Instructions to change fuel maps on 14CUX Griffith, Chimaera

Instructions to change fuel maps on 14CUX Griffith, Chimaera

Author
Discussion

blitzracing

6,387 posts

219 months

Monday 1st September 2014
quotequote all
rangerovering said:
My own car has a 3.5 converted to serp and 14cux, currently with a 1992 3.9 Range Rover ECU. The idle control is not brilliant, when starting the engine fires then almost instantly dies. Second turn it fires and manages to gain control and the rpm is higher for a few seconds then drops down as per it should do.

Cheers
Use RoverGauge to check the stepper position when cold, or physically remove it and check its fully wound back- its normally wound back until it hits the end stops and you should hear a buzz as it does this when the ignition goes OFF. It should then wind down to say 30% to maintain a stable idle once the engine is started- so it sounds like it might not be fully retracting to start with, so when you start the motor its starting with reduced airflow until the stepper is opened up a bit.

rangerovering

8 posts

114 months

Tuesday 2nd September 2014
quotequote all
danbourassa said:
rangerovering said:
The Land Rover Discovery used a 3.5 14CUX engine for a short while so I intend to get a look at the map in that chip ASAP.
I wasn't aware of this and I would love a copy of the binary when you get it.

rangerovering said:
Its a 5AM AFM and the temp sensor is new genuine Land Rover.
I wasn't suggesting that the temp sensor was bad. Chilling it is just a convenient way to enrich the mixture for trouble shooting purposes.

Keep us posted on your progress.
Ah my mistake, I will see what I can do re temp sensor.

Yes I will get a copy of that code up when I manage to track an ECU down, they are not uncommon but there are not many left.

Incidently this morning she started and held idle fine, it was moving in slow traffic where it huted up and down a bit and dropped quite low.

stevesprint

Original Poster:

1,114 posts

178 months

Wednesday 17th September 2014
quotequote all
As it's now the first anniversary of this thread and we can now do most things an engine tuner would hope for with the 14CUX it seems a good time to publish much simpler instructions with as little computer jargon as possible.

I’ve listed the 14CUX tuning options now available with the simpler instructions than I first published a year ago to update the tune chip. You no longer have to understand the complexities of the chip or correct the checksum as the free program called TunerPro does it all for you, but you’ll need a Windows laptop, RoverGauge cable and a chip read/writer for about £30. In fact that's all you need if you just want to copy a chip to convert a Land Rover ECU to TVR spec, use my modified tune chip files or make simple changes like idle speed, rpm limiter etc.

I’ve eliminated my slow speed shunting by tweaking the top row of the fuel table and merging Land Rover’s final program code called ‘Operation Pride’ with my Precat fuel data. Plus I’ve also resolved the fuelling issues above 5,000 rpm and enabled my MIL lamp. You can download my modified tune files for standard Griff Precats with Land Rover’s improved program code with or without the extended fuel table to 6200rpm. I’ve already had positive feedback from one owner who said “I’m impressed version R3652 is much smoother and is what I’ve always been looking for”.

I’ve also described the gadgets and procedure to self remap on your local rolling road or from data logs created while driving safely on the road.

If like me you are determined to remap the 14CUX yourself or use Land Rover’s improved code then please feel free to use my instructions and downloads at http://www.stevesprint.com/remap-14cux

Should you have any questions or suggestions to improve the instructions then please do not hesitate to contact me, but please remember I’m not in the motor trade or computer industry and I’m doing this for fun in my limited spare time.

I’m sorry I can’t offer any tuning advice on individual modifications as re-mapping the fuel table is still best left to the engine tuning experts like Mark Adams and Jools (Spitfire4v8). At the very least I hope this 14CUX effort will help reduce the number of 14CUX casualties like Hoover.

Good Luck, Steve Sprint

Edited by stevesprint on Thursday 25th September 23:03

TV8

3,118 posts

174 months

Wednesday 17th September 2014
quotequote all
Great work Steve.

I know there are lots of people who have contributed in many ways, but your summary takes me back to when I bought my car in the second half of 2013 and realised the map was wrong! This thread was very timely for me and much appreciated although it was Jools who did the work!

davep

1,141 posts

283 months

Sunday 19th October 2014
quotequote all
Steve, I'm really looking forward to the imminent resumption of all things 14CUX; I wonder if there's been any developments on the other side of the pond!

danbourassa

246 posts

136 months

Tuesday 21st October 2014
quotequote all
Dave, there's been plenty going on here, although almost none of it has been 14CUX related so I'm also looking forward to things picking up again. The closest thing to 14CUX has been Colin's diagnostic software for the Rover MEMS engine management as used on his Mini with single point fuel injection. This is another use for the FTDI serial interface cable. Only this time, no line inversion or resistor needed. And the connector is still available!

http://alum.wpi.edu/~colinb/mems_interface.html

As for 14CUX, it was nice to end on a high note with Steve's discovery of the overrun value before we took a break for the summer. Hopefully, there are more discoveries yet to come.

PeteGriff

1,262 posts

156 months

Tuesday 21st October 2014
quotequote all
stevesprint said:
As it's now the first anniversary of this thread and we can now do most things an engine tuner would hope for with the 14CUX it seems a good time to publish much simpler instructions with as little computer jargon as possible.

I’ve listed the 14CUX tuning options now available with the simpler instructions than I first published a year ago to update the tune chip. You no longer have to understand the complexities of the chip or correct the checksum as the free program called TunerPro does it all for you, but you’ll need a Windows laptop, RoverGauge cable and a chip read/writer for about £30. In fact that's all you need if you just want to copy a chip to convert a Land Rover ECU to TVR spec, use my modified tune chip files or make simple changes like idle speed, rpm limiter etc.

I’ve eliminated my slow speed shunting by tweaking the top row of the fuel table and merging Land Rover’s final program code called ‘Operation Pride’ with my Precat fuel data. Plus I’ve also resolved the fuelling issues above 5,000 rpm and enabled my MIL lamp. You can download my modified tune files for standard Griff Precats with Land Rover’s improved program code with or without the extended fuel table to 6200rpm. I’ve already had positive feedback from one owner who said “I’m impressed version R3652 is much smoother and is what I’ve always been looking for”.

I’ve also described the gadgets and procedure to self remap on your local rolling road or from data logs created while driving safely on the road.

If like me you are determined to remap the 14CUX yourself or use Land Rover’s improved code then please feel free to use my instructions and downloads at http://www.stevesprint.com/remap-14cux

Should you have any questions or suggestions to improve the instructions then please do not hesitate to contact me, but please remember I’m not in the motor trade or computer industry and I’m doing this for fun in my limited spare time.

I’m sorry I can’t offer any tuning advice on individual modifications as re-mapping the fuel table is still best left to the engine tuning experts like Mark Adams and Jools (Spitfire4v8). At the very least I hope this 14CUX effort will help reduce the number of 14CUX casualties like Hoover.

Good Luck, Steve Sprint

Edited by stevesprint on Thursday 25th September 23:03
Having been hooked by this thread from the beginning and having several long chats with Steve (SteveSprint), I am now running his latest R3652_430_6200 on a spare 14CUX on my Griff 4.3. What a lovely map. The low down fuelling glitches have gone along with the shunting in 4th at low speeds! I can now drive in 4th at 20mph and accelerate without the shunting and hesitation. As my interest is now 'piqued' I am keen to monitor the new map and report accordingly. Keep up the good and valuable work guys. Regards, Pete

stevesprint

Original Poster:

1,114 posts

178 months

Thursday 23rd October 2014
quotequote all
davep said:
Steve, I'm really looking forward to the imminent resumption of all things 14CUX; I wonder if there's been any developments on the other side of the pond!
Thanks and well done for breaking the ice, I've since been asked offline if we can apply the same smoothness improvements to a 500.

I already have a winter wish list of 'nice to have' enhancements, but it seems more important we turn our attention to improving the smoothness and shunting on the cat cars. Although I’m currently busy fixing all the parts that broke on my Griff during the summer I'll always make time to knock up a Griff 500 prom based Land Rovers final code to see if it helps.

PeteGriff said:
Having been hooked by this thread from the beginning and having several long chats with Steve (SteveSprint), I am now running his latest R3652_430_6200 on a spare 14CUX on my Griff 4.3. What a lovely map. The low down fuelling glitches have gone along with the shunting in 4th at low speeds! I can now drive in 4th at 20mph and accelerate without the shunting and hesitation. As my interest is now 'piqued' I am keen to monitor the new map and report accordingly. Keep up the good and valuable work guys. Regards, Pete
Thanks everyone for all the positive feedback, it's amazing, much appreciated and spurs me on to working on the 500 versions. Let's not forget it was a team effort, so may I forward your thanks the whole team.

It’s actually helpful having the positive feedback as it confirms we are making good progress and that it’s safe for Dan to start dropping the older revisions to make further developments for Dan less hard.
Thanks again
Steve

stevesprint

Original Poster:

1,114 posts

178 months

Monday 27th October 2014
quotequote all
Jools,
Due to popular demand I’ve knocked up a Griffith 500 tune chip for a standard 500 to run with or without lambda control. It’s based on Land Rover’s final firmware called Operation Pride and I’ve copied the standard 500 fuel data in to the CAT and non CAT maps. I’ve also stopped the overrun fuel cut below 4,500rpms so you have the options to lean off the overrun mixture to induce some overrun snap crackle and pops. I was pleasantly surprised the overrun popping worked on my 430 even with lambda control. You can see from the RoverGauge logs the ECU runs open loop above 3,400rpm and then takes 3 seconds to hit 100% lambda trim.

As the 500 cat and non-cat maps run ok-ish on my 430 I’ll email you a copy for further testing and it will be interesting to see if it helps reduce the shunting.
Cheers, Steve.

spitfire4v8

3,990 posts

180 months

Tuesday 28th October 2014
quotequote all
Great to see you refining things further steve. I must admit i've not had a lucas car in for ages on which to try any of your recent changes on. you're well ahead of the game compared to what's happening here !

Podie

46,630 posts

274 months

Tuesday 28th October 2014
quotequote all
spitfire4v8 said:
Great to see you refining things further steve. I must admit i've not had a lucas car in for ages on which to try any of your recent changes on. you're well ahead of the game compared to what's happening here !
wavey

Let us know if you want the Griff back. smile

Still tempted to see if the injectors make a difference!

FlipFlopGriff

7,144 posts

246 months

Tuesday 28th October 2014
quotequote all
Podie said:
wavey

Let us know if you want the Griff back. smile

Still tempted to see if the injectors make a difference!
Which injectors did you put in?
FFG

Podie

46,630 posts

274 months

Tuesday 28th October 2014
quotequote all
FlipFlopGriff said:
Podie said:
wavey

Let us know if you want the Griff back. smile

Still tempted to see if the injectors make a difference!
Which injectors did you put in?
FFG
None yet! Joo had to peg it back a bit on the standard injectors as it was running too lean at the top end.

stevesprint

Original Poster:

1,114 posts

178 months

Wednesday 29th October 2014
quotequote all
Mark and I have proved the CO trim effects the AFR up to 3,000+ rpm when running the non-cat map (similar to lambda control) and also when switching from non-cat to cat map the Lambda long term trim is reset according to the CO trim value as follows:

CO Trim 0.05 volts



CO Trim 1.25 volts



Co Trim 2.5 volts



Furthermore we also know the CO trim and Lambda long term trim are stored in the same memory address which proves the CO trim is adjusting the lambda long term trim.

Dave
Please can you help me understand, in the code, how the Lambda long term trim/CO trim is taken into account when calculating the injector pulse width when running the non-cat map. I’m sure you realise I’m ultimately trying to prove copying a cat map to a non-cat map is an effective way to turn off automatic Lambda control and instead have manual lambda control.


Dan,
It would be good to hear from you if you are around but David did say he wanted something to get his teeth into. smile
Cheers for now, Steve

Edited by stevesprint on Wednesday 29th October 20:55

danbourassa

246 posts

136 months

Wednesday 29th October 2014
quotequote all
Steve, normally you should not be able to control the long term Lambda trim this way. In open loop, software measures the CO trim voltage and stores the same value in both memory locations, however nothing is done with the values. In closed loop, the CO trim voltage is not measured since the channel 9 measurement is missing from the ADC control list. This frees up the two memory locations for use as long term Lambda trim. In closed loop, software calculates the long term odd/even trim values independently and stores them in the two memory locations.

If you are able to control the long term Lambda trim this way, it may be that you have a mixed set of data in your fuel map data structure, such as, closed loop map with open loop ADC table.

blitzracing

6,387 posts

219 months

Thursday 30th October 2014
quotequote all
stevesprint said:
Mark and I have proved the CO trim effects the AFR up to 3,000+ rpm when running the non-cat map (similar to lambda control) and also when switching from non-cat to cat map the Lambda long term trim is reset according to the CO trim value as follows:

CO Trim 0.05 volts



CO Trim 1.25 volts



Co Trim 2.5 volts



Furthermore we also know the CO trim and Lambda long term trim are stored in the same memory address which proves the CO trim is adjusting the lambda long term trim.

Dave
Please can you help me understand, in the code, how the Lambda long term trim/CO trim is taken into account when calculating the injector pulse width when running the non-cat map. I’m sure you realise I’m ultimately trying to prove copying a cat map to a non-cat map is an effective way to turn off automatic Lambda control and instead have manual lambda control.


Dan,
It would be good to hear from you if you are around but David did say he wanted something to get his teeth into. smile
Cheers for now, Steve

Edited by stevesprint on Wednesday 29th October 20:55
There is one PH,er who said this happened on his car, but Ive not managed to reproduce it so far at least- As this wont need to fire the engine up for this, I will double check however.

Dan or Steveon a completely different bit, I was thinking about the expected values shown by the AFM on RoverGauge for the Lucas 20 AM or Bosch. Now as the voltages are different, the direct readings will be different, but would the Linear values be nearer what you would expect for the 5AM- ie would the direct to linear calculations work out correctly for different AFMs?- I cant see this happening as the response curves are not the same. I assume therefore that when remapping for the bigger AFM simply involves getting the multiplication factor to best suit full power, and then simply skewing the map to compensate for the different response?

blitzracing

6,387 posts

219 months

Thursday 30th October 2014
quotequote all
danbourassa said:
In open loop, software measures the CO trim voltage and stores the same value in both memory locations, however nothing is done with the values.
That cant be the case, as it directly affects the mixture at low RPM?

danbourassa

246 posts

136 months

Thursday 30th October 2014
quotequote all
blitzracing said:
That cant be the case, as it directly affects the mixture at low RPM?
Sorry, I should have said "nothing is done with the values by software".

danbourassa

246 posts

136 months

Thursday 30th October 2014
quotequote all
blitzracing said:
Dan or Steveon a completely different bit, I was thinking about the expected values shown by the AFM on RoverGauge for the Lucas 20 AM or Bosch. Now as the voltages are different, the direct readings will be different, but would the Linear values be nearer what you would expect for the 5AM- ie would the direct to linear calculations work out correctly for different AFMs?- I cant see this happening as the response curves are not the same. I assume therefore that when remapping for the bigger AFM simply involves getting the multiplication factor to best suit full power, and then simply skewing the map to compensate for the different response?
The calculation to linearize the AFM is complicated but it's also important to get right since software uses the linearized value and not the direct value. A few months ago I bought a used Sagem 20AM to investigate the difference between the 5AM and the 20AM. My intent is to connect them both in series and run varying amounts of air through them while taking readings from both. If the differences are minor, I may be able to tweek the existing equation to get it close enough for the 20AM. Otherwise, there are math tools on the Internet to create best-fit equations for curves. Colin has used them so I might hand the data over to him to get the new equation.

But I may have a better way to handle this. I'm convinced that it's just through luck or accident that the 14CUX seems to function at TVR's 6000 plus RPM redlines. I have been looking for ways to save processing time in the code and since math is not this processor's strength, this is a good area to improve. Converting the linearization function from a calculation to a look-up-table will help. This is a trade off between processing time and memory use. There is not enough unused ROM memory to use a LUT for the whole 10-bit range, but there IS enough to use a LUT for the top half. And that is when the RPMs are highest.

Hoover.

5,988 posts

241 months

Thursday 30th October 2014
quotequote all
Still all over my head...BUT both Mark and Steve have really got the bit between their teeth, and their enthusiasm is over whelming.... had a good chat with both at open day last month and they were like two kids in willy wonkas chocolate factory,,,,, grinning from ear to ear wanting, to talk n share all that they know smile.... keep the work up ((albeit no good to me at present hehe)