[jponline77]

Edit: Please see the following link for detailed instructions on how to code Anti-Dazzle high beams properly: NGHB How To_F25.pdf

The following thread has been kept to show the history of the investigations:

I have coded the Anti-Dazzle high beams by removing 5AP and 8S4 from VI and coded FRM and KAFAS2 modules.

However, after trying it out on dark roads, I found people were flashing their lights at me. So, it seems there is something I am still missing from the coding. After some searching I found this:

Edit: Updated URL to latest to avoid confusion http://f80.bimmerpost.com/forums/att...8&d=1442033030

It seems that I need to code a few more things to get this to avoid glare. Does anyone have experience with this? This guide is for the F30, does anyone know if it works for the F25?

Thanks,

JP

[jponline77]

I got this working for the F25. I loosely followed the guide with some F25 specific twists. I will try to write something up and post it soon.

JP

[jponline77]

I sorted it out... The original post I had followed some instructions where I just updated the FA for the FRM and KASAF modules to remove 5AP and 8S4. This seemed to allow the high beams to come on based on tail lights but the proper "tunnel" wasn't obviously formed.

In the morning after, I was basically able to figure out how to do step 2 for the F25. I updated the FA on both TMS and LHM as well as CAS and FRM. As well, I programmed all the M1-15 values on both LHM to init_F025_ECE instead of US.

I wasn't able to do step 3. F25 does not have FEM_BODY but it has the options required for step 3 in the FRM. However, there aren't a lot of options for settings there and no European defaults.

I have tried it out now and am pretty happy with the performance. How important is step 3?

[dmnc02]

Originally Posted by jponline77

After some searching I found this:

http://www.bmwesys.com/guides/Euro%2...20How%20To.pdf

First, just for the sake of avoiding any confusion, that PDF is taken from this post and it is not the latest version.

Originally Posted by jponline77

I sorted it out... The original post I had followed some instructions where I just updated the FA for the FRM and KASAF modules to remove 5AP and 8S4. This seemed to allow the high beams to come on based on tail lights but the proper "tunnel" wasn't obviously formed.

In the morning after, I was basically able to figure out how to do step 2 for the F25. I updated the FA on both TMS and LHM as well as CAS and FRM. As well, I programmed all the M1-15 values on both LHM to init_F025_ECE instead of US.

It is not that easy: see this post.

[jponline77]

dmc02,

Thanks for the help. I guess life is never that easy. It seems to work but I think you mentioned in another post that it can be hard to tell with oncoming traffic because their glare makes it difficult to tell if it is working perfectly. I definitely didn't get any flashers and it creates a good clear tunnel for cars I am trailing.

However, I don't think my low beams are really doing as much movement as I hear I would expect. So, I suspect there is more work that needs to get done to get the full effect.

Do you have access to the actual European default values for the F25? I can get my FA and SVT info tonight I hope, once I get my PC back up and running.

BTW, I may seem incompetent because I can't get a simple PC utility installed... However, I am fairly technically competent as I design the integrated circuits that go into cars like this... I'm just used to working on Linux and not on these crappy Windows machines.

[dmnc02]

Originally Posted by jponline77

dmc02,

Thanks for the help. I guess life is never that easy. It seems to work but I think you mentioned in another post that it can be hard to tell with oncoming traffic because their glare makes it difficult to tell if it is working perfectly. I definitely didn't get any flashers and it creates a good clear tunnel for cars I am trailing.

However, I don't think my low beams are really doing as much movement as I hear I would expect. So, I suspect there is more work that needs to get done to get the full effect.

Do you have access to the actual European default values for the F25? I can get my FA and SVT info tonight I hope, once I get my PC back up and running.

BTW, I may seem incompetent because I can't get a simple PC utility installed... However, I am fairly technically competent as I design the integrated circuits that go into cars like this... I'm just used to working on Linux and not on these crappy Windows machines.

I did check the European default values for the F25 some time ago at the request of somebody on the other forum. However, I would like to check them again to confirm (which is why I asked you for those two files from your car).

As far as the LHM ECUs are concerned, the differences are below (Left is the US value and Right is the ECE value): you can see that just setting everything to init_ECE_F025 is NOT the correct choice, for the reason mentioned in the post I previously linked.

LHM [43]:

M1 Left: init_US_F025 = FA FA 00 00 C8 C8 00 Right: init_ECE_F030 = FA 64 00 00 FA FA 00
M2 Left: init_US_F025 = FA FA 00 00 C8 C8 00 Right: init_ECE_F030 = FA 64 00 00 FA FA 00
M3 Left: init_US_F025 = FA FA 00 00 C8 C8 00 Right: init_ECE_F025 = FA 00 00 00 FA FA 00
M4 Left: init_US_F025 = FA FA 00 00 C8 C8 00 Right: init_ECE_F030 = FA 00 00 00 C8 C8 00
M14 Left: init_US_F025 = FA 4B 00 7D C8 C8 FA Right: init_ECE_F025 = FA 00 00 FA FA FA 00
M15 Left: init_US_F025 = FA 4B 00 7D C8 C8 FA Right: init_ECE_F025 = FA FA 00 FA FA FA FA
ERR_AL Left: init_ON = 01 Right: initwert = 00
ERR_FL Left: init_ON = 01 Right: initwert = 00

LHM [44]:

M1 Left: init_US_F025 = FA FA 00 00 C8 C8 00 Right: init_ECE_F025 = FA FA 00 00 FA FA 00
M2 Left: init_US_F025 = FA FA 00 00 C8 C8 00 Right: init_ECE_F025 = FA FA 00 00 FA FA 00
M3 Left: init_US_F025 = FA FA 00 00 C8 C8 00 Right: init_ECE_F025 = FA FA 00 00 FA FA 00
M4 Left: init_US_F025 = FA FA 00 00 C8 C8 00 Right: init_ECE_F030 = FA 00 00 00 C8 C8 00
M6 Left: init_ECE_F030_RL = FA FA 00 00 C8 C8 00 Right: init_ECE_F030 = FA FA 00 00 FA FA 00
M7 Left: init_ECE_F030_RL = FA FA 00 00 C8 C8 00 Right: init_ECE_F030 = FA FA 00 00 FA FA 00
M8 Left: init_ECE_F030_RL = FA FA 00 00 C8 C8 00 Right: init_ECE_F030 = FA FA 00 00 FA FA 00
M9 Left: init_ECE_F030_RL = FA FA 00 00 C8 C8 00 Right: init_ECE_F030 = FA FA 00 00 FA FA 00
M14 Left: init_US_F025 = FA FA 00 7D C8 C8 FA Right: init_ECE_F025 = FA FA 00 FA FA FA 00
M15 Left: init_US_F025 = FA FA 00 7D C8 C8 FA Right: init_ECE_F025 = FA FA 00 FA FA FA FA
ERR_AL Left: init_ON = 01 Right: initwert = 00
ERR_FL Left: init_ON = 01 Right: initwert = 00

[jponline77]

Finally got my computer updated... They must have put a lot of crap on the PC when I bought it because, now that I wiped it, it runs noticeably faster... It's actually a usable computer. Anyway, E-SYS works now.

The coding you listed above for the M* parameters turns out to match that which I coded before. This is because the init_ECE_F030 match the init_ECE_F025 parameters. So, I didn't need to make any changes there. I did change the ERR_AL/FL values though. Any idea what those are for?

I'll take it for a test run later tonight. I emailed you my FA_SVT files in case you are able to look for other places where I might need to make changes.

Thanks again.

[jponline77]

Test run went well. It could have been my imagination but it seems that the lights were moving quicker tonight. So, maybe the ERR_AL / FL allow the lights to move quicker. Anyway, I captured some video. It's on an iPhone 6, so it doesn't do it justice but you can see the effect.

It really is an amazing feature. It's sad that we have to hack the car to get it working. I'll add the video after I edit it, maybe later tonight.

[dmnc02]

Originally Posted by jponline77

The coding you listed above for the M* parameters turns out to match that which I coded before. This is because the init_ECE_F030 match the init_ECE_F025 parameters. So, I didn't need to make any changes there.

It appears that either new literal values have been added to the CAFDs or the mapping in E-Sys Launcher has been corrected. There were no "init_ECE_F025" values for several Mxx functions in the CAFDs I originally looked at (CAFD_000010BA_001_007_000 and CAFD_000016BF_001_007_000), which were from a 2014 X3. That is why I asked to see you FA and SVT files. Your CAFDs (CAFD_000010BA_001_008_051 and CAFD_000016BF_001_008_051) are newer.

I did change the ERR_AL/FL values though. Any idea what those are for?

The ERR_AL/FL values determine what happens if an error is detected in the Mxx values: they are inconsequential as long as long as no errors are detected. There is no reason to change them.

Originally Posted by jponline77

Test run went well. It could have been my imagination but it seems that the lights were moving quicker tonight. So, maybe the ERR_AL / FL allow the lights to move quicker.

The lights are not moving quicker, but they appear to, because shaping of the light beam is occurring at the same time due to the change in the Mxx parameters.

[jponline77]

I decided to not post the video I took. It's just not good enough quality to do the effect justice. I'll post the BMW video here for people to view.

I will try to write up the coding method clearer so I will be able to remember in a year or two when my dealer does a maintenance wipe of my coding. I'll post it here so others can use.

I find it hard to believe that I'm the first person to code this properly for X3. It's not like this is new... I would be interested to hear a reply from someone who has this coded up.

[dmnc02]

I was really hoping to see your video. The reason I did not post the coding after the trial with the 2014 X3 is that I was not convinced from the video the owner sent me that the tunnel was working as it should. Although I have not checked if there has been a change in the part numbers for the headlights or ECUs, my hope was that things might be different with newer cars.

Hopefully, somebody will eventually post a video. Both the F80 and the F15 forums have videos that are not post-processed of NGHB in action on US cars.

[jponline77]

dmnc02,

I guarantee you it's working... There is no doubt. I am at about 2 hours of testing on dark rural roads going at over 70km/h and the effect is obvious and clear. I haven't been flashed once yet (with the previous coding method I was flashed six times in the period of 30 minutes). However, if you don't take my word for it, I'll be getting a Go Pro over Christmas and should be able to get some better video for you in the new year. My iPhone video won't convince you anyway and probably just cast more doubt.

JP

[dmnc02]

Originally Posted by jponline77

dmnc02,

I guarantee you it's working... There is no doubt. I am at about 2 hours of testing on dark rural roads going at over 70km/h and the effect is obvious and clear. I haven't been flashed once yet (with the previous coding method I was flashed six times in the period of 30 minutes). However, if you don't take my word for it, I'll be getting a Go Pro over Christmas and should be able to get some better video for you in the new year. My iPhone video won't convince you anyway and probably just cast more doubt.

JP

I have no reason to not trust what you say: so I will assume this is mission accomplished and look forward to seeing it in action whenever you or somebody else have a chance of posting a video.

[jponline77]

OK, I realized my DSLR can take good video, so I mounted it on a tripod and went for a drive. I got some better video. Still not as good as seeing the effect in real life but the point is made. Another 45 minutes of driving with no flashers.

Check it out.

[dmnc02]

Everything looks good in the video, but if you have a chance post also some footage of you following another car with the tunnel active.

[jponline77]

One thing that I don't like is that when an on-coming car passes and the left beam turns off, it will not turn on again unless there are no more on-coming cars even if they are very far away. You can see this in the video at about 24 seconds when the house lights trick the car to turn off the left high beam and it doesn't come back on until all the cars have passed. Do you know if there is a parameter for that that I can adjust?

As for following, I was kind of avoiding that scenario when I was taking the video. The camera doesn't have good contrast so the best way to see the effect is to go from full high beam, to blocked out, to passing car, to full high beam. This is easier to capture with on-coming cars. I'll see if I have time to go for another run tonight to capture following a car.

[jponline77]

Took a long video last night but camera crashed right at the end and lost the footage. I'll have to try again tomorrow night. One thing that I did notice while shooting though is that my low beams do not move up and down at all. If cars are far away, the tunnel is nicely formed around them but when I am close, the low beams are about mid-way up their cars. left an right side is well defined but you can't see a full tunnel around the car because the bottom of the tunnel is half way up their car. I don't think I'm glaring them but I see in the videos of these features that the low beams move up and down with the cars. Is this something you've seen before?

It's too bad I didn't get footage because there is a good shot in the fog of this effect as I was going up a hill. The tunnel was formed nicely but when he went up the hill and down the hill the tunnel did not track vertically, and it did glare him at one point in the transition very briefly but this was a big hill with big transitions.

Do you know what parameters might impact low beam vertical tracking?

[jponline77]

I found in my old recordings a period where I was trailing someone at a bit of a distance. I added it to the end of the video and uploaded it. It should start at about 4 minutes 30 seconds.

[dmnc02]

Originally Posted by jponline77

One thing that I did notice while shooting though is that my low beams do not move up and down at all. If cars are far away, the tunnel is nicely formed around them but when I am close, the low beams are about mid-way up their cars. left an right side is well defined but you can't see a full tunnel around the car because the bottom of the tunnel is half way up their car. I don't think I'm glaring them but I see in the videos of these features that the low beams move up and down with the cars. Is this something you've seen before?

That is in line with what I saw in the video of the 2014 X3.

It's too bad I didn't get footage because there is a good shot in the fog of this effect as I was going up a hill. The tunnel was formed nicely but when he went up the hill and down the hill the tunnel did not track vertically, and it did glare him at one point in the transition very briefly but this was a big hill with big transitions.

Agreed, too bad, as a bit of fog makes it very easy to see exactly what the lights are doing.

Do you know what parameters might impact low beam vertical tracking?

The stepper motors for the headlights are controlled by TMS. I will check if there are any differences in there.

Originally Posted by jponline77

I found in my old recordings a period where I was trailing someone at a bit of a distance. I added it to the end of the video and uploaded it. It should start at about 4 minutes 30 seconds.

I do not see the added footage.

[jponline77]

Did you have a chance to check the TMS? I didn't get out last night. Christmas party prevented it. Hopefully, tonight.

Thanks,

JP

[dmnc02]

Originally Posted by jponline77

Did you have a chance to check the TMS? I didn't get out last night. Christmas party prevented it. Hopefully, tonight.

Thanks,

JP

I did not notice anything that I believe is relevant in TMS, but you might want to try changing the following two parameters in FRM:

• LAMP_MAP_PARA_SATZ_09
  USA: BV_F025_wert_01__BFD_US = 0D 12 0B A0
  ECE: BV_F001_wert_99 = 00 00 00 00
• LAMP_MAP_PARA_SATZ_10
  USA: BV_F025_wert_01__BFD_US = 0E 05 0B A0
  ECE: BV_F001_wert_99 = 00 00 00 00

[RaslDasl]

I coded my 2016 successfully. I basically followed the instructions to remove the two options that turn off VLD and anti-dazzle and then VO code every section but the engine.

VLD is very easy to check in your garage or other dark area. Turn the fog lights on and off and you should see the low beams adjust to a new pattern.

The tunneling feature of anti-dazzle is nearly impossible to see without fog. I do definitely see the left beam turning off for oncoming traffic and see the right shoulder well lit when following.

I get flashed once in a while but I attribute it more to the fact that all 4 headlights are on, both top and bottom, than to drivers actually being blinded. Between the high beam assist (which still turns the high beams fully off in many conditions) and the anti-dazzle I'm sure I'm "dazzling" a lot fewer drivers than I would doing my high beams manually.

The VLD is actually quite good. If you pay attention to the beam pattern you will see it adjusting based on speed. At slower speeds it throws a wider pattern and at higher speeds it projects further down the road. Plus it also does the different pattern with fogs.