|
|
|
|
|
|
BMW Garage | BMW Meets | Register | Search | Today's Posts | Mark Forums Read |
|
BMW 3-Series (E90 E92) Forum
>
N54 Throttle Sensitivity Discovery thread
|
|
05-02-2019, 01:16 AM | #1 |
Brigadier General
1492
Rep 3,148
Posts |
N54 Throttle Sensitivity Discovery thread
A really nice discovery has been made for N54, basically we have found that the throttle sensitivity in these cars are artificially neutered in the tune, and we can have a sharp throttle response in these cars finally.
A few years ago, I drove a 1M and thought the engine response felt great compared to my 135i. Very sharp and responded very quickly. Didn't think too much of it at the time, and attributed it to the lightened flywheel that the 1M is equipped with. Have been fortunate to have driven various 1Ms and they always felt different, but never really thought about it much. My friend PeterY and I performed a manual conversion to his 335i, and fitted a twin disc clutch with lightened flywheel. I expected the flywheel to give him similar results on his car to a 1M, but was sorely disappointed, it felt dull and barely any better than my car. Since then, we managed to get the 1M IKM0S rom working on his car, and the throttle response was transformed, matching the 1M. Basically confirming the dull throttle response is software related! In general, I find the throttle response of the N54 cars to be pretty bad, you can stomp on the throttle briefly and barely anything happens. Turns out this is a rubbish thing that BMW puts into the car to make the car feel less jerky and feel smoother. Downside is, it totally ruins the throttle response. The current publicly available XDFs are derived from leaked files from the manufacturer, called OLS files. This is basically an (almost) full definition file for the parameters in the DME's rom, and you can modify these to change various things. I found a pre-production OLS floating around the internet, which really opened up my eyes as to how much development a BMW goes through! Unfortunately, these definitions point you to data in the provided preproduction rom, and not the ones in our cars (such as IJE0S and I8A0S). After knowing its a software controlled thing, I went into the OLS file and found some tables that seemed like they could be what we were looking for. There is a bunch of tables called "correction" under the folder "SYS_Determination_of_accelerator_pedal_value: "_Determination of accelerator pedal value" Unfortunately this all had different data than our production roms so searching the byte pattern for this data in IJE0S was unfruitful. Luckily, there is an unrelated table that is next to the throttles that IS the same for all versions, so I could use this to locate the throttle tables I was interested in IJE0S, I8A0S and IKM0S! What is interesting is that these throttle tables are zeroed out in IKM0S (The 1M rom), and only for manual transmission. This really pointed me towards these throttle correction tables responsible for the sharp thottle response of the 1M. I set these tables to zero on my IJE0S rom (135i manual transmission), and immediately felt a sharp throttle response as I reversed out of my driveway. Since then I've been driving with this change for a few weeks and really enjoying it. The throttle is sharp and very responsive, it was a little bit tricky driving smoothly initially, but after a few days have no issues. Very happy with the change and has completely transformed the car. I have also made this change on an I8A0S manual transmission vehicle, which threw some check engine lights regarding torque monitoring. I applied a custom linear throttle map that seemed to fix this, but the owner states that the error has come back, just quite infrequent. This Torque monitoring fault means that the throttle pedal is disabled until you restart the engine, so this needs to be fixed before the discoveries are released for that rom. I would also caution anyone wanting to try this out on IJE0S DCT or AT cars, as we have only tested this on a small number of manual cars so far. Here's a discussion of the tables and what they do: Basic correction - These tables pertain to each transmission and what mode theyre in. They apply when the car is in gear. A value of 0 means no correction and is the most responsive, higher values means more lag/filtering. Unknown what negative values do. Clutch intervention correction - These apply when the clutch pedal is pressed in, so this table zeroed out really help with throttle blips when heel and toe downshifts. Not sure when this would apply for AT and DCT Gearshift intervention correction - These are zero on the I8A0S and IJE0S roms already, not sure if any variants have values in these tables. Basic correction Vehicle speed - This is a throttle correction vs speed, at low speeds the default values are very highly filtered, but since the above tables are in effect when the car is in gear or when you're clutched in, this table only affects neutral gear. So this table really just helps you rev the car out in neutral with a sharp throttle response. I thought this would possibly make gearshifts lsightly more responsive but doesn't seem to affect normal driving apart from neutral. Additive correction in reverse - These tables are zeroed on manual transmission cars, on DCT and auto there's values, I assume this induces a bit of filtering/lag when you're in reverse. Hopefully these will be added to the public xdfs soon and be implemented as a flash time option with MHD Tuning Disclaimer: use at your own risk, I am only posting the IJE0S file, as we are still trying to prevent torque monitoring errors on I8A0S. Torque monitoring errors will probably happen if you are not using a linear throttle, if this happens the throttle pedal will become unresponsive, you need to shut the engine off (hold down the button until engine switches off, you need to hold down the button for 1-2 seconds until the engine shuts off when the car is moving, then you can immediately depress the start button to resume) You'll need to have a custom tune to be able to try these out, do not contact me about how get this on your car for you, your best bet is just to wait until you can flash this with an MHD tickbox option. Big thanks to the support of my N54 circle of friends PeterY and titium Socket and others, would not have been possible otherwise. Attached below is the xdf I have created that includes these tables, enjoy UPDATE: Here is the link to the XDFs for IJE0S, I8A0S and IKM0S https://www.dropbox.com/sh/e4d01nmbs...WqCVls7BfBPmEE
__________________
Last edited by vtl; 05-30-2019 at 12:41 AM.. |
Appreciate
17
|
05-02-2019, 01:51 AM | #2 |
Major
127
Rep 1,054
Posts |
Yoda and PeteY
Well done guys been great seeing you guys discover this from the sidelines. Look forward to seeing you guys soon on our next drive day and porting it to N55 :P |
05-02-2019, 02:15 AM | #3 |
Lieutenant
320
Rep 568
Posts
Drives: BMW M4 DCT (F82)
Join Date: Jan 2012
Location: Melbourne
|
Kudos to vtl for persisting and finding those tables in IJE0S and I8A0S, I was just content with running IKM0S bin from the 1M and calling it a day. vtl is always happy to discover and share info publicly without looking for any personal gain or reward, instead of hoarding information or asking for money for every string of information like others do. People like him help platforms develop much quicker!
Flashing IKM0S from the 1M to a standard 335i was the eye opener that the laggy numb throttle response of the standard 135i/335i was electronically induced. My car was transformed right afterwards, I still remember how great that first drive felt We've tested the newly discovered throttle correction tables on IJE0S and I8A0S, and they're both performing like the IKM0S on the 1M. So no need to go through the pain of migrating to the IKM0S bin if you don't want to. We don't get INA0S cars here in Australia, but it'll just be the same concept and procedure except for the memory addresses where the tables are stored of course. Links to the original discovery on my buildthread from a couple of weeks ago: 1- https://www.e90post.com/forums/showp...0&postcount=12 2- https://www.e90post.com/forums/showp...1&postcount=17
__________________
Current: F82 Mineral White M4 7DCT
Gone but never forgotten: E92 Sparking Graphite 335i (N54) 6MT build thread. |
Appreciate
1
Aus335iguy650.00 |
05-02-2019, 03:18 AM | #4 | |
Lieutenant
143
Rep 417
Posts |
Nice work vtl and all involved
Quote:
|
|
Appreciate
1
Aus335iguy650.00 |
05-02-2019, 03:21 AM | #5 | ||
Lieutenant
320
Rep 568
Posts
Drives: BMW M4 DCT (F82)
Join Date: Jan 2012
Location: Melbourne
|
Quote:
__________________
Current: F82 Mineral White M4 7DCT
Gone but never forgotten: E92 Sparking Graphite 335i (N54) 6MT build thread. |
||
Appreciate
1
Aus335iguy650.00 |
05-02-2019, 03:29 AM | #6 |
Brigadier General
1492
Rep 3,148
Posts |
I8a0s definitely needs more testing. I haven't released my i8a0s xdf for this reason. You could in theory just port the ije0s changes over yourself of course.
The torque monitoring limits need to be raised (I think MHD tunes already have this). I've left my torque monitoring tables stock and have no issues on ije0s. Should be an easy fix but have not got solid testing on it. I've only got 2 guys running it on i8a0s and they aren't really putting many KMs on them. Peter was just saying the tables work as intended and really sharpen up the throttle response on i8a0s
__________________
Last edited by vtl; 05-02-2019 at 06:21 AM.. |
Appreciate
2
Aus335iguy650.00 TiAgAu142.50 |
05-02-2019, 03:51 AM | #7 |
Private
39
Rep 74
Posts |
This modification doesn't apply to my car but I have to say I'm glad there are people out there like you doing this kind of thing and sharing it.
thank you and keep it up!
__________________
2007 E92 328i coupe. 19" staggered VMR v710, front: 19x8.5 et35 continental 235/35 19, rear: 19x9.5 et33 continental 265/30 19; 1 inch front lowering perches; Msport rear bumper; roof spoiler; trunk lip; Bavsound UF110; carbon fiber wrapped interior trim; 3 stage BMW intake manifold; aFe magnum force stage 2 si pro 5R cold air intake; active autowerkes catless header; active autowerkes gen 2 signature exhaust; BPC stage 2 header tune
|
Appreciate
2
Aus335iguy650.00 vtl1492.00 |
05-02-2019, 05:10 AM | #8 |
Major General
898
Rep 6,285
Posts |
This is really awesome work guys!
MHD Tuning should incorporate this as a checkbox feature in their maps! |
Appreciate
2
Aus335iguy650.00 Works29.50 |
05-02-2019, 05:17 AM | #9 | |
Colonel
650
Rep 2,139
Posts
Drives: 2009 e92 335i MSport DCT
Join Date: Jul 2015
Location: Sydney Aus
|
Quote:
I also think that the people involved in this should be rewarded in some way. I understand that it was VTL and PeterY Titium and of course Socket ... anyone else? I’d like to pitch in a couple of dollars to say a genuine thanks for helping the platform move forward. GREAT WORK PEOPLE!
__________________
When you doppelkaplung you doppel your fun.
|
|
05-02-2019, 05:30 AM | #10 | |
Private First Class
30
Rep 93
Posts |
Quote:
I'm so happy that this came out around a week after I posted my frustrations to the 135i throttle behavior on I8 ROM: https://www.e90post.com/forums/showthread.php?t=1607537 Can't wait to try this out!
__________________
E82 N54 hobby project - FlexFuel by Motiv/MHD - Stage1 turbos + FBO - ST XTA w/ AD08R
|
|
Appreciate
2
DR-JEKL897.50 riczon.335i48.00 |
05-02-2019, 06:48 AM | #12 |
Lieutenant General
4525
Rep 11,898
Posts |
great write up mate
|
Appreciate
0
|
05-02-2019, 07:03 AM | #13 |
Brigadier General
1492
Rep 3,148
Posts |
Attached is my throttle map (torque request high and low)
I found that a perfectly linear throttle map was quite hard to drive smoothly with the new throttle correction tables, tried the 1M map which was easier to drive but still found it a tad jerky. I further reduced the sensitivity of the torque request map below 50% pedal and find it a bit better. I also flattened (linearised) out the table so it's not all boosted at low rpms. This works well for manual, personal preference though. I dont believe DCT and AT cars can get away with a linear throttle if you want a D mode that works well (feels sluggish and has poor shift logic). If anyone with DCT and AT could provide feedback that would be appreciated.
__________________
|
05-02-2019, 08:29 AM | #14 | |
Private First Class
30
Rep 93
Posts |
Quote:
I'm assuming that both were MSD81's? If I'd go for a MSD80 -> MSD81 upgrade, how much of a pain in the ass it would be to go from a IJE0S to IKM0S and after this throttle behavior can be patched in to both IJE0S and later on to I8A0S, is there anything to be gained from migrating to IKM0S at all?
__________________
E82 N54 hobby project - FlexFuel by Motiv/MHD - Stage1 turbos + FBO - ST XTA w/ AD08R
|
|
Appreciate
0
|
05-02-2019, 02:52 PM | #16 | |
Colonel
650
Rep 2,139
Posts
Drives: 2009 e92 335i MSport DCT
Join Date: Jul 2015
Location: Sydney Aus
|
Quote:
I do know that it can’t be done with my skill set (let’s call it that) lol |
|
Appreciate
0
|
05-02-2019, 04:45 PM | #17 | |
Brigadier General
1492
Rep 3,148
Posts |
Quote:
Basically with stock torque tables and standard throttle mapping, it would throw the code after 30-120 seconds of driving. Switching to a custom linear throttle mapping helped (can drive for weeks before it pops up again) but I feel this just masked the issue. Should be an easy fix, is probably just the torque max difference table in the xdf (Ive got this stock on my car and the test I8A0S car). I think the ots maps already have these tables raised anyway? |
|
Appreciate
0
|
05-02-2019, 06:25 PM | #18 |
Captain
647
Rep 779
Posts |
If it occurs at part-throttle, especially in higher gears, it probably is the max diff table if it hasn't been raised (unless I8A0S has something else going on?). Should be a 2048 limiter. I could trigger like clockwork a few years ago with 40-50% sustained throttle in 6th or 7th gear from too much request. Monitor A/B were already maxed, raising max diff resolved it.
DCT won't be able to zero out and I doubt AT will either, but room for some improvement and keep it driveable. I tested a few changes today on INA0S DCT and still needs some tweaks. I flashed to a stock tune for it and even some small tweaks on stock request went from dead area (completely obvious in the stock rom Aus335iguy ) to traction control at the same place in 1st, so adjustments need to be baby steps and would probably still be trouble/fun on an actual tune. I'll try to get basic shapes, but it's probably going to need to be adjusted for driver preference/tune. I'm big a fan of the 1M request tables, work just fine on DCT, but doing this on the stock 335is tables for now. I'll add the tables I have to a an XDF and send them if you want jyamona , but won't post because I haven't tested any outside of DCT base and speed. The rest will need to be vetted. |
Appreciate
0
|
05-02-2019, 06:49 PM | #19 | |
Brigadier General
1492
Rep 3,148
Posts |
Quote:
Why do you say it wont be able to be zeroed out on Auto/DCT? Do you mean it'll be jerky or cause problems with the TCU? Auto/DCT already have these throttle correction tables a lot lower than MT. If you have to preserve a non linear throttle on DCT/AT and zero out the throttle correction tables then yes it could indeed be hard to drive smoothly (although AT has a torque converter which dampens things out anyway) I think an adjustable slider in MHD would probably be the best idea for people who find it a bit too jerky, I found I have adjusted my driving to have more finesse on the throttle. If anyone associated with N54 development needs the XDF for I8A0S just PM me and ill send it across
__________________
|
|
Appreciate
0
|
05-02-2019, 06:56 PM | #20 |
First Lieutenant
69
Rep 310
Posts
Drives: 2007 BMW 335i M-Sport E90
Join Date: Jul 2012
Location: land of the long white cloud
|
Does COBB AP (sport and aggressive map) also control the throttle response map?
In general driving in D will be slower response, but when you use the manual M or Sport mode, the response is better. I can only imagine how awesome and responsive it must feel with manually adjusting the throttle response map. Keeping in mind though, too harsh of a throttle response would put more wear and tear on your gear box internals. BMW on purpose make the throttle response slower for a (comfort mode) and b longevity of the internals. |
Appreciate
0
|
05-02-2019, 08:28 PM | #21 | |
Captain
647
Rep 779
Posts |
Quote:
1st on DCT is already a bit of a party from a stop at 50% of the stock INA0S base values, stock speed corr, 1M torque request (way less than stock 335is) and not a lot of throttle. It will come together, but will take some finessing. If anyone does haul off and just 0 out the base corr on a DCT, PLEASE vid the first drive No idea on the 2048 fix, that's an MT issue I've paid little attention to. Not sure what conditions the 2D59 tripped on that car, but if it hit a limp mode with it, my experience is that it would be a 2048. I could pop them all day long at PT with too much request and not enough monitor space, that code never showed up without a 2048 or vice versa. There are PV monitors, but not sure if they are only for cruise control, plausibility or are in play here at all, but it would be simple to verify if it's a 2048 or not by logging. Use a stock request table, stock max diff and A/B monitors or however he was hitting them with ease before, let it trip and see what the log says. Slider would be a good idea and probably necessary Glad to know I'm not the only one that spends nights going cross-eyed staring at it, nice job! INA0S DCT tables: |
|
Appreciate
1
vtl1492.00 |
05-02-2019, 08:48 PM | #22 | |
Brigadier General
1492
Rep 3,148
Posts |
Quote:
I would think the shifts would be fairly unaffected on DCT/Auto, these correction tables only seem to affect pedal changes, and so if you are holding your foot down constant and flick the paddles should operate the same. Good info regarding hunting down the 2D59 faults From my testing, speed correction seems to do nothing except for when the car is in neutral. So this table doesn't seem to be as useful as the name implies. If you test this by throwing it in neutral while the car is at speed, the throttle response is very heavily filtered. If you do the same test at higher speeds (pop the car into neutral at 180kph) the values are lowered and you get a bit better throttle response. This speed table is left standard on the 1M, and when I zeroed this out I noticed no change except for the throttle response vs speed in neutral only.
__________________
|
|
Appreciate
0
|
Bookmarks |
Thread Tools | Search this Thread |
|
|