[xj-s] Adaptive idle fueling trim ...?

In reply to a message from sbobev sent Fri 20 Sep 2013:

Definitely a frustrating system. It uses many of the logic
controls we see in more ‘‘modern’’ vehicles, but it sacks the ability
to tell us what is wrong via the OBD plug.

I’ve learned quite a bit from you guys…now that I’m home, I’ve
got the book (both of them!) out and will start catching up on how
the EFI thinks.

I find it interesting that the early fuel controls required an idle
trim. It has to be mixture only, as the idle speed is set manually
on these cars…with no adjustment of the air, other than the
canister purge and the AC supplemental air solenoid. (sorry…need
to study to get the nomenclature back). I would have to think this
is because they knew the O2 sensors would not be hot enough to work
right at idle, before they added the O2 heaters?

So, if the air is constant, our 6.0 roll has to be mixture
related. Gotta go study!–
John. '95 XJS 6.0L convertible. Southlake, TX
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from sbobev sent Fri 20 Sep 2013:

‘‘They provide scanned images, not text, so copy and paste doesn’t
work – I have to re-type everything.’’

Steve,

If you have Adobe Acrobat, convert the page with OCR (optical
character reader), then you can copy and paste as usual.

I don’t know if Adobe Reader will do that - if it has an OCR
function, try that.

AE–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from CJ95 sent Sat 21 Sep 2013:

John,

Someone mentioned that the earlier ECU’s (up thru the 26 I believe)
operate open loop in idle - thus the idle trim to clean up the idle
emissions.–
lockheed 92 XJS Cpe/97 LT1 Miami FL/ 96 XJS Cv 4.0 Austin TX
Austin, TX, United States
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

I find it interesting that the early fuel controls required an idle
trim. It has to be mixture only…

Correct. One thing I’ve never been able to ascertain is whether it
is literally idle only, or whether the ECU interpolates a straight
line between there and the first off-idle point on the fuelling map.
It probably doesn’t make much difference. It may be worth noting,
though, that I learned later that early EFI systems – on cars in
general, not just the XJ-S – used most of their limited fuel mapping
capacity in the near-idle range; apparently it’s far easier to map
the high power and WOT regions, you can just plot a few points and
let the interpolator take care of the rest. The area around idle is
so difficult, though, that getting a manual transmission with one of
these early EFI systems can be a serious mistake – as my family
found out with a Dodge they bought that was really difficult to drive
slowly into a parking space. With an A/T, nothing of importance
happens below 1500 rpm anyway, and what does is smoothed by the
torque convertor, so you can get away with a little weirdness in the
throttle response.

I would have to think this is because
they knew the O2 sensors would not be hot enough to work right at
idle, before they added the O2 heaters?

That, plus the fact that trying to run closed-loop at idle tends to
result in an unstable idle even when the O2 sensors are working.

So, if the air is constant, our 6.0 roll has to be mixture
related. Gotta go study!

Remember, it could also be ignition timing related. The mid-80’s
models used a vacuum regulator to hold the vacuum advance fixed at
idle specifically to avoid idle instabilities. You might wanna try
connecting a timing light and watching what the timing actually does
during this “roll”.

– Kirbert

// please trim quoted text to context onlyOn 21 Sep 2013 at 7:34, CJ95 wrote:

In reply to a message from Kirbert sent Sat 21 Sep 2013:

I have read through the 6.0 fuel section. It states that the idle
trim adjustment is applied across the entire map. Still reading…–
The original message included these comments:

Correct. One thing I’ve never been able to ascertain is whether it
is literally idle only, or whether the ECU interpolates a straight
line between there and the first off-idle point on the fuelling map.


John. '95 XJS 6.0L convertible. Southlake, TX
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from CJ95 sent Sun 22 Sep 2013:

Correct. That is because the trim is applied to the strategy
not to the idle at that specific moment in time. This is what
I meant about changing the baseline, it essentially moves the
entire 16 x 16 matrix.

Cheers,

Allan–
The original message included these comments:

I have read through the 6.0 fuel section. It states that the idle
trim adjustment is applied across the entire map. Still reading…


1994 XJS 4.0 2+2 Convertible 5 Speed Manual
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

Correct. That is because the trim is applied to the strategy
not to the idle at that specific moment in time. This is what
I meant about changing the baseline, it essentially moves the
entire 16 x 16 matrix.

Wow. Since the issue is supposedly to deal with engine wear, perhaps
Jaguar figured out how engine wear impacts the entire fuelling map
and then uses this adaptive idle strategy – which it says is engaged
only at idle – to shift the entire map in a direction that helps
correct for engine wear across the board.

– Kirbert

// please trim quoted text to context onlyOn 22 Sep 2013 at 13:00, AllanG wrote:

In reply to a message from Kirbert sent Sun 22 Sep 2013:

Or, maybe it was just easier to apply it that way with the
ECU processor available in the early '90’s. It’s a very
small adjustment compared to the ±20% the O2 sensors
provide.–
John. '95 XJS 6.0L convertible. Southlake, TX
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from CJ95 sent Sun 22 Sep 2013:

The adaptive idle fuel trim is able to provide +/- 20% trim to
the nominal injector pulse duration.

Cheers,

Allan–
The original message included these comments:

Or, maybe it was just easier to apply it that way with the
small adjustment compared to the ±20% the O2 sensors
provide.


1994 XJS 4.0 2+2 Convertible 5 Speed Manual
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from sbobev sent Fri 20 Sep 2013:

OK this is what I think happens in the first 1-3mins or so in
regard to fueling. It would be easier to draw it as a flow
chart than write it out but here goes :slight_smile: Sorry if it’s a bit
long and covers stuff you already know.

Irrespective of open or closed loop operation, the Fuel
Injection ECM uses the same basic multi stage process to
control injector pulse duration for any operating condition.

Primary control is established based on engine load (MAP
sensor) and engine speed (from the Ignition ECM). These
values are mapped to a 16 x 16 matrix (256 sites) that resides
in EPROM. If, as is likely the mapped value is between sites,
a two dimensional interpolation is used to calculate the
correct pulse duration.

This value is then subjected to a secondary control that
provides correction factors based on coolant temp, throttle
position and movement, engine speed and exhaust oxygen content
(closed loop only) These correction factors allow the ECM to
utilize the correct strategy for idle, after start enrichment,
warm up, acceleration enrichment, deceleration leaning and
full load enrichment.

Once the correction factors have been applied, the resulting
injector pulse duration is further corrected for battery
voltage using a programmed voltage correction strategy. (This
is necessary because the time to lift the injector pintles
increases as voltage decreases.)

The final pulse duration is then outputted and all the
injectors are triggered simultaneously.

Right, back to the first few minutes of engine run :slight_smile:

After every engine start, air injection is enabled by the ECM
in either a cold or hot start mode. Fueling is in open loop
when air injection is enabled. Air injection also always
runs if coolant temp is below 47C. The ECM uses a strategy
based on number of injector pulses vs coolant temp to turn air
injection off. For hot starts, if the fuel rail temp is above
70C when air injection is enabled, the hot start switch
actuates a valve to purge the fuel rail (Combat vapor lock)

When the ECM disables air injection, fueling goes into closed
loop mode. The O2 sensor in the exhaust sends a feedback
voltage to the ECM which then further corrects the pulse
duration with the goal of maintaining an air/fuel ratio of
14.7:1 (Lambda = 1). Closed loop is temporarily cancelled by
the ECM if full load enrichment or deceleration fuel cut-off
is required.

When fueling is in closed loop mode, the ECM closes the purge
valves allowing vapor to flow from the charcoal canister to
the inlet manifold. The rate of flow is electrically
controlled by the ECM.

Also, as described above, the ECM performs an idle fuel trim
adaptation routine when preconditions are met. These values
are stored in RAM and are preserved provided battery power is
available. A break in power will force the trim to be reset
to factory programmed values until another adaptation cycle
takes place and then the car is driven at least 100 yards.

Cheers,

Allan–
The original message included these comments:

I’ve (reluctantly) accepted it for what it is, but I am
still trying to figure out what the ECU does 1-3 minutes
after start-up. During that period, I have some complaints.
One of the hypothesis I’ve got is that the fueling is
incorrect for some reason.


1994 XJS 4.0 2+2 Convertible 5 Speed Manual
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

Or, maybe it was just easier to apply it that way with the
ECU processor available in the early '90’s. It’s a very
small adjustment compared to the ±20% the O2 sensors
provide.

Do those cars still go into open loop at WOT? If so, having this
“adaptive idle” scheme occasionally revise the BASELINE map that the
car runs on would be handy indeed, keeping it optimized as the engine
wears. Yeah, during regular cruise the O2 sensors are gonna control
the mixture anyway, but when you stomp it you don’t want it running
too lean or too rich.

– Kirbert

// please trim quoted text to context onlyOn 22 Sep 2013 at 18:44, CJ95 wrote:

In reply to a message from Kirbert sent Mon 23 Sep 2013:

Yep, as mentioned in my post above, closed loop is cancelled
at WOT. (Jaguar’s term is Full Load Enrichment as it is load
dependent and not just based on the TPS)

Closed loop is also disabled when the overrun fuel cut off is
activated.

Cheers,

Allan–
The original message included these comments:

Do those cars still go into open loop at WOT? If so, having this
‘‘adaptive idle’’ scheme occasionally revise the BASELINE map that the
car runs on would be handy indeed, keeping it optimized as the engine
wears. Yeah, during regular cruise the O2 sensors are gonna control
the mixture anyway, but when you stomp it you don’t want it running
too lean or too rich.


1994 XJS 4.0 2+2 Convertible 5 Speed Manual
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from AllanG sent Sun 22 Sep 2013:

Thanks Allan:
This is the most complete ‘short’ description of the 36CU
operation I’ve seen.

I do have a question for you. Below you state that the ECU
closes purge valves to allow canister regeneration in closed
loop?
The way you worded it (closes purge valves) implies the
opposite – purging is happening in OPEN loop. CLOSED loop
enabled would mean no purge. Do I have it right?

I have got an off-list email from Mr. Bywater, who explained
to me that the stand-alone 45-timer unit was phased out with
the introduction of the 6.0L (VIN188004) and the 36CU
switches to CLOSED loop (irrespective of engine temperature
or O2-sensor input) exactly 1 minute after start-up.

You also touched upon the voltage correction, the issue I
was going to bring up next. Can you elaborate, please?

My car sits in the garage most of the the time and the
battery is always connected to a trickle-charger. Hence,
every time it starts from cold, the battery will be at 100%.
I wonder if this is what confuses the ECU.

Apparently, I’ve got a condition that tricks the ECU and it
trims the fueling too much during the first a couple of
minutes. Having ruled out TPS, CPS, CTS, ATS, flywheel
sensors, fuel pressure, freshly tested injectors, and
perhaps more, I wonder if EVAP function or the above-
mentioned voltage correction is what I need to check.

Best regards,
Steve–
The original message included these comments:

OK this is what I think happens in the first 1-3mins or so in
regard to fueling. It would be easier to draw it as a flow
chart than write it out but here goes :slight_smile: Sorry if it’s a bit
long and covers stuff you already know.
When fueling is in closed loop mode, the ECM closes the purge
valves allowing vapor to flow from the charcoal canister to
the inlet manifold. The rate of flow is electrically
controlled by the ECM.


'95 XJS V12 6.0L, saphire/tan
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from AllanG sent Sun 22 Sep 2013:

Allan:
Given that the RAM is volatile and power break erases
everything, this must mean that if I develop a condition of
some kind, related to this subject (e.g. the speculated
‘overcorrection’ due to adaptive idle fueling trim), then, the
behavior of the engine will be different if started after
disconnecting the battery, right?
This should be easy to check…
Steve–
The original message included these comments:

Also, as described above, the ECM performs an idle fuel trim
adaptation routine when preconditions are met. These values
are stored in RAM and are preserved provided battery power is
available. A break in power will force the trim to be reset
to factory programmed values until another adaptation cycle
takes place and then the car is driven at least 100 yards.


'95 XJS V12 6.0L, saphire/tan
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from sbobev sent Mon 23 Sep 2013:

Yes, once you restart the engine after reconnecting the
battery, you will probably notice that the engine revs quite
high, perhaps 1200 - 1500 rpm then slowly lowers to the normal
idle speed.

For the adaptive idle fuel trim, the final values will not be
written to RAM until the car has been driven at least 100
yards at speeds over 3 mph. (It’s probably held in a temporary
memory location until that time.) This may be the case with
other parameters also.

Cheers,

Allan–
The original message included these comments:

Given that the RAM is volatile and power break erases
everything, this must mean that if I develop a condition of
some kind, related to this subject (e.g. the speculated
‘overcorrection’ due to adaptive idle fueling trim), then, the
behavior of the engine will be different if started after
disconnecting the battery, right?
This should be easy to check…


1994 XJS 4.0 2+2 Convertible 5 Speed Manual
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from sbobev sent Mon 23 Sep 2013:

You’d think so but no :slight_smile:

The purge valve is a solenoid operated valve that is normally
open, so ‘closing’ the valve will allow vapor to flow. The
amount of flow is controlled by a pulsed signal from the ECM

Cheers,

Allan–
The original message included these comments:

I do have a question for you. Below you state that the ECU
closes purge valves to allow canister regeneration in closed
loop?
The way you worded it (closes purge valves) implies the
opposite – purging is happening in OPEN loop. CLOSED loop
enabled would mean no purge. Do I have it right?


1994 XJS 4.0 2+2 Convertible 5 Speed Manual
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from sbobev sent Mon 23 Sep 2013:

Steve,

I don’t know how to do a multi part response so I have had to
make separate posts to your questions, sorry about that.

Yes as I mentioned earlier the 45 sec timer was replaced by
ECM control for the 6.0. I’m not sure about the 1 min cutoff,
my understanding from the Jaguar documentation is that the ECM
switches off air injection using a strategy that looks at
injector pulses versus coolant temperature.

In terms of battery voltage correction, the fact that your
battery is always 100% will have no effect. The ECM just uses
a fudge factor to compensate for the actual battery voltage to
determine the pulse duration. For instance if the strategy
says that at 14V no correction is needed, then the calculated
pulse duration will go unchanged. If however, the battery
voltage dropped to say 13.5V then the ECM would apply a
correction based on the voltage correction strategy to
increase the pulse duration in order to maintain the required
fuel flow.

The reason this is necessary is that the rate at which the
injector pintle opens is directly related to voltage. If the
voltage decreases the pintle will open more slowly and less
fuel would flow for a given pulse duration. The ECM
compensates for this by increasing the pulse duration.

Sorry for being wordy but it’s sometimes hard to find an easy
way to explain things :slight_smile:

Cheers,

Allan–
The original message included these comments:

I have got an off-list email from Mr. Bywater, who explained
to me that the stand-alone 45-timer unit was phased out with
the introduction of the 6.0L (VIN188004) and the 36CU
switches to CLOSED loop (irrespective of engine temperature
or O2-sensor input) exactly 1 minute after start-up.


1994 XJS 4.0 2+2 Convertible 5 Speed Manual
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from sbobev sent Mon 23 Sep 2013:

Possibly. During open loop operation the valves should not
allow vapor to flow to the manifold. If it did, that could
weaken the mixture and cause a stumble because the ECM would
have no way of correcting for it.

The valve is controlled by a series of pulses from the ECM.
The pulses vary in duration increasing as engine load and
speed increase. They should be about 10msSec duration at idle
with a periodicity of about 145mSec.

Cheers,

Allan–
The original message included these comments:

perhaps more, I wonder if EVAP function or the above-
mentioned voltage correction is what I need to check.


1994 XJS 4.0 2+2 Convertible 5 Speed Manual
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from AllanG sent Mon 23 Sep 2013:

Allan:
Are we talking about the ‘Rochester’ valve here or the two
valves that are plumbed to both banks?
Closing the Rochester valve in order to purge the canister
is in order, otherwise the manifold vacuum will be sucking
the gas tank and collapsing it (well-documented issue)
I’ve had my gas tank lid open/closed during various test and
have not seen the ‘‘whoosh’’, so, I believe my EVAP system is
OK.
Is there are way to check the solenoids with a simple VOM?
Steve

PS. If you are curious, you can read my other posts under
the topic ‘‘6.0L idle woes…’’ Therein, I described a test,
whereby the ECU is put in forced open-loop operation by
removing the O2-sensors feedback. The idle speed increases
by 50-100 rpm, which I attributed to the slightly richer
mixture and more air into the manifolds through the carbon
canister vent… did I get this wrong too?–
The original message included these comments:

You’d think so but no :slight_smile:
The purge valve is a solenoid operated valve that is normally
open, so ‘closing’ the valve will allow vapor to flow. The
amount of flow is controlled by a pulsed signal from the ECM
Cheers,
Allan


'95 XJS V12 6.0L, saphire/tan
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only

In reply to a message from AllanG sent Mon 23 Sep 2013:

Allan:
Related to what I said above. There is no stumble in open
loop. I think the ECU is tricked into doing something
excessive just as it transitions from open to closed loop.
Steve

PS From your signature I gather you are a 6-cylinder owner.
How do you know so much about the 36CU operation?–
The original message included these comments:

Possibly. During open loop operation the valves should not
allow vapor to flow to the manifold. If it did, that could
weaken the mixture and cause a stumble because the ECM would
have no way of correcting for it.
The valve is controlled by a series of pulses from the ECM.
The pulses vary in duration increasing as engine load and
speed increase. They should be about 10msSec duration at idle
with a periodicity of about 145mSec.


'95 XJS V12 6.0L, saphire/tan
–Posted using Jag-lovers JagFORUM [forums.jag-lovers.org]–
–Support Jag-lovers - Donate at http://www.jag-lovers.org/donate04.php

// please trim quoted text to context only