untilnow
Full Member
Karma: +2/-3
 Offline
Posts: 90
|
 |
« Reply #345 on: September 22, 2024, 05:04:12 AM »
|
|
|
The _slope_ of the line reporting torque needs to match the _slope_ of the actual torque. That is, if torque increases by 2x from point A to point B, the reported value also must increase by 2x or DSG is sad. Which usually means reported torque needs to be a line, or pretty close to a line, unless you're running your setup in some terrible inefficient area.
If you scale the whole line (ie underreport torque by 30% ACROSS THE BOARD), things continue to work as long as you re-adapt the DSG, although there is usually no reason to do this if you are competent and/or have access to a DSG ASW torque cap patch. If you flatten the top off (like most clueless tuners), everything breaks.
The absolute numeric values are not important on DQ250, the slope/gradient is what is important. The way the clutch model on DQ250 works is basically by picking two data points and drawing a line to determine the relationship between clutch engagement and pressure. The values that are used to draw the line are not important so long as everything is scaled to fit on that same line.
If I want to under report torque by 30% I could multiply the load axis of IOP by 1.43 (1/0.7)? Isn't this the same egregious act as changing MDNORM? |
|
|
|
|
Logged
|
|
|
|
untilnow
Full Member
Karma: +2/-3
Offline
Posts: 90
|
|
« Reply #346 on: September 29, 2024, 07:55:46 AM »
|
|
|
Yes, but "treat" is not the right word. 100% is MDNORM.
Also, the issue is that a lot of the stuff in the ECU uses torque internally, such as idle pid, friction torque etc.
Once you change your 100% to something else, you have to change everything else too.
It's kinda like 5120, just for torque.
DQ250 also can't really read higher than 630nm.
If I scale the KFMIOP load axis up so it is reporting less torque but maintaining linearity (I'm thinking *1.5 to go from 150max to 225max) surely the same rescaling of dashpot, idle PID etc is required? |
|
|
|
|
Logged
|
|
|
|
|
prj
|
|
« Reply #347 on: September 29, 2024, 08:20:42 AM »
|
|
|
surely the same rescaling of dashpot, idle PID etc is required? Yes |
|
|
|
|
Logged
|
|
|
|
untilnow
Full Member
Karma: +2/-3
Offline
Posts: 90
|
|
« Reply #348 on: October 02, 2024, 08:24:36 AM »
|
|
|
Yes
Okay am I right here: LLR = Idle Control Reserve; for managing extra demands of the engine at idle (power steering, A/C etc) DASHPOT = Manages throttle angle smoothing torque transitions Friction Torque = Manages the constant innate resistance of the engine itself in turning over LSD =  Please help I've found quite a few maps relating to these functions and am in the process of making a comprehensive list, translating from German to English is helping me find the maps that include torque. Any help would be great  |
|
|
|
|
Logged
|
|
|
|
|
prj
|
|
« Reply #349 on: October 02, 2024, 10:34:31 AM »
|
|
|
That isn't an exhaustive list, not even close.
Every single torque value needs rescaling in ECU to do it properly, regardless if it's single value or a map axis/value.
You're not going to do that manually. A2L is needed and then some coding to automate the process.
|
|
|
|
|
Logged
|
|
|
|
untilnow
Full Member
Karma: +2/-3
Offline
Posts: 90
|
|
« Reply #350 on: October 04, 2024, 05:19:08 AM »
|
|
|
That isn't an exhaustive list, not even close.
Every single torque value needs rescaling in ECU to do it properly, regardless if it's single value or a map axis/value.
You're not going to do that manually. A2L is needed and then some coding to automate the process.
I've spent about 10 hours working on this since my last post as I really want to do this! I've been using chatgpt (going round in circles) to help me parse the attached file. (Won't let me attach for some reason) https://file.io/2W7WJjC0N1LlAs I understand it; in attempting to find "where torque is" I can use: 1. The scaling factor rel_uw_b100 as we know torque is a 0-100 value 2. The word "moment" 3. a chatgpt interpretation of the maps descriptions I'm 99% sure there is a better way of parsing this file lol. Please Mr PRJ, throw me some bread crumbs! |
|
|
|
« Last Edit: October 04, 2024, 05:22:00 AM by untilnow »
|
Logged
|
|
|
|
elias
Full Member
Karma: +21/-3
Offline
Posts: 80
|
|
« Reply #351 on: October 04, 2024, 11:39:48 AM »
|
|
|
I can give you some breadcrumbs:
1. First get a A2L. this is a manual process as it involves a lot of asking ppl and ass-kissing the right ones. Someone may have one for this or a similar ECU
2. Then you could write some WINOLS LUA script which will remap all the values as you wish.
(You could do it with some code which will parse a2l, get all maps out and remap them, but WINOLS would be easier)
|
|
|
|
|
Logged
|
|
|
|
untilnow
Full Member
Karma: +2/-3
Offline
Posts: 90
|
|
« Reply #352 on: October 05, 2024, 05:37:08 AM »
|
|
|
I can give you some breadcrumbs:
1. First get a A2L. this is a manual process as it involves a lot of asking ppl and ass-kissing the right ones. Someone may have one for this or a similar ECU
2. Then you could write some WINOLS LUA script which will remap all the values as you wish.
(You could do it with some code which will parse a2l, get all maps out and remap them, but WINOLS would be easier)
1. I have the a2l, that's the attached file. 2. I know lol, I'm asking how to identify which maps need changing. Thanks for the help though bro |
|
|
|
« Last Edit: October 05, 2024, 06:50:57 AM by untilnow »
|
Logged
|
|
|
|
|
Dave9n3
|
|
« Reply #353 on: November 23, 2024, 07:14:44 AM »
|
|
|
Does anyone know the function of the map 'Maximale Synchrondrehzahl-Differenz bei NAb-Sensorausfall ' (Ges_NSyncDiffMax_ko) 0x6a3bc in my file
From a rough translation i *think* it is the max allowable difference between input shaft speed and output shaft speed when the NAb sensor fails, NAb (Abtriebdrehzahl?) I presume is the output speed sensor/s.
Reason for asking is i currently have output speed sensor faults in the TCU like lots of other people with ME7 + DQ250 setups, which seems to be fixed by changing the ABS module to a MK60 and recoding, unfortunately not an option on my car. I'm wondering if this differential limit may be the reason why when I go for a higher rpm downshift, the clutch seems to drag almost forcing itself to synchronise the speeds before it fully engages (works better if I press the throttle slightly before attempting to downshift). I'm just looking to get the downshifts the same as a native dsg car, i know that a throttle blip function is not possible without lots of work.
|
|
|
|
« Last Edit: November 23, 2024, 07:51:33 AM by Dave9n3 »
|
Logged
|
|
|
|
MirXas
Full Member
Karma: +4/-7
Offline
Posts: 93
|
|
« Reply #354 on: February 01, 2025, 04:55:43 AM »
|
|
|
Anyone know how to smoothen clutch engagement during downshifting with brake pressed or just simply coasting to a stop ? Its probably hardware/solenoids related cause it shouldn't do the jerking motion when coming to a stop, but maybe there are some downshifting torque gradient maps that could help cure this hw issue ?
|
|
|
|
« Last Edit: February 01, 2025, 06:39:19 AM by MirXas »
|
Logged
|
|
|
|
untilnow
Full Member
Karma: +2/-3
Offline
Posts: 90
|
|
« Reply #355 on: July 27, 2025, 06:11:56 AM »
|
|
|
Torque = (miist_w - mdverl_w)*MDNORM
miist_w has a maximum value of 100%.
If your MDNORM is 350, then you can never go over 350 even if mdverl was 0%.
In case of DSG KFMIRL and KFMIOP don't have to be just linear, KFMIRL actually has to have correct miist_w (inner torque) filled for every operating point.
Can be done on dyno, preferrably an engine dyno, but normal one works too if you record mdverl_w at same time.
And if you are capping out MDNORM, then you need to increase it, but that means rescaling the whole map.
Now, if you increase it a little, you won't have a lot of problems, but if you increase it a lot (and rescale IRL/IOP of course), then all your idle control, DASHPOT, ARMD and LSD is gonna go to shit. Have fun calibrating that.
If your KFMIRL is linear as you say (and looks like a line, not a curve), then you have no problem yet.
Yes, I'm still on this lol, I just had a thought, surely if I rescale KFMIOP and MDNORM by 125% - (150max → 187.5max and 400nm → 500nm respectively) - then the reported torque to idle/DASHPOT/ARMD/LSD would be less on the 0-100 torque value but that would correspond to the correct and original LOAD value just like before the rescaling, keeping things happy. The analogy being; preparing someone to lift something by telling them how heavy it is in effort from 0-100% (reported torque). If 50% effort requires 30kg of force (load) and we increase the 100% effort scaler (MDNORM) by 1.25 then now 100% = 75kg and also rescale the force to effort convertor (KFMIOP) so that 30kg = 40% effort the person will have the same relative output as below even though the information was arbitrarily different. What am I missing here? |
|
|
|
|
Logged
|
|
|
|
untilnow
Full Member
Karma: +2/-3
Offline
Posts: 90
|
|
« Reply #356 on: July 27, 2025, 07:05:10 AM »
|
|
|
I now realise that above is only applicable to things that don't have torque thresholds.  The amount of files I've seen with bent torque that come from commercial tuners, does anyone tune these cars properly? |
|
|
|
|
Logged
|
|
|
|
vr5turbo
Newbie
Karma: +0/-0
Offline
Posts: 1
|
|
« Reply #357 on: September 24, 2025, 01:35:29 PM »
|
|
|
Can someone help me or give me a tip or maybe even change something? I've already adjusted a lot of things but it looks like I forgot to adjust something. The car isn't releasing the full clutch pressure of 18 bar and isn't reading the 600 Nm that the control unit is outputting. Can someone help me?
|
|
|
|
|
Logged
|
|
|
|
|
Dave9n3
|
|
« Reply #358 on: May 31, 2026, 10:20:38 AM »
|
|
|
I have recently taken some CANbus logs directly from the powertrain bus, filtered/sorted them to be easily viewable in ecuxplot. I'm chasing a slow/delayed downshift issue with me7.5 ecu running DSG.
I logged a med9.1 car also to compare, what I can see is that the MED9.1 car (2.0tfsi BWA) sees a spike in miistc just as the shift is initiated by b_gsch and after that wkba increases. the result is the engine speed begins to rise shortly after shift initiation by the downshift paddle. You can see the me7.5 car there is a delay whereby engine speed does not increase at all, and continues to fall. The TCU seems to be waiting for something, and then eventually executes the downshift but after a wait time. In the screenshot I have aligned the logs so that shift initiation lands at the same sample on both, to make it easy to compare.
Does the med9.1 car have true ZG schaltung (zwichengas) active? or is it more of an assist, raising miist which allows the TCU to use the Zug patch to shift rather than the schub path, as I notice in TCU the torque ramp rates are different for zug vs schub. From the maps i've seen the TCU seems to use schub where engine torque is 0 or below, and uses zug when engine torque is above 0.
I know I will not be able to have true throttle blip with me7.5 whereby engine responds to the nsync request from the gearbox, but just removing the dead time/delay where the TCU hangs with me7.5 is my aim.
would really appreciate some info from those who are well versed with DQ250 and its communication with the ECU.
|
|
|
|
|
Logged
|
|
|
|
hbfreak93
Newbie
Karma: +0/-3
Offline
Posts: 7
|
|
« Reply #359 on: June 04, 2026, 10:46:01 AM »
|
|
|
Anyone got for this file Damos?
|
|
|
|
|
Logged
|
|
|
|
|
