stoich clc operation sm2 1.94

[roberto arano]

i noticed something; (o2 is wideband/linear input)

if i change the stoich fuel ratio to something like 16:1 (under Base settings)
and i have clc enabled it will show a target afr of 16:1 during clc when on a 16:1 target site on the AFR table, but will really only trim to 14.7 still!

(i did not get into lean highway mode to test it yet, regardless it should target my stoich value of 16.)

this is strange, because the rest of the operation is normal
so i tested it with clc disabled, and it will as normal open loop ,give me a mixture of 16:1 as i expect. Also if i have a AFR target leaner than stoich (leaner than 16:1) it will as expected only attempt 16:1 or richer(not in lean highway mode).

is this a bug

[Ian]

I did a reply saying that would be the expected operation with just the target displaying incorrectly but while previewing it I thought of another implication that I will need to check out first.

[roberto arano]

o.k,

just to clarify;

i have the stoich set as 16:1 under base settings

i have the afr target (at the site/load cell) in use at 16:1

and the target on the monitor is shown as 16:1 (also the "tick/hash" mark on the bar graph), but it is apparent that it is still trimming to an actual target of 14.7:1 as seen by the actual AFR readout on the monitor. So the target is being displayed as expected, but the CLC is not internally using that as the target, it's using 14.7 as stoich instead of the actual stoich which was selected to be 16:1 by the user under the base settings.

[roberto arano]

any progress on this yet?

[Ian]

I don't quite recall the unedited first post but changing the Stoic value parameter will certainly change things in the Open loop calculations. However, changing the Stoic parameter is a separate issue to what the target displays and what it actually does.

Buzzard kindly did extensive testing today with an SM2/1.94 and we have to examine the log in more detail. However, it appears at least that the target AF does not display correctly in some circumstances. What it may be showing is the CLC target (even with CLC disabled) rather than the Open loop target. But we have to examine the data some more.

[roberto arano]

i did some more testing with 1 .94 chip;

the problem can be broken down into two parts: (summary)
-with 16:1 set as stoich under base settings
-16:1 will be used correctly as the stoich target on the AFR target table, i.e values equal to this willl be run in closed loop, richer values won't and leaner ones only when in lean hwy mode (i didn't actually confirm lean hwy mode yet)
-however the 1.94 will only actually really trim in clc to 14.7 still even when it selects a stoich target (and displays it ) as 16:1
- so the error is not in the displayed target at all, it's in the target it is actually USING


(detail)
with clc operating, no lean highway the problem is that the ecu does not correctly accept 16:1 as stoich even if 16 is defined as stoich under base settings. SO what happens is this: even if you have 16:1 (supposed to=stoich ,as set) set in the AFR target table, the ECU shows 16 as the target on the monitor, but it is actually trimming ,still to 14.7 in CLC. It is behaving as though it still thinks 14.7 is stoich therefore it will only trim to 14.7 because 16 is leaner, this would be expected if stoich was 14.7, but in this case it is an error because stoich is was set as 16.

So it is actually displaying the correct target AFR, but can only ever trim to 14.7 (disregarding lean highway mode for now)

So when one set the stoich=xx.x under base settings, it will use that AFR as the correct displayed target. Which will now let the target be numerically greater than 14.7 even when not in lean highway mode (as it should). It will not effect what the ECU actually trims to (it's not using the displayed target to actually trim to). That is the error.

if this is really confusing i can take some screen shots

either way i doubt it will be fixed as it would require a new chip?

[Wroom]

Are you using a wideband lambda sensor with linear input, and that it is correctly configured in the SM2?

If you are using a standard lambda sensor, or really any sensor without linerarization box the lambda sensor will show either very rich or very lean outside a very narrow area around lambda 1, (AFR 14.6 to 14.7), and the CLC algorithm will do what it is supposed to - Regulate the long term integral lambda reading from the sensor to lambda 1 by applying an adaptive correction value plus a regulating curve that proportionally strive towards this integral lambda 1 with a 50/50 duty cycle.

This behaviour is determined by the sensor characteristics.

So, unless you are using a properly configured wideband lambda sensor with linear input to the SM2 you will not get CLC regulation to any other value than lambda 1.

Also, the lean highway mode will only lean out when it receives expected feedback of the AFR corrections from the lambda sensor.

Connecting a standard, nonlinearized lambda sensor, will keep the lean highway mode off, and the fuelling will stay at lambda 1 on any site that is leaner than lambda 1.

[roberto arano]

yes, i see what you are saying, but that is not what i am talking about;

maybe i will take a screenshot, it seems to be hard to understand unless you see it.

yes, it's a linear input wideband, not a narrowband. the difference is important as the stoich point is configured by the user as 16 NOT 14.7 under BASE SETTINGS MENU

everything work as normal (lean hwy,etc) if using a "normal" stoich setting of 14.7 UNDER BASE SETUP

it's just the odd case of if stoich is set to 16:1 (not gasoline stoich), then it does not trim to stoich it trims to 14.7 STILL!! probably no one has tried this as it is is unusual and therefore have not seen the glitch.

i think this is an undiscovered error as no one has had a need to try it and maybe i will just forget about it as it's not going to kill me!!! ;-)

be observant l if you try this, the monitor (and AFR table ) target will show 16:1, but it will actually still trim to 14.7. This is entirely different that if 14.7 was set as stoich and it had a lean highway target of 16 but not in hwy mode, then the lean target would be 16 on the AFR table, but the current target would be shown as 14.7 until in lean hwy mode. To even illustrate it in another way, if stoich=16 and lean hwy enabled, but not in lean hwy mode yet one would expect to be able to CLC 16:1 when not in lean hwy mode, and AFR's leaner than 16 when in lean hwy mode.

[Ian]

I don't have the full answer yet but it appears at least that the software is not reporting the correct Target AF in some circumstances. Buzzard reported that when he put 16 in the Target table the Target showed 14.7 but activating Autotune tuned to 16 even though 14.7 was displayed. This was with the Stoic AFR set at 14.7.

The Target AFR is derived by the software for Autotune and display purposes, it is not passed up the data stream from the ECU. Further testing is required to ascertain if something else is astray other than the display.

[roberto arano]

in my scenario the target is displayed correctly (16:1) , it just doesn't trim to it, only to 14.7

meaning the target numeric and target tick marck is at 16:1 but the readout of the AFR show it is CLC trimming very well to 14.7

[Chris]

in my scenario the target is displayed correctly (16:1) , it just doesn't trim to it, only to 14.7

meaning the target numeric and target tick marck is at 16:1 but the readout of the AFR show it is CLC trimming very well to 14.7

The target is displaying 16:1, because you have 16:1 set as the stoic af. If you set the stoic af to 14.7, when the ecu operates on cells with 16:1target, the target display is 14.7.This is not normal, at least as compared to smc chip versions. Does this match what you are seeing ?

Chris

[roberto arano]

The target is displaying 16:1, because you have 16:1 set as the stoic af. If you set the stoic af to 14.7, when the ecu operates on cells with 16:1target, the target display is 14.7.This is not normal, at least as compared to smc chip versions. Does this match what you are seeing ?

Chris

yes this is true for my sm2 1.94, but my understanding is that the above scenario is normal. The ecu should never target AFR's that are leaner than stoich UNLESS in lean hwy mode. For example even if the afr cell=16, it should display a target=stoich (14.7), because it is not in lean hwy mode yet. When it is in lean hwy mode it will target 16.

(this reply was assuming 14.7 as set stoich)

also today i tried a set stoich on the richer side of 14.7, and had the same kind of error, it would only truly trim to the 14.7:1 stoich.

i guess the bottom line is, if it is a glitch, then it would require a chip to fix and it's unlikeley to be used anyway since this is an old model anyway and an unusual need to boot, but is is interesting.

maybe i will go try this on another sm2 i have (same chip version) and compare.

[Chris]

The ecu should never target AFR's that are leaner than stoich UNLESS in lean hwy mode. .

That is not my understanding, to me the ecu should suppy the afr set in the target table, be it 12:1 or 18:1

Chris

[Wroom]

The ecu should never target AFR's that are leaner than stoich UNLESS in lean hwy mode.
The ecu should target AFR's leaner that stochiometric if it is in non CLC, non LH mode, or in LH mode and has adapted to that lean value.

I am not clear on if it should target an AFR leaner than stochiometric when in CLC mode.

[roberto arano]

The ecu should target AFR's leaner that stochiometric if it is in non CLC, non LH mode, or in LH mode and has adapted to that lean value.

I am not clear on if it should target an AFR leaner than stochiometric when in CLC mode.

yes i believe it is supposed to target lean AFRs when in LH mode and in CLC mode, that is why a linear o2 is mandatory for it's operation. Also i beleive that is what Ray Hall explained to me years ago, Ian?