PCMtec

[K4fxd]

The right click hide all except for changed from stock helps.

I would not like to see things go away. Never know when or if that table might help in some situation.

Sponsored

 

[K4fxd]

By having all these tables I figured out I was populating the op stability table wrong. I got rid of a slight stumble I have been fighting with lock outs.

 

[K4fxd]

Welp I found the downside of having all these tables. First mistake was changing several things at once. Anyway I flashed the tune and my idle would randomly drop to 500 then raise back to normal. There was no pattern and the only thing I saw in the log was the ign timing changing. So scratched my head for a while re-flashed and continued to scratch my head.

Went and flashed an older finished flex tune and the problem went away.
This is where I'd like the compare to be a bit easier, but I suppose things like this will get me to learn this system better.

Up side is maybe I can figure out why some people get this floating idle.

 

[K4fxd]

I cannot figure this out. When I compare the tunes, one has idle messed up one is rock stable, nothing stands out. There are only 4 things different.
1 is I raised the idle RPM from 700 to 800.
Spark is different but only at WOT
I changed a shift point in the trans.
I changed the OP intake cam timing at WOT.

I even changed the idle speed back and it still floats.

Now I understand Lund telling people it is mechanical. But I know this is tune related because the revision before this one runs perfectly.

 

[Rolls]

Remember that there is KAM and learning values. When you say one runs perfectly, try flashing it in and leaving it for a week and see if it still runs perfectly. It may "learn" the issue.

Also vice versa, you typically want 80kms of varied driving (eg basically have all your emissions codes passed) before all the learning has completed. Only then would I judge the idle quality.

Also FDRS has some fantastic diagnostic procedures, same with Forscan. Buy a 2 day copy of FDRS and run through all the diagnostic procedures.

You can also datalog all the learning values, if you post on our forum I can give you the datalogger IDs of the idle learning errors accumulators. This will help tell you if you have a mechanical issue.

if you can post the two tunes up on our forum, we can also have a compare and see if there is anything unmapped that is changed (eg something outside the calibration area). I assume these are identical strategies/operating systems?

 

[engineermike]

 

[Rolls]

Love it!

 

[bankyf]

Question for anyone who has messed with injector data in PCMTEC. ID refers to, and HPT has a table called Injector Offset Multiplier vs Rail Temp. The only thing I can find in PCMTEC is Injector offset correction as a function of temperature. HPT is a multiplier and PCMTEC is a correction in time(us). I would assume this accomplishes the same thing, but I believe the stock file has a multiplier of something other than one at most temperatures in HPT, but PCMTEC shows 0us correction for all temps on the stock file. Am I missing something, or just overthinking this?

Edit: After digging further, Injector Slope vs. fuel rail temp is showing all 1's for my stock calibration in HPT and ranges from .965-1.057 for the stock calibration in PCMTEC, where Offset multiplier mentioned above has all 0us(I believe the equivalent to an offset of 1) in PCMTEC and ranges from 1-1.120 in HPT. Could these tables be defined incorrectly by either PCMTEC or HPT?

 

[K4fxd]

@Rolls

Might be faster to ask this on PCM's forum.

 

[Rolls]

Check your decimal places setting. Press + on the scalar in case it has rounded it to 0 for display due to being a very small number. If you post on the forums we can find the scalar/table for you.

We have seen errors in all major tuning companies templates before, we also are not immune though we do a lot of runtime checking for tables (eg we validate the internal structs to ensure equations/addresses are what the binary says they should be). This runtime validation is only possible on tables though.

I would think it is likely you are either looking at a different table, the units are different, or decimal places are set too low (double click to edit the value and see if you see a very small number like 0.0000000123 etc). A mapping error would be easily determined via comparing the same table/scalar in say 10 stock files and seeing if they are all approximately the same. If mapped wrong you would usually see complete garbage like 459879437598 for the number, you will also get a yellow error in the log saying "unusually high/low value, binary validation error". If the error is bad enough the table/scalar will be automatically hidden.

We manually audit these reports as well to do our best to validate all our templates are correct. As there are 100,000s of parameters we pick the top 10,000 most commonly tuned parameters to check by hand. Checking all is obviously not feasible due to the time required however I believe we got to a much greater extent to checking our templates vs anyone else with the runtime checking catching anything out.

You'll see it on files read from HPT very occasionally, the axis is all messed up or you get a table that is 400x400 long or full of random looking numbers where they have mapped the struct address instead of the data address.

 

[engineermike]

Just fyi, pcmtec got my os working. I did a test write in 19 seconds and I included the optional tcm write. I think it’s a few seconds faster when only writing engine parameters.

 

[shogun32]

where they have mapped the struct address instead of the data address.

Ahh classic C mistake. And I thought all you wizbang kids were using the latest hotness: Go or Rust.

 

[K4fxd]

Just fyi, pcmtec got my os working.

Which OS are you using?

The write speed is incredible. It is dependent on internet.
I did a flash away from wifi and it did a full write, about 1 min 30 seconds.

 

[engineermike]

Which OS are you using?

The write speed is incredible. It is dependent on internet.
I did a flash away from wifi and it did a full write, about 1 min 30 seconds.

Mine is ZRBJ5. I have an early 2018 kit and Whipple apparently did some really weird stuff to make it work. Like it’s a control-pack os apparently.

 

[Rolls]

Mine is ZRBJ5. I have an early 2018 kit and Whipple apparently did some really weird stuff to make it work. Like it’s a control-pack os apparently.

Yes for anyone interested it appears they disabled checksumming and used a control pack base. Or at least decided to omit a bunch of functionality/checks that are normally in the file.

A lot of the large tuning shops tend to use a small list of OEM strategies when tuning Whipple/Roush vehicles as once they have done a few all the quirks are known to them and they can get a reliable result.

If you are using files like a whipple OS, and then comparing to an OEM OS, you'll likely find you cannot copy tables over and get the same result due to some fairly dramatic changes they have made to how the HDFX is configured.

Personally I would recommend if tuning these vehicles that you use the original factory OEM operating system and spend the time to make that work, once its done you can re-use these as base files on other vehicles.

Sponsored