Relearning the ECU during logging data with Freescan ?

[MarkL]

Hi,

I have adjusted the TPS and will have to swap my battery again. Voltage rarely exceeds 11.3 in the cold and drops to 7 volts when cranking but it starts the engine flawlessly at around 10 degrees Celsius.

Upon swapping the battery for safety I will reset the ECU again just as documented in the manual. I had already done this a few weeks before.

But now I would also like to hook up the new ALDL interface in the process just for curiousity.

Will this be bodging the learning cycle ?

I guess I will be logging at least half an hour starting from idling to a warmed up engine in the garage via moderate driving on the road up to some more spirited runs with little boost.

Regards

Mark

[fjmuurling]

The "learning" of the ECU willl  not be affected if you hook up the ALDL inteface with Freescan/Espritmon for monitoring. Just don't reset anything with the software (BLM/Error Codes/Idle Air Controller).

Freescan or Espritmon are great tools for showing what is going on in the ECU / Car.

[CHANGES]

17 hours ago, MarkL said:

Hi,

I have adjusted the TPS and will have to swap my battery again. Voltage rarely exceeds 11.3 in the cold and drops to 7 volts when cranking but it starts the engine flawlessly at around 10 degrees Celsius.

Upon swapping the battery for safety I will reset the ECU again just as documented in the manual. I had already done this a few weeks before.

But now I would also like to hook up the new ALDL interface in the process just for curiousity.

Will this be bodging the learning cycle ?

I guess I will be logging at least half an hour starting from idling to a warmed up engine in the garage via moderate driving on the road up to some more spirited runs with little boost.

Regards

Mark

 

Hi Mark,

During the tuning process I undertook on the 910 engine , we found deficiencies in the std TPS.    The data we downloaded showed the values changed with temperature..   What we found was the fully closed value (in our case) of 28  would increase to 31 when engine at working temp..   Because I was running stand alone ECU i could monitor this in real time.  At first i was assumed it was the TPS at fault.  As a result i purchased another, the results were worse...  Because this replacement TPS was after market i assumed that was poor quality so purchased an OEM unit.. Result no change..  I eventually had 5 brand new OEM unit to test...... The results were very poor , in fact the one I had originally fitted was the best of the bunch. some went as far as 5 points out...  

So how and why did this become such an issue for me and not so much for std Esprit. ?  

When setting up My stand alone ECU we set the fully closed at 28 .  When engine temp raised and effected the closed value the ECU assumed the throttle was open by the new value introducing fuel and tuning modes accordingly. This in turn effected the HC's  etc which the ECU also tried to compensate.  The result was a poor inconsistent tick over..   The only immediate solution was to write in an error factor to the mapping..  We set it up as throttle closed between > 28 and < 31, this gave an error factor of 3 point at any one point but addressed the problem and the engine ticked over as it should and produced the 400+ bhp..

As with all things on things on this project I was not going to leave it like that...  The main reason for this was the error factor through the throttle range ..    To memory the range was 28 to 96  witch gave a scale of 68 with a 3 point error,  this means at any one point the TPs could be 4.4% out     

 I addressed this info with an informed person who told me such an error has been programmed into the standard ECU to compensate for the discrepancies.. .    Which is why the std Esprit did not show and issues.. 

This does not change the fact of a tolerance in the set up circa 5% error on tuning values because of the TPS  .   So we eventually fitted a contactless  TPS which gave NO  fluctuations in value at all...  The results were instant, you could hear and feel the difference at tick over, with a far smoother progression during revving..   The tuning data feed back improved dramatically and so completed that part of the package..  

The reason for detailing this is so you can understand when adjusting your std TPS what you are dealing with and that some may be worse than others, a minor adjustment may be more significant than you would realise..  The tick over on some Esprit 4 pots is better than others although they are all a bit lumpy , I expect the TPS  set up along with built in error tolerance has a lot to do with it..  

Dave   

[MarkL]

Dave,

thank you for the interesting explanations. I am about to investigate furhter on the Esprit ECU. However I currently have a constant 0.51 Volts TPS throttle closed reading. It did not vary over 3 separte extensive data logging runs in the course of one day starting 'cold' around 10 degrees Celsius. Do I understand you right measuring .28 to .31 volts at the TPS ?

Can one clean a TPS or is it a sealed unit ? I have made good experiences with potentiometres from old CRT monitors I had been restoring using special purpose electronic cleaner.

Inspecting my logs I don't find uneven TPS variations which could explain any surge on low speed and revs cruising in town I am currently experiencing.

There are still a few things to check in the very detail: spark leads, some more vacuum connections. At a first glance they all were OK.

Maybe just setting a target idle slightly above 1000 rpm instead of slightly below might cure it.

 Mark

[CHANGES]

3 hours ago, MarkL said:

thank you for the interesting explanations. I am about to investigate furhter on the Esprit ECU. However I currently have a constant 0.51 Volts TPS throttle closed reading. It did not vary over 3 separte extensive data logging runs in the course of one day starting 'cold' around 10 degrees Celsius. Do I understand you right measuring .28 to .31 volts at the TPS ?

 Mark,

No the 28 and 31  were just scale settings within the ECU on the throttle fully closed ..  96 was throttle fully open giving a scale range of 68 increments for tuning .... When we moved to the more advanced Penny and Giles contacless unit the scale was 22 closed and 115 fully open , this gave us a 93 increment scale to tune from..  This was not only consistent but a lot finer scale allowing a more refined tuning parameter..  I do not have the voltage at my upper and lower points , but the standard setting should be in the technical files, I am sure @sailorbob or someone will step in with the std reading you should be within.. make sure the TPS is set correct first. 

4 hours ago, MarkL said:

Can one clean a TPS or is it a sealed unit ? I have made good experiences with potentiometres from old CRT monitors I had been restoring using special purpose electronic cleaner.

No its sealed.. 

4 hours ago, MarkL said:

Inspecting my logs I don't find uneven TPS variations which could explain any surge on low speed and revs cruising in town I am currently experiencing.

That will be because the variations will be lost within the tolerance allowed within the std ECU ..   You would need to power up the TPS in isolation  and measure the values at set points, then change the temp to that of engine bay and re-measure,  that should expose the variation..    As far as the surge in low speed revs. could be because the TPS is set just outside one of the tolerance points at closed..  or a poor unit that exceeds the tolerance within the ECU ..  Try micro adjustments or a new TPS. 

4 hours ago, MarkL said:

There are still a few things to check in the very detail: spark leads, some more vacuum connections. At a first glance they all were OK.

These always give problems if not perfect..      

[sailorbob]

The Service Notes state the TPS should be adjusted to give 0.66 to 0.70 volts when closed but in the typical data and fault finding sections it states 0.45 to 0.70 volts. I aim for about the middle of the latter voltage range.

In determining whether the throttle is closed, the ECU uses  4 no. TPS % values (0.8%, 0.4%, 1.2% and 0.4% again). The particular value to use is selected according the vehicle speed (3 mph) and the state of a conditional flag. A further percentage (3.5%) is added to the selected value if the throttle jack is in operation. The percentages should avoid any TPS jitter in normal operation.

All example values are taken from the stock SE code.

[MarkL]

Ok, I see . If you describe 28 or 31 as scale settings from the ECU you are most probably pointing to to some raw TPS data the ECU code deals with.

I already have noticed that there is some more logic behind that. I had a deeper look at a Freescan log from a time before I had adjusted my throttle pedal stop. The 27 year old throttle cable had prolongued and nobody ever had touched this.

The TPS-%-value already had reached 100% during test runs although in ignition on position, no engine running, I was only able to get 93% at maximum static reading.

Prior to the TPS adjustment I had reached 4.04 volts regarded dynamically as 100% . After the adjustment maximum voltage was 4.31 volts. I.E. the ECU at least compensated for the saggy cable prevented full mechanical opening by the first learning procedure.

My EBPV is fixed mechanically in open position. The throttle jack vacuum line is removed and closed by a T-piece loop as recommended. Thus in a warmed up state of engine there should not be any disturbances frrom that direction.

I just add a most current log here. Maybe some one sees any peculiarities. I don't.

But one lesson is learned. A new ECU training is scheduled. Maybe the weather gets better and I pay a visit to the barn garage my Esprit is living in. I just don't only want to swap the battery and check for spark cables and vacuum pipes in order to leave the place without driving this lovely machine.

Mark

freescanlog4.xlsx

[sailorbob]

Ignore what I said above about determining a closed throttle state, upon closer inspection that bit of code looks to be associated with BLM cell selection.

[MarkL]

I did not understand it anyhow :-)

[CHANGES]

23 hours ago, MarkL said:

Ok, I see . If you describe 28 or 31 as scale settings from the ECU you are most probably pointing to to some raw TPS data the ECU code deals with.

Mark, 

I may not have explained my situation that clearly.   I am not using the stock ECU.. As a result I can look at exact figures without being effected by any tolerances written in the program to cater for discrepancies within the TPS..    This is how we were able to identify the fluctuation in the TPS..  .

What i was trying to point out is the std TPS fluctuates and the std ECU caters for this within its program,  it will not be identifiable by your down load data as the ECU will only show a a given value within the tolerance at any set point within the TPS position as per the mapping installed..   It is only when it goes outside of these that you will see it..      The problem is the TPS signal fluctuates with temperature so is inconsistent, you will need to gather data at time of problem.  Even then it may be that it is just on the edge of tolerance producing the tuning issue possibly combined with other factors.   Micro adjustments of the TPS position may elevate your problem .. 

It is also worth noting that excepting the Throttle is fully open because the TPS say's 100% is not enough..  You really should access the throttles and set/balance them all for full range with the peddle first , insuring fully open and closed are obtained.   Then set the TPS position accordingly..    You are very limited to what else you can do..      

A lot of this is why I and others have gone with a stand alone ECU system.  We can isolate these areas and adjust the ECU to the TPS values which is so much more accurate..    These were the 28 -31 figures I was referencing .   Sorry if i confused your thinking..

Dave              

[Chillidoggy]

Is there an idle throttle setting/measurement? I haven't looked in the Service Notes, but I recall attempting to adjust this in with the throttle cable adjuster in the past (the one below the plenum outlet, and suspect it also has a bearing on the issue. By this, I mean is there a fixed measurement for the throttle plates which is adjusted first, then the TPS is set up afterwards?

[MarkL]

I have not found a reference for a throttle adjustment measure. But I presume mine is ok. My Freescan setup on Windows XP was buggy. I could not write any useful logs. They all ended with 0kb file size.

I also wanted to setup the TPS voltage a little different than 0.51 volts at idle to check for influences on this surging problem at cruising speed. Perhaps the TPS is worn in the most used places.

That would become an indicator for a new one. However the voltage display in Feeescan is not very cooperative. Even a new driver setup did not help. It reacts very very retarded and mostly shows only

part of the display. No good starting point for a physical TPS adjustment. Hence I tried Espritmon4.

Does Espritmon4 offer voltage readings on screen? I could only find percentage values.

 

Edited May 20, 2019 by MarkL

[Vulcan Grey]

There's only so much you can adjust with the throttle.  You can adjust the pedal range, the cable slack (the cable must have some slack when the engine is fully warmed up, or it can pull on the throttle and cause higher revs), and you can adjust the base voltage of the TPS at idle (engine off) by rotating the TPS when the screws are loose.  By the way that should be between .45-.65V.

The ECU will control the idle via the IAC (Idle Air Controller stepper motor). DO NOT try to adjust the idle by adjusting the screws on the throttle butterflies!

I looked at your data, and while it was not an ideal run for data, it does look like your IAC counts are too high at idle.  That can be a cause of poor running when starting cold (stalling), or it could be a problem with surging at low throttle openings.

You should start logging freescan from a cold engine, start the engine without touching the throttle or anything, and let it idle to a coolant temp of 82C. Let it stabilize there.  The IAC should come down to between 20-40 at idle with a fully warmed engine. If it is higher, it'll cause poor drivability.

Your MAP should be = 1/2 Baro at Idle, and MAP = Baro when engine off.  If not look for leaks.

IF your IAC is not between 20-40 at idle, then you might need to adjust the minimum air rate bleed screw (not the butterfly plate balance adjuster!) in very tiny increments (like 1/64th of a screw driver turn at a time) while watching the IAC on freescan to get it to stabilize between 20-40.

 

[MarkL]

I have only just adjusted the pedal range to get a 100% TPS reading with ignition on. Slack seems still ok when engine is warmed. I have IAC readings around 60 after 10 minutes of the engine running to normal temp from cold. 

Where is this minimum air rate bleed screw loacated ?

I only know of a throttle plate stop screw from the service notes section EMH.3 page 18 . And reading on page 17 I get confirmed what you have been describing. Thank you for pushing me into that direction.

Besides I already have I new intake plenum cover gasket handy if I will inspect the throttle plates and bores. However I don't expect much grime to clean away at 40000 kms. I had other old cars around 100tkms and more which had hardly any residue in it. I am curious what I will be finding in there.

 

 

freescanlog_cold.xlsx

[Vulcan Grey]

GREEN arrow under the protective aluminum cap. You can turn the flat head screw within the cap, don't try to remove the cap.

DO NOT turn the screw at the RED Arrow.  That is the throttle butterfly adjustment, it is done at the factory to match the flow of the 1&2 to the 3&4 cylinders.

[MarkL]

Great advice here, thanks.

I have cleaned all throttle plates and bores thoroughly. Nothing extraordinary was in there just soft oily blackish residue.

After adjusting the main throttle stop I now have an IACV count around 40. And I feel the surging has  lightened.

Prior to a final evaluation I would like to reset the ECM again.

Will I have the same effect as like the standard procedure of disconnecting the battery etc. when I just reset the BLM and IACV via Freescan ?

Isn't there further variables stored the ECM has to relearn anew ?

 

BTW meanwhile I have also changed the spark leads and wonder about the differences of coil terminal order if comparing them to my service notes EMH section from 1988.

Did they change the layout of the coils package within the SE ? The appear changed over: from bulkhead to rear I have first 1-4 and then 2-3 coil. Normal SE ?

 Or is this already S4 layout ?

Edited May 27, 2019 by MarkL

[Vulcan Grey]

disconnecting the battery will also wipe the BLM. 

Never disconnect the battery or the ECU fuse while the keys are in the ignition!!!

[MarkL]

I am just too lazy to ever disconnect the battery, wait 15 minutes, discharge residual voltage from the systems and reconnect. I sometimes like software ways of doing one or another task for me :-)

At least one mysteria remains:

Where is that throttle position sense value stored in NVRAM ? Maybe it is not at all just because the ECM is auto-adjusting from my observation. Even with a bad mechanical adjustment the logs throw 100% dynamic readings in spite of static 93% e.g. You just loose definition in steps between lowest and highest TPS value, and of course a little real throttle plate opening. If that matters ,hehe.

Does'nt Freescan's reset of BLM and IACV anything untouched ? Isn't there more a 'hard' reset as per battery can handle ?

Edited May 29, 2019 by MarkL

[Vulcan Grey]

The throttle 0% position is saved at startup IIRC, if you have your foot on the throttle slightly, it'll set that as 0% I believe. If you push the throttle 100% though, it go into clear flood mode.

There is no NVRAM. 

The BLM is the long term fuel trim, resetting it in freescan is the same as disconnecting the battery.  The integrator is the short term fuel trim.

The IAC is a stepper motor and must find home each time the car is started. It may lose home, and be reset in freescan. From what I understand resetting the battery may not mean that the IAC is homed correctly. Resetting it via freescan is a better bet.