A bitter Story & the end of my DCT Project

This is the only project that I relied entirely on a vendor to deliver and their promise for “OEM driveability”. Some more experienced may laugh at this – you can’t get OEM driveability from an aftermarket controller. Well back then, I believed what I was told and I was convinced that the product would perform as advertised. Boy was I wrong…

It all started February 2020 when I started looking into DCT Conversions for my BMW E46. Following some research, friend’s recommendation and direct contact, I decided to go with “Vendor #1” for controlling the M4 DCT gearbox. Back then there were 2 options available. Since I already had the gearbox delivered – an 8000mile unit out of a crashed M4 – I ordered the rest of the package which included:

• Gearbox Control unit
• Wiring Loom
• Modified Mechatronics unit (core exchange)
• Tuning Service
• Installation / Configuration

After some back and forth communication, I received the hardware. At the same time, the car was ready with a fully built engine, forged internals, supercharger, ECU mapping, etc so timing was spot on. Took the mechatronics unit and the wiring to the shop to have them installed in the gearbox and complete the remaining things. Sorted some inconsistency in the documentation regarding soldering on the solenoids and we proceeded to close everything, fill up with oil and handle it to the electronics shop for their part.

Luckily for me, the owner of the gearbox controller was visiting Cyprus for tuning sessions, and it was a good opportunity to also tune mine. I witnessed the 1st start of the car gently rolling away from the lift and leaving with them for tuning. Due to work commitment (I live around 80km away from that shop) I left.

At night I got a call that the car had broken down and was also misfiring (!!!!). Had it towed to the shop that did the engine build to find out what was wrong. Turned out 2 things: belt tensioner had snapped, took the oil filter housing with it, the water pump pulley and shredded the belt (which was brand new). I was buffled as to what may have caused this problem. The hydraulic tensioner was also brand new along with bearings etc, all genuine BMW. Cause found out later on: excessive engine jolts due to very bad gear changes caused the tensioner to give away.

Misfiring was not actually engine misfiring, it was ABS system intervention due to a checksum / issue on the canbus. I used the E46 M3 cluster and its SMG gear indicator uses the same address as the non-m E46 ABS system. So with current controller config, I couldn’t use the gear indicator or I would had to swap to a stock non-m auto cluster.

The Traction control was permanently lit. At first I was told it’s not related with the gearbox controller, but as soon as I unplugged it, the traction control was back to normal. Investigation showed it was an error in the Canbus protocol implementation on the gearbox controller, something that as of Aug. 2021 was (and probably still is) unfixed. As I couldn’t work without a gear indicator, I went ahead and installed an OLED panel with its own canbus controller and programming to show gears. Once the previous canbus settings were removed, the traction control also came back normally.

Unfortunately, cruise control never worked again, as the engine ECU was expecting gear change input. Not a fault of the gearbox controller but a rather upsetting thing as I really liked cruising on 7th gear with cc.

We got the car repaired mechanically, and I went to drive it back to the local “base” for further tuning. I could not make a single takeup without banging gear in and stalling. I spent a good 2 months in a field with gravel trying to sort this. The car refused to take up normally and would just bang gear consistently regardless of the settings I changed. This with Firmware 1.0.9b1 which was supposed to be a good base. Here’s the odd, if I rolled back to previous firmware, 1.0.8, the car would behave normal and could be tuned. I was never given an explanation why this happens. Vendor’s remote diagnosis was that I had a sticky solenoid. But a firmware fixed that?

October 2020:

After a couple of hours tuning with local gearbox controller dealer/tuner, we got the car to upshift properly (forget about takeups or downshifts now). Then gearbox started slipping all gears. I thought it was due to overheating, but even letting it cool would not fix it. Loaded the 1.0.9b3 firmware and only 4th gear slipped for some reason after. Also upgraded the cooling system of the DCT.

November 2020:

Over aweekend I decided to see if I could fix the slipping issue, what a rollercoaster. Started with firmware upgrade which in changelog did have some better strategies in clutch control. Me thinking I am updating from 1.0.9B1 to 1.0.9B4 should be something straight forward..

Result was: 

1. Unable to engage gear

2. Gearbox controller unresponsive

3. Joystick unresponsive after 1st command.

I’ve spent my good weekend switching between B1 and B4 firmwares (all of them on 1.0.9). I then discovered a discrepancy between firmware files and changelogs of the vendor which lead me to conclude that:
1. Apparently there was a buggy firmware file available whereas it was not mentioned in changelog.

2. I was switching between 2 problematic firmwares (1.0.9B1 & 1.0.9B4). 

3. I was tuning with utility written for 1.0.9 for older firmware. I was told the gearbox controller came with 1.0.9b1 whereas it came with 1.0.8. This now explains why some features did not apply even when selected, possibly the slipping issue, the odd behavior after a while etc.

3.3 As soon as I made the above correlation and reported the issues I found in 1.0.9B1, more people came and agreed they had same issues. Vendor connected remotely via anydesk on my car (I had to hang an old wifi access point down a tree to get signal where my car landed) and 3h troubleshooting, they released newer firmware 1.0.9Bxx. 
1.0.9Bxx solved all issues but apparently as I had essentially gone from 1.0.8 to 1.0.9 this rendered the WHOLE tuning done over the past weeks completely useless. I am back to zero. I’m willing to swallow this if the new firmware once tuned does solve the slipping gear issue.

November 15:

Long story. Car on stable firmware 1.0.8 as I couldn’t make the car takeup on later versions. Modified my configuration to match another identical E46, and lo and behold, car drives normally apart from an issue, a slip issue (2/4/6 gears only). Some detailing on takeup tables to match my setup and would be perfect.

I asked vendor about the slip issue publicly. My question on the forum was deleted because i “don’t know what I want”, I should use their beta 1.0.9B3/4/5 firmwares if I want to get the job done and help more instead of being negative or swear (yea i said the word fuck.. (which appears a couple of times on various posts)), as if I am part of the dev team or beta tester. 

They did however connect remotely to check (discussion about negativity was on PM). They immediately flashed their beta version and started from there. Modified all existing config, going through many many valve tests, calibrations, downloading new firmwares over and over again, adjusting here and there etc. This over 6 whole hours of Anydesk session. I have never seen my drivetrain abused so much. Each change they did either froze the gearbox controller, or stalled the engine or didn’t engage gears at all. My driveway appeared as if I was doing test drives over it from the skidmarks.

Their conclusion is that my DCT is faulty, not the control unit or Firmware. It has stuck solenoids, and valve proportionality issues. I simply asked: how is it even possible to work just fine on 1.0.8 though? Was asked for proof without letting me do it for some reason.. They tried to apply my config on their beta again, and after another hour, closed the session with “you can still tune it even if it’s problematic”. The session ended with the car not being able to move or engage gear at all. I struggled 3 whole days to get it moving (not being negative yea) with 100% fail.

Not having this BS, I went and flashed backdated firmware 1.0.8, loaded my config and car behaves and works normally (apart from the slip issue). A friend of mine sent his config (S54 AWD Turbo) which I applied on my setup, and guess, NO MORE SLIPPING!

At this point I deducted that the firmware is too unstable for my taste to do any meaningful work on it. What worries me most is that people with the same firmware experience different  issues/bugs. For some, car won’t engage gear, on others it will stall, on others it will go unresponsive, etc. 

When I kept discussing the issue over the group private chat mode, vendor insisted that my anti-stall and clutch pressure tables were wrong. Lets not forget the fact that they remotely diagnosed my DCT as faulty a few weeks ago. After I went public and more users said the same thing happened to them, they miraculously “located the issue”. So myself as an idiot, wasted 2 full days changing settings in hope of fixing it.

From late November 2020 until January 2021 I spent countless hours, from my own family time, trying to figure out whats going on and tune this POS. All holidays, vacations, road trips, track days etc cancelled and I had to ground myself home or get a f*** rental to move. Bear in mind this is my daily car. Vendor did attempt to help me in some occasions during that time over Anydesk session, but the first thing that was done was to load the latest firmware, essentially bringing back the clutch bangs. Reponse: Tune it.

January 2021:

Car has been sitting again for day#4 as vendor is still trying to figure out their issues. With the latest firmware, after adaptation, the takeup would be simply on/off, completely ignoring the pressure buildup in the takeup table. Vendor again insisted the DCT has faulty solenoids, which are stuck and are activated suddenly instead of gradually. 

After a 3h anydesk session (stopped only because my laptop ran out of battery) trying to troubleshoot, apparently my DCT doesn’t have faulty solenoids anymore, having demonstrated to all of their team that car works fine with older firmwares, valve tests and also according to a different tech person there in their own support. I am particularly frustrated that immediately they point to faulty hardware on the client side rather than anything else. Hardware that is not easy or cheap to replace. Sure you are aware of this and the fact that solenoids can’t be found as spare parts.

They redid the adaptation with 2 different parameters:1. Stone cold DCT, which according to support officer #1, it is fine (support officer #2 says only with warm) 2. Adaptation with their internal utility
Clutch takeup suddenly behaved perfect with the latest firmware. I could start and move the car very smoothly, better than an actual manual clutch.

I’ve also sort of discovered another “issue” which definitely seems on the software side.. scenario:Go from N–>1 is perfect, go back to N and then again to 1 and controller engages 1st thus stalling engine. Restart engine and it’s fine again for the 1st attempt. WTF. The above may happen on 2nd,3rd,4th, etc attempt. Only the 1st start from boot behaves normally. 

As a last attempt, as i did not want to have the slightest possibility of having provlem hardware, I actually went ahead and dropped the gearbox and replaced the clutch pack (€1800), installed a new oil pump and valvebody, new oil and new gaskets/seals. This set me back around €2600 in total. At the same time, the newer firmware came out with “features” and “bugfixes” which I promptly loaded while the gearbox was being repaired. Valvebody was sent to independent specialist and came back as “no issues”. Clutch basket had worn inner clutches, which may or may not explain the slipping. I should had left the preloaded firmware and test so that I could see whether it was a gearbox problem or firmware problem. Unfortunately I did not.

How do you like your toast? Crispy, with a chance of over-slipping.

February 2021 – August 2021

Went through various firmware revisions with my forever co-driver

1.0.10

Forgot what happened here.. New tuning suite I think..

1.0.11

Clutch slams introduced. “Workaround” to lower clutch pressure to essentially have 100% slip when you’re off the throttle on 1st and R gear. The car would be very unpredictable when this setting was active. New tuning suite would mix long ratio with short ratio gearboxes with the relevant side effects. Try driving with the SR setting using a LR gearbox.

1.0.11 tells you to install 1.0.10……

1.0.12

This firmware fixed the clutch slams and introduced the Clutch Pulsating issue. On takeup, the clutch would act like a vibrator on 415v. The pulsation killed the whole takeup experience. At first it was suggested to rise the takeup RPM as a fix, which not only did not work, but it was a fucking stupid suggestion, or a bandage being more polite.

All of them with their fair share of issues. But being fair, it was a continuous improvement. On each and every release, the firmware and the utility became more refined and stable. However, far away from a reliable daily driver.

Paddles decide not only not to work, but to also reverse their function (i.e. upshift instead of downshift) randomly. Support team points towards bad wiring or failed resistor or not their issue as command does not reach gearbox controller. Cause: Uncrimped cable in their own made and supplied harness.

1.0.13

This firmware was supposed to fix the pulsating issue per public release notes, which did not. It also brought other issues, such as automatic mode not working when you wanted to select vehicle speed instead of RPM for shifting, plus miscalculation in the vehicle speed, causing the car to downshift and stall engine whilst being stationary. Also the park lock control was broken. It would randomly lift the park lock, so better not park downhill and use that.

Looping through gears forever because it translates SPEED-based cells to RPM, i.e. i have upshift at SPEED 20,30,40,50,80,100,120 and gearbox controller these values are RPM, and as they’re obviously below idle rpm (800) it goes into an infinite upshift loop.

1.0.14

Released and FINALLY the pulsation issue fixed however the other issues remained. Plus the car would not pass adaptation due to “noise in the wiring” – which was made and supplied by the vendor. After several anydesk hours, car passed adaptation, and was able to drive. The adaptation and anydesk was done before the official release to the end-users, myself accepted to try this after invitation from vendor. However, 1.0.14 would also corrupt the configuration and made the gearbox controller unresponsive or act in an unexpected way. Of course, you stop the car in front of another one, and it decides to engage 1st instead of… not? and car leans forward… This release also made the gearbox controller not being able to work with USB only but had to be ignition powered with 12v to work.

Test-N2

Released to sort the USB communication issue and shifter light not working. I was asked to test this once again. Result instantly bricked my gearbox controller so a 10-minute job ended up being a 2h as I had to remove the gearbox controller from the car, open it up and use the bootloader option to allow me to re-flash the previous working firmware.

1.0.14-1

Released to sort the above issues. Sorted indeed, but now again, temperamental park lock, and Reverse does not engage on 1st attempt. Blinks but on 2nd attempt, it works. This attributed to the miscalculation of the speed from the gearbox controller and essentially kicking in the VSS safety I put not to engage R when going over 1km/h. Thing is, car was PARKED and as still as king Tutankhamen inside the pyramid of Giza.

Conclusion: All of them with their fair share of issues. But being fair, it was a continuous improvement. On each and every release, the firmware and the utility became more refined. However, far away from a reliable daily driver.

Firmware Test-N7 – FINAL BLOW

After tracing the issue, I got a message from the support team to try that firmware which resolved the problems above, and also included enhanced adaptation. Stocked that I would finally drive like a normal human being. Yes, all worked, for 2 consecutive days. Started car on Day #3 (Sept. 10), put it to M, and I hear a grinding sound on Gear #1. WTF. Retry, and engaged gear normally for the rest of the time. I reported this to support on Friday without any response. Yes, Friday and I complain for no response, BECAUSE, when you use Facebook for support without official support hours or anything in between, I expect reply wherever and whenever there’s an issue.

Sunday, Sept. 12, 2021. I had an urgent appointment, went to leave my house, and on R attempt, it grinded gears again and Reverse would not engage. Other gears just fine. Subsequent R attempts would result in a normal gear engagement feel but car would not move, despite of pressure buildup. Reset gearbox controller and again grinds gears and doesn’t move. At this point it was a turning point. Maybe turning point at 99% but when I had to call a TAXI it was a game over.

On the coming Monday, car was reverted back to a trusty 6-speed manual. The Gearbox Controller, DCT and the DCT wiring loom were removed and are up for sale. (Upd. Oct: Controller and wiring SOLD). The remaining parts (shifter, paddles, power supply, cooling, etc) I kept them either packed or in car for the next project, which doesn’t include this gearbox controll vendor or DCT.

Upon removal, wiring loom plug was full of oil. I blame the poor quality. aka el-cheapo.

My conclusion, summarized in a picture of the aftermarket gearbox controoller and DCT is all about: