Is MAT correction possible with MAF tuning?

[shauer]

I am noticing some temperature related tuning issues with my MAF setup. It appears that whatever temperature-based correction the MAF sensor is performing is not completely handling the requirements of my engine. My baseline tuning has been performed between 60 and 70 deg F. This morning was cold (32 deg F) and my engine was running between 5% and 10% lean. I have noticed this several times over the past few weeks.

I know the MAF sensor is supposed to compensate for intake air temp, but I want to be able to manually correct a little bit on top of that. Is the MAT correction curve feature available when using MAF? (http://www.msextra.com/doc/ms2extra/MS2-Extra_Conversion.htm#matcor)

From what I understand, MAF tuning ignores the MAT sensor in the fuel calculation, correct?

My MAF sensor is a Bosch HFM-2 sensor that appears to function correctly in all other aspects, it just is not correcting for MAT as much as I need it to.

[jsmcortina]

That's quite interesting.

My first thoughts are a couple of possible scenarios.
- air density alone isn't the temperature related effect you are seeing
- the MAF (plus transfer function) isn't reporting the real mass air flow to Megasquirt.

James

[shauer]

Yes it is quite interesting and confusing from my point of view. The MAF sensor does indicate slightly higher readings when the ambient air is cold as expected. I can easily see this at idle. My issue is that I would need to correct my calibration curve well beyond the point needed for the same exact mass flow reading when the ambient air is warmer if you follow me. For example, I may idle at about 3.5 g/s at 70 deg F but may idle closer to 3.7 - 4.0 g/s at 32 deg F but I run lean. The calibration curve values between 3.7 and 4.0 g/s would need to be increased to the point that I would be running noticeably rich at a warmer ambient air temp.

I am also in the process of getting another MAF sensor to try. Based on the data sheet, the Bosch HFM-2 has an ambient temperature sensor (thin film resistor) incorporated into the MAF circuit. I guess it is possible I have a bad sensor but it works reliability in all other aspects of operation.

There is no direct data provided in the data sheet regarding temperature response but the data sheet does have an accuracy spec of +/- 4%.

I guess my original question still stands, is the manual MAT correction curve feature available in MAF mode? If the sensor has an accuracy spec of +/- 4% and we are using old, used sensors, then why are we assuming the sensor will be 100% accurate under all operating conditions and not require some manual correction? I'm trying to search and see if OEMs sometimes add some additional correction to the raw MAF reading in their ECUs but it is difficult to locate such information.

[techsalvager]

Yes it is quite interesting and confusing from my point of view. The MAF sensor does indicate slightly higher readings when the ambient air is cold as expected. I can easily see this at idle. My issue is that I would need to correct my calibration curve well beyond the point needed for the same exact mass flow reading when the ambient air is warmer if you follow me. For example, I may idle at about 3.5 g/s at 70 deg F but may idle closer to 3.7 - 4.0 g/s at 32 deg F but I run lean. The calibration curve values between 3.7 and 4.0 g/s would need to be increased to the point that I would be running noticeably rich at a warmer ambient air temp.

I am also in the process of getting another MAF sensor to try. Based on the data sheet, the Bosch HFM-2 has an ambient temperature sensor (thin film resistor) incorporated into the MAF circuit. I guess it is possible I have a bad sensor but it works reliability in all other aspects of operation.

There is no direct data provided in the data sheet regarding temperature response but the data sheet does have an accuracy spec of +/- 4%.

I guess my original question still stands, is the manual MAT correction curve feature available in MAF mode? If the sensor has an accuracy spec of +/- 4% and we are using old, used sensors, then why are we assuming the sensor will be 100% accurate under all operating conditions and not require some manual correction? I'm trying to search and see if OEMs sometimes add some additional correction to the raw MAF reading in their ECUs but it is difficult to locate such information.

yes they do still keep IAT around for spark and fuel triming. upload a log showing your issue.

[jsmcortina]

I'm pretty sure that all of the airdensity calcs are disabled as MAF was (incorrectly) assumed to accommodate all of that.

Maybe I'm jumping to conclusions, but it would seem that your findings could also lead us to an explanation of why the speed-density air-density calcs "don't work" - because they are only including the ideal-gas law of the air density - not other factors in the engine that are also impacted. (Such as the cooling effect due to fuel vaporisation or heat injection from the manifold.)

James

[S.Bretz]

Those are just a few areas for error James...as Im sure you know.

OEMs haven't dialed the equations in all the way either. That is why they monitor the o2 and have short and long term fuel trims to help compensate.

Heck...just changeing the gas could mess could slightly alter the a/f at cold operating temps with all the different detergents and junk added these days. I'm sure those additive alter the reed pressure slightly and thus the volitility/vaporation rate, which would mess with the a/f ratio.



To the OP...have you messed with the geometry of the MAF in any way...ie, shaved or removed and air foil or damaged the screen on the inlet? Any tampering with the airflow will cause different readings to the element. Small changes could effect the airflow at differently at different densities resulting an error in tuning. It would be more evident if it was tuned at one extreme ( really hot) then expected to operate at the other (really cold).

[shauer]

I did another test today. I can post log files if you really want them but there isn't that much to really show except the amount of EGO correction required.

I drove into work this morning, 40 degrees F. I ran VEAL to get within a few percent on my tune. EGO correction was running between +/- 5% depending on where I was running on my MAF curve.

I drove home this afternoon, 70 degrees F. I kept the VE table where I set it in the morning. I was running rich across the entire operating range. EGO correction was pegged at -5% (my current limit) the entire drive and I was still not reaching my leaned-out cruise AFR completely.

Could I have a flaky MAF sensor or some layout problem in my intake? Sure I could, but it would be really convenient if I could compensate for it with the non-linear MAT correction curve.

I did an experiment and tried the non-linear MAT correction and it had no effect as expected due to MAF tuning. Is it possible to enable just this manual correction curve while in MAF mode? You obviously don't want the entire MAT algorithm to be enabled as the MAF sensor is supposed to be compensating for ideal gas law but some amount of manual correction capability would be welcome.

S. Bretz -

My MAF sensor installation is 100% custom as my car was originally produced with Bosch kjet mechanical injection. I have a custom made intake plenum connecting 40mm ITBs. I have not modified the actual MAF sensor at all. I have been as careful as I could to place the MAF sensor in a similar way as I have seen OEM installations. I'm sure it is not completely ideal and could be contributing to the non-linear behavior I am seeing. For the time being I am using EGO correction with my wideband O2 to correct the mixture. I need between 5% and 10% correction by the looks of it.

[techsalvager]

yeah I would like to look at the logs for other things as well.
Plus you have a N\A vehicle so I would like to see how its working on your N\A car.
Thanks

I take it you are still using mafmap mode for fueling?

[muythaibxr]

Yeah, it seems like something might be wrong there because as I understand it, the MAF output is supposed to "just work" as far as air density due to temperature change goes. I kinda wonder if your MAF is broken somehow.

As far as air density in speed density mode goes, we're researching ideas to try to make that better.

Ken

[billr]

Well, no doubt this is going to "ruffle some feathers", but... I think MAF is an expensive farce. For all the reasons mentioned above, a pound of air isn't processed the same in an engine at low rpm and high MAP as it is at high rpm and low MAP. The VE table that we tune isn't just accounting for the actual cylinder-filling characteristics, but is also trimming for the other factors.

[elaw]

Just to toss out my two cents, and I think I've said a lot of this here before, but I've seen the same thing as the original posted. Funny thing is, I've experienced it both with speed-density and MAF. With my setup it seems most pronounced at idle - at mid and high load the tune seems pretty stable. But at idle it seems like I have to retune seasonally.

One thing that could be at least a partial explanation in my case is I'm using the factory PCV setup on my engine. So at idle, there is some "stuff" (air, combustion products, whatever other gases are in the crankcase) coming into the engine via a path that bypasses the MAF sensor. I'm also using evaporative canister purge, but that's not active at idle. Canister purge is one place I could easily see a seasonal variation, as there will be much more evaporation from the fuel tank in warmer weather.

[shauer]

Just to toss out my two cents, and I think I've said a lot of this here before, but I've seen the same thing as the original posted. Funny thing is, I've experienced it both with speed-density and MAF. With my setup it seems most pronounced at idle - at mid and high load the tune seems pretty stable. But at idle it seems like I have to retune seasonally.

One thing that could be at least a partial explanation in my case is I'm using the factory PCV setup on my engine. So at idle, there is some "stuff" (air, combustion products, whatever other gases are in the crankcase) coming into the engine via a path that bypasses the MAF sensor. I'm also using evaporative canister purge, but that's not active at idle. Canister purge is one place I could easily see a seasonal variation, as there will be much more evaporation from the fuel tank in warmer weather.

My variations are also most pronounced in the low-flow regions of operation as well. Mainly idle like you mentioned and also during cruise where the revs are higher but the throttles are almost closed. Under mid/high load, it appears very stable across the temperature range. Just a thought, can air temperature effect the turbulence of the air entering the MAF? I know turbulence in the intake will have a more pronounced effect in the lower-flow operating regions.

I will be the first to state that MAF is not perfect but I can say that I have run SD, AN, blended, and ITB modes on this engine and it has never run better than it is now with MAF. MAF has solved most of my problems caused by ITBs and a high-overlap / high-duration cam.

I do not believe for a minute that the MAF sensor can correctly compensate for all temperature related factors across all types of installations without a little help. The MAF sensor can only correct for what it can see. Other temperature related effects outside the "view" of the MAF sensor still need to be accounted for. I'm not sure what the best solution is but was thinking that enabling the manual MAT correction curve was as good a place to start as any.

[sjf911]

I would think the thermodynamics of airflow over a hot wire would dictate that hotter air of the same mass flow would cool the wire less than cold air of the same mass flow yielding leaner AFR's at hotter temps. Do we know if any OEM's use temperature correction?

[prof315]

Well, no doubt this is going to "ruffle some feathers", but... I think MAF is an expensive farce. For all the reasons mentioned above, a pound of air isn't processed the same in an engine at low rpm and high MAP as it is at high rpm and low MAP. The VE table that we tune isn't just accounting for the actual cylinder-filling characteristics, but is also trimming for the other factors.

It certainly doesn't ruffle my feathers. As a mechanic with close to 30 years experience I have learned to loathe MAF. It's a great idea....... in theory. Reality says they tend to be more than a little twichy, easily subject to contamination by oil, dirt and dust that can cause either inaccurate measurement of the air flow or (ususally) failure. Even granting the fact that far more of the cars I work on are equipped with MAF vs MAP, the failure rate is well over 10 to 1 proportionately in favor of a MAF failure.

[S.Bretz]

I would think the thermodynamics of airflow over a hot wire would dictate that hotter air of the same mass flow would cool the wire less than cold air of the same mass flow yielding leaner AFR's at hotter temps. Do we know if any OEM's use temperature correction?

The MAF's i deal with on a daily basis ( Cadillac, Saturn, sometimes hummers and buicks) have the IAT incorporated into the MAF assembly. Though, I have not looked into the operation so I can't tell you if they use a seperate sensor or not. I do know that the first or last wire on the maf connector (I tihnk its yellow) has something to do with it. They tend to break right at the connector for some odd reason and the ECM throws a inlet air temp code and the reading with the tech2 scanner will read -40, which leads me to assume its a seperate sensor, but I cannot be sure at this time with no investigation into the fuctionality of the circuit.

[muythaibxr]

OK. I will re-allow MAT correction with MAF on an optional basis.

It is in the TODO for 1.1.

Ken

[Paul_VR6]

The VWs I deal with all have maf and a temp sensor. Sometimes separate, sometimes as part of the maf sensor but output on a separate wire. Its hard to say if its used for non fuel related calcs (iat timing retard, etc) or its to complement the airflow sensor.