Live mapping 850CSi

[John in DC]

Thanks for the info John, very interesting. Was your car an Auto or manual ?.

My twin turbo car was an automatic (slushbox) but I've seen the same 4200 rpm dip on dyno sheets from V12 E31's from all over (including the 6 Speeds) so definitely not an isolated incident. Witness MattSimis's dyno sheet here: http://forums.bimmerforums.com/forum...3#post25157533

Regarding the EML bug, I originally suspected that it was related to a faulty temperature sensor but the problem persisted even after swapping in a new one. My current thoughts are that the EML code that triggers WOT depends on a temp value that is too moderate and that changing the reference trigger temp would likely fix it.

I saw above that you've disassembled the EML binary and are tracing through it so perhaps you'll see something there that could be the culprit.

Rgds...

[Omega man 1969]

I looked at Mattsimms sheet and decided the car had slipped on the rollers.....it sure looks like that's what caused the dip. Interesting that you have seen the same dip on other cars.....I know Auto always give huge inflated torque curves as the torque converter does it stuff.....you can only compare the last part of an auto box dyno curve to a manual one for that reason.

I will look into the temperature maps and see what I can find.

[John in DC]

I looked at Mattsimms sheet and decided the car had slipped on the rollers.....it sure looks like that's what caused the dip. Interesting that you have seen the same dip on other cars.....

Definitely didn't slip the rollers.

I've seen this phenomenon during dyno pulls on numerous different E31's over the years. Sometimes it's subtle. Sometimes it's pronounced. Example: The prototype 6-Speed used for development of our supercharger kits showed the same dip back in 2009...

http://www.itsystems.net/PersonalPic...rcoDyno01a.JPG

MarcoDyno01a.JPG

As I said, the datalogging sessions I did confirm that it's EML WOT signal related. Now just need to figure out why.

Your efforts will go very far into answering this.

Rgds...

- - - Updated - - -

And here is another example from just a few months back.

http://www.itsystems.net/PersonalPics/850iA/EMLBug.jpg

EMLBug.jpg

Unfortunately, this particular dyno operator used road speed interpolated to RPM (ie didn't use the tach signal) - - thus why the "dip" appears closer to 4500RPM than the typical 4200RPM that I've seen so many times before.

Rgds...

[yamoda]

Interesting stuff but the more I read about the engine management in our cars the more I want to rip the lot out and put in a modern standalone system. Can this be done? Will it affect the other speed related systems like AHK, Servotronic etc?

[Omega man 1969]

I see your point John. Look at my dyno sheet



no problem on any of the pulls I did. It's confusing, as once the pedal is down to the floor lambda is switched off at all times, meaning wot signal is present at all times .....I know this for sure as I have 2 wideband displays sat next to me.....and they lock to around 12.3:1. At all other pedal positions lambda is swinging it one way then the other to keep it CAT safe.

could it be ignition timing or mixture related ? We're they running standard factory maps?

- - - Updated - - -

Interesting stuff but the more I read about the engine management in our cars the more I want to rip the lot out and put in a modern standalone system. Can this be done? Will it affect the other speed related systems like AHK, Servotronic etc?

Don't do that ! :-)) this setup is excellent, it has the scope to be programmed the same as any aftermarket ECU but will be a million times more reliable. Aftermarket ECU's can only dream of reliability records that could match Bosch ECU's, why more people don't pull them from old cars and reprogram them for their own modified car project is beyond me. Instead they use practically "homemade" aftermarket setup s that have a whole host of bugs/glitches. That being said, there are some very good high end aftermarket ECU's out there .....Motec being one of them.

You can disable the EML....a guy on the forum did so but using older style DK's from an e32 which has mixture screws and turning off the EML feedback in the DME software. Other than some mild gains in the mid range I can't see any need to do that.

Once I figure out if road speed is related to throttle opening later today then that can be disabled if need be.

- - - Updated - - -

Definitely didn't slip the rollers.

I've seen this phenomenon during dyno pulls on numerous different E31's over the years. Sometimes it's subtle. Sometimes it's pronounced. Example: The prototype 6-Speed used for development of our supercharger kits showed the same dip back in 2009...

http://www.itsystems.net/PersonalPic...rcoDyno01a.JPG

MarcoDyno01a.JPG



http://www.itsystems.net/PersonalPics/850iA/EMLBug.jpg

EMLBug.jpg
...

If I hadn't seen my dyno sheet I would say a possibility would be a harmonic pulse In the intake manifold at a given RPM, I've seen that on a few cars. On the 2nd shot is the AFR correct ? 10.8:1 for the entire pull? Surely not ?

[John in DC]

could it be ignition timing or mixture related ?

Not ignition timing or mixture related. Here's why I think this: 1) If you connect a datalogger to the WOT signal wires that run from the EML to the DME's, you will see that the "dip" occurs at the exact moment that the WOT signal latch to ground occurs. This means that the phenomenon is caused by the EML; and 2) The EML has a small number of inputs - - RPM, pedal angle, temp, etc - - none of which is timing or mixture related.

Were they running standard factory maps?

I've seen this phenomenon on '91-'93 normally aspirated cars, Dinan TT cars, and several of my customer's cars alike - - all of which were running stock EML. The DME chips don't seem to matter as the aforementioned had a mix of stock BMW or tuner-supplied DME fuel/spark maps. IIRC, BMW corrected the bug in late '93 / early '94 timeframe. There were some whispers about this on Roadfly 8-10 years ago but the details escape me.

Don't do that ! :-)) this setup is excellent, it has the scope to be programmed the same as any aftermarket ECU but will be a million times more reliable.

Agreed. The stock Bosch M1.7 engine management will do most everything you need. Leave it alone.

Once I figure out if road speed is related to throttle opening later today then that can be disabled if need be.

Curious to know what your investigation reveals.

If I hadn't seen my dyno sheet I would say a possibility would be a harmonic pulse In the intake manifold at a given RPM, I've seen that on a few cars.

I considered that also but the datalogging we did made it clear: The dip occurs at the moment that the WOT signal is sent from the EML to the DME's. To confirm this on my twin turbo car, I built a circuit that forced WOT signal to the DME's independently of the signal from the EML. Result is that the 4200RPM dip went away.

On the 2nd shot is the AFR correct ? 10.8:1 for the entire pull? Surely not ?

Pls disregard as AFR wasn't connected for that run.

[Omega man 1969]

Well that answers all my questions :-) thanks! It would be nice to have a collection of EML files to compare.....a 91 and a 95 EML file. I could compare them and see what changes were made by BMW. I could run some more dyno pulls with different EML files to see if the "dip" happens, only takes a second with the emulator to swap files.

I have a collection of BMW files but unfortunately no date on them, I guess the hardware/software number would be an indication of date it's just knowing where and when it changes.

[Fatandre]

So much tech talk. At first it was confusing, but now I am starting to understand some of it. Threads like this are like gold to the community.
Wish I could give something from me to this thread

[John in DC]

Well that answers all my questions :-) thanks! It would be nice to have a collection of EML files to compare.....a 91 and a 95 EML file. I could compare them and see what changes were made by BMW.

You've got mail.

I could run some more dyno pulls with different EML files to see if the "dip" happens, only takes a second with the emulator to swap files.

No need to do this on the dyno. Just connect a datalogger to capture RPM and the WOT signal(s) from the EML (pin#5 for cyl 1-6 or pin #33 for cyl 7-12) and run it down the street. It will be very clear if the "dip" is there because it won't latch to ground until 4200 RPM's.

Rgds...

[Omega man 1969]

Ok will do, alternatively I could just watch my 02 displays.

I tried to do some more testing today but the EML gods weren't with me :-) first I couldn't find a long enough USB cable to allow my laptop to connect to the EML and sit on the passenger seat, then when I finally found one the EML started acting up......the car was revving on its own and the idle was stuck around 2000 rpm....the only way I managed to get home was by not using the accelerator pedal and using the brake to stop the revs....luckily the brake switch signal to the EML made the revs drop to idle .....it was a tricky drive :-)


At first I though it must have been related to me removing the pedal linkage and changing the stop point but after hooking up DIS I could see the pedal degrees at idle shifting by itself.......tuner pro confirmed this as I could see the data trace marker not on any of the idle (0 value) cells in the throttle table.



I removed the EML and pulled it apart, I thought maybe electrolytic problems but they all checked out ok...then I found the culprit .....water on the underside of the PCB, it cleaned up ok so I will try again tomorrow.

[Fatandre]

How did water get there?

[Omega man 1969]

How did water get there?

Andre I hate you for asking that question :-) I washed the car this morning and the cover was off the ECU box......I'm guessing that was how it got in there . It was unlucky.....or maybe lucky that the moisture was across the pedal sensor pin 7 and some other signal pins.

I repaired Ola's EML while I had my soldering iron out, I've fitted my CSi chip in it to test it out.

[Fatandre]

Well, there is a short distance between hate and love so I guess you are close to loving me

[Omega man 1969]

Well, there is a short distance between hate and love so I guess you are close to loving me

Is that what you say to all those Icelandic beauties ?

The EML is back and behaving itself now. Hopefully things will go better tomorrow.

[Fatandre]

Doesnt matter how cheesy it is as long as it works

Good that you got it all sorted out.

[Max Lumens]

IIRC someone once solved the 4200 RPM problem using a MAP sensor to activate ground when you stomp the pedal and vacuum drops???

[John in DC]

IIRC someone once solved the 4200 RPM problem using a MAP sensor to activate ground when you stomp the pedal and vacuum drops???

Yup. That was me. But the circuit was designed to work under boost conditions only (so wouldn't be applicable for normally aspirated cars).

Rgds...

[Omega man 1969]

Ok I'm sat in the car after an hour drive. As far as my 94 CSi is concerned WOT can be achieved regardless of road speed in any gear. I couldn't find any map in the EML that moved in relation to road speed...only RPM and pedal angle. I can see from my O2 gauges that it goes to WOT map as soon as I press the pedal to the floor.

John has kindly given me some EML files to look at, I will see if I can see any differences in the earlier software.

I'm glad to say my EML is working perfectly again :-)

- - - Updated - - -

How about your old CSi exhaust system ? ;-) BMW quoted £7,400 for a new one !!!

So much tech talk. At first it was confusing, but now I am starting to understand some of it. Threads like this are like gold to the community.
Wish I could give something from me to this thread

[Fatandre]

- - - Updated - - -

How about your old CSi exhaust system ? ;-) BMW quoted £7,400 for a new one !!!

My CSI had a custom one all ready. The UK weather had done its thing to the OEM one.

[Omega man 1969]

John, I know how you yanks have been emmisons mad for many years......do you think BMW could have tweaked the EML binary on the US cars ? Keeping you guys lambda controlled for as long as possible.

[John in DC]

I guess it's possible but given that mattsimis' Euro car exhibited the same phenomenon I would say it's not probable.

[Max Lumens]

Yup. That was me. But the circuit was designed to work under boost conditions only (so wouldn't be applicable for normally aspirated cars).

Rgds...

Maybe there's a market opportunity for you to make one for the stock setup, non FI??? Especially if it turns out to affect early US cars, which are the most abundant.

[Omega man 1969]

Maybe there's a market opportunity for you to make one for the stock setup, non FI??? Especially if it turns out to affect early US cars, which are the most abundant.

Easy enough to stick a later EML chip in an early model car. Chips are cheap enough and it only takes a few seconds to program one.

- - - Updated - - -

I guess it's possible but given that mattsimis' Euro car exhibited the same phenomenon I would say it's not probable.

Hmmm I forgot about Mattsims car..........I have a copy of his DME files but unfortunately not the EML. That would have been perfect as I can easily compare his CSi EML file to my own and pick up the differences straight away. It's not so easy with the Auto file as the code is different in places and offset.

I guess it could be poorly adjusted pedals only allowing the throttle map to reach a certain row when floored.....that way it would need to reach 4200rpm before the pedal map had a WOT value in its cell. I would imagine there is a constant for the WOT pedal value somewhere in the code....its possible that was changed in later software to allow for more tolerance in pedal values being on the low side. I tested my car and it switches to WOT at 3 volts.

- - - Updated - - -

Actually it is possible. On several screens in DIS you can configure your own set of data to display (select reading and click right arrow, repeat for other readings and finally confirm with another right arrow click). Here's the PWG potentiometer angle and voltage and the VL signal in one screen:



... and just a fraction further:



The voltage jumps from 2.90 to 2.92 - there's no value in between (I used a screw mechanism to adjust the pedal in a controlled an accurate way). To conclude; I guess it's safe to assume the VL pedal angle is hardcoded to approximately 60° for a manual S70 (and probably M70 as well).

Revtor I will upload an auto file to my car while I have DIS hooked up today and see at what voltage WOT is enabled.......John was kind enough to let me have an old "dip present" EML file so I will see if that is any different.

Thanks for the DIS screen tip....I didn't know that was possible.

- - - Updated - - -

Ok looking at the part numbers. On the EML case you will see some stickers.

Larger black sticker with Bosch written on it is the Bosch hardware number.

White sticker with a 7 digit number at the top is the BMW hardware number

On the EPROM inside there is a 10 digit number, this is the software number.


These numbers are stored in the EPROM, on the file John sent to me the software number was:

228735 7375 (has the 4200rpm problem)

228735 5924 ( is from Ola's 850 AUTO ) This is an European car....the software is almost identical apart from a few lines of code and is suffixed by F31 at the end of the software......Johns has F32 at the end.

[John in DC]

Maybe there's a market opportunity for you to make one for the stock setup, non FI??? Especially if it turns out to affect early US cars, which are the most abundant.

Swapping in an EML chip without the bug would be easier/cheaper/faster.

[Omega man 1969]

ok I loaded the file that John gave me into my EML



here is the sport program switch screen.....the sport mode button on the CSi operates this as expected.



Even at 3.5 volts 75.5 degrees VL signal was still off!



VL (WOT) signal switch on at 3.56 volts 75.97 degrees, also activating the kickdown feature.







So it would seem that pedal adjustment is crucial in obtaining WOT....I tried several other EML files including an Alpina file and they all behaved in the same way. The Bentley manual says Automatic transmission pedal at full throttle 3.7-3.8 volts and manual transmission 3.1-3.3 volts.