading was more like a reasonably constant 0.85 which I though was strange because it should have been moving around a bit more with more or less throttle . I stopped a few times to lean out the light load round town revs areas and all that did was lose torque and drivability . It then occured to me that I should turn off the oxygen sensor feed back so that I'd get a better indication of alterations I was making and what numbers the wide band showed when it felt good - or better anyway . MAJOR difference , I was able to easily get the cruise mixtures more like 0.95 Lambda and it felt better everywhere . The mixtures moved round at varying loads more like I think they should .

Mt inexperience with a PFC is showing here so can someone tell me if they are not set up to deal with E70-85 as far as closed loop mixture triming goes ? Is it possible to change settings to make it work in closed loop with suitable E70-85 cruise mixtures ?

I know some people here bag the humble Power FC and if I was starting with a factory ECU I wouldn't buy one now , it is what I own and if I can make it work for a roadie less to spend . I too was recently keen to have a flex fuel setup but to be honest I don't think I want to go back to petrol and Caltex Eflex is available from three servos not too far away . Gary and I looked online and if I had to drive to Melbourne its aavailable in Goulburn and a few places between there and Melbourne . I'd carry a jerry just in case but I could also stick a petrol tune back in it to get to the next Eflex location .

I would be interested to hear the sorts of Lambda numbers people use that work in RB25DETs though high load tuning should really be done on a dyno with an experienced operator IMO - controlled conditions .

I'd also like to hear about ethanol content sensors because knowing whats in your fuel system gives you the ability to "trim" the ethanol content if necessary to try to keep it consistant . The sensor Holden are using on the late Eco Nizi Door looks interesting and looks like it has 8 or 10mm barbs straight through it , different style to the Siemens "tugboat" one .

Exhaust temp probes are also not a bad idea and I'd like to know knoe about whats available and where poeple are mounting their probes .

Also speaking of probes Bosch has had a later LSU 4.9 02 sensor out for a while which is supposed to be more accurate reliable and heat up faster from cold . I want to talk to Tech Edge about setting up my wide band to use them .

Eventually I want Insight to tune my car properly but I'd like to lean a bit more about getting my car in a reasonable state myself . If it starts easily drives nicely and gets acceptable consumption i'll be happy - then get it dynoed .

Anyhow if others can share their E70 tuning findings others may get on the band wagon a bit more easily .

Cheers A .

[zebra]

Did you change the TE Wb settings for ethanol or leave it pre configured (petrol)

I have aTE 2J1 and from memory you can change that stuff

But if tuning by lambda works than that is good to know

[paulr33]

the pfc (and stock ecu) assume you are using normal 98 penis fuel

in datalogit you can set from memory the target af/r or co-efficient for the closed loop operation

you will need to adjust this if using different fuel

its cos its using the 98 penis fuel target settings and trying to aim for that

[paulr33]

as in its trying to aim for the ideal a/fr for std 98 penis fuel

each fuel has its own ideal stoichmetric

no idea what it is for E85 but just enter this value into the PFC and you should be ok

[paulr33]

[GTSBoy]

The oxy sensor has no idea what fuel you are using and has no care for what the stoich air fuel ratio is. It just senses lambda, and the ECUs just work on Lambda. Any time you see an AFR coming out of a wideband or an ECU, it is just lambda multiplied by some factor, and it completely arbitrary, and it makes no difference. If your actual stoich AFR is 1000:1 (to pick a bullshit number out of the air) and the ECU/wideband thinks it is 14.7, then at lambda 1 it will tell you that your AFR is 14.7, when it is really 1000. Simply doesn't matter, and shouldn't bother you unless you cannot think in terms of petrol AFRs (which would be daft, because until recently that's all any of us have thought in.)

My take on what is happening in the report above is that your PFC appears to be trying to run lambda 0.85 in closed loop, which is crap. Either the oxy sensor is not working properly or the ECU is acting strange. It should, of course, be aiming for lambda 1 as an average.

Another though is that you cannot aim for any lambda other than 1 (the above behaviour notwithstanding) so if you want to tune for leaner cruise mixtures (than 1) then you have to disable feedback. That is unless you adjust the wideband's output to have an offset to fool the ECU into fueling closed loop to a leaner target.

[paulr33]

yeah or maybe thats the go

the 02 closed loop is for out of the box tunes

if you can be bothered tuning the fuel map on light cruise at 80kmh/ and 100 and 110kmh for example then youll probably yield better fuel economy than the std 02 closed loop defaults. i think from memory on my 33 i had afr's of around 15 at light cruise or even 16 and with the std 02 closed loop it would alwyas around around 14 or just under and stay there

[zebra]

Ahhh that makes sense...kinda

to used to thinking Gasoline AFR's and not lambda

so just to clarify, Lambda of 1 on any fuel is the cleanest (ala full) of the fuel?

So, will E-flex make best power at Lambda 1 or something like 0.85? or richer again?

[GTSBoy]

No, you can't even think of it like that. Lambda 1 just means that exactly the correct amount of fuel to burn with the induction air is being injected. That means, no excess air, and no extra fuel. Stock engine management aims for lambda 1 (14.7:1) on petrol because that is realistically the only AFR that has been possible to aim for (given how narrow band oxy sensors actually operate). Until recently that is, where more and more cars have a lean cruise mode that may be using a wide range sensor, or just more computational power being used against a narrow band. Doesn't matter for this discussion anyway.

You can't run lambda 1 on load, because with petrol it tends to result in excessively high temperatures in the combustion chamber. That leads to detonation. So your choices are to either run less fuel, or more fuel, because either will lead to lower combustion temperatures. But engines are limited by how much air they can breathe. So if you add less fuel, you end up making more power. So that is the main reason why it is much more common to aim for richer mixtures. (the other reason is that oxides of nitrogen increase at leaner mixtures, and NOx is harder to get rid of than CO). With petrol in a heavily loaded engine (as in turbod) then you need to have petrol mixtures down in the 12:1 or less range to control temperatures and detonation.

Ethanol is much more resistant to detonation, so you will be able to run leaner a mix with less risk of damage, or you can take advantage of the extra margin of safety by running more timing, or more boost. It is probably smarter to use a bit from column A and a bit from columns B & C as well, as there are limits to how much advance is actually useful, how much boost you can realistically make from any turbo and how much extra temperature you can accept from leaning the mixture.

So the answer is no, there is no "best AFR" to run. There is a good range and that range is likely somewhat wider than on petrol. For cruise, it still makes sense to run lambda 1 (or leaner if you can) and on power you're still going to have to aim for maybe 0.85 or a little bit higher. For reference, on petrol, lambda 0.85 is about 12.5:1, which is a little too lean for hard working engines.

[zebra]

I know that, I'm talking purely about e-flex/e85

So let me re-phrase

Because e85 needs to run much richer than petrol

Would lambda 1 in the sense of e85 be that same as lambda 1 on petrol and how does the o2 sensor know you are on ethanol, which is why I asked if disco changed the settings on the tech edge to suit the ethanol fuel

[GTSBoy]

No, still thinking wrong.

E85, E70, Ewhatever, do NOT need to run richer than petrol. You need to use more fuel (measured by volume or mass, doesn't matter), but you don't run it richer. Richness is only relative to lambda 1.

I'll stop this here, and write another post that might take longer, and I migth get interrupted (helping kids with homework right now).

[zebra]

Ah, my bad writing richer, I did mean volume (drinking and watching movies ATM lol)

So tune as per usual (using lambda) but will need more injector to get the mixtures correct

So I confused myself for no reason

[GTSBoy]

Petrol. Stoichiometric air-fuel ratio is 14.7:1 on a mass basis. You need to inject fuel so that there is 14.7 times as much air going into the engine as fuel in order to reach lambda 1. If you inject extra fuel, so that the ratio is 12:1, then you are running at lambda 0.81. You can consider that to be a rich mixture.

E85. Let's say that the Stoich AFR is 9.7:1 (I haven't looked it up, just run with that number for the moment - it doesn't matter). Again, that's on a mass basis. So, to hit lambda 1 you need to inject fuel so that there is 9.7 times as much air as fuel. That means you are injecting a lot more fuel, but the mixture is not richer than for petrol. The mixture is still exactly lambda 1. 100% the right amount of fuel for the right amount of air. If you run at lambda 0.81 then you'll be running at an AFR of 7.9:1. That's rich. That's exactly as rich as when running petrol at 12:1. There is no difference.

There may be differences in how the engine responds to running at the same lambda value between the two fuels. You might ping if you try to run hard on petrol at 13:1. But on E85, it might just run withotu pinging (and it will therefore make more power as a result too).

Get it?

-edit. Yup, you go it.

Edited August 25, 2012 by GTSBoy

[zebra]

Yeah got it

Was over thinking it

[discopotato03]

My 02 feedback injector correction is set up the same as in Pauls pic above ie 1.047 whatever that means - assumes 1.047 Lambda .

My probe is the LSU 4.2 pn 0 258 007 200 and has not seen a lot of use . I think once or twice I started up in the past on petrol without it being up to temp but otherwise it should be ok . It came out of the dump last night and is very clean from Eflex compared to Ultimate 98 .

Anyhow the plan is to talk to TechEdge about converting to the more current LSU 4.9 probe and I may end up down their way next week if they can make it happen quickly . These later probes are supposed to operate differently and heat up faster than earlier ones like mine .

Something I should have mentioned was that my dump pipe only has provision one probe so the PFC with feedback on would have been looking for something that wasn't connected while the Tech edge was operating .

Anyway with feedback off the TE shows sort of what I'm expecting to see even though I've no way of knowing if my probe is good . Something definately changed when the feedback was disabled and the Lambda numbers indicated this .

Also have to reset the LCD screen so it wakes up in Lambda rather than AFR , easy when you know how .

Might get out on the F6 in a while where I can do 110 up some reasonable hills and see how the upper middle load range mixtures are .

Need to research EGT gauges/probes and look into those Kmon knock sensor gadgets but those were pretty exy from memory .

More later , cheers A .

[discopotato03]

The expressway Lambda readings were still around the 0.95 dipping to 9 and at times 8.5 on boost .

I'm really not sure if these are accurate but driving it feels fine , if anything smoother than on BP98 .

I think some of my low speed drivability issues could be a combination of the IAC valve trying to overcompensate for very small throttle openings . ATM I have it set at 850 revs , was 950 to allow for the lightened flywheel but it can probably come down to 700-750 if that cuts some of the IAC motor agro . Also I suspect the throttle plate may not be closing fully always or the TPS is a bit screwy when the throttle is shut . At least the closed throttle fuel cut off seems to work because the mixture reading goes off the scale then just says lean . My display is the small LCD backlit one like I think Guilt Toy had when he was here .

My TechEdge is a WB3A2 and to reset it it for a Bosch LSU 4.9 probe it needs an internal jumper change and I think a firmware reflash . I'll talk to TechEdge tomorrow and hopefully get down to them and get it sorted sometime this week . Its pointless trying to tune with a wideband that tells lies . TechEdges site shows two LSU 4.9 probes and they are basically the same except for the cables lengths .

My 4.2 probe is what they call the Bosch 7200 (0 258 007 200) and their 4.9 is a 7123 or 7025 (0 258 017 123 / 0 258 017 025) .

By the way what sort of ignition timing are people using with their RB25s ? I'm mainly interested in the light to medium load ranges because I think its riskiy tuning higher load ranges without a chassis dyno - for all concerned inc the engine .

I'm not using anything over 40 and thats at reasonably light loads , all my high load ranges are set to 25 to be safe with limited intercooling and tuning inexperience . Boost is whatever a GTRSs actuator is and from memory the Apexi elec valve was set to 12 or 13 pounds back in the ceramic turbine days .

Cheers A .

Edited August 26, 2012 by discopotato03

[rob82]

That variable is to tell the PFC when to go open loop. Ie when the values in your fuel map are over 1.047 it goes open loop and multiplies the theoretical pulse width by 1.047....

All of what gtsboy has said about lambda is correct.... I wouldn't usually run too much extra light load timing obviously under load you can add extra ignition advance. From 98 to e85 on a knock limited engine I've found an extra 8 degrees to be the magic number but will probably be less on a rb25.

Also I've had no issues running closed loop mixtures on e85. You sometime need to retard your injection event due to the increased vaporization. Unfortunately you don't have access to injection phasing with a PFC.

Edited August 26, 2012 by rob82

[discopotato03]

Hmm injector settings over the set value 1.047 , will look into that .

Edit - looked at that thread Jessie did at Celica.org and he reckons PFC feedback works so poorly that he leaves it off permanently .

I reset my idle speed to 700 and have less agro at very low speeds , however the tacho doesn't always show ~ 700 so with have to look at cleaning the throttlebody and find out how to check the TPS .

The low fuel light is just starting to show when pulling up so my first full tank of Eflex has got me almost 400 Km . My PFCs maps are better than when I started tuning with this stuff so with luck the next tankfull should go a bit further . The best I think I did on 98 was about 420 but the level never got down to where it is now - never saw the fuel light before ...

I must grab one of those ethanol content test tube things soon , cheers A .

Edited August 26, 2012 by discopotato03

[wolverine]

2.44am?!?!

I can just picture Disco driving a late night train with one hand on the smart phone posting at SAU thinking about tuning E70-85.

[discopotato03]

Actually I was not long home from a trip on the rails and a smart phone I don't have . The laptop and 3G internet plug in go out on the tracks often .

Been in touch with TechEdge and they need my controller box and cable to switch to the later Bosch LSU 4.9 oxygen sensor . The box gets a jumper change and firmware reflash and I'd say the later sensors cable uses a different plug socket .

I'm going to change fuel filters today just so I know it won't be an issue in the near future and I'll open up the used one to see iff emptying the tank pumped through any crud .

Wheres everyone getting their ethanol "test tubes" , cheers A .