Karter16's Silbergrau E46 M3 Journal

[Slideways]

Hmm, what happens if the thermostat is stuck open and it never reaches 80C? I really should get to fixing that

[karter16]

It was raining heavily today so I drove to work and took the opportunity to run some logs.

I've written previously about the CSL MAP sensor and its function. As a quick recap it performs 2 functions during normal operation. One is to act as a kind of running offset between calculated actual conditions and the AlphaN map, for this reason the MAP integrator value rf_psau_i is stored as a adaption in EEPROM. The second thing is does is act more instantaneously to address variations in realtime.

It's fairly commonly known that the integrator is limited to +-2.5%. As I've noted before it's actually +-2.5% of max RF, which in the range the integrator is active in is actually more like +-10% of current RF.

I've previously hypothesised that in normal operation the integrator doesn't bounce off its limits anyway, I just haven't actually got around to posting any evidence of this. So here we go

This is a normal drive home from work. rf_psau_i integrator is always applied to final RF, but the integrator is only updated once engine temp reaches 80 degrees, hence the straight line at the start of the log.

What we can see more generally though is that once the TMOT condition is reached the integrator is very active. We can also see that it hasn't got anywhere near its +-2.5 limits, sticking between +1.094 and -1.205.



I thought another way to understand how much of the time the MAP sensor is actively correcting RF would be to look for the percentage of time it's changing. So here we go:

Once TMOT=80 is reached in the log above there are 78,859 records captured at approximately 100Hz frequency.

Of the 78,859 records, 41,655 of them have an rf_psau_i value that differs in value from 10 records (100ms) earlier. By this measure the MAP integrator is active approximately 52.8% of my drive home. If we extend the measure to look where it differs in value from 100 records (1 second) earlier, the figure rises to 64.4%.

In short - for round town and highway commuting the MAP sensor is going to have a significant effect :-)

[karter16]

Life continues to be busy, have managed to make some progress in a few places.

I've finalised the design of the Gauge.S enclosure and done a final test assembly (see photos below). I have the final prints too ready to be sanded and painted.







I've also been putting the dev version of Community Patch V2 to full use. I've been doing more work on understanding the behaviours of the cylinder air mass calculations specific to the CSL and looking into a couple of interesting curiosities. I've been maxing out the 3 custom CAN messages as well as pulling across slave-only parameters. A lot of this investigation I'm doing would be pretty much impossible without the flexibility of being able to pull, at high data rate, the specific variables I need.


I also received my second-hand E86 front triangulation braces that I ordered while I saw them for cheap. Looking forward to getting heinzboehmer's front triangulation brace together and installed when I get the chance!

[karter16]

Boom! AVAN1_SOLL (target exhaust vanos angle), which is a Slave-only variable, logged at 100Hz over CAN



Now to build out a bunch of test cases and work my way through them...

[0-60motorsports]

Glad you are feeling better!

[karter16]

Spent the last week in bed with Covid, and only just re-emerging now.

Directly before I went down I completed building out the generalised configurable CAN message functionality for inclusion in the next version of Community Patch. This functionality is an extension of the custom program ROM I've been running for the last few months.

• Up to 3 configurable CAN messages for a total of 24 bytes of variables of your choice at 100hz.
• Ability to push up to 16 bytes of Slave-processor-only variables to the DPR (shared RAM) to make them available to then expose in the configurable CAN messages. This is something new I've been working on and means there's now no restriction on what variables can be exposed to CAN, if it exists in the DME you can push it on the CAN bus.
• Configuration validation on DME startup for both CAN message config and Slave-to-DPR config.
• New functionality off by default, must be configured to enable it (means backward compatibility with any 0401 Tune file)

If I get a chance this weekend I plan to flash the program and test it out. I'll then need to flash a bunch of different tune files to validate the various combinations of config validation, etc.

This feature is only fully useful to others when available alongside a list of DME variables. Obviously we don't know all of these, but I'm intending to extract a list from the disassembly, along with descriptions where we have them, that people can refer to. With that and some decent instructions it should be a fairly useful feature.

[0-60motorsports]

With my DME recovered I was able to finally finish off the last little bit of something I've been working on for a while: MSS54HP CSL 0401 Community Patch Binaries

These binaries wrap up a bunch of QoL improvements into a single package. I was very close to having this done when I messed up a memory reference and that's what bricked my DME. With my DME recovered I've been able to complete testing on this, very pleased to have this done and out there.

As noted in the post I need to swap over to the CSL boot loader and test that version before I make it available. I'm extremely confident it is correct as it's the exact same as the version for the 2300 boot loader minus the additional changes needed to get the 2300 boot loader working, but unless someone volunteers to test it it'll need to wait until I get a chance.

Thats awesome! Thank you for doing that mate.

[karter16]

With my DME recovered I was able to finally finish off the last little bit of something I've been working on for a while: MSS54HP CSL 0401 Community Patch Binaries

These binaries wrap up a bunch of QoL improvements into a single package. I was very close to having this done when I messed up a memory reference and that's what bricked my DME. With my DME recovered I've been able to complete testing on this, very pleased to have this done and out there.

As noted in the post I need to swap over to the CSL boot loader and test that version before I make it available. I'm extremely confident it is correct as it's the exact same as the version for the 2300 boot loader minus the additional changes needed to get the 2300 boot loader working, but unless someone volunteers to test it it'll need to wait until I get a chance.

[karter16]

After a bit of a prolonged shambles the second USBJTAG uLink NT FINALLY arrived yesterday, so this afternoon after work I put together an adapter cable and BDM'd the DME to recover it. Aside from Microsoft causing me grief all was straightforward and I can now connect to the DME with INPA. I restored to the base CSL program and tune so tomorrow morning I'll flash my latest program and tune files and get up and running.

Why did this take so long?

I ordered the first uLink NT from USBJTAG at the beginning of November. about two weeks later I stopped seeing any updates on tracking. I chased them up (6 weeks after ordering) and was told to be patient, chased them again (10 weeks after ordering) asking for a resolution and they simply never replied. I generally don't like doubling down on a poor experience but in this case didn't have too many alternative options so I ordered a second one which thankfully made it here. I still have had no resolution from them regarding the first one that went missing in transit.

The device and software is great, very easy to use and now that I have it I'm very happy with it, but I do advise caution/realism around the shipping side of things. Hopefully mine is an outlier experience in this regard and others don't encounter it as this is a very simple and clean solution for those of us with ST eeproms.

[davidinnyc]

Does this help in the meantime? https://cad.onshape.com/documents/40...f5b6a01462b639

I might need to borrow this…

[karter16]

Nice job! But maybe you will have the same problem as I do, you can't properly open the false floor now, because it could get stuck against the bracing.
Only a problem if you often want to access the tools below or store some stuff between there.
But I will have to make bigger cutouts sometime because it's little bit annoying.

Yeah it's a good point - I accounted for this enough with mine to mean I can lift the floor up enough to get the tray out without having to remove the floor, but you can't tip it all the way up to the trunk lid. Good point though to remember this when doing your cutouts.

[rtrules]

Nice job! But maybe you will have the same problem as I do, you can't properly open the false floor now, because it could get stuck against the bracing.
Only a problem if you often want to access the tools below or store some stuff between there.
But I will have to make bigger cutouts sometime because it's little bit annoying.

[Bry5on]

Wow that came out miraculously well. Nice job.

[karter16]

Life has been super busy since August or so. Over the summer break I was mostly prioritising spending time with family and getting things done around the house. This weekend though my wife and kids have been busy with some friends which has given me some free time.

I've been avoiding doing the trunk interior reinstall after doing the rear RACP brace because I keep changing my mind on the best way to approach it. Finally decided on the following approach.

1: Modify the stock trunk false floor to fit around the RACP brace.
2: Make a cutout in each of the side forms just large enough for the brace to protrude through (e.g. side form is installed after the brace is installed), with a vertical cut to aid installation. As far as I can tell from looking at photos this is how BMW did it with the GTR Strassenversion so figured I'd go with the same approach.
3: Cutouts in the trunk carpet around the vertical tubes for the rear points of the RACP brace.


First up - the false floor. It's made of a composite shell with a rubber foam core, with a felt-type material over the whole thing. I wanted to make it look as original as possible so experimented with separating the felt layer with a blade and the like. It proved to not be possible to separate the felt that way, so I came up with a truely dumbass idea.



Prior to this I made a cardboard copy of the false floor and used it to verify the exact shape I needed to cut out of the false floor. I then cut out a shape 10mm smaller, and used a Dremel cutting disk which had 10mm of cutting depth, to slot out the foam and then grind out the composite shell leaving only the felt. I'm as surprised as you are that this actually worked, although I still think it's a dumb idea. I managed to execute this without any mistakes, and was then able to glue and fold over the felt. I then trimmed the edge the following day. If you look at it closely you can tell it's been modified, but from a few feet away you can't tell. I'm pleased with how it turned out. I considered various options like making a replacement false floor, but to get something lightweight enough was getting pretty expensive for something I didn't feel that strongly about, and still wasn't going to have all the perfect form shapes for mating with the polystyrene tool tray like the original has.

Apparently I didn't take any close up photos, but here's a photo of it installed, which ironically doesn't really show all the effort I went to with felting the edges of my cuts:



This morning I moved on to the side forms. I'd previously bought a replacement LHS side one that doesn't have the space for the CD changer etc, as I don't need it now I have the Xtrons head unit and I'd rather have the additional trunk space.

As I mentioned above I spent some time months ago looking at photos of the GTR and figuring out what they had likely done. I think they simply made a cut, and then shaped out around where the tube of the brace protrudes through. I decided the same approach was good enough for me. I didn't want to go the route of having a cut out large enough to install the form first and the brace second - I wanted it to look as OE as possible.

Fortunately the luggage loops provide an excellent reference point for measurements as the side forms have cuts in them where the loops protrude through. I carefully measured everything up, under cutting to begin with. A few back and forths and I had them fitting very nicely indeed!







The forms are pretty stiff, so once installed properly it all fits back into place (you can actually see in these photos that the RHS one isn't quite aligned, there was a cable running by the luggage loop that seems to have been misrouted by someone in the past and was making the side form sit slightly funny, relocated that and it looks good. I'm very happy with the outcome - this worked out how I was envisioning - fortunately my execution was good enough!

Then finally it was on to installing the side trays and the trunk liner.

Despite a lot of measuring I managed to cut the lining so that it is about 1/3 of an inch too far forward. I rechecked my measurements afterwards and they still seem good to me, so seems to be a case of just the way it lies doesn't quite match up. To be honest you guys and myself are the only people who will ever notice, but I will pick up a replacement liner when I next see one and will redo this to satisfy my own unrealistic standards of myself when it comes to doing things perfectly that I've never done before.


Here is the finished result:









Still to do:

- Trim the rear bench seat sound mat that you can see protruding through.
- Put the caps on the rear vertical pipe sections of the brace.
- Install the push pins things that keep the side forms in place.
- Clean the carpets now that I've finished messing around with install.

[karter16]

I'm super happy with how these turned out!

Laser-cut genuine Alcantara Soft (9002 Anthracite).