Immo3 mfa cluster fuel consumption reading adjustment

2.2TDI · May 28, 2020

Has anyone who has an mfa cluster or done the swap on their car managed to change the mfa accuracy without touching IQ? I know pre 02 cars, which happen to be Immo2 have channel 3 available in adaptation under 17-instruments, but Immo3 clusters don't

I'm sure somewhere in the eeprom dump it can be adjusted, just haven't managed to find where. Since my car is tuned, the fuel consumption is grossly optimistic. I just filled up and hand calculated I got 5.6l/100km and the mfa read 4.3..that's almost a 25% error which is alot... I've seen 10% as a standard, but 25 is nuts...

It's not a big deal, but it kind of renders part of the mfa useless if the fuel consumption is off by that much

Nero Morg · May 28, 2020

You could try a different bin dump. I don't remember which one I used, but the cluster is always within 2-3mpg of my scangauge II. Also, I have a FIS cluster from a W8, don't know if that makes it more accurate or not.

2.2TDI · May 29, 2020

To be honest I don't think a different dump will make a difference, as the byte address where fuel consumption is located is probably the same in every dump

gmenounos · May 30, 2020

See here:
https://forums.tdiclub.com/showthread.php?t=427636

2.2TDI · May 30, 2020

gmenounos said:
See here:
https://forums.tdiclub.com/showthread.php?t=427636

Thanks, I'll give it a try, cheers

2.2TDI · Jun 8, 2020

gmenounos said:
See here:
https://forums.tdiclub.com/showthread.php?t=427636

I think this may have done the trick, from the short trip I've done so far the consumption seemed on par, but I'll know more accurately after a full tank which will be in some weeks to come

Nuje · Jun 11, 2020

On the IMMO3 cluster, editing byte 0x145 is where you can change the fuel economy reading on the MFA.

Default is 64 (100 in decimal); IIRC, changing to 69 (decimal 105) results in it reading 5% more fuel consumed. Changing to 5F (decimal 95) would make it read 5% less.

2.2TDI · Jun 12, 2020

Yea that's what I changed, I put mine to 73 which is 10 percent higher, but it was reading 20% lower before the change so I'm trying 10% as a baseline to see what the difference will be now after a full tank

Nuje · Jun 12, 2020

So, just to be sure I have it correct.
Let's say it had been (incorrectly) reading an average of 5.0L/100km; by changing to 73, it's now reading 5.5L/100km, right?

(I always get question myself as to what upping the byte value does....extra confused because while our civilized "L/100km" measure makes the larger number indicate more fuel is being used, the barbaric mpg measure that most on the forum use makes the larger mpg number indicate less fuel is being used.)

2.2TDI · Jun 12, 2020

Nuje said:
So, just to be sure I have it correct.
Let's say it had been (incorrectly) reading an average of 5.0L/100km; by changing to 73, it's now reading 5.5L/100km, right?

(I always get question myself as to what upping the byte value does....extra confused because while our civilized "L/100km" measure makes the larger number indicate more fuel is being used the barbaric mpg measure that most on the forum use makes the larger number indicate less fuel is being used.)

Yea the larger number increased the consumption, theoretically 64 to 73 would be 10% so yes 5.0 would turn to 5.5 but I have no way of knowing exactly

But to put it into perspective, my mfa read 4.3 on a full tank and my hand calculated was 5.6, which was a gross error of 23%. Currently I'm showing 5.0 or 5.1 on this tank which already a lot better of a correction
Probably in about a week or so I'll fill up and I'll see how much error there still is

2.2TDI · Jun 18, 2020

So my error this time around was only 11% versus 23ish on my previous fill, which makes sense because I only changed it from 100 to 115, so now I'll change it to 125 which is 7D in hex and that should do the trick hopefully

Nuje · Jun 18, 2020

Thanks for reporting back on the effect. Good to know there seems to be a linear direct relationship.

2.2TDI · Jun 18, 2020

Nuje said:
Thanks for reporting back on the effect. Good to know there seems to be a linear direct relationship.

No problem, i'll report back after this tank to see if 125 gets it close to hand calculated... I expect it should, but i'm also wondering if there is a limit on how much you can correct. I know when I had my mk7, via vcds (or obdeleven in my case), you could only change the mfd fudge factor within 10% (or 15%, can't remember now)... that wasn't a problem though as most mk7's read by about 5-10% lower then actual

This is nit picking small details at the end of the day, but if you like having stats at your fingertips I guess every 0.1 L/100km counts

Nuje · Jun 18, 2020

Yeah - now that you mention it, I seem to recall a limit of +/-15....even with the Mk4.

I'll be interested to see if further adjustment makes any impact.
(Which is purely academic at this point for me; with my MFA in the "new" '03, my first fill-up calculated out to 5.6L/100km on paper, while the MFA reported 5.4L/100km. It was a "small" fill-up, so I'm gonna wait another couple thousand km before making any adjustments, if any.)

2.2TDI · Jul 7, 2020

So filled up today, hand calculate 5.97L/100km, MFA 5.4, so still a 10% error... maybe it'll learn over time but I doubt it... still better then 25% error though

Nuje · Jul 7, 2020

So, looks like the MFA adjustment is still capped at +/-15%.

Couple fill-ups in, looks like my car is under-reporting about about 12%, so i should be able to get it dialled in correctly. Probably going to wait until I get the .205 injectors and tune, though - I don't like repeatedly writing the EEPROM (unless there's a way to over-write just one byte that I'm not aware of...?).

2.2TDI · Jul 7, 2020

Nice, at least you can get it almost exact. Funny that on my previous golf, the mfa was pretty accurate versus my hand calc... don't understand why this one is so off but anyways, this is a small issue to complain about

And as far as I know, you always have to upload the entire eeprom file to the cluster again, annoying but that's the name of the game with old technology. Like you, I had a MK7 golf TSI, and coding that thing was so easy, did everything off my phone with obdeleven...

Nuje · Jul 7, 2020

You mentioning "old technology" made me think...going even older (DOS), I wonder if VDS-PRO would work to just change one byte (?).
I mean, we're changing one byte at a time in activating the hatch-unlock and key fob window roll-down, so....if we can access that module with VDS-PRO, it should work.

(I think my afternoon activities / priorities just got re-arranged.

)

2.2TDI · Jul 7, 2020

Nuje said:
You mentioning "old technology" made me think...going even older (DOS), I wonder if VDS-PRO would work to just change one byte (?).
I mean, we're changing one byte at a time in activating the hatch-unlock and key fob window roll-down, so....if we can access that module with VDS-PRO, it should work.

(I think my afternoon activities / priorities just got re-arranged. )

Indeed that is really interesting thought. I've never used VDS PRO but have heard of it. Let us know how it goes, i'd be curious to know

gmenounos · Jul 7, 2020

Nuje said:
I don't like repeatedly writing the EEPROM (unless there's a way to over-write just one byte that I'm not aware of...?).

These things are spec'd for 1,000,000 writes: http://ww1.microchip.com/downloads/en/DeviceDoc/21202j.pdf
(I don't know if that's the exact 2402 part that is used, but I've looked at 3 different datasheets and they're all 1 million writes).

Also, there is a way to just update certain bytes in the cluster's EEPROM but I've only seen it used when VAGTacho and similar tools update the odometer. It could definitely be used to change the fuel consumption fudge factor, I just don't think any program currently supports just updating that memory location.

2.2TDI · Jul 13, 2020

well, a little better. 5.5 MFA versus 5.89 hand calculated, so who knows, maybe the cluster does adjust, will just have to continue monitoring and seeing the progression

Nuje · Jul 13, 2020

gmenounos said:
These things are spec'd for 1,000,000 writes: http://ww1.microchip.com/downloads/en/DeviceDoc/21202j.pdf
(I don't know if that's the exact 2402 part that is used, but I've looked at 3 different datasheets and they're all 1 million writes).

Yeah - it's not that I was worried about the durability of writing to the EEPROM repeatedly, more just the hassle/waiting to read then write all 2000 bytes every time I want to make a little tweak.

I did try connecting with VDS-PRO the other day, and while VDS would connect to module 46 (Convenience - which is what I use it for), and Module 01 (engine), I couldn't get it to connect to Module 17 (Instruments).

Nuje · Jul 18, 2020

Update: I *was* able to access the Module 17 - Instruments with VDS-PRO, by changing the baud rate; upped it to 10K and got in. And then also successfully logged in with security access of my cluster's PIN / SKC (obtained previously with VAG-TACHO).

Unfortunately, still couldn't do anything. Tried reading EEPROM bytes and it would seem to connect for an instant, and then immediately pop up an error message ("unable to connect at this time" - or wording to that effect).

So, I think I'm gonna say that making a byte-by-byte adjustment in the cluster EEPROM with VDS-PRO...probably not gonna happen.

gmenounos · Jul 20, 2020

Last weekend I did some reverse engineering to figure out how VDS-PRO does its thing so that I can create a tool that doesn't require booting into DOS. I'm getting pretty close...

What I know so far: The cluster communicates at 10400 baud and the CCM uses 9600. Unlocking the CCM for EEPROM read/write just requires logging in with 19283 first. Gaining access to EEPROM read/write for an instrument cluster is much more involved. There's a challenge/response (also known as Seed-Key) between the tool and the cluster that needs to happen first. Unfortunately the algorithm used to calculate the response from the challenge is undocumented. If you had access to that algorithm, you could 100% perform a single-byte read/write to the cluster's EEPROM.

Nuje · Jul 20, 2020

I didn't know the how/why, but my experience totally supports what you just laid out.

Maybe 20years after these cars went out of production here, we'll figure out all of the tricks to access the cluster.

Immo3 mfa cluster fuel consumption reading adjustment

Veteran Member

Top Post Dawg

Veteran Member

Vendor

Veteran Member

Veteran Member

Top Post Dawg

Veteran Member

Top Post Dawg

Veteran Member

Veteran Member

Top Post Dawg

Veteran Member

Top Post Dawg

Veteran Member

Top Post Dawg

Veteran Member

Top Post Dawg

Veteran Member

Vendor

Veteran Member

Top Post Dawg

Top Post Dawg

Vendor

Top Post Dawg