Live mapping 850CSi

[John in DC]

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.

What is the production date of mattsimis' CSi vs yours? We should try to get a copy of his EML so we can perform a HexDiff against yours. Would be very curious/telling if they were identical, no? If that were the case, it would probably point us back to something input related - - which doesn't leave much as there aren't that many inputs to the EML. Here is an excerpt from the ETM that contains EML pin assignments:

ETM92_EML_Pin_Assignments.pdf
http://www.itsystems.net/PersonalPic...ssignments.pdf

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.

Very interesting theory re: constant for the WOT pedal value somewhere in the code. Does your analysis of the code suggest this or you just going on a hunch?

Rgds...

[Omega man 1969]

I'm not sure of the year, mine is a 94. Thanks for the link John....interesting the Traction uses a 100hz square wave signal to reduce throttle angle.....that could be handy for you......fit some knock sensors and trigger a simple circuit to inject 100HZ into the EML...fool proof protection :-)

Re the WOT value.....As the manual and Auto expect different voltages as WOT but have the same pedal/rpm map layout there must be another map somewhere holding that value. I did find a 9 x 1 map that was active in sport mode....all its cells have the value 182.....I'm beginning to think this may be the one. I will change the values and see if the WOT switch angle changes with it. Basically the 9 cells related to engine RPM and I think the value of that cell may tell the EML what value to accept as WOT.

[John in DC]

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.

Good stuff but I suggest that we should compare apples to apples and stick to 6-speed EML's only for now. I think the second EML file I sent you was from a 6-Speed. Can you confirm?

I'm not sure of the year, mine is a 94. Thanks for the link John....interesting the Traction uses a 100hz square wave signal to reduce throttle angle.....that could be handy for you......fit some knock sensors and trigger a simple circuit to inject 100HZ into the EML...fool proof protection :-)

Clever !


Rgds...

[Omega man 1969]

Could be onto something here, a shot of my CSi 9 x1 map that I mentioned.
CSI PEDAL CONSTANT .jpg

Now a shot of the same map that I found in the 850 auto file you gave me. It was a different location but I found it searching the descriptor.

850 AUTO WOT CONSTANT .jpg

I reckon that is why you can't get WOT until 4200 RPM.

Note...I haven't filled in the RPM values on those charts but 4200RPM is where the auto map hits 182 :-)

John the second file you sent was also an Auto EML file.

[chadderz2010]

If I can help out in anyway just shout me ( PM ) in South Yorkshire UK 850CSI 1993 CC88004.

Regards
Chadderz

[Omega man 1969]

If I can help out in anyway just shout me ( PM ) in South Yorkshire UK 850CSI 1993 CC88004.

Regards
Chadderz

A copy of your EML file would be most helpful, thanks for the offer.

I plan on sticking together a 3 chip package for the CSi, 1 tweaked pedal response EML chip and 2 x tuned DME chips.......really just the price of the chips and a couple of pounds to cover the cost of the kit I've used.....Probably around the £50 for the set sounds fair.

[Fatandre]

Omega. It sounds good.
Your presence here to the community is amazing and just this thread has opened my eyes.

[chadderz2010]

A copy of your EML file would be most helpful, thanks for the offer.

I plan on sticking together a 3 chip package for the CSi, 1 tweaked pedal response EML chip and 2 x tuned DME chips.......really just the price of the chips and a couple of pounds to cover the cost of the kit I've used.....Probably around the £50 for the set sounds fair.

On its way just a soon as I work out how to do it and as for the kit sounds more than fair.

Chad

[Omega man 1969]

I'm still working on this but I'm having problems with the checksum. If anyone reading this knows how to increase the checksum value if all the free space is filled with "FFFFFFFF"'s then please chime in. It's easy to reduce the value of the blank space if the table values are increased but if I want to set a table to all zeros the I can't see a way to increase the free space to keep the checksum the same.

If all else fails I will have to find the checksum routine in the code and remove it. I'm thinking that DIS will report a bad checksum and give an error code.....if I can find that error code within the EProm then it may point me in the right direction.

[Omega man 1969]

Right, the sun is out and time to get back to sorting the car for the summer. Thanks to 8tech for supplying me with a complete set of air ducting I can now fit the new CSi nose. I'll post some pictures once completed.

[Fatandre]

Finally. I have been waiting all winter :-P

[fstenb]

Well this was usally done simple as an adress n + (n+1) sum. However, when you pass sum FFFFFFFF the counter will overflow and set n+1 to the rest value of the integer sum. If you max a counter to FF and add 1 the result will be 1,0000,0000.

ok, i'll try. a 2 bit register could have 00,01,10,11 ->0,1,2,3. a binary addition with (01+11)=00 beqause the result is 100,(4). the 1 will be an overflow bit outside the register. Convert all your hex into binary and add them together. Don't add the register with the checksum. The checksum could be 1,2 or 4 registers, depending on the hardware.
Sorry can't be more precise unless i have the hardware in front of me. Maybe a photo could help. Not into ripping my own out.

// Freddy

[Omega man 1969]

Well this was usally done simple as an adress n + (n+1) sum. However, when you pass sum FFFFFFFF the counter will overflow and set n+1 to the rest value of the integer sum. If you max a counter to FFFF and add 1 the result will be 1+overflow. // Freddy

Could you clarify this please.

- - - Updated - - -

On its way just a soon as I work out how to do it and as for the kit sounds more than fair.

Chad

On a side note, Chadderz2010 read his EML Eprom for me and surprisingly it was identical to my EML, I was expecting it to at least have its own serial number like the DME's have.

- - - Updated - - -



One freshly painted CSi spoiler .

[Fatandre]

This one is going to look spanking on your car. Put it on and I demand a photo session!!

[Grush]

Could you clarify this please.

- - - Updated - - -

On a side note, Chadderz2010 read his EML Eprom for me and surprisingly it was identical to my EML, I was expecting it to at least have its own serial number like the DME's have.

- - - Updated - - -



One freshly painted CSi spoiler .

Did you order it new from BMW or removed the original and send it to be repainted? I`m looking for a P/N for this part, new. Realoem just gives me the main upper bumper P/N, and I have no idea if it comes with the spoiler or not.

[Omega man 1969]

Yes I ordered it from BMW UK. Cost was around £550 including VAT. To have it Painted was another £130

[Omega man 1969]

Thats more like it.

[Fatandre]

Yeah. Looks good now. We need more.
From the sides and so on

[Omega man 1969]

- - - Updated - - -





- - - Updated - - -

[Fatandre]

Looks ace Omega!!

[Omega man 1969]

Thanks....It wasn't easy putting it all back together. The most difficult part was sliding the bumper back on and not having the rubber seals on the side of the front wings move out of position, luckily the Earl was home to give me a hand. The parts 8tech supplied cleaned up like new, I fitted the underside tray back on as well just to make sure it was back to how it should be.

[Fatandre]

So since you got that sorted whats the plan?

[Omega man 1969]

[Fatandre]

Dont know if you dont need the OEM side skirts

[Omega man 1969]