PDA

View Full Version : AFR READOUT Flashscan Vs LM1



johnv
August 7th, 2005, 08:28 AM
Hi all,

I have an LM1 wideband and find the AFR readings from the controller display are quite steady while the readings in flashscan tend to jump around alot.
I have set it up using the LM1 pids in flashscan, AFRs at idle seem pretty close between the two, but flashscan reads about 4-5 points higher at WOT.

Any thoughts !!
Cheers
John

mistermike
August 8th, 2005, 01:55 PM
The innovate WB's have a very fast response time. So fast that they can differentiate individual exhaust pulses. My LC-1's have an internal setting to dampen the response. It might be worth investigating the inner workings of the LM-1 to see if the response time of the built in display is different than the analog output.

Highlander
August 8th, 2005, 03:00 PM
I have seen this, but on logged data.. the display is damped.. it will not update as fast as the data logger does. but would like to see more experiences on this.

Steve Bryant
August 9th, 2005, 05:45 AM
You actually need some damping or averaging effect (like the averaging function in the maps) in order to make the AFR data meaningful. For instance in you log AFR for 50 kPa and 3,000 RPM's, in the map function and you have an event count of 50 (i.e. n = 50); you might have an average of 14.83, a minimum of 14.22 and a maximum of 15.96. It's not the max or the min or any specific event that you need to look at, it's the average, which is running a little lean compared to Stoichiometric. If you are doing speed density tuning and your goal is Stoic, you need to increase the VE cell value proportionately according to the average.

Steve

johnv
August 11th, 2005, 01:40 PM
So as long as we get a high enough event count the average will be accurate enough for SD tuning ?

Highlander
August 11th, 2005, 03:47 PM
I am having problems with sd tuning since i get it all near 1.0 but when i go back to CL i get all screwy.

Steve Bryant
August 11th, 2005, 03:59 PM
So as long as we get a high enough event count the average will be accurate enough for SD tuning ?

Yes, that's correct. It is important to try to get steady state readings as much as possible where the AFR for a given cell is logged with a fairly constant MAP and RPM instead of a highly transitory state where you are really changing the throttle position and RPM's rapidly.

Steve

Steve Bryant
August 11th, 2005, 04:07 PM
I am having problems with sd tuning since i get it all near 1.0 but when i go back to CL i get all screwy.

When you say that you are getting the AFR all near 1.0, I don't exactly understand what you are meaning. Also, when you go back to closed loop, do you have the MAF enabled or disabled?

Here is the way that I have found to work the best:
1. Retard the timing to minimize Knock issues.
2. Disable Closed Loop, MAF and DFCO.
3. Log data for as many cells as possible in the VE Table.
4. Adjust the VE values according to the proportional error that the measured AFR is from Stoic (14.68).
5. Repeat steps 3 and 4 several times until the AFR is starting to accurately give measured AFR's that approximately equal commanded AFR.
6. and after, I'll write some more on this tomorrow.

Highlander
August 11th, 2005, 04:44 PM
1.0 i mean ben factor


I do not use the maf... yet...

I am doing everything that way.. either my afr readings are incorrect in flashscan which i am checking ... OR.... my stock O2 sensors are shot...

will check that out tommorrow...