Datalogged AFM values (g/s)

FK2 Civic Type R 2015-2017, FK8 Civic Type R 2017-2021, FL5 Civic Type R 2023+ and Integra Type S 2024+
Post Reply
FK2Flashpro
Posts: 6
Joined: Tue Jul 21, 2020 1:05 am

Datalogged AFM values (g/s)

Post by FK2Flashpro »

Are the datalogged AFM values (g/s) based on a sensor, or is it calculated, or read off the AFM table in the uploaded calibration?

I had assumed the g/s values are read off the AFM table, but looking at my datalogs, it doesn't seem so:
e.g.
From datalog: 280g/s @ 4.82V
From calibration: 260g/s @ 4.82V

I'm running FlashPro v185 for FK2 and FlashPro Manager v3.6.1.

Thanks.
User avatar
Hondata
Site Admin
Posts: 10434
Joined: Thu Aug 15, 2002 12:13 pm
Location: Torrance, CA
Contact:

Re: Datalogged AFM values (g/s)

Post by Hondata »

The ECU might compensate the sensor reading.
Hondata
FK2Flashpro
Posts: 6
Joined: Tue Jul 21, 2020 1:05 am

Re: Datalogged AFM values (g/s)

Post by FK2Flashpro »

Hi, I have another datalog where AFM values will suddenly jump and stay at 303g/s. This happens from around 4.94 (AFM.V) for every run. Is this hitting a limit, hardware issue, or is my MAF really flowing 303g/s?

AFM.jpg
AFM.jpg (52.83 KiB) Viewed 6170 times
User avatar
Hondata
Site Admin
Posts: 10434
Joined: Thu Aug 15, 2002 12:13 pm
Location: Torrance, CA
Contact:

Re: Datalogged AFM values (g/s)

Post by Hondata »

You'd have to ask Honda. I can't explain everything in their ECU.
Hondata
FK2Flashpro
Posts: 6
Joined: Tue Jul 21, 2020 1:05 am

Re: Datalogged AFM values (g/s)

Post by FK2Flashpro »

Understood, although I doubt Honda/Bosch will provide more details than this forum 😊
jpierro79
Posts: 15
Joined: Fri Mar 27, 2020 1:57 pm

Re: Datalogged AFM values (g/s)

Post by jpierro79 »

A MAF sensor and any other sensor for that fact can not read accurately near its limit. 5volt sensors the closer you get to 5 volts the worse the accuracy. That is why you see fluctuations. A MAF sensor reads by heating a wire and the more air goes over the wire the more it requires power to heat it which in turn increases the value the ecu sees. When you are nearing the limit of the Hotwire you will get large fluctuations.
The reason for your difference in grams a second is a simple one. The air temp colder or hotter is compensated to correct fueling. The MAF values are only there to get it close while the Bosch ecu corrects at full closed loop and the value is altered to actual grams instead of expected at a set voltage.
All cars iat temps and ect temps hold sway on airflow values.
FK2Flashpro
Posts: 6
Joined: Tue Jul 21, 2020 1:05 am

Re: Datalogged AFM values (g/s)

Post by FK2Flashpro »

That makes perfect sense, thank you!
Centripetal
Posts: 14
Joined: Sun Mar 15, 2020 11:04 am

Re: Datalogged AFM values (g/s)

Post by Centripetal »

This is all speculation because I did not work for Honda. In automobiles, analog sensors have diagnostic levels. With a 5V sensor like the MAF, generally the diagnostic levels are < 0.5V and > 4.5V. This is part of the safety system design (ISO-26262). If the sensor measurement wire (analog voltage) faults to ground, the voltage will be less 0.5V. If the sensor measurement wire faults to either power supply line, the voltage will be greater than 4.5V. While the diagnostic to catch the later fault is disabled, there is likely some logic in there to go to a safe value so more fuel is added. This measurement is filtered, which is why you see a few measurements between where it pegs at 303 g/s,

The fluctuation reasoning is without merit. The sensor is supplied 12V so that does not happen. Consequently, the reason it has less precision (not accuracy) is the slope of the line at that point. Think of it like steps. Each step you have is a measurement of height. Steep steps will have less steps and therefore less measurement points of height. Whereas, if you have gradual steps, you will have more steps for the same height and therefore more measurements (higher accuracy).
FK2Flashpro
Posts: 6
Joined: Tue Jul 21, 2020 1:05 am

Re: Datalogged AFM values (g/s)

Post by FK2Flashpro »

Centripetal wrote: Wed Feb 03, 2021 7:20 pm This is all speculation because I did not work for Honda. In automobiles, analog sensors have diagnostic levels. With a 5V sensor like the MAF, generally the diagnostic levels are < 0.5V and > 4.5V. This is part of the safety system design (ISO-26262). If the sensor measurement wire (analog voltage) faults to ground, the voltage will be less 0.5V. If the sensor measurement wire faults to either power supply line, the voltage will be greater than 4.5V. While the diagnostic to catch the later fault is disabled, there is likely some logic in there to go to a safe value so more fuel is added. This measurement is filtered, which is why you see a few measurements between where it pegs at 303 g/s,

The fluctuation reasoning is without merit. The sensor is supplied 12V so that does not happen. Consequently, the reason it has less precision (not accuracy) is the slope of the line at that point. Think of it like steps. Each step you have is a measurement of height. Steep steps will have less steps and therefore less measurement points of height. Whereas, if you have gradual steps, you will have more steps for the same height and therefore more measurements (higher accuracy).
Interesting points and I too wondered if ecu logic is a factor, as much as hardware limitation and temps will be the obvious ones.

Here's a run which shows the AFM.v (blue) hitting 5V limit just before 17:27, which led to a big step up in AFM g/s (red). It stayed at 5V limit, but the g/s is not pegging to a fixed/safe max value and A/F is also trending lean... possible that the ecu switched to a different logic for measuring load?

datalog.png
datalog.png (18.18 KiB) Viewed 2704 times

One thing sure though, the AFM g/s values below 5V will indeed vary slightly from calibration values, based on the temps.
Centripetal
Posts: 14
Joined: Sun Mar 15, 2020 11:04 am

Re: Datalogged AFM values (g/s)

Post by Centripetal »

FK2Flashpro wrote: Sat Feb 06, 2021 9:13 am Interesting points and I too wondered if ecu logic is a factor, as much as hardware limitation and temps will be the obvious ones.

Here's a run which shows the AFM.v (blue) hitting 5V limit just before 17:27, which led to a big step up in AFM g/s (red). It stayed at 5V limit, but the g/s is not pegging to a fixed/safe max value and A/F is also trending lean... possible that the ecu switched to a different logic for measuring load?

datalog.png

One thing sure though, the AFM g/s values below 5V will indeed vary slightly from calibration values, based on the temps.
If you share the calibration and datalog, I can comment a little more. It looks like your g/s edges up to the pegged value. Just like previous poster... As I said before, the transition from real measurement to pegged value is likely from filtering. A square wave with a low pass-filter will have a slope instead of straight shoot up.

Yes, the g/s will vary from temperature changes. That's why the MAF has a temperature measurement in the same place. Ideal gas law reveals that for a given volume and pressure there will be more gas mass at lower temperatures: n = PV/RT .
FK2Flashpro
Posts: 6
Joined: Tue Jul 21, 2020 1:05 am

Re: Datalogged AFM values (g/s)

Post by FK2Flashpro »

Centripetal wrote: Sun Feb 14, 2021 10:00 am
FK2Flashpro wrote: Sat Feb 06, 2021 9:13 am Interesting points and I too wondered if ecu logic is a factor, as much as hardware limitation and temps will be the obvious ones.

Here's a run which shows the AFM.v (blue) hitting 5V limit just before 17:27, which led to a big step up in AFM g/s (red). It stayed at 5V limit, but the g/s is not pegging to a fixed/safe max value and A/F is also trending lean... possible that the ecu switched to a different logic for measuring load?

datalog.png

One thing sure though, the AFM g/s values below 5V will indeed vary slightly from calibration values, based on the temps.
If you share the calibration and datalog, I can comment a little more. It looks like your g/s edges up to the pegged value. Just like previous poster... As I said before, the transition from real measurement to pegged value is likely from filtering. A square wave with a low pass-filter will have a slope instead of straight shoot up.

Yes, the g/s will vary from temperature changes. That's why the MAF has a temperature measurement in the same place. Ideal gas law reveals that for a given volume and pressure there will be more gas mass at lower temperatures: n = PV/RT .

Here's the datalogs for 2 runs up the same gentle slope. I prefer not to share the entire calibration though. Notice that in WOT2, there are still incremental g/s values when AFM.V is pegged to max:

WOT1.fpdl
(18.4 KiB) Downloaded 102 times
WOT2.fpdl
(19.44 KiB) Downloaded 116 times

The low pass filter will send a clean max V signal to the ECU. It is up to the ECU to interpret and churn out the g/s values as a step-up, slope/ramp, or both. This interpretation at max V and possible other sensor inputs (MAP?) are what I'm after...

Appreciate your inputs.
Post Reply