Pages: [1]
Author Topic: M3.8 HFM2 to HFM5 MAF conversion  (Read 604 times)
marantzvieta
Full Member
***

Karma: +3/-1
Offline Offline

Posts: 84


« on: June 11, 2025, 07:29:52 AM »

Hi all,

As the title says I am working on a HFM2 to HFM5 maf conversion in a M3.8.3 ecu PN 06A906018CG for a 1.8t DBC. Found out that early VR5 engines came out with M3.8.3 ecu (071906018) and HFM5 2,75" maf from the factory, so we have a OEM starting point to make the conversion.

I first made the pinout and new connector (already covered on other posts in the forum, wiring is ok with 12v, extra 5v pin added, ground, and info), so this is correct.

Then copied the original MAF curve from VR5 ecu to 1.8t ecu and started the engine, but it ran badly.

I suspect there are other things to be copied from original VR5 map like KHFM and MLOFS (MLOFS for sure, FR says to use MLOFS 0 for HFM2 and MLOFS 200 for HFM5), but I belive KHFM will mess up load values and timing, so maybe only MLOFS is needed?

Is there anybody with experience in this? Just asking before trying to run it again because in my M3.8.2/3 damos MLOFS is not listed and I have to guess it's location based on deduction, and I don't want it to not run properly and not know if it's because I need more tweaking or because the MLOFS location deduction is not correct.

I believe MLOFS is in 0646A in the ori file I uploaded as I have seen in a VR6 M3.8.1 file that MLOFS is quite near KHFM. Anybody can confirm?

Thank you so much for your time.

Best regards,
« Last Edit: June 11, 2025, 10:13:36 AM by marantzvieta » Logged
gremlin
Hero Member
*****

Karma: +216/-10
Offline Offline

Posts: 731


« Reply #1 on: June 12, 2025, 03:51:20 AM »

I believe MLOFS is in 0646A in the ori file I uploaded as I have seen in a VR6 M3.8.1 file that MLOFS is quite near KHFM. Anybody can confirm?

This assumption is unlikely. In firmware 071906018, the similar cell  near KHFM is also equal to zero.
Logged
marantzvieta
Full Member
***

Karma: +3/-1
Offline Offline

Posts: 84


« Reply #2 on: June 13, 2025, 12:37:32 PM »

This assumption is unlikely. In firmware 071906018, the similar cell  near KHFM is also equal to zero.

Hi Gremlin, thank you for having a look at it. I belive we may not have the same address for KHFM because in my case, 07387, i have no zeros around it. Also I do not have a DAMOS or map pack for 071906018, so my KHFM is an assumption too.

I have deducted that a value of 200 in hex may look like C8 00 from changing the value 0 to 200 in MLOFS map in the VR6 M3.8.1 damos, so looking in the VR5 file for C8 00 I have found the first one seen in the picture, but sincerely, I am not sure it may be correct...

I am a bit lost on this, does anybody have a bit more insight?

Thank you so much

Best regards

Logged
gremlin
Hero Member
*****

Karma: +216/-10
Offline Offline

Posts: 731


« Reply #3 on: June 16, 2025, 05:05:27 AM »

The original HFM calibration map in firmware 071906018 already contains negative values ​​(reverse flow detection).
For this calibration method, MLSOF is not required and is set to zero.
Logged
marantzvieta
Full Member
***

Karma: +3/-1
Offline Offline

Posts: 84


« Reply #4 on: June 16, 2025, 07:27:24 AM »

The original HFM calibration map in firmware 071906018 already contains negative values ​​(reverse flow detection).
For this calibration method, MLSOF is not required and is set to zero.

Gremlin you are very right, I had the assumption but you confirmed it. It is curious because M3.8.1 damos, MLOFS is present and has value 0, and offset in MLHFM map is 0 too. But in M3.8.2 there is no MLOFS in damos but offset is -200 in MLHFM although the sensor is HFM2, just like the 071906018 software we are looking at that has HFM5 sensor.

So, once we know we do not need to touch MLOFS and I have the HFM5 MLHFM data copied into the 1.8t file:

- I suppose I need to adapt KHFM from the VR5 bin to the 1.8t, correct?  Will I need to redo any maps? Or load values will be correct for the new MAF with its KHFM?

- Using stock 1.8t injectors may I need to adapt FGAT0 too?

My idea is to be able to use the HFM5 on a stock setup as a base. Only with MLHFM the setup ran very very rich and lumpy.

Thank you so much for your input, I really, really appreciate it.

Best regards!
Logged
marantzvieta
Full Member
***

Karma: +3/-1
Offline Offline

Posts: 84


« Reply #5 on: June 22, 2025, 09:04:03 AM »

Hi all,

I am still dealing with this. I do not understand why once we give the ecu the correct MLHFM (maf linearisation) for the new maf, it is not able to work as it does with the old maf.

I mean: given the already well known formula to calculate theoretical load Tl = Q/(n*Ki) where Q is airflow in Kg/h, n is engine speed in RPM and Ki is the injector constant, why if I change the MAF and use it's good known MLHFM from same ecu type (M3.8.3), the ecu is not able to make the calculations correct?

From this thread: http://nefariousmotorsports.com/forum/index.php?topic=16964.0, our collegue Kompiesto said he had to tune FGAT0 and KHFM too, but considering the formula, it does not make sense.

I did the test using the VR5 KHFM and increasing FGAT0 by 50% and the engine ran way better but not perfect, hence why this still makes no sense to me.

Hope somebody can provide some input. Thank you.

Best regards
« Last Edit: Yesterday at 12:40:39 AM by marantzvieta » Logged
Pages: [1]
  Print  
 
Jump to:  

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Page created in 1.487 seconds with 15 queries. (Pretty URLs adds 0.54s, 0q)