setting lambda delay
Moderator: LT401Vette
-
- Super MS/Extra'er
- Posts: 1021
- Joined: Sun Jun 27, 2004 12:04 pm
- Location: Long Beach, CA - USA
Re: setting lambda delay
To add to this - I need to figure out something because when I try and use autotune around idle, it gets caught in an oscillation and really starts messing up the VE table. I might change the difficulty to hard or very hard and see if that helps things out...
-
- Super MS/Extra'er
- Posts: 12731
- Joined: Sat Jul 16, 2005 8:07 am
- Location: Moorseville, NC
- Contact:
Re: setting lambda delay
Usually when I see occilations at idle, it is the spark table that needs some adjustment. Does the advance need some flattening around the idle point?I need to figure out something because when I try and use autotune around idle, it gets caught in an oscillation
Phil Tobin
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
-
- Super MS/Extra'er
- Posts: 1021
- Joined: Sun Jun 27, 2004 12:04 pm
- Location: Long Beach, CA - USA
Re: setting lambda delay
nope - it's flat. I'll play with it some more and see if I can get it to do a better job...
Re: setting lambda delay
Hi Phil,
I had some time to loose (or not finally) this afternoon so I decided to take a look of my lambda delay values.
The car is a inline 4 cylinder 1400cm3 turbocharged. The engine is transversal and since there is not much place under the bonnet and between the engine and the chassis, the turbo
is on the gearbox. After the turbo, the downpipe goes back to the rear of the engine to catch the exhaust under the car. So the exhaust system is a bit long. The lambda sensor is about 15-20 cm after the turbo. Intake and exhaust are on the same side of the cylinder head. The throttle is 60mm, yeah a little big but is is very fine
The lambda controler is an NGK Powerdex and I suspect latency from him. Here is why.
What do you think about those value ? I didn't try yet those value with the live analyser.
Some value were extrapolated. I am not sure at all for the WOT delays.
Those pics are from the log file.
I had some time to loose (or not finally) this afternoon so I decided to take a look of my lambda delay values.
The car is a inline 4 cylinder 1400cm3 turbocharged. The engine is transversal and since there is not much place under the bonnet and between the engine and the chassis, the turbo
is on the gearbox. After the turbo, the downpipe goes back to the rear of the engine to catch the exhaust under the car. So the exhaust system is a bit long. The lambda sensor is about 15-20 cm after the turbo. Intake and exhaust are on the same side of the cylinder head. The throttle is 60mm, yeah a little big but is is very fine
The lambda controler is an NGK Powerdex and I suspect latency from him. Here is why.
What do you think about those value ? I didn't try yet those value with the live analyser.
Some value were extrapolated. I am not sure at all for the WOT delays.
Those pics are from the log file.
Re: setting lambda delay
And the log
Re: setting lambda delay
Couldn't upload the whole log, it says file is to big, maximum allowed size is 716 KiB
-
- Super MS/Extra'er
- Posts: 12731
- Joined: Sat Jul 16, 2005 8:07 am
- Location: Moorseville, NC
- Contact:
Re: setting lambda delay
Seeing the real data is helpful, but lot's of pictures is good.
One thing, when the event occurs, like an quick throttle change, there is often a movement for the Lambda sensor where it goes up, then down. I usually look at that first move as the point to measure to figuring the later down turn is a real lean/rich period that could be worked out with some more AE tuning.
it looks like you are on the right track to me.
Are you really seeing that big a delay at the high map / low RPM sections? To get good data there you probably need a manual trans, I have an automatic with a slippery converter, so I can't make my engine run there
One thing, when the event occurs, like an quick throttle change, there is often a movement for the Lambda sensor where it goes up, then down. I usually look at that first move as the point to measure to figuring the later down turn is a real lean/rich period that could be worked out with some more AE tuning.
it looks like you are on the right track to me.
Are you really seeing that big a delay at the high map / low RPM sections? To get good data there you probably need a manual trans, I have an automatic with a slippery converter, so I can't make my engine run there
Phil Tobin
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
Re: setting lambda delay
Thanks Phil,
I was able to rar the log
I will test those values tomorrow and see if it changes something or not. Yes I have a manual gearbox and yes of course with automatic trans it may be hard to go where you want on the map.
Benjamin
I was able to rar the log
I will test those values tomorrow and see if it changes something or not. Yes I have a manual gearbox and yes of course with automatic trans it may be hard to go where you want on the map.
Benjamin
Re: setting lambda delay
I have also tried to determine the correct lambda delay by using table switch. But I'm not really sure which point I should use from the log. I've drawn lines on the log for every log entry, first just before PW jump. Line 2 is first higher PW and already at the 3th line AFR is starting to fall. Engine is Ford Cosworth 2,0L 16V, injectors approximately 10cm before sylinder, stock exhaust manifold with 20-25cm pipes to turbo and LC-1 12cm after turbo in 3" pipe. And how can you determine lags under 100ms because the time between log entrys are a bit over 100ms??
Here are the log point's in attached log for table swithes if somebody could tell me what is the right delays in these situations?
657,953
669,078
690,734 (screen capture point)
716,797
734,784
740,187
849,500
Thank you.
Here are the log point's in attached log for table swithes if somebody could tell me what is the right delays in these situations?
657,953
669,078
690,734 (screen capture point)
716,797
734,784
740,187
849,500
Thank you.
-
- Super MS/Extra'er
- Posts: 12731
- Joined: Sat Jul 16, 2005 8:07 am
- Location: Moorseville, NC
- Contact:
Re: setting lambda delay
That graph raises an interesting point.
Your WB sensor is not only giving a delay to react to the event, but also has a delay in converging on the new value. Many WB controllers have an averaging or smoothing feature. For Example on an LC-1 there is a response speed setting allowing for:
Instant
1/12 second
1/6 second
1/3 second
What this does is average over the period selected. The value here is that by averaging over the period it makes the AFR more friendly to read with human eyes. So if you have a gauge in the car or viewing a log file the line will be smoother. However, from the ecu's or VE analyze stand point it is adding additional delay for input data and a sluggish reaction.
So in answer to your question. I would probably look at the half way down the slope as the delay putting you at the 4th or 5th line, but if you can disable the smoothing feature of your controller to see sharp movements that would be preferable. Leave the smoothing function for if you are just using it to drive a gauge.
Your WB sensor is not only giving a delay to react to the event, but also has a delay in converging on the new value. Many WB controllers have an averaging or smoothing feature. For Example on an LC-1 there is a response speed setting allowing for:
Instant
1/12 second
1/6 second
1/3 second
What this does is average over the period selected. The value here is that by averaging over the period it makes the AFR more friendly to read with human eyes. So if you have a gauge in the car or viewing a log file the line will be smoother. However, from the ecu's or VE analyze stand point it is adding additional delay for input data and a sluggish reaction.
So in answer to your question. I would probably look at the half way down the slope as the delay putting you at the 4th or 5th line, but if you can disable the smoothing feature of your controller to see sharp movements that would be preferable. Leave the smoothing function for if you are just using it to drive a gauge.
Phil Tobin
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
Re: setting lambda delay
I actually don't have any averaging on the innovate LC-1 and the lambda Lag Factor is 60 with mega. Is this too much?
I think a part of the AFR delay comes from physics because at idle the flow in the engine is very low so at first a lot of the extra fuel sticks to walls or not...? Because the delay is quite different at higher revs for example the first log point at 3000rpm and 38kpa.
But for normal testing the delays should I use reference the point the first higher PW point and delay point where the AFR has gone full way down or up or as you said before the first AFR movement?
Thank you.
I think a part of the AFR delay comes from physics because at idle the flow in the engine is very low so at first a lot of the extra fuel sticks to walls or not...? Because the delay is quite different at higher revs for example the first log point at 3000rpm and 38kpa.
But for normal testing the delays should I use reference the point the first higher PW point and delay point where the AFR has gone full way down or up or as you said before the first AFR movement?
Thank you.
-
- Super MS/Extra'er
- Posts: 12731
- Joined: Sat Jul 16, 2005 8:07 am
- Location: Moorseville, NC
- Contact:
Re: setting lambda delay
Very true. If it was all electrical or the WB controller, the delay would be more or less a constant. What the Lambda Delay table really needs to be is the aggregate delay realized in the afr TS sees. I would still be inclined to start with the mid point of that reaction or lean toward later closer to the point of stabilization. It would be interesting to see a graph just like that with the PW increasing for a number or records then brought back down, see if it still climbs toward the stabilization point before curving back to the original stabilization point.think a part of the AFR delay comes from physics
It might be worth playing with the Lambda lag to see if that makes it react quicker. That may very well have impacts on the firmwares correction algorithm too though.
Phil Tobin
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
Re: setting lambda delay
I played a bit more with the delays, now the lag factor is 100 so no lag with MS. Noticed that over 2500rpm any kpa the lag is quite low, hard to decide it from the logs, but as you can see at idle the lag is wery big going both ways.
-
- Super MS/Extra'er
- Posts: 1331
- Joined: Mon May 21, 2007 8:36 am
Re: setting lambda delay
60 is probably too much. 90 would probably get rid of 99% of the "noise", but really 100 would be best and work just fine.Perkele wrote:I actually don't have any averaging on the innovate LC-1 and the lambda Lag Factor is 60 with mega. Is this too much?
I think a part of the AFR delay comes from physics because at idle the flow in the engine is very low so at first a lot of the extra fuel sticks to walls or not...? Because the delay is quite different at higher revs for example the first log point at 3000rpm and 38kpa.
But for normal testing the delays should I use reference the point the first higher PW point and delay point where the AFR has gone full way down or up or as you said before the first AFR movement?
Thank you.
I don't really like the MS averaging anyway, it's too slow and it's not linear. I've wanted to try setting the LC-1 to 1/12th of a second and see how it works out.
Re: setting lambda delay
ok, so im trying to set my delay, and im a lil lost..
so heres a place where my PW dropped, as well as my afr.
and heres the afr drop
so would i put ~700 ms in the delay?
here are what my delays are, i didnt set them up, must be defualt
my afr lag is 99
so heres a place where my PW dropped, as well as my afr.
and heres the afr drop
so would i put ~700 ms in the delay?
here are what my delays are, i didnt set them up, must be defualt
my afr lag is 99
-
- Experienced MS/Extra'er
- Posts: 301
- Joined: Tue Jul 27, 2010 6:15 am
Re: setting lambda delay
I have similar HIGH lag watching the logs (around 700ms at idle to 1500rpm, then I don't know engine isn't tuned yet), I need to experiment with lag settings. Will post news next days.
I think that LC-1 is 1 meter far from exhaust valves. Is this enough to justify such high lag? I may need to set it to "instant" to see if they change.
I think that LC-1 is 1 meter far from exhaust valves. Is this enough to justify such high lag? I may need to set it to "instant" to see if they change.
Robert
MS3x 1.4.1 | TS 3.0.28
MS3x 1.4.1 | TS 3.0.28
Re: setting lambda delay
Hi.
My sensor is in the downpipe ~1m from the head. I am using a Tech Edge WBO unit and noticed delays of about 600-650ms at low load/rpms.
For my application the default delay was far too low either. No prob at WOT and high RPMs, but at idle for example this didn't work at all.
Thomas
My sensor is in the downpipe ~1m from the head. I am using a Tech Edge WBO unit and noticed delays of about 600-650ms at low load/rpms.
For my application the default delay was far too low either. No prob at WOT and high RPMs, but at idle for example this didn't work at all.
Thomas
-
- Experienced MS/Extra'er
- Posts: 301
- Joined: Tue Jul 27, 2010 6:15 am
Re: setting lambda delay
nice to know, at least now I know I'm not the only.
But can't work on the car until three or four days, so maybe you'll post more updates and results.
Thomas what is lag from your logs at high rpm, since you have the sensor at my same distance?
(just to know, since my engine isn't tuned and I'd like to use VE analyzer live to tune it.)
But can't work on the car until three or four days, so maybe you'll post more updates and results.
Thomas what is lag from your logs at high rpm, since you have the sensor at my same distance?
(just to know, since my engine isn't tuned and I'd like to use VE analyzer live to tune it.)
Robert
MS3x 1.4.1 | TS 3.0.28
MS3x 1.4.1 | TS 3.0.28
-
- Super MS/Extra'er
- Posts: 12731
- Joined: Sat Jul 16, 2005 8:07 am
- Location: Moorseville, NC
- Contact:
Re: setting lambda delay
It would probably be god to have some sort of wizard to set the default delays.
Phil Tobin
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
EFI Analytics, helping to simplify EFI
Next Generation tuning software.
Supporting all MegaSquirt versions and firmwares.
http://www.TunerStudio.com
http://www.efiAnalytics.com/MegaLogViewer/
Support the firmware running your engine:
http://www.msextra.com/doc/donations.html
Re: setting lambda delay
Hi.
Here are my lambda delays. I didnt figure them out 100%. Just increased the defaults until it was roundabout working ;-)
My "Wizard" was MLV (:
Thomas
Here are my lambda delays. I didnt figure them out 100%. Just increased the defaults until it was roundabout working ;-)
My "Wizard" was MLV (:
Thomas