I just came from a dyno-run, and my real world A/F was by about 15 % off.
The question is ->
If I were to change K by 15% would that offset the error?
So in other words:
If I change K by 15%, will A/F change by 15% as well?
(I will also to double check with my wideband ..)
Is K proportional to A/F Error?
Moderator: Matt
Re: Is K proportional to A/F Error?
very important, before you make any changes: where did they measure the AF ?
and to answer your question regarding K: for the Z32 ECU, the K value is not a linear variable in regards to AF...so changing K by x%, usually does not result in a x% change in AF, unless the change is (very) small
and to answer your question regarding K: for the Z32 ECU, the K value is not a linear variable in regards to AF...so changing K by x%, usually does not result in a x% change in AF, unless the change is (very) small
Re: Is K proportional to A/F Error?
Hello Eric,
The A/F was measured at the end of the exhaust, but (nonetheless) seemed to be pretty accurate.
(so we had 14.7 while the Lambda-Sensor was in the equation)
Since you say K is not applied linear (I thought K in a way gives the ECU the 'size' of the injectors).
What other factors are there?
The A/F was measured at the end of the exhaust, but (nonetheless) seemed to be pretty accurate.
(so we had 14.7 while the Lambda-Sensor was in the equation)
Since you say K is not applied linear (I thought K in a way gives the ECU the 'size' of the injectors).
What other factors are there?
Re: Is K proportional to A/F Error?
there are lots of things that can throw a perfect K off balance...
incorrect injector latency and/or battery correction values.
enrichment table(s)
fuel map(s)
and mechanical issues:
faulty AFM
vacuum/boost leaks
misfires (also the ones barely noticeable)
exhaust leaks
(old/worn) engines that have a lot of blowby
etc.
Measuring the AF at the end of the exhaust is far from ideal...and if the car has a catalyst it's completely useless to measure at the exhaust end..
At idle it may look okay, but as rpms get higher and exhaust gasflow increases, AF figures will start to become less accurate (leaner).
if you have a good funtioning narrowband O2 sensor, you can look at the AF-alpha correction values that can be read through consult. (not sure if nistune can also read these values)
if these values are within -5% to +5% at several rpms, you have a near perfect K-value.
if values are outside +15% or -15%, you need to make some tweaks, but first make 100% sure it's not a mechanical issue.
values of +25% or -25% means your K-value is completely off.
incorrect injector latency and/or battery correction values.
enrichment table(s)
fuel map(s)
and mechanical issues:
faulty AFM
vacuum/boost leaks
misfires (also the ones barely noticeable)
exhaust leaks
(old/worn) engines that have a lot of blowby
etc.
Measuring the AF at the end of the exhaust is far from ideal...and if the car has a catalyst it's completely useless to measure at the exhaust end..
At idle it may look okay, but as rpms get higher and exhaust gasflow increases, AF figures will start to become less accurate (leaner).
if you have a good funtioning narrowband O2 sensor, you can look at the AF-alpha correction values that can be read through consult. (not sure if nistune can also read these values)
if these values are within -5% to +5% at several rpms, you have a near perfect K-value.
if values are outside +15% or -15%, you need to make some tweaks, but first make 100% sure it's not a mechanical issue.
values of +25% or -25% means your K-value is completely off.
Re: Is K proportional to A/F Error?
Hmmm
Sure, mechanical issues can always throw off AF.
But in my 'case' I would dare ruling these out for the moment.
(engine is low milage, and always treated nicely)
Also my car has no cat, and what we have seen from the dyno's AF-sniffer looks
consistent on other cars.
I might suspect incorrect voltage correction/latency but don't know if that could have such a profound effect.
(I am using the standard RB25 injectors which seem to be the same as in the ZX300TT.)
I am not familiar with AF-alpha correction values .. but it sounds interesting.
Is this something used for closed loop? (possibly indicating how much the ECU has to alter fuel supply to accommodate Lambda 1?)
Would be good to have in Nistune.
(My Lamda Sensor is brand new)
Sure, mechanical issues can always throw off AF.
But in my 'case' I would dare ruling these out for the moment.
(engine is low milage, and always treated nicely)
Also my car has no cat, and what we have seen from the dyno's AF-sniffer looks
consistent on other cars.
I might suspect incorrect voltage correction/latency but don't know if that could have such a profound effect.
(I am using the standard RB25 injectors which seem to be the same as in the ZX300TT.)
I am not familiar with AF-alpha correction values .. but it sounds interesting.
Is this something used for closed loop? (possibly indicating how much the ECU has to alter fuel supply to accommodate Lambda 1?)
Would be good to have in Nistune.
(My Lamda Sensor is brand new)
Re: Is K proportional to A/F Error?
>> (I am using the standard RB25 injectors which seem to be the same as in the ZX300TT.)
they have the same flow, but different latency...also the battery correction value is different between the 2
>> Is this something used for closed loop? (possibly indicating how much the ECU has to alter fuel supply to accommodate Lambda 1?)
yes, it's the correction value/percentage as used by the ECU to get stoich AF.
it (self) learns this by monitoring the narrowband signal and is constantly updated.
However, the alpha correction value is only reliable with a good functioning standard O2 sensor
they have the same flow, but different latency...also the battery correction value is different between the 2
>> Is this something used for closed loop? (possibly indicating how much the ECU has to alter fuel supply to accommodate Lambda 1?)
yes, it's the correction value/percentage as used by the ECU to get stoich AF.
it (self) learns this by monitoring the narrowband signal and is constantly updated.
However, the alpha correction value is only reliable with a good functioning standard O2 sensor
Re: Is K proportional to A/F Error?
Thanks,
Matt said this ALPHA value is avialable (need to test this)
My Lambda Sensor is brand new, so worth a shot I guess.
Good that you mention the injectors.
RB25 (GTST) and ZX300TT seem to use the same injectors (as for flow).
I noticed that the Voltage Correction and Latency is different though.
And I thought that this was due to the two different ECUs using different calculations?
Or would you think that RB25 and ZX300TT share the same 'routines' ?
I would'nt have thought that the injectors are actually different?
Matt said this ALPHA value is avialable (need to test this)
My Lambda Sensor is brand new, so worth a shot I guess.
Good that you mention the injectors.
RB25 (GTST) and ZX300TT seem to use the same injectors (as for flow).
I noticed that the Voltage Correction and Latency is different though.
And I thought that this was due to the two different ECUs using different calculations?
Or would you think that RB25 and ZX300TT share the same 'routines' ?
I would'nt have thought that the injectors are actually different?
- Attachments
-
- RB25Z32.jpg
- (29.76 KiB) Downloaded 3028 times
Re: Is K proportional to A/F Error?
>> I would'nt have thought that the injectors are actually different
No, the standard Z32 TT injectors in the early models (Nistune type2 board) are one of a kind and not used in any other Nissan model.
The injectors used in the later model Z's (Nistune type3 board) are different and more common and used in several Nissans/Subarus/Mitsus.
>> And I thought that this was due to the two different ECUs using different calculations?
no, DTS and TS14V have the same unit for every Nissan ECU i have seen
No, the standard Z32 TT injectors in the early models (Nistune type2 board) are one of a kind and not used in any other Nissan model.
The injectors used in the later model Z's (Nistune type3 board) are different and more common and used in several Nissans/Subarus/Mitsus.
>> And I thought that this was due to the two different ECUs using different calculations?
no, DTS and TS14V have the same unit for every Nissan ECU i have seen
Re: Is K proportional to A/F Error?
Alright, got it ..
So I will stick to my RB25 values for latency and voltage correction.
(But flow seems to be 370CC for these)
So I will stick to my RB25 values for latency and voltage correction.
(But flow seems to be 370CC for these)
Re: Is K proportional to A/F Error?
Eric wrote:there are lots of things that can throw a perfect K off balance...
if you have a good funtioning narrowband O2 sensor, you can look at the AF-alpha correction values that can be read through consult. (not sure if nistune can also read these values)
if these values are within -5% to +5% at several rpms, you have a near perfect K-value.
if values are outside +15% or -15%, you need to make some tweaks, but first make 100% sure it's not a mechanical issue.
values of +25% or -25% means your K-value is completely off.
Hello Eric/Matt,
I have tried to log theses Alpha Values with Nistune.
I warmed the car up, and held the throttle at 2000rpm / 3000rpm.
(I hope this is a valid procedure)
The values for ALPHA/learn stayed as per attached picture:
(Sorry, I posted the image twice)
Re: Is K proportional to A/F Error?
95% selflearn means -5% fuel to get stoich at that particular rpm. (~2000)
as said before, -5% is very good...but selflearn levels may vary over different rpms, so just rev it between 1500-4000 or so and see/log what the selflearn level(s) are and if the average is between -5% to +5% I would leave the K-val as is.
as said before, -5% is very good...but selflearn levels may vary over different rpms, so just rev it between 1500-4000 or so and see/log what the selflearn level(s) are and if the average is between -5% to +5% I would leave the K-val as is.
Re: Is K proportional to A/F Error?
Thanks Eric,
These levels already reflect the 10% increase I did on K.
I will do as you suggested and test on other RPM-Levels.
This seems to be an interesting procedure to test your injector setup ...
These levels already reflect the 10% increase I did on K.
I will do as you suggested and test on other RPM-Levels.
This seems to be an interesting procedure to test your injector setup ...
Eric wrote:95% selflearn means -5% fuel to get stoich at that particular rpm. (~2000)
as said before, -5% is very good...but selflearn levels may vary over different rpms, so just rev it between 1500-4000 or so and see/log what the selflearn level(s) are and if the average is between -5% to +5% I would leave the K-val as is.