CanBus temp guage

[ss2115]

Was doing some research on hacking the Canbus system and came across and interesting mod for the E46's where the OEM temp gauge is designed to point upwards in the 12 o'clock position for a very wide buffer reading of 75 - 115degrees.
The E46 owners were not happy with this and someone researched a iBus hack that reduced the 12 o'clock position to buffer 90 - 100 degrees only and nicely spaced either side over the remainder of the gauge with 115 on the beginning of the red mark. Its a simple matter of changing some of the numbers in a string of an i-bus code, but I don't know enough about the codes or interpreting to work it out myself.

I've done some searching to see if anyone's done the same for our E39's but cannot find anything similar.
I've looked through the i-bus code of my E39 IKE using PASoft but the codes don't coincide with the E46 codes.

Anyone clever done this or perhaps have read about it somewhere for the E39 and can point the way with links?

[geargrinder]

Our clusters contain the code for the temp buffering. I did some hacking around with the more easily implemented coding for the cluster (there's some tables you can work with) and got some slightly better temp resolution. Because of reasons, you can't fully fix everything with just coding. For example there's still 'time buffering' so the gauge doesn't change rapidly, and thats in programming not in coding values. Then there's an issue w/ coding bins and a limitation having to do with 8-bit scaling that means you can only get it so good but basically I had my cluster setup so that "normal" was more or less on the left hash, and that indeed you could see more granular temp changes sooner vs the stock "needle will only suddenly move at the same time the engine is melting down".

I haven't re-applied that coding to my new M5 cluster since I got it, I do have an action item to get around to that... todays a crappy rainy day maybe I'll do it...

[DustenA]

Our clusters contain the code for the temp buffering. I did some hacking around with the more easily implemented coding for the cluster (there's some tables you can work with) and got some slightly better temp resolution. Because of reasons, you can't fully fix everything with just coding. For example there's still 'time buffering' so the gauge doesn't change rapidly, and thats in programming not in coding values. Then there's an issue w/ coding bins and a limitation having to do with 8-bit scaling that means you can only get it so good but basically I had my cluster setup so that "normal" was more or less on the left hash, and that indeed you could see more granular temp changes sooner vs the stock "needle will only suddenly move at the same time the engine is melting down".

I haven't re-applied that coding to my new M5 cluster since I got it, I do have an action item to get around to that... todays a crappy rainy day maybe I'll do it...

I read his post and was going to link your thread, but you covered it.

[ss2115]

Our clusters contain the code for the temp buffering. I did some hacking around with the more easily implemented coding for the cluster (there's some tables you can work with) and got some slightly better temp resolution. Because of reasons, you can't fully fix everything with just coding. For example there's still 'time buffering' so the gauge doesn't change rapidly, and thats in programming not in coding values. Then there's an issue w/ coding bins and a limitation having to do with 8-bit scaling that means you can only get it so good but basically I had my cluster setup so that "normal" was more or less on the left hash, and that indeed you could see more granular temp changes sooner vs the stock "needle will only suddenly move at the same time the engine is melting down".

I haven't re-applied that coding to my new M5 cluster since I got it, I do have an action item to get around to that... todays a crappy rainy day maybe I'll do it...

Did you say there was a table available. Do you have a link please.
And which line of code for our IKE is the temp guage please?
In regards the normal position, from your own playing around is it possible to still have about 95 deg in the 12 o'clock position. I quite like the normal being the guage "straight up". Better at a glance and either side tilt attracts immediate attention.

[DustenA]

Did you say there was a table available. Do you have a link please.
And which line of code for our IKE is the temp guage please?
In regards the normal position, from your own playing around is it possible to still have about 95 deg in the 12 o'clock position. I quite like the normal being the guage "straight up". Better at a glance and either side tilt attracts immediate attention.

http://www.bimmerforums.com/forum/sh....php?t=2056901

[ss2115]

http://www.bimmerforums.com/forum/sh....php?t=2056901

Thanks you for the speedy reply.
I'll have a read through.
Cheers.

[geargrinder]

Did you say there was a table available. Do you have a link please.
And which line of code for our IKE is the temp guage please?
In regards the normal position, from your own playing around is it possible to still have about 95 deg in the 12 o'clock position. I quite like the normal being the guage "straight up". Better at a glance and either side tilt attracts immediate attention.

Its slightly complicated and I'll have to refresh myself on the details quite frankly - it was quite a while ago when I effed with it. It's not quite as easy as a simple single "temp degrees -> gauge angle" table. There's at least 2-stages of signal-to-gauge coding.

IIRC one of the tables is something like 'incremental movement from last position' or something like that, but then there's another problem like (and again... IIRC from weak memory...) something like since the 'bins' are 8 bit you can only 'add' 256 between each 'bin', capping how steep you can make the angle. Vaguely something like that. Can you tell I don't recall entirely?

[ss2115]

Its slightly complicated and I'll have to refresh myself on the details quite frankly - it was quite a while ago when I effed with it. It's not quite as easy as a simple single "temp degrees -> gauge angle" table. There's at least 2-stages of signal-to-gauge coding.

IIRC one of the tables is something like 'incremental movement from last position' or something like that, but then there's another problem like (and again... IIRC from weak memory...) something like since the 'bins' are 8 bit you can only 'add' 256 between each 'bin', capping how steep you can make the angle. Vaguely something like that. Can you tell I don't recall entirely?

Ha Ha.
Yeah - it seems clear it was a past project achieved and then forgotten.
The E46 forum thread I found only changed a few numbers in one line of code. http://forum.e46fanatics.com/showthread.php?t=1013600. It seemed to be well researched.
Despite the E46 and E39 being different (I believe the E39 is less CanBus than an E46) would it not still be a single line of code of 16pairs? You suggest there's more to it.

[ss2115]

Ha Ha.
Yeah - it seems clear it was a past project achieved and then forgotten.
The E46 forum thread I found only changed a few numbers in one line of code. http://forum.e46fanatics.com/showthread.php?t=1013600. It seemed to be well researched.
Despite the E46 and E39 being different (I believe the E39 is less CanBus than an E46) would it not still be a single line of code of 16pairs? You suggest there's more to it.

Okay - forget my last question. I've read through your old post and its pretty detailed. much of it over my head in regards the actual coding. I was hoping to just copy a code into PASoft or NCS and write it and have a temp guage that sat in the middle but for a narrower range and gave me a clue if the engine was getting hotter rather than 'bang - its too late'.

Why though do you have the normal range as around 105 - 110?
I know the BMW's always run hot but I didn't think it was quite that hot. I always thought it was around the 95, 98 mark?
My car has the electric thermostat but to my knowledge it works normally mechanically and only has the electric solenoid to push it open in anticipation of increased heat such as sudden flooring of the accelerator or heavy load conditions.

[DustenA]

Okay - forget my last question. I've read through your old post and its pretty detailed. much of it over my head in regards the actual coding. I was hoping to just copy a code into PASoft or NCS and write it and have a temp guage that sat in the middle but for a narrower range and gave me a clue if the engine was getting hotter rather than 'bang - its too late'.

Why though do you have the normal range as around 105 - 110?
I know the BMW's always run hot but I didn't think it was quite that hot. I always thought it was around the 95, 98 mark?
My car has the electric thermostat but to my knowledge it works normally mechanically and only has the electric solenoid to push it open in anticipation of increased heat such as sudden flooring of the accelerator or heavy load conditions.

The 540 runs over 100c most the time

[ss2115]

The 540 runs over 100c most the time

So going back up to an earlier reply of yours, you mentioned you had "normal" temp on the left hand marker of the usual operating range.
So if I used your code, my 'normal' might be hanging the needle a bit lower still - probably just under the left hand operating marker?

[geargrinder]

OK. I actually dug up my old files on this (thanks for posting link Dusten I didn't have that handy either!) And am working on another set of values.

Just realized I need a new calibration for 88C t-stat now anyway, and maybe I can get a better "centered" calibration for OEM guys. Will make a new post & update later.

Oh - and yeah, ... No. The OEM MAP t-stat is basically a mechanical 110C almost all the time. It can drop to "like an 85C" momentarily but in reality that has very little impact on the median temp. So 105-110 is a good approximation.

[geargrinder]

Just updated that old thread w/ some new values I'm messing with.

[ss2115]

Just updated that old thread w/ some new values I'm messing with.

Thanks.
I'll switch to that thread and monitor it.
Cheers

[DustenA]

Looking forward to a ncs method. I don't have pa soft