[OpenFlash]

Originally Posted by Joshboody

can't open the log files... should excel work? Do you need procede software?

You can view them in Excel if you'd like. Or you can view them graphically with the Procede logger (which is a part of the user software). All you need to do is uncompress the .zip attachments. Then they should be in .xls format.

Shiv

[Joshboody]

nevermind, I use Jzip and not winZip... that's the difference believe it or not

[OpenFlash]

BTW, thanks to everyone for keeping this thread civil and informational. It's nice to see both "camps" honestly interested in the data. Looking forward to more constructive input. Hopefully more people (including BMS) will do similar comparative testing/data collection in the near future.

Shiv

[Renus]

Push push, nice job Shiv ^ ^

Kenji

[Big Tom]

Some Vbox comparsions between the tunes would be interesting.

[R1000K3]

It will be done as soon as the current glacial period is done with

[335 Freak]

Originally Posted by Bimmer335i07

I have to agree with Mike. Not many members are running the 18ohm boards. If you want this to be a fair comparison, please use the 24ohm board.

Anybody that purchases a standard JB3 2.0 or lower will get the 24ohm board.

As a JB3 2.0 Beta Tester of myself, I'm kind of surprise the guy who gave up the JB3 2.0 was using 18onm board!! I thought Tery/Mike recommended all 2.0 Beta testers to use the 24 ohm board for testing the updated beta firmwares for more accurate result through our data logging returned to Terry for making the analysis in straighten out the final mass production version of 2.0 firmwave!! Should this be a fair data logging comparison for JB2.0 till it produced its final produtional firmwave while V4 does?

Keep up the good work, Terry/Mike!!

[Sal@AUTOcouture]

Originally Posted by 335 Freak

As a JB3 2.0 Beta Tester of myself, I'm kind of surprise the guy who gave up the JB3 2.0 was using 18onm board!! I thought Tery/Mike recommended all 2.0 Beta testers to use the 24 ohm board for testing the updated beta firmwares for more accurate result through our data logging returned to Terry for making the analysis in straighten out the final mass production version of 2.0 firmwave!! Should this be a fair data logging comparison for JB2.0 till it produced its final produtional firmwave while V4 does?

Keep up the good work, Terry/Mike!!

IIRC, it was a 1.47 that was updated to 2.0 via chip burner

[Graystone]

Civility is nice to see. Two great minds combined ,..30psi stock turbos.....food for thought.. then we could all have JBishu 3v's
Seriously Would love to see some of the more educated experienced bumbles to circle and explain with lines and pretty colors the thinks we laymen are missing.
Tks to all for info...

[kjrulz]

Good thread thus far.

[Joshboody]

how about maintaining boost between shifts with 6MT... looks like JB is better at this?

It would be interesting to see these logs with hotter temps... JB's timing curve looks better then expected. v4 is at 100% correction right?

[cstavaru]

I see that JB3 reaches 16.5psi while Procede reaches 18psi in the datalogs ? Isn't 18psi a bit too much ?

[OpenFlash]

Originally Posted by Joshboody

how about maintaining boost between shifts with 6MT... looks like JB is better at this?

It would be interesting to see these logs with hotter temps... JB's timing curve looks better then expected. v4 is at 100% correction right?

Both tunes were run at their default settings. This means 100% ignition correction for the Procede. The Procede needs to be shifted a bit faster in order for it to naturally hold more boost between shift. The JB3, on the other hand, looks like it applies an active bogfix that clamps the Map sensor ouputs to a low value briefly just after the shift (sew the drop in debug word 6 just post-shift) similar to what we did with v3. And yes, it holds good boost between shifts.

Both cars are pretty much getting the best possible condition as far as knock tendency goes since intake temps under boost are so low (close to the 50F ambient temp. Judging by the 3rd gear 2500-7000rpm logs, the Procede car would benefit (gain power) by a reduction of ignition correction. And the JB3 would benefit (more ignition advance stability) by adding some ignition correction. This condition would likely change in hotter conditions or with the FMIC upgrade.

Shiv

[OpenFlash]

Originally Posted by cstavaru

I see that JB3 reaches 16.5psi while Procede reaches 18psi in the datalogs ? Isn't 18psi a bit too much ?

Look at the raw data because I don't think you are reading the peak values correctly on the graph. It's a bit tough since the y axis scaling (although held constant for comparison purposes) is a bit cramped.

Shiv

[RBC]

I sure would like to maintain more boost between shifts...

Is there any reason why Procede isn't doing it?

Many users say that JB3 is stronger up top... This may be the reason why...

[OpenFlash]

Originally Posted by RBC

I sure would like to maintain more boost between shifts...

Is there any reason why Procede isn't doing it?

You just have to shift faster with the Procede for this to happen. Dropping upshift softening to 0 also helps by eliminates some of the throttle closure that occurs during the shift. We may go back to applying an active "bogfix" if we can keep the post-shift boost pressure stable. With it, boost tends to get a little stable after the shift due to the sudden change in perceived DME boost. Ultimately, I really wanted to stay clear of band-aids with v4 and keep the actual boost control logic simple and fast acting.

Many users say that JB3 is stronger up top... This may be the reason why...

The reason is ignition advance. Out of the box with default settings, the Procede retards 3-5 deg of timing under full boost. In some cars (mainly those running in good conditions, with good gas and with upgraded FMICs), this is unnecessarily conservative. In lesser modified cars, this setting is close to spot on.

For those in the former case, you can get this extra power back by just dropping ignition correction a bit. Dropping it to 0 (from 100% default) would give you JB3 like timing curves which are advanced but, at times, quite unstable. Setting it to a middle ground (50-70%) gives you the same max timing values but with improved consistency.

Shiv

[Kelvin1000]

Looking at these logs is great and gives you a good idea about the operation of both of these tunes. I have to say that they both look pretty good and they are both getting the job done nicely (though with a different approach).

The Procede seems to be more refined but its funny how the JB3 has a simple approach but manages to run so well in so many conditions. The boost through the shift (Bogfix) works nicely too.

Shiv,

Why does the JB3 place so much emphasis on IAT versus the Procede? Terry always said that IAT had a big influence in everything the JB3 was doing...

Originally Posted by shiv@vishnu

Lots of data. Default user adjustable settings on both cars. No gain changes, no ignition correction changes, no user torque changes.

Jb3 v2.0 is running Map7 map (2-17)
Procede v4 is running Stg 3 map (2-23)

Stg 3 modified 6MT (FMIC, catless dps and cat-back w/cat deletes)
61k miles
91oct fuel

3rd gear WOT pulls (2500-7000rpm)

JB3 v2.0



Procede v4



Also attached is the raw datalog files. So you can view the runs however you want, with whatever scaling you want.

Next up is 3rd to 4th gear WOT datalogs.

Cheers
shiv

[huyner328]

Possible to hook up a 3rd tune to this jerry rig?

[._.]

Originally Posted by Kelvin1000

The Procede seems to be more refined but its funny how the JB3 has a simple approach but manages to run so well in so many conditions. The boost through the shift (Bogfix) works nicely too.

There is an elegance in simplicity, many coders have a hard time doing it. Why write hundred's of lines of code when 10 accomplish relatively the same thing. Other than the extra really cool features the Procede has .

[cn555ic]

Originally Posted by R1000K3

It will be done as soon as the current glacial period is done with

Tell me about it, in the USA, northeast region its been snowing like crazy...Some places have recorded over 60 inches of snow in February!!! Global warming...psst

[supracg]

Originally Posted by r45t4m4n

There is an elegance in simplicity, many coders have a hard time doing it. Why write hundred's of lines of code when 10 accomplish relatively the same thing. Other than the extra really cool features the Procede has .

I don't think there is any evidence that the PROcede has more lines of code than the JB3 does. I do agree with you in that the PROcede has more hardware parts than the JB3, but that doesn't necessarily correlate to significantly more code; actually, where the JB3 lacks in hardware, Terry makes up for in code/algorithms (ie. estimating DME target load).

My 2 cents, Mike/Shiv feel free to correct me if im wrong.

[OpenFlash]

Originally Posted by supracg

I don't think there is any evidence that the PROcede has more lines of code than the JB3 does. I do agree with you in that the PROcede has more hardware parts than the JB3, but that doesn't necessarily correlate to significantly more code; actually, where the JB3 lacks in hardware, Terry makes up for in code/algorithms (ie. estimating DME target load).

My 2 cents, Mike/Shiv feel free to correct me if im wrong.

As for as tuning-related code goes, v4 is far simpler than v3. And v3 is very similar, in concept, to jb3 2.0. The reasons v4 is able to trim so much fat is that it doesn't have to "model" anything. There is no need for all these watchdogs that we had to apply with v3 to keep the car running well. With v4, we decided to make better use of the all hardware that the Procede packs. This meant that we increased the number (and quality) of inputs. With more info, the Procede can make better decisions quicker. And require far fewer lines of code to do it. And you can see the benefits of this just be comparing boost logs and DME boost error graphs. And that is just when testing relatively simple-to-control full throttle/max load conditions. The partial throttle/dynamic test results will be another story,

Ultimately, the Procede may have more total code since it also offers a bunch of diagnostic and convenience features.

Shiv