E38 Economy techniques

[macca_779]

Ok guys I'm after some tips from people that have got some good ideas to improve economy on E38 ECM equiped cars. From what I can see and the little experience I have with tuning these computers (only tuned 1 E38 car), there isn't much available for us to tweak. Things like DFCO which are only touched on in the cal. I don't see anywhere where I can alter TPS and mainly MAP activate point. Also is there a planed update to include a lean cruise or some other way to alter cruising AFR's, in a semi open loop fasion. I know I could do a work around by forcing OL and tricking the ECM to going lean while cruising, but I don't really want to go down this route. So far the only way I've been able to improve economy on these cars is to optimise timing. It works but there is better gains to be had if I could alter fueling.

[The Alchemist]

I sympathise, I feel your pain macca :)
I've done 2 E38's with another 2 booked in this week. Having good results with timing, as you do, but fuel is another story. What I have found is that the MAF low is slightly lean resulting in positive fuel trims. Following the logic with the LS1's regarding fuel trims applied to full load fueling, I trim the low MAF until I get slightly negative fuel trims. Not sure if this actually helps though. Perhaps someone else would care to comment.

Mike

[macca_779]

I noticed the same trends with positive trims. I scaled both MAF tables up 1% to improve it slightly as it was consistently hoping around 3% LTFT. It didn't get it down to zero but as it was a street tune I didn't want to screw with the tables too much without a wideband on board. Plus I was hoping economy wouldn't get worse even with more accurate MAF tables.

[GMPX]

I don't know the answers for you guys, but it's a bit dissapointing they have gone from mid 12L/100 (LS1) to 14L/100 (L76).

Cheers,
Ross

[The Alchemist]

I was rescaling the MAF today on a Caprice and found that even when the MAF HZ stayed constant the LTFT would change from 0.8% to 3.9% on a MAP cross over point at 45KPA while cruising at 1100rpm ( Auto).
I guess the VE constants are having an effect as well.....food for thought?
Also was doing full load runs (knock sensors turned off by setting the "disable rpm" to 1000rpm)with knock phones and a tech watching the exhuast for detonation and tuned the full load timing. Those knock sensors can now be detuned to register real knock.
This particular car was getting about 11 to 12 l/100km on the instant readout at 100km/hr.
Mike

[macca_779]

I was rescaling the MAF today on a Caprice and found that even when the MAF HZ stayed constant the LTFT would change from 0.8% to 3.9% on a MAP cross over point at 45KPA while cruising at 1100rpm ( Auto).
I guess the VE constants are having an effect as well.....food for thought?
Also was doing full load runs with knock phones and a tech watching the exhuast for detonation and tuned the full load timing. Those knock sensors can now be detuned to register real knock.
This particular car was getting about 11 to 12 l/100km on the instant readout at 100km/hr.
Mike

Don't suppose it would be too much to ask to get a look at those knock settings would it. I've de-sensitised mine by scaling up the tables to be pretty much a factor of 8 where they were a factor of 2 which follows on from the lower mg cells (BTW what the hell does mg stand for), this was under the advice given to me from a mate. Alas I have no real data or long term testing to back this up, so something more solid would be great.

[The Alchemist]

basically I set up a log to scan all the knock sensors individually, rpm, TPS, load in g/s , knock retard and load up the car either on the dyno or on the road after full load tuning has been completed and I am satisfied that the engine is not knocking. When the scan tool logs knock you check the knock data to see which sensor is playing up. I have found that 1 or 2 sensors are always a bit twitchy which pulls timing for the whole engine ruining on road performance. The way I interpret those 8 knock tables is as follows:
Say the knock occurs at a load of 0.5g/s at 3200rpm on knock sensor "c" I go to the "c" knock map at the closest rpm to 3200 and go to the load point of 500mg/s. I'm assuming this is meant to be 0.5g/s. Now say this sensitivity is set at 2.399, I have found that tweaking this to as little as 2.6 or 2.8 is enough to shut the offending sensor up and stop false knock. Its a bit of a process as you can have noisey knock sensors at different rpms or loads and I find that once you've got rid of most of the false knock this way you get a much better on road throttle response and torque which is what you want the customer to notice. I'm on to my 4th E38 so far and the results have all been the same although the knock sensors that need tweaking vary from car to car, this process seems to work well.
The Caprice gained a solid 15rwkw and about 7% torque across the entire rpm range It was bog standard with no alterations. An amazing car to drive!!!

[macca_779]

Thanks for the tip mate. I'll have a go at working on that. Appears that I may have detuned them a bit too much going by your experiences. Although in saying that, I've found that they still remain quite active when I tried my first spark map in one. They certainly don't like anywhere near the kind of timing that has worked well for me on LS1's.

[The Alchemist]

well I have found that they like 2 or 3 degrees more timing than the typical LS1 engine at all rpm points during a full load run, infact at 6200rpm I run 26 to 27 degrees and the engine loves it. With an LS1 22 or 23 degrees is the limit on our fuel at 6200rpm. Of course the knock sensors at this rpm go nuts regardless if its knocking or not which is why I check first then adjust knock sensors later. Works a treat.


Mike

[oztracktuning]

It would be good to be able to somehow change the AFRs at cruise to something in the 15s. I played with the 02 voltage switch points today, I sent them right down to 0.2V and it didnt make any difference.

Are there any other possibilities?

[macca_779]

One idea I had with this Steve is to go OLSD and fudge the Virtual VE to force it leaner. Its something I'd do on my own car if I owned an E38 but haven't had the chance to try it on one yet. Or even SD for that matter. Idealy I don't like this approach but alas we have no alternate as yet that I know of

[oztracktuning]

I had the E38 yesterday in open loop. The problem is balancing any drift with weather condition variations in open loop. If it was my own car i could get enough logging to rely on it. Maybe if the owner had a wideband installed permanently it would be OK.

[ringram]

well I have found that they like 2 or 3 degrees more timing than the typical LS1 engine at all rpm points during a full load run, infact at 6200rpm I run 26 to 27 degrees and the engine loves it. With an LS1 22 or 23 degrees is the limit on our fuel at 6200rpm.
Mike

On this topic I understand the combustion chambers have been designed to require more timing so that emissions are reduced. I guess more time for completing burn.

[ringram]

Hey Oz, if you check out a NB graph you will see it sticks near stoich until almost 0 volts or 1v when it swings right away.
http://www.ztechz.net/sitebuildercontent/sitebuilderpictures/narrowbando2.jpg
Maybe you need to aim even lower?
Also what about adjusting the O2 trim rates so that if its over it swings faster to the lean side and when its lean it moves only really slowly towards the rich side of the switchpoint?

Also are you guys saying there is no longer a lean cruise mode?
If so I suspect this is because leaner mixtures produce larger amounts of NOx emissions and given the VE has worse economy than the earlier models I wouldnt be surprised if they did remove lean cruise altogether.

Ross do you know if this is the case?

[macca_779]

Ringram Holden hasn't used lean cruise since the VZ LS1. It was still available in the cal though and enabled easily. Unfortunentely with the new ECM's there is no provision for lean cruise.

[ringram]

Ok makes sense. I bet its NOx emissions.
So fuel economy is unimportant. Which reinforces my personal view that nobody should care about CO2, just the toxic polutants..
Im sure open loop can come to the rescue.

[Highlander]

any prgress??

[swingtan]

Hi Steve,
I mentioned the O2 switch points in another thread a fair while back. I can't make that table work no matter how much I change the switch points. I gave up on it in the end .

To get a "pseudo" lean cruise, I've tried running OL and found it works pretty well. You can command a leaner mixture in the light cruising cells and richen up again where it's needed. I'm going to try a long trip later this week to see how it goes.

Simon.

[The Alchemist]

Hey Swingtan, how did that work out for you ?
Mike

[gmh308]

Hey Swingtan, how did that work out for you ?
Mike

Ditto ^^^

[oztracktuning]

Its working well for mine. The 6L Twin Turbo made it 500miles on a tank of Gas. It averaged about 31mpg for a whole tank cruising on our Sydney to Melbourne "Hume Highway".

[swingtan]

Hmmm... where's the thread miner smiley.....

Well, since my last post there has been a lot of changes made. I did the "log trip" but I was towing a big trailer the whole way. From memory I was getting around 13L/100KM but given the weight of the trailer, I was running in PE mode quite often. Remaining in an OLSD tune did help a bit though.

When we got back from holidays, I put in a 220/224 @ 114 cam with 0.550" lift. After tuning it up and looking at a few different issues, I ended up turning the MAF back on and tried OL-MAF for a while. This saw the best economy figures I've ever seen in my car with 8.6L/100km

http://farm4.static.flickr.com/3462/3398392156_29def37d1b_o.jpg

I did have an issue with leaning out when coming to a stop so I turned on the O2's and am now running back with a CL-MAF tune. It's been very good since then though I'm not sure I'd repeat the above figures.

I've been talking to Jezza about the O2 switch point thing though. We believe that there is a table / parameter missing that has a sanity check on the O2 switch points. I've tried altering mine to below 100mV and saw no change in average O2 voltage when in CL. Maybe this is something the developers might like to look at....... In the mean time, I'd be tempted to remain in CL around town and then go OL if I do a trip.

Simon.

[oztracktuning]

My data from 30hrs of travelling Sydney to port campbell return and throw Heathcote in is that the E38 has incredibly stable AFRs when in open loop in all weather conditions and altitudes. Mine is operating nicely and giving very similar afrs in the bulk of cells every time its logged. Plus that is the race tune as well when on pump gas.

[ringram]

MAF or MAFLESS Ozt?

My SDOL tune is a little wibbly, but Im running 70lb injectors! So she goes a bit rich when transitioning around idle.

Im hitting the dyno tomorrow with the excel sheets and intend to go SDCL long term. (Russo otrcai)

[swingtan]

I found that you could get near perfect AFR's for constant or slowly changing loads when running OLSD. However, getting the dynamics correct for a "single" tune, is impossible ( running a separate "daily tune" and "race" tune you can get an acceptable compromise ). The problem is that the dynamics in the E38 are all calibrated back to MAF airflow. If you turn off the MAF, the ECM can no longer adjust the dynamics referenced to the air flow. So you can really only just set it to max and hope to adjust the other parameters to give an acceptable performance. This is really noticeable in a manual, between gear changes, an auto will not be effected as much. Because of the dynamics issues, you will never get a really smooth WB AFR reading when running OLSD, if the throttle changes, the dynamics will throw off the AFR's

[gmh308]

Hmmm... where's the thread miner smiley.....

Well, since my last post there has been a lot of changes made. I did the "log trip" but I was towing a big trailer the whole way. From memory I was getting around 13L/100KM but given the weight of the trailer, I was running in PE mode quite often. Remaining in an OLSD tune did help a bit though.

When we got back from holidays, I put in a 220/224 @ 114 cam with 0.550" lift. After tuning it up and looking at a few different issues, I ended up turning the MAF back on and tried OL-MAF for a while. This saw the best economy figures I've ever seen in my car with 8.6L/100km

I did have an issue with leaning out when coming to a stop so I turned on the O2's and am now running back with a CL-MAF tune. It's been very good since then though I'm not sure I'd repeat the above figures.

I've been talking to Jezza about the O2 switch point thing though. We believe that there is a table / parameter missing that has a sanity check on the O2 switch points. I've tried altering mine to below 100mV and saw no change in average O2 voltage when in CL. Maybe this is something the developers might like to look at....... In the mean time, I'd be tempted to remain in CL around town and then go OL if I do a trip.

Simon.

Excellent! 8.6L/100. Thats almost 33mpg Imp, 28mpg US.

What AFR to get this ? :)

[JezzaB]

I feel there is a "Lean Limit" table missing in the E38s. There is one in the E40 computers.

Jez

[swingtan]

Excellent! 8.6L/100. Thats almost 33mpg Imp, 28mpg US.

What AFR to get this ? :)

Over the 30km trip I did this test on, I averaged 15.3:1 and spark averaged 40'


I feel there is a "Lean Limit" table missing in the E38s. There is one in the E40 computers.

Jez

Exactly right. That and the table to set up the "Air flow modes" to fine tune the O2 switch points.

[gmh308]

Over the 30km trip I did this test on, I averaged 15.3:1 and spark averaged 40'



Exactly right. That and the table to set up the "Air flow modes" to fine tune the O2 switch points.

Was the 30km trip undulating or flattish?

Isn't the air flow mode and O2 data already in B1515/16?

Cheers.

[swingtan]

the trip was down the SE Freeway here in Melbourne, so not particularly hilly.

Map of route (http://maps.google.com/maps?f=d&source=s_d&saddr=Ferntree+Gully+Rd&daddr=Brandon+Park+Dr+to:Magid+Ave,+Wheelers+Hill+ Victoria+3150,+Australia+to:-38.079312,145.43375&hl=en&geocode=FQ6jvf0d4P6mCA%3BFbOZvf0dXgSnCA%3BFeKVvf0d R_-mCClTCnPJZxXWajE5YMZ7nONddA%3B&mra=pe&mrcr=1,2&mrsp=3&sz=13&sll=-38.074041,145.458126&sspn=0.067837,0.147972&ie=UTF8&ll=-37.99021,145.291443&spn=0.27166,0.852814&t=p&z=11)

The tables B1515 and B1516 are for referencing the O2 switch points to the air flow mode. There are no tables to define what the ariflow mode actually refers to. If you check an LS1 tune, there is a table ( B4107 ) that defines what each mode actually is in real air flow ( though it's called CL Mode ). The E38 does not have a similar table, it's pretty safe to say there are still undiscovered tables in the E38 tunes.

Simon

[Highlander]

changing the switchpoints... does little to nothing on the E38s... can you show your differences by changing this?

[swingtan]

Correct, we know it makes no difference, hence the mention of the "O2 switch point sanity check" pid that we are missing. If we can get access to the PID, then we may be able to make some changes for lean cruise CL operations.

Simon

[hymey]

I would probably try setting up like steve does. Raise OL tables and match them together and change stoich aswell increase these to something around 15.3:1 this is what the lean limitation is in the e38 its the OL tables then try reducing the switch point to around 100mv and see what happens then.

[Ira]

I know this has nothing to do with EFILive, but I've written all the embedded code to make a proper digital dash gauge for an LC1 and adding an adjustable narrow band out where you could set the switch point on the fly would be a rather trivial addition. Conceptually, that would allow lean cruise on any vehicle as it would let you change what the ECM sees as 14.7. The issue might be how it messes with trims, but other than that it ought to work. I offered to do it for some of the hydrogen generator increases mileage folks last year, but never heard back.

Ira

[swingtan]

Bringing this up again...... because I have a new PB for economy. Last week I went as a parent helper on my daughters year 6 camp. So I thought I'd see what I could get on a bit of a drive. Starting from Yarragon and driving about 20km to just past Moe (http://maps.google.com/maps?f=d&source=s_d&saddr=Yarragon+VIC,+Australia&daddr=Princes+Fwy&hl=en&geocode=FRIOuf0dJNW0CCnHwgMuYaIpazF66x_BwIkitA%3BF X8quf0ddC24CA&mra=dme&mrcr=0&mrsp=1&sz=12&sll=-38.193095,146.16678&sspn=0.139501,0.295258&ie=UTF8&ll=-38.198799,146.176186&spn=0.13949,0.295258&t=h&z=12), I got the following.

http://farm3.static.flickr.com/2771/4373991003_3a8e46b314_o.jpg (http://www.flickr.com/photos/swingtan/sets/72157608442344351/)

After that I hit some road works and a couple of bigger towns so that figure rose to 8.7l/100km when averaged over 150km. But I'll still claim that 8.1 for my record ;)

Back in town this week I've managed the following....

http://farm5.static.flickr.com/4068/4390205035_28d8925eab_o.jpg (http://www.flickr.com/photos/swingtan/sets/72157608442344351/)

That's after 3/4 of a tank, driving 15km to and from work for 5 days and a trip across town to Thornbury (http://maps.google.com/maps?f=d&source=s_d&saddr=Glenfern+Rd&daddr=-37.854796,145.170593+to:Victoria+Rd,+Thornbury+Vic toria+3071,+Australia&hl=en&geocode=Fcawvf0dXgSpCA%3B%3BFcHWv_0du7mkCCnHUmzhe0 TWajGGXTsjAXIb8Q&mra=dpe&mrcr=0&mrsp=1&sz=11&via=1&sll=-37.823887,145.154114&sspn=0.280412,0.590515&ie=UTF8&ll=-37.823345,145.171967&spn=0.280414,0.590515&t=h&z=11) last night. I'm pretty happy given I can still do 0-100kmh in under 5 seconds. All this with the MAF and CL still enabled.

Simon.

[gmh308]

Bringing this up again...... because I have a new PB for economy. Last week I went as a parent helper on my daughters year 6 camp. So I thought I'd see what I could get on a bit of a drive. Starting from Yarragon and driving about 20km to just past Moe (http://maps.google.com/maps?f=d&source=s_d&saddr=Yarragon+VIC,+Australia&daddr=Princes+Fwy&hl=en&geocode=FRIOuf0dJNW0CCnHwgMuYaIpazF66x_BwIkitA%3BF X8quf0ddC24CA&mra=dme&mrcr=0&mrsp=1&sz=12&sll=-38.193095,146.16678&sspn=0.139501,0.295258&ie=UTF8&ll=-38.198799,146.176186&spn=0.13949,0.295258&t=h&z=12), I got the following.

http://farm3.static.flickr.com/2771/4373991003_3a8e46b314_o.jpg (http://www.flickr.com/photos/swingtan/sets/72157608442344351/)

After that I hit some road works and a couple of bigger towns so that figure rose to 8.7l/100km when averaged over 150km. But I'll still claim that 8.1 for my record ;)

Back in town this week I've managed the following....

http://farm5.static.flickr.com/4068/4390205035_28d8925eab_o.jpg (http://www.flickr.com/photos/swingtan/sets/72157608442344351/)

That's after 3/4 of a tank, driving 15km to and from work for 5 days and a trip across town to Thornbury (http://maps.google.com/maps?f=d&source=s_d&saddr=Glenfern+Rd&daddr=-37.854796,145.170593+to:Victoria+Rd,+Thornbury+Vic toria+3071,+Australia&hl=en&geocode=Fcawvf0dXgSpCA%3B%3BFcHWv_0du7mkCCnHUmzhe0 TWajGGXTsjAXIb8Q&mra=dpe&mrcr=0&mrsp=1&sz=11&via=1&sll=-37.823887,145.154114&sspn=0.280412,0.590515&ie=UTF8&ll=-37.823345,145.171967&spn=0.280414,0.590515&t=h&z=11) last night. I'm pretty happy given I can still do 0-100kmh in under 5 seconds. All this with the MAF and CL still enabled.

Simon.

13.8 around town! Very impressive for commuter traffic.

[swingtan]

To be fair, I think the trip computer does some funny stuff. I get the feeling that....


The average speed is calculated over the time since the last reset of the tirp computer.
The average economy is a moving window that averages figures over something like the past 30min.


Before the trip across town, I was sitting on 14.7l/100km which was just the trips to / from work. The trip to Thornbury was done in 34'C heat though with the aircon running the whole way.

Still, I'm pretty happy with those figures.

Simon.

[stigmundfreud]

all just having a quick look at a tune on a friends cam'd car which is suffering from economy beyond what is expected from just the cam.

Without a wb fitted there is little to present log wise but noticed that the switch point for b1515 is set to 600mv rather than 450mv (which is what it was set to before cam). Now the tune shop that fitted the cam have been fiddling around as the car was stalling a lot but looking at a side by side tune before and after cam there is little difference (other than slight maf table, spark and kr burst adjustments). Not convinced the tune is anywhere near as optimum as could be. Tune is basically flashed, then feedback (no logs) given back to Aus, update tune sent etc and repeat.

Now having seen reports that adjusting the switch point has no effect, why would hsv themselves be adjusting b1515 to 600mv? It will surely be used elsewhere for calculating feedback from the o2's.

The tune was flashed with hp tuners but was based on the stock o/s, are there any issues in reading a tune from hp tuners like this?

[swingtan]

HSV and/or HPT may have the switch point calibration access. Going from memory, the O2 switch point is comprised of...


B1515 / B1516 : O2 Sensor Rich/Lean Threshold
An "Airflow Mode" table that does not yet exist for the E38 (it would be something like B4107 in the LS1 controler).
A hard limit setting for the minimum lean value in the tune (not available in the E38 calibrations that I have seen). Setting the stoich setting to say 15.5:1 seems to have no effect on the changing of the O2 switch points.


That being said, I've looked at some log data and my stock NB-O2's are generating



771mV @ 14.5:1
107mV @ 14.9:1
36mV @ 15.5:1
23mV at 16.3:1


So the "switch" is extremely fast on these O2's. Chances are that even setting the switch point at 36mV will do very little as the O2's themselves will be the limiting factor. Due to their switching characteristics, the average AFR they will end up returning is probably still very close to the 14.7:1 std.

the HSV's running switch points of around 600mV is probably an attempt to run marginally richer when cruising. This might be for power, exhaust note or even a smoother running motor. It's probably not goign to make a huge difference though.

Simon

[stigmundfreud]

I'd imagine its part of the cold stall issues he was having - the tune looks pretty stock with no changes other than to timing and knock burst. I'd have thought a more dialed in ve table of some sort but nothing. It even has an osram otrcai so there is more potential with a decent tune, car runs good but the fuel use seems excessive more than seen on other similar cam setups (with more customised tune).

Think the best bet for him is to look at borrowing my scanner and talking to the likes of oztrack etc. Thanks for the pointers too, I take it hp have a few more tables open to them than we do?

[swingtan]

If it's an E38, then check the idle tuning thread. there is a huge amount of info there.