When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Long story short - I've suspected I might have a gasket leak around the turbo downpipe for a while. I fitted my new ESL ECU on Saturday, before doing a 50 mile round-journey, so although I haven't done any proper logging yet, I've had a look today at the RAM tables, to see how the ECU has learned to adapt itself. I'm currently running the standard 8S map, and doubt I'll attempt any proper DIY mapping until I've been on the ESL course next month.
Anyway - looking at the Fuel Trim table below (closed loop, presumably), if I'm reading it correctly hopefully confirms my suspicion I've got a gasket leak somewhere. My thinking is that the lambda sensor is after the suspect gasket and if there is a leak, air is likely to be being drawn in, before the sensor, adding air at that point and making the ECU think it's running lean (when it's actually not) and therefore having to add more fuel via the fuel trims.
Before I start pulling the turbo off to replace the (original 17 year old) gaskets, does that sound feasible and/or is there anything else I should check/log first?
EDIT: Should probably add that engine wise, mine is a pretty standard '97 Turbo 2000. The only things that have been changed are the basics - a Cosworth panel filter, silicone intercooler Y-pipe, removed middle cat (still got standard downpipe & cat) and Mongoose back-box.
Last edited by ben.harris; 20 December 2015 at 09:31 PM.
generally an exhaust leak would be letting gasses out rather than drawing air in - though i have seen both. Most likely you have an air intake leak somewhere if the corrections are positive so get a tin of EZstart and spray around the connectors and intake piping from the MAF all the way to the throttlebody - 90% of the leaks I've seen will be in this area.
If you view the corrections as raw values how do they look? It could also be the extra airflow is due to the uprated panel filter and the silicone y pipe. Do you have a wideband to check the on boost fuelling?
That makes sense. I've tried the soapy water thing before, without finding anything. I'll get some EZStart and see what I can find. Presumably I'll hear the idle pick up if there's a leak present?
This was my first look at the ESL software with it connected to the car, so I didn't think to check the raw values, but I've still got it on the screen on my laptop, so I'll check when I get home and post the results here.
I don't have a wideband yet, but I'm going to get one and have spent the last few days looking at all the available options, trying to decide what buy. I'm liking the look of the PLX DM-6 as it has narrowband simulation and once fitted, I can combine it with a boost sensor and any other sensor I feel like at a later date without having to mess around with my interior.
ohhh very swanky, I use the MTX-L as it's a very fast update and the price was good too. AEM Uego also seems to make decent logs so any of those should fit the bill rightly. I wouldn't be too worried about the narrowband output as it rarely works as you hope it will (the ecu is calibrated for the scale the stock sensors uses not just the min/max values but the curve between them). Just stick it in where the rear o2 goes and the readings will be grand.
Soapy water is usually good for finding where air is getting out, EZstart is handy to see where air is getting in. As you say just listen for it plopping through the exhaust, I like to bring the revs up a touch to so you might need a glamorous assistant or a screwdriver to wedge the throttle body open.
Thinking about it, could a the breathing mods I've done really make as much as a 9% increase in airflow - assuming I'm reading the fuel trim table correctly? It sounds too much to me, but if it's feasible, then that could well explain it!
Mine's an early(ish) UK classic, with only one lambda sensor location up at the top of the downpipe, directly behind the turbo, hence the reason for needing the narrowband simulation. I was hoping that the narrowband simulation would allow me to get away with just replacing the stock lambda and I'd read that the PLX narrowband simulation was pretty good. If I end up needing to get another bung welded onto the downpipe to run a separate sensor, I'd want to take the opportunity to fit a sports cat downpipe at the same time, but as it stands at the moment, that would just exaggerate the current fueling trims even further!
EDIT - I should probably add that I'm simply trying to make sure that the engine is all in good health and a 'known' state, before I start having a go at tuning it properly. The IAM had increased from 8 to 16 and it hadn't registered any knock anywhere at all, so all good there.
Last edited by ben.harris; 21 December 2015 at 10:36 AM.
there's nothing to say that the stock MAF scaling was accurate to begin with though, most cars will have some amount of correction on them so if you check for leaks and find nothing then I wouldn't stress too much about it unless the numbers start to change a lot or your fuel readings are messed up.
The PLX narrowband output may be good, but you'll still have to scale it in the ECU to make sure the readings are accurate. The alternative is just to have a 2nd bung added at the top of the downpipe and keep the standard sensor where it is with the new wideband beside it.
If your not getting any knock then that's always good news though.
I think I might be getting confused... I thought narrowband was a standard measurement of the time difference based on oscillations between 0 - 0.9v, with the time difference being used as a comparison to determine if the AFR is greater than 14.7 or less than 14.7. I understood it to work as follows:
if the time spent > 0.45v is exactly the same as the time spent < 0.45v then the AFR is stoich.
if the time spent > 0.45v is greater than the time spent < 0.45v then the AFR is rich.
if the time spent < 0.45v is greater than the time spent > 0.45v then the AFR is lean.
If the above is indeed correct, how does the ECU scale for a narrowband input, and how can this be changed? I've not come across any settings in the ESL software for scaling yet. Is it not a case of being either "rich, lean, or stoich" based around time spent above/below 0.5v? The only issue I can potentially see is if the frequency of the oscillations from a simulated narrowband are too quick for the ECU to pick up on in which case I've read that a capacitor can be used to 'smooth' the signal.
Have I misunderstood how this works, or am missing something obvious!?
Hope this doesn't come across as argumentative - it's not supposed to - I'm just trying to understand how it works! I guess I'll know more once I've got the wideband connected up and properly checked for leaks!
Last edited by ben.harris; 21 December 2015 at 04:23 PM.
Reason: fixing typo!
no i get ya mr - the front sensor isn't quite narrow band it reads a good range of mixture/voltage and the ECU uses that to determine what AFR your at. The newage cars have a lookup table that compares voltage to o2 readings so if you change the sensor you have to rescale that. Even tho 0v and 5v may read the same the inbetween readings may differ by quite a bit (usually noticed by installing an ebay sensor and it reading too rich/too lean).
The narrowband sensor is usually the rear o2 which I don't even think you have - newage cars use it for confirming the cat is working as expected and the 32bit cars are very picky about it.
It's quite possible the classics have just a narrowband sensor, but I always though it was a semi wideband like the newage cars.
I'm pretty certain the classics just have a Narrowband. I found a post from Andy from ESL a while back where he said that his demo car has a permanent wideband running as a simulated narrowband, so if it's good enough for him....
I've just been out to the car and retrieved the most recent fuel trim data, both in converted and raw mode. Looks like it's chucking in a bit more fuel across quite a lot of the range now. This is after approx 100 miles of driving and the ECU adapting. As I won't get a chance to try and track down any possible leaks until the weekend, I'm going to try and get a proper log of my journey to work tomorrow to see if it highlights anything out of the ordinary.
Mine is the opposite,I have a different set-up tho,running mafless and full decat,cone filter but my fuel trims are taking small amounts of fuel out,I am running a bit rich in closed loop and switching off closed loop the AFR gauge confirms I'm running rich.
I've got a feeling my maf scaling needs a little work But don't want to mess with it yet until I full understand about re-scaling(it does look a little odd but was done by a 'pro' mapper so maybe it's how it should be?)
My AFR readings on the MTX-L i use are spot on , so im confident my fuelling is good hence never worried about the fuel trims , my MPG is good enough 26 mpg etc,so never had a reason to worry about them, maybe i should ha, i dont really understand what they mean, they start off as Red cells and at 1.00 , Some have stayed the same, some have gone green with readings of say 1.10, some go the other way, reading 98.4 in red
as i say, whats good and whats bad in the fuel trims lol ?
My fuel trims are adjusted in the same fields as yours, im just re-learning the car so mine are a bit behind but will see if i can get a screenshot
I thought they were normal to be fair lol
I'm assuming that those are the main closed loop areas. The non-adjusted areas are probably when it goes open loop, which is why there is no adjustment there. I'd be very interested to see what yours look like though. My understanding is that the values are the percentage of increase that is being applied - A cell reading 1.10 means it is delivering 110% of the fuel, meaning a 10% increase.
Interesting to hear that yours are similar values to mine. Are you running a standard map, or have you tweaked it? My finding are all based on the 8S map supplied by Andy with no changes (yet!).
Last edited by ben.harris; 21 December 2015 at 09:27 PM.
My fuel trims all at 0
Wots coarse fuel show
If use wideband 0 5v then have no closed loop fuelling
If u change the maff scale then load changes so map be out
Wots all your maff scales set as at maffless seen couple done by mapper and scaling been 6000 different
Last edited by Jaysz; 21 December 2015 at 09:36 PM.
I'm assuming that those are the main closed loop areas. The non-adjusted areas are probably when it goes open loop, which is why there is no adjustment there. I'd be very interested to see what yours look like though. My understanding is that the values are the percentage of increase that is being applied - A cell reading 1.10 means it is delivering 110% of the fuel, meaning a 10% increase.
Interesting to hear that yours are similar values to mine. Are you running a standard map, or have you tweaked it? My finding are all based on the 8S map supplied by Andy with no changes (yet!).
I didn't think the fuel trim table has any cells for open loop as the ecu doesn't adjust fueling in open loop.
Originally Posted by Jaysz
My fuel trims all at 0
Wots coarse fuel show
If use wideband 0 5v then have no closed loop fuelling
If u change the maff scale then load changes so map be out
Wots all your maff scales set as at maffless seen couple done by mapper and scaling been 6000 different
I will get some pics of my maf table Jaysz,i don't think its been rescaled correctly tho as in the logs the highest i can get the load is like 54 max,surely i should be hitting over 60s tho?
My fuel and ignition tables have been scaled to 62 engine load.
I didn't think the fuel trim table has any cells for open loop as the ecu doesn't adjust fueling in open loop.
Possibly not, but the highest cells on there are 4800 RPM and I'd be surprised if it's still in closed loop at that point. I'm only guessing though. None of this is based on any kind evidence!
Possibly not, but the highest cells on there are 4800 RPM and I'd be surprised if it's still in closed loop at that point. I'm only guessing though. None of this is based on any kind evidence!
I thought it was a combination of RPM and Engine load to when it moves to open loop,i could be wrong tho? My fuel trim table has a different RPM range,i thought they would be the same but it must be different on version and model i guess?
here's mine.
The manual shows maff scale that's maff calibration my car seems to go out of closed loop any revs without much throttle
My load scaling could also be improved makes 62 load at 56% throttle manual shows try get load 24 per bar which must mean atmos mines at 30 so much of map needs to be tweaked if change load so I left mine
The manual shows maff scale that's maff calibration my car seems to go out of closed loop any revs without much throttle
My load scaling could also be improved makes 62 load at 56% throttle manual shows try get load 24 per bar which must mean atmos mines at 30 so much of map needs to be tweaked if change load so I left mine
AH i see mate,yes so stock maf scale is 22137
and this has been changed to 14000 by the mapper.
Maybe the same mapper?
I'm just a bit worried that the maf calibration is not as it should be? It doesn't look right at 4.38v but maybe this is how it's supposed to be when mafless??
I'm not confident as there were a few odd things done by the mapper,like fitting the IAT sensor wrong and not setting any fuel or ign compensations,plus the car was detting it's **** off and ran cr@p in cruise,oh and was overboosting to fuel cut lol
Not what i would expect from a 'PRO' mapper
My maff calibration wasn't changed at all see yours has in places but not at full voltage
Tricky to know what's right when so different would think scaling should be progressive without random numbers
Manual doesn't say anything about mafless calibration for mafless
Sure u find more than that
Last edited by Jaysz; 21 December 2015 at 11:13 PM.
just checked my maf cal too and thats unchanged, maf scale is at 19000, believe its from the same tuner, at 1 bar im around 62 load, however, on my old ecu i started from scratch when i switched to maffless, and using the maf scale i aimed to get the load simular to when it was running maf, 1 bar around 42 load iirc
on another note, my fuel trims never change from the default 1.0, im running a wideband with 0-1v output, i can see the ecu making trims, but never registers on the table!
Last edited by sean turbo2000; 21 December 2015 at 11:16 PM.
Hmm strange how my maf calibration has been changed then?
Need to do some more reading.
I believe that as you're running narrowband simulation that the ecu thinks you're running 14.7 all the time in closed loop so this is why your fuel trims will never change from 1.0.
I was going to run narrowband simulation with my LC-2 but have an extra bung in the dp so no need to,in the innovate software you can calibrate the 0-1v Output
These are my current fuel trims, its still learning tho as i rese tthe ECU as im still tweaking the MAP, its not a standard map, far from it, im running bigger TMIC, cone filter, yellow injectors, 3" exhaust, newage coils, standard FPR, aftermarket boost solenoid, later spec MAP sensor, but still running standard MAF.
with a speed density setup the maf scale shouldn't be such a big deal - the ECU is reading from the fuel table RAW values so as long as your MAP sensor values are correct and the wideband is happy your good to go.
it might be the MAF scale was adjusted on ossett2k's to dial in injectors before switching to speed density mode?
sean_turbo2000 I think you have the right idea trying to get the load values back to where they where on MAF when switching to SD - certainly this is what I try and do on a newage car when switching over to speed density as it usually confirms you have the right injector flow and latency values in place. Keeping load the same means your load based ignition timing (and AVCS on STI's/newer WRX's) stays the same so less re-working of things to do for you.
I must make a note to check on the training day what the CL/OL transition values are on the ESL boards as I can't see any settings for it in the roms so it must be static somewhere. On a newage car you can set it by throttle/rpm/load/vehicle speed as well as the fuel targets so you have to balance all the values to get a decent cruise.
20 December 2015, 09:12 PM
ESL Fuel Trims - Gasket leak?
