john banks

Since the K&N won't fit with the APS FMIC, I thought I would fit this instead, since I am running out of options which will not toast the MAF sensor.

Impressed with the fit and finish and the vibration protection for the MAF sensor.

Sound is good too - not too intrusive like a 57i although my exhaust is louder than previously. I also gain on top end boost by about 1 PSI.

However, the drive was really odd. Lots of +ve airflow correction at closed loop (over 20% in places), and at warm idle about +3% average compared with previously -3% on the airbox. At 5 PSI it was incredibly lean and you could actually feel the car losing torque as it came out of closed loop. Very weird.

I did logs before on the airbox to try and get a rough calibration at idle, 50 mph cruise, 5,10,15,20 PSI in various gears, and have spent the evening going through them all.

The long and the short of it is a 0.3V offset in the MAF voltage at high loads, 0.2 V at moderate loads, and 0.5-1 V at low loads. I will rescale the MAF calibration map to see if my base map can become sensible again.

Like others I found it was about 2 AFR points leaner.

I think it will be a good setup, but will need a lot more work to get a base calibration map for the MAF sensor in this setup. Presumably the more laminar airflow causes the under-read.

WREXY

Hi John,

Excellent stuff m8! I mean, that you are getting the APS kit sorted with your car, just in time for the mapping session with my car. Glad your going with the APS kit and hope it works out safely.

Have you compared my logs to your logs I sent you the other day, with the APS kit which I already had on the car and the VF23 which I fitted? If so, were they similar?

Cheers,

Wrexy.

[Edited by WREXY - 9/15/2002 1:34:20 AM]

Bob Rawle

John, that stacks up with what I have established, another P1 yesterday with a massive offset in maf output, I think this confirms that the CAK is not suitable for any car running maf without a significant remap. Its not due to a more laminar flow though as far as I can tell to date, I am hoping I can test it specifically for flow distribution in the relatively near future, the maf output "jitters" significantly indicating turbulant flow not "laminar" flow. Since changing mine to the HKS I have leaned the cruise/low flow areas but had to richen up the medium/high flow areas, also maf output is significantly less "jittery" and much more stable at any given air flow.

It can me mapped around and you shouldn't need to rescale maf more than normal, but then obviously the fuel map needs specific attention almost everwhere.

It is important for users to be aware that once the ecu has been remapped to suit this filter then no other can be used without another significant remap. (or blowing the ecu back to std if reverting the car) So once on then don't remove without discussing with someone (your tuner) first.

john banks

Thanks Bob. I'll have a look through and look at the smoothness. Am I hearing you right than on full load the HKS was leaner?

The ideal I suppose would be that APS would supply a calibration map from a flow bench saving us doing their R&D

The appeal to me of changing the MAF calibration is that this would hopefully make it hit all the right load zones on a map I know is pretty near, so it would also sort the ignition, and is solving the problem at source - ie the MAF calibration in the new "system". If this doesn't work then I will attack all the zones individually.

Converting the MAF calibration map to decimal and plotting it gives an exponential curve (well you can flatten it by taking logarithm anyway all the way from no airflow at 1.02V). Looking at other MAF sensors on the internet the same appears to be the case ie. MAF is proportional to X^(Y(voltage output-nil flow offset)) where X and Y are constants.

The offsets I am getting appear to be zero at no airflow rising linearly with voltage to 0.3V at high airflow but I need to collect some more intermediate load points to compare.

At least I can use the airflow correction from the lambda sensor to give me an idea when I am getting near for the closed loop zones.

Do others find that the standard setup runs mainly negative air flow correction at idle and cruise? (So a failsafe if the lambda sensor goes)

[Edited by john banks - 9/15/2002 10:59:14 AM]

Bob Rawle

Yes, needed more fuel at higher air flows, that could be a function of my specific setup (bigger injectors perform differently to smaller ones over a range of duty cycles for example)or even just filter area differences.

I see typically the airflow range compressed into the more linear part of the maf curve, 0.3v offset on the steep part of the curve is a lot of air flow.

Pavlo

I wouldn't be surprised if there isn't some pulsing effects to the long uninterupted pipe.

I wonder if you are not seeing air moving back and forth in the tube and MAF sensor, which of course is going to be unstable, and very variable with RPM.

One way to perhaps smooth it out would be a small resonator box somewhere in the system. Or possibly a small annular lip before the MAF, similar in function to a rear lip spoiler, creating a fixed point of seperation, regulating the flow.

The instability sounds like vortices being shed off the main bend, which would be cyclic +ve and -ve flow in the MAF sensor. The stock airbox, and even a 57i would have a more direct flow, possibly with with more turbulence, but in a predictable nature, giving poorer flow but with a constant nature.

All just supersition, but I am sure some of it applies.

Paul

john banks

On my car at least, the MAF signal seems pretty smooth - like the original goes pretty smoothly up with only the very odd drop and then rise again.

Working on a few curves. Graphs below. Series 1 is the new calibration curve Series 2 is the original. MAF voltage scale is 10mV/unit. Load scale is 0-65535 (16 bit). I have also shown them logged - I took the points I had measured the deficits and filled in with a straight line in between on the log graph and then converted back. Will let you know the real results on the car - theory is all well and good! At least it should be driveable now I hope.

This is the original calibration curve logged - so it seems exponential all the way through after the no flow areas below 1V (100 on X-axis).



New vs old MAF calibration - difference looks very small - it isn't - about 20% fuel in areas!


...and logged:

john banks

Works great. Closed loop and open loop were fine and going through a mountain of logs suggest it is just about right. Nice and smooth now - low down torque recovered. Very much fun at the top end Now I will just pray daily for my MAF sensor

If anyone wants the calibration map let me know and I'll email it to you.

[Edited by john banks - 9/15/2002 5:12:14 PM]

WREXY

Excellent stuff John! Bring on Saturday.

As for your MAF going on you, mine has lasted with the APS CAK and stock ECU for over 18000km and still going strong.

Cheers,

Wrexy.

[Edited by WREXY - 9/15/2002 5:48:52 PM]

Bob Rawle

I'm refering to instability about "points" not the general curve, the data log won't show it to you. These are millivolt differences. Noticeable difference between filters though and I've now seen the same thing on two other cars in addition to mine. I run with an additional afr meter in the car in addition to the Lambdalink, it covers a wide range and is very sensitive to high afr's, you can see this meter jittering as if it was runing closed loop at 0.3 bar boost. The lambdalink just expands two led positions on the extra meter into its 9 led's.

Several induction kit manufacturers supply an aluminium "gasket" that is shaped with minute tabs to ensure the airflow thro the map is as representative as possible. they "steady" the air flow, I also believe that reversion pulses back into the maf due to the filter intake pipe design are causing it to misread, all three cars so far have generated historic maf connection fault codes with this filter, on my car these codes stopped appearing when I reverted to the HKS.

Regarding tha actual amount of air flowed then you can only rely on the oxy sensor reading as the maf output is so different, thats what I based my lean/rich observations on. Until some proper flow tests are done then its impossible to be precise.

WREXY

Agree with Bob about the MAF sensor appearing as a historic trouble code every so often on Delta Dash, i.e. if that's what he means. I get them too. But I've seen them when I've either undone the airfilter MAF housing, twice, and again the other day after the VF23 was fitted, although I had not been near the airfilter that time.

Perhaps it was coincidence and I thought it was due to the above reasons. Maybe it just keeps coming up. I'll need to keep an eye out.

Cheers,

Wrexy.

Bob Rawle

Wrexy, yes spotted with Deltadash.

john banks

Thanks for the pointers gents.