Karter16's Silbergrau E46 M3 Journal

[karter16]

I'll let you walk your way there!

In addition, my euro mullet ended up with several little custom changes in the end, rather than pulling straight from euro maps (US fuel likely cause). It'll be more rewarding for you to chase it yourself I think.

❤️
For sure, ya don't give me the answers! - I'll keep posting here with updates as I work stuff out. Writing things out helps me spot where I'm doing dumb things!


Sent from my iPhone using Tapatalk

[Bry5on]

I’ll let you walk your way there!

In addition, my euro mullet ended up with several little custom changes in the end, rather than pulling straight from euro maps (US fuel likely cause). It’ll be more rewarding for you to chase it yourself I think.

[karter16]

There are a few extra tables that needed dialing in for me but that may be in part because the US fuels are different.

I'm curious - aside from KF_TI_N_RF (which I missed out of the list in my post above) the only other Injection/Ignition tables I see differences in are in the warm-up/cat heating/secondary air pump tables for calculating the warm up factors (which I'll not bother with for now).

I do note that in the US map KF_RF_N_AQ_REL and KF_TIENDE_N_RF differ to the CSL map, however these two are the same between my Euro tune and the CSL tune.

Or I'm missing something 😅

[ugaexploder]

Hey - the CSL rail part is now, I believe, ready to go. I haven't published the CAD file yet as a bit reluctant to put it out there without having proven it first (I've test fitted it and all seems well, but haven't proven it in operation yet). If you're keen to give it a go let me know and I can send you the CAD file :-)


Sent from my iPhone using Tapatalk

sweet! i’m in no rush at the moment to install the csl intake so i’ll wait to hear your results with them first to see if you decide to make any improvements or revisions. really excited to see the results and the iterating work with the tune

[karter16]

always love the updates, what's happening with the csl rail part by the way?

Hey - the CSL rail part is now, I believe, ready to go. I haven't published the CAD file yet as a bit reluctant to put it out there without having proven it first (I've test fitted it and all seems well, but haven't proven it in operation yet). If you're keen to give it a go let me know and I can send you the CAD file :-)


Sent from my iPhone using Tapatalk

[ugaexploder]

always love the updates, what’s happening with the csl rail part by the way?

[karter16]

Happy iterating! Once you're in the weeds, keep posting your notes here regarding where it works well and where it doesn't. There are a few extra tables that needed dialing in for me but that may be in part because the US fuels are different.

Thanks! I'm looking forward to it! Yes will definitely keep this updated with how I go - just need the rest of the parts to arrive now!


Sent from my iPhone using Tapatalk

[Bry5on]

Happy iterating! Once you’re in the weeds, keep posting your notes here regarding where it works well and where it doesn’t. There are a few extra tables that needed dialing in for me but that may be in part because the US fuels are different.

[karter16]

This weekend I spent a bit of time finishing up the first cut of my CSL clone tune.

Essentially I'm making my own clone along the lines of the approach Bryson has taken with his Mullet Tune, but doing it myself for the entertainment/learning value. All credit to Bryson for pioneering this approach, I'm basically just building my own version of what he's already done. If you haven't read up already I highly recommend following along in his journal.

When I do the swap, I'm going to do it in two stages.

Stage 1: Flash the car with the base CSL tune. This will give me a baseline datum point and is a known configuration to start with and ensure that the install has been successful.

Stage 2: Flash the car with V1 of my customised tune. This is a starting-point tune which blends Ignition, Injection and VANOS tables from my current Euro M3 tune with the CSL tune. The hope is that this will be an improvement on the base CSL tune and a starting point for iterative fine tuning.


(I'm getting in to a bit of detail now, so those who know more about this than me will be able to point out my mistakes :-))


Scope

• For a vehicle with MSS54HP DME (not modified for flap)
• Vehicle is running a Euro tune
• Vehicle has a CSL airbox conversion using a MAP sensor (OE 1.0 Bar so no conversion factor required)
• Vehicle is MSS54HP so does not require IAT conversion factor

For Stage 1 I'm starting with Terra's 0401 PD31 Full binary. This binary is a modified version of the 0401 full binary with support for the non-CSL boot loader, error codes, etc. I'll be loading this full binary at the time that I do the install. I'll then be flashing the CSL tune (partial binary) as well. There are a number of changes to be made to the original 0401 PD31 partial.

1. Change 0xE002 from 00 to 01 to make the partial work with Terra’s modified full binary (support for the non-CSL boot loader)
2. Enable Alternator light over canbus
3. Set non-CSL cam offsets
4. Set CSL intake flap control to always be open
5. Disable DTC 124 (CSL flap)

I used TunerPro to make the above changes to the partial, and now have this saved ready to go when I do the airbox install.


For Stage 2 I then took the partial I'd just generated and then added the modified ignition, injection and VANOS maps that I'd prepared. The preparation of these maps were of course the bulk of the work.

The merging of the maps is an interesting maths challenge. For this first version I took the following approach:

1: Take the standard M3 and CSL tables and interpolate both tables to allow for direct comparison.
2: Compare the difference between the tables (allowing for any conversion adjustments between the two)
3: I then manually applied a mask to blend the 2 tables together into the output table.
4: We then finish up with a nice optimisation problem. Finding the x-axis points that best describe the table as a whole and thus minimising the error. The M3 tune tables tend to have more x-axis points in the lower half of the RPM range and the CSL tune tends to have more in the upper half. This of course means that you can't just use all of the x-axis points from both maps. Ideally I would dust off my very rusty Matlab skills and write something to do it, however for now I simply added some calculations in the sheet to identify absolute change, rate of change, inflection points, etc. and then rank them by importance, this made it fairly straightforward to more manually identify the most important x-axis values to bring across.


The VANOS maps in particular are quite fascinating. There is a very significant difference in the lower RPM range between the M3 Euro and CSL tunes (example of the difference in the exhaust maps below). It's easy to see why cars with standard cams struggle so much in the lower RPM ranges with the CSL map.



Methodically working through these maps took me a reasonable amount of time (although to be honest a lot of that was diving into researching various topics to help make sure I understood correctly.

The end result was updates to the following maps:

KF_TI_N_RF_VL: Fuel Injection Full Throttle
KF_TI_N_RF: Fuel Injection Duration
KF_TZ_VL: Ignition Full Throttle
KF_TZ_GRUND: Ignition Ground
KF_TZ_MIN: Ignition Minimum
KF_TZ_LL: Ignition Idle
KF_EVAN1_SOLL: Intake Target
KF_EVAN1_SOLL_KATH: Intake Target Cold
KF_EVAN1_SOLL_DMAX: Intake Target Max Deviation
KF_AVAN1_SOLL: Exhaust Target
KF_AVAN1_SOLL_KATH: Exhaust Target Cold

KF_TZ_MIN and KF_TZ_LL are the ones that concern me most. For both I've erred on the side of taking the values from the CSL map, as for KF_TZ_MIN at least they appear to be more conservative. and for KF_TZ_LL it is different enough that I can't be sure there aren't changes in the 0401 program that interpret this map differently. I'll find out I guess when I get the tune running.


It's worth noting that once I have the tune up and running I'll also be doing the VE map tuning process described here.


Oh and the Haimus Racing snorkel arrived today. Well packaged and in a lovely soft bag. It seems to be a fairly nice piece. You can certainly see it's 3D printed when you look at it closely, and there is evidence of printing errors in places. The exterior has been post-processed to a fairly nice finish, and unless you look closely it looks good. It also feels nice and light for its size, but seems sturdy. All in all I'm ambivalent about it given the price (especially with shipping to NZ) (that said I haven't actually looked into how much it would cost to get one printed from CAD file), but it will look pretty nice in the engine bay I think once all is done.

[karter16]

This weekend I made a start on preparation for installing the 6 point rear brace. The plan is to get as much pre-work done as possible so that the process of dropping the subframe and bolting through can be as straightforward as possible.

I got this done in a few hours here and there across Saturday and Sunday in between adventures with Mr 7, Mr 3 and Mrs Karter16.

First up is to remove the interior:



I need to get the isopropyl alcohol out and clean up the rear seat and prep it for repainting some rust spots, but I won't do that until the heavier prep-work is done.

Before I got started with drilling I wanted to place the brace in the car just to see, with my own eyes, the placement of the mounting points etc. I had all the measurements from Marin at Yurkan Cages, but figured seeing it with my own eyes would help avoid any mistakes.

I wanted to make sure though that I didn't damage the paint on the brace, so spent some time first taping up all of the contact surfaces:




I carefully dropped the brace into place (obviously sitting higher than it will eventually with the floor sheet metal in the way. But this meant I could sight the approximate locations and reassure myself about the approach.




Note: You don't see it in the photos because I removed some of it to get better lighting for photographs, but I was very careful to ensure that interior surfaces were covered and protected when drilling, cutting, sanding, etc. It's really worth taking the time to protect everything well.


First up to prepare for cutting the top layer of sheet metal is removing the two little brackets that sit atop the locations where the front mounts will go. The diameter of the hole for the spot weld is about 6mm, and I found a 7mm bit was necessary to drill out the weld entirely.






Note that the brackets are stuck on as well (presumable to hold them in place before being spot welded). I needed to carefully encourage them to lift off with a flathead screwdriver (which I taped with masking tape to protect the metal surfaces).




With these out of the way I then moved on to cutting the holes for the rear two mounts. I carefully measured and remeasured several times to make sure I was in exactly the right spot. I used a centre punch (always use a centre punch) to mark the point and then drilled out with a 3mm bit. I had some old, unused, BMW M TwinPower Turbo 10W60 which I used for lubricating on the drill bits and hole saw (only ever 10W60 for my car 😂).




Note: With things like this it's worth taking the time to clean your drill bits after each cut, that way they're clean for the next cut and aren't flinging more mess around than necessary.

I then drilled out to 6mm before making the final cut with the 44mm hole saw.




You can see in this photo that there was some overspill from when the BMW RACP "foam" resin procedure was done. I've only just realised I didn't take a photo, but the way I cleaned this up was to carefully down down as far as I dared with the hole saw (not wanting to damage the next layer of sheet metal) and then use a chisel to very careful chip away at the resin bit by bit. By carefully doing this I was able to completely clear it from the vicinity of the hole without causing any damage to anything else.

Side note: calling this stuff "foam" does a disservice to people's understanding of it. The stuff is much better described a plastic-like resin - It's absolutely solid stuff.

I then did the other side (the resin had only just domed up from underneath so didn't need to clean it up like on the left hand side. Pretty soon I had the rear points taken care of:






Then it was on to the front points, which are certainly the harder pair to do. I realised, as I went to write this, that I didn't take any photos of this in progress, probably because I was cramped in the back of the car and carefully concentrating on what I was doing. Essentially though it's a case of carefully measuring in from the sides and then siting the pilot hole in the right place. In order to do this correctly I used a flat edge on the top of the RACP and measured out and down to find the bottom edge of the pocket. I then split the distance in two horizontally and dropped down to get the centre point at which to drill the pilot hole. I then drilled this out to 6mm and "verticalized" it to guide the hole saw. I then cut down vertically from the top using the hole saw, and then cut out the bottom of the pocket with a cutting disk on the Dremel (highly recommend having/getting a Dremel if you're taking this job on).

Working carefully resulted in this (wife's beauty product as a conveniently sized stand in for 40mm tube steel):



After some careful work on both sides (including removing burrs with the sanding attachment on the Dremel) I then took the time to use a magnet to remove all trace of the cutting, including from inside the cavities. Handy tip - if you use a medical disposable rubber glove as a cover for your magnet it makes cleaning up easy. Each time the magnet is "full" just invert the finger of the glove and move on to the next finger, after 5 goes throw the glove away and grab the next one.




The front now looked like this:




I then dropped the brace back in to check the fit. Fitted into place first time! (obviously still sitting higher as it's resting on top of the rear shock mount nuts).






Having reached this point and confirming the fit I then took the brace back out and cleaned up the cuts and area around them (inside and out) with isopropyl alcohol, before applying a coat of kill rust undercoat to the exposed edges of the cuts. I have an enamel top coat in the same colour as the e-coat BMW used which I will apply as well, but only after confirming the fit once the new extended subframe bolts are in and I can check the fitment exactly. Until then the undercoat will protect the edges from rust.


I'm glad that I decided to tackle this part of the project separately, as it took a number of hours to do and means the top side is ready to go when we drop the subframe.


Oh, and in an entirely expected and predictable turn of events, the top side of the front right RACP mount has cracks in it. I'm unsurprised given the cracks that we welded on the bottom side when Dad and I did the reinforcement plates. We'll stop drill the cracks and weld prior to dropping the subframe. I wish it hadn't been quite so long between doing the plates and getting to the top side, but such have been the realities of dealing with medical kids. On the plus side the car has done very few km's since the plates were done!

[karter16]

My son and I spent some time this morning removing the rear bumper and having a look at why it sits so badly and with a view to affixing the replacement speed clip brackets to the kevlar support.



Unfortunately the reason is a little more serious than I was expecting. The rear bumper had been replaced prior to my ownership due to a collision (something I was aware of when I bought the car). I suspected that the reason that the bumper didn't sit well was because it wasn't installed properly. This was indeed the case, with the foam screws missing, etc. and the lower brackets of the rear bumper cut up, presumably to make the diffuser fit. I was expecting to have to get these items sorted and plastic weld the bumper, but I wasn't prepared for the fact that the kevlar support is damaged (in hindsight I should have realised it likely would be). The kevlar support has several folds/tears toward the centre of the piece which has broken the resin. This in turn appears to have allowed, over time, moisture to absorb into the kevlar weave. As a result the support appears to be beyond practical repair. Also as an integral part of the collision system I'm keen not to skimp on it.

The red outline in the photo below shows the affected section.




In the meantime we have reinstalled the bumper as-is, while I source a new kevlar support and a couple of other bits and pieces. Oh, and it's time to replace the exhaust hangers again...

Aren't old cars just the best!

[karter16]

Gave the car a quick wash this afternoon and was finally able to put the door trim back on the driver door. I took the trim off several months ago to adjust the drivers door window, but the rubber grommets on the clips were too hard and brittle to be able to use again, so had to wait for replacements from Schmiedmann (I ordered a significant amount for the future). Anyway, here the trim is with new clips ready to go back on.



And reinstalled.

[karter16]

Hell yeah, excited to see the results of the tuning. Don't forget the flap tables if you end up getting an aftermarket snorkel.

I just set all the flap opening tables to 100 rpm with Bryson's tune and it's been working great

Yeah I'm looking forward to having all the stuff here so that I can really get into it. Am doing what I can in advance in terms of getting my calculations in order etc. :-)

Good reminder thank you on the flap tables. I shall definitely make sure I don't forget them

[karter16]

Haimus snorkel is a good alternative. The only annoying thing is that the partition in the center opening for the upper and lower air feeds needs to be removed. They really should print it without that part.

Thanks yes I saw you mention this on someone else's thread - will definitely keep this in mind if I go the Haimus route!

[heinzboehmer]

Hell yeah, excited to see the results of the tuning. Don't forget the flap tables if you end up getting an aftermarket snorkel.

I just set all the flap opening tables to 100 rpm with Bryson's tune and it's been working great