Auto upgrade from build 72 to build 81 kills serial wideband.

[kp]

I recently ran an online update on build 72 and it updated it to build 81 (scanner). Ever since the serial wideband quit working on pass through with the V2. Uninstalled it and downloaded 72 and it worked fine again.

On build 81 the PID is there as before but it will just switch between 2.56 and 0.00 randomly and doesnt change no matter what the wideband reads. The wideband works fine on the V2 and if I uninstall build 81 and go back to 72.

Not sure whats going on..

[mr.prick]

Try build 82

[kp]

I may be missing something but where do you get 82 from, the main download is build 72 and it auto updates to 81 and then it says no new updates.

[gmperformancecentre]

do a auto update and you should have build 82 after updating
build 82 came out a couple days ago

[kp]

do a auto update and you should have build 82 after updating
build 82 came out a couple days ago

I did, like 5 times, still have build 81 on both my laptop and desktop.

[joecar]

With build 82 can you see the serial AFR pid...?

[CV8IIM6]

I upgraded to build 82 and have the same problem.

The AFR display on the V2 is working but the data log shows 0.00

The PID is still visible and validated

[mr.prick]

I had this happen to me the other day with the RAFIG PIDS.
I had drove for 2 hours to my destination with no problem,
I left my PC in hibernation and after I started logging on the way back home
I noticed the PIDS were not showing. any data.
Weird.

[joecar]

Can you post your log files where you see serial AFR reading 0.00.

[Blacky]

I did, like 5 times, still have build 81 on both my laptop and desktop.

Use Notepad to open the file:
\Program Files\EFILive\V7.5\EFILiveV7.original

Check the value of VersionOriginal

[OriginalSetup]
VersionOriginal=82

If it is set to 82, then the auto updater thinks you already have 82 installed so it won't download/install the current 82.

To force it to re-download and install 82, change that value back to 81 and retry the Check for Updates option.

Regards
Paul

[kp]

Use Notepad to open the file:
\Program Files\EFILive\V7.5\EFILiveV7.original

Check the value of VersionOriginal

[OriginalSetup]
VersionOriginal=82If it is set to 82, then the auto updater thinks you already have 82 installed so it won't download/install the current 82.

To force it to re-download and install 82, change that value back to 81 and retry the Check for Updates option.

Regards
Paul

That worked, I have no idea why my two computers said 'VersionOriginal=82' and had build 81 on them.

I'll go try the wideband now

[kp]

Still no go, wideband just sits and flashes between 0.00 and 2.56 with build 82, thought maybe it was an issue with the new Atom processor so I drug out my older laptop.

I tried it and it worked fine with build 80 that was on it, let it update to build 82 and it stopped working also.

On both of them if I uninstall 82 and put 72 back in they both work fine.

Here is a log of it not working, car isnt running and its just in free air, but it should read 25.0 on the AFR. If I run the car it will still read 0.00/2.56.

[Blacky]

That worked, I have no idea why my two computers said 'VersionOriginal=82' and had build 81 on them.

A few days ago, I incorrectly posted the build 82 update with build 81 files in it. That was "live" for about 30 minutes until the problem was detected and fixed. If you downloaded/updated in that 30 minute window then you would have that situation.

Regards
Paul

[Blacky]

Still no go, wideband just sits and flashes between 0.00 and 2.56 with build 82, thought maybe it was an issue with the new Atom processor so I drug out my older laptop.

I tried it and it worked fine with build 80 that was on it, let it update to build 82 and it stopped working also.

On both of them if I uninstall 82 and put 72 back in they both work fine.

Here is a log of it not working, car isnt running and its just in free air, but it should read 25.0 on the AFR. If I run the car it will still read 0.00/2.56.

Are you logging using black box logging or pass-through logging to the laptop?

Regards
Paul

[kp]

Are you logging using black box logging or pass-through logging to the laptop?

Regards
Paul

pass through, black box works fine.

wait, the wideband displays fine on the V2, I havent made a black box log to see if it reads ok in build 82.

[joecar]

In Log0000.efi all the pids have "zero" values... the data is all some non-value.

[kp]

The car isnt running right now, I noticed the wideband not working yesterday morning after I ran the auto upgrade when it was still together and running and couldnt figure out why so I put 72 in and it worked fine, didnt savee any logs when it was running when I was trying to figure out what was going on.

If I put build 72 in the wideband displays fine without the car running like this log. I put the sensor alone in another running car's tailpipe, the previous log was done exactly the same way, no matter what the wideband reads on the V2 the computer shows 0.00/2.56 and thats it. I had to get the heads off so its not running right now.

[CV8IIM6]

Can you post your log files where you see serial AFR reading 0.00.

Here's an E38 log file from tonight.
V2 displayed correct AFR.
Logging in pass-through mode.

[Wolfie]

Update ate my AFR...
I have the same problem.
My PLX gauge reads correctly,
Flashscan unit reads correctly,
but the computer just sits on 0.00

It worked fine till the update.
I just now changed the EFILiveV7.original back to 81, and
let it update agan. and have the same problem.

[joecar]

Cross reference thread: showthread.php?p=88802 (http://forum.efilive.com/showthread.php?p=88802)

[mr.prick]

Plug in one of the AD inputs and see if you get serial WBO2 to display.
My issue is passthru showing incorrect data or no data with out one of the AD plugged in.
If someone wants to configure my options.txt to show AD volts and serial WBO2 in BBL
then you will see the difference between BBL and passthru.4857
This is definitely a bug in build 82, 81 worked fine.

:doh2:
My apologies for suggesting the OP to update.

[Blacky]

Make sure your FlashScan device is running firmware 2.06.06 or later
Regards
Paul

[swingtan]

So are you saying there are issues with AD WB as well as Serial?

I just did 2 logs, one on build 82 and one on build 67 ( the only other old version I currently have ). As has already been indicated, 82 does not display correct WB data while 67 works fine. I also tried opening the B82 log in the B67 Scantool and the data was exactly the same. So the issue is with the actual recording of WB data, not the display of the data. It is not WB brand specific either, as I tested using a TE and others have used a PLX.

With B82, there was no change from the car being off, not started for 16hours and then starting the car. This should have given the largest change in WB signal, from almost free air to 2:1 starting mixture. The logged data didn't change at all.

Logs are attached.

Simon.

[Blacky]

Is the correct wideband type selected in the FlashScan Setup menu?
Regards
Paul

[swingtan]

Blacky,

I just swapped an old version of "obdiiV7.dll" in to the current working dir and passthrough WB logging works again. I think there is a misconfiguration in the setup data for that file that is reading the wrong data from the USB frames ( big guess I know ).

Here's a log I just did, only swapping....

odbiiV7.dll ver: 7.5.5.82 Modified: Monday, 2 March 2009, 20:12:04

with

odbiiV7.dll ver: 7.5.4.66 Modified: Friday, 8 August 2008, 16:13:32

Simon.

EDIT: This may break a lot of other stuff as this DLL appears to control how data is read from the USB frames. I wouldn't go swapping files around unless Blacky gives the OK first. I only did this as a trouble shooting exercise.

[Blacky]

Can you do the same thing with build 82 and build 66, only this time as well as saving the log files, can you also generate and save an error report after each 5-10 second log?
Please exit and restart the scan tool software just prior to starting each log, that way the error reports will be cleared before you start.

The error trace will show me the exact data being sent by FlashScan, which I will attempt to match up with the data in the *.efi log files. That way I may be able to find out which data EFILive is "seeing" instead of the correct wide band data.

Thanks
Paul

[Blacky]

You can try this version of the V2.06.06 firmware with the build 82 release. I've kept the firmware version number the same so you can revert back to the release V2.06.06 if required.

After programming, the firmware date should show as March 01, 2009.
The released V2.06.06 firmware has a date of Dec 14, 2008.

Regards
Paul

[Blacky]

Ok, I think I have found the problem. The diff command was my friend :) I compared the source code from build 81 to build 82...

In the change from build 81 to build 82 there was a code change in the routine that extracted the external data from the buffer:

Original code: ExternalPID_Index*16
New code: ExternalPID_Index*sizeof(WORD)

Time for me to go back to programming 101... :(
Sorry for the screw-up, I'll get a new release posted tonight.
Anyone care to post what the code should have been changed to?

I still don't understand why it is working here... but I will get to the bottom of it.

Regards
Paul

[Blacky]

I still don't understand why it is working here... but I will get to the bottom of it.

Hmmm, I was running build 60 on the machine I was testing on - I thought it had the latest build 82...

Its not been a good day so far...:shock:

Regards
Paul

[Blacky]

Hmmm, I was running build 60 on the machine I was testing on - I thought it had the latest build 82...

Its not been a good day so far...:shock:

Regards
Paul

The development environment was pointing to build 60 as the host application when testing the obdiiv7.dll - instead of build 82. Unfortunately, build 60 was the build that was regression tested prior to releasing build 82, so build 82 never got tested before release.

Attached is an updated obdiiv7.dll (still called build 82) that should fix the external PIDs and serial wideband PIDs problem.
Just copy it into the V7.5 folder, overwriting the original build 82 version.

I'll release an official and tested update tonight (it will be build 83).

Regards
Paul

[swingtan]

Sorry, about the MIA Blacky, I just had a guy come round to pick up all the old Grey motor gear. Good work though, as usual ;)

Simon

[joecar]

I'll try...



...
New code: ExternalPID_Index*sizeof(WORD)
...
Should that have been:

ExternalPID_Index*sizeof(Int16)

Or something like:

ExternalPID_Index*sizeof(ExternalPID_Array[1])

[Blacky]

I'll try...

Should that have been:

ExternalPID_Index*sizeof(Int16)

Or something like:

ExternalPID_Index*sizeof(ExternalPID_Array[1])

It should have been (sizeof(WORD)*8), because the value is supposed to be the bit size (not the byte size). I changed it to sizeof(WORD) which is only 2 (bytes) and so each value that was being extracted was only 2 bits apart instead of 16 bits apart.

That's assuming sizeof() *always* returns the byte count of a field where a byte is *always* 8 bits - which I am sure is the case on all Intel based Windows systems.

Regards
Paul

[joecar]

Ah, yes of course, now I see clearly... I don't know what previously happened to me... :chair: ...Ross's programming joke was still with me...:hihi:

Yes I agree, sizeof(WORD) and sizeof(INT16) are 2 and not 16...

lol, we both fell into the same hole...:doh2:

Does that mean if bit widths are being used a lot, then a macro like this may come in handy:

#define bit_sizeof(x) (8*sizeof(x))

[kp]

Ok, I think I have found the problem. The diff command was my friend :) I compared the source code from build 81 to build 82...


FWIW build 81 was broke for me using the serial wideband, 80 worked fine.

Unless the build 81 I had was really build 82 :)

[Blacky]

FWIW build 81 was broke for me using the serial wideband, 80 worked fine.

Unless the build 81 I had was really build 82 :)

There was a mixup for about 30 minutes when build 82 was posted and it was labeled build 81. The best thing that can happen is for build 81 and 82 to die quietly and I'll get build 83 out tonight that will fix the problem.

Regards
Paul

[kp]

There was a mixup for about 30 minutes when build 82 was posted and it was labeled build 81. The best thing that can happen is for build 81 and 82 to die quietly and I'll get build 83 out tonight that will fix the problem.

Regards
Paul

Thanks for getting it fixed, of course I ended up in that 30 minute window lol.

[Chevy366]

Was going to do some AutoVE logging last night saw the thread and canceled the session , maybe tomorrow night .
Great work on a fix .
Old Sig :Always be wary of the Software Engineer who carries a screwdriver.

[Blacky]

Was going to do some AutoVE logging last night saw the thread and canceled the session , maybe tomorrow night .
Great work on a fix .

Your signature couldn't be more accurate :) except to say, beware of the programmer using a power saw - I managed to cut through the power cord last time I used a power saw:help2:.

In other news a year or so ago, I managed to end up in A&E requiring stitches after watching the (professional) guys remove tiles with a jack hammer from our bathroom which was being renovated. It looked like so much fun, jack hammer verses porcelain tiles, what could be more fun that that? My wife told me not to touch the equipment, she knows me and power tools don't co-exists happily, but while they were having lunch I snuck in and had a go. Wife refused to take me to A&E, I had to convince the builder to drive me there. Got back, hand all stitched up and they were taking to a marble bath with a sledge hammer. Man I had to have a go of that!!! Well, one lump of marble in the face and one black eye later I called it quits for the day. As I've always said - I'm just a programmer.

Regards
Paul

[Chevy366]

Your signature couldn't be more accurate :) except to say, beware of the programmer using a power saw - I managed to cut through the power cord last time I used a power saw:help2:.

In other news a year or so ago, I managed to end up in A&E requiring stitches after watching the (professional) guys remove tiles with a jack hammer from our bathroom which was being renovated. It looked like so much fun, jack hammer verses porcelain tiles, what could be more fun that that? My wife told me not to touch the equipment, she knows me and power tools don't co-exists happily, but while they were having lunch I snuck in and had a go. Wife refused to take me to A&E, I had to convince the builder to drive me there. Got back, hand all stitched up and they were taking to a marble bath with a sledge hammer. Man I had to have a go of that!!! Well, one lump of marble in the face and one black eye later I called it quits for the day. As I've always said - I'm just a programmer.

Regards
Paul

Yeah , worked in the construction field for a while , and would get calls from ladies asking if I could fix or redo what their husbands had started or finished , I have to say Software Engineers were the worst . :grin: JK
I had a wealthy client , say to me one time , "I wish I had your talent with tools" , I looked at him and said funny I wish I had your talent for making millions .

Gil

[joecar]

Gil,

The software engineer also has software versions of screwdrivers, power tools and other heavy equipment available... and they can be just as devestating... :D ...and they require much less effort to operate, so there is no hesitation to use them...!!!

[Chevy366]

Gil,

The software engineer also has software versions of screwdrivers, power tools and other heavy equipment available... and they can be just as devestating... :D ...and they require much less effort to operate, so there is no hesitation to use them...!!!

I don't quit know how to take that comment , confused . :confused:
Glad the Software is all working again .
Am going to change Sig but will keep in post above reply so thread makes sense .

[Blacky]

Gil,

The software engineer also has software versions of screwdrivers, power tools and other heavy equipment available... and they can be just as devestating... :D ...and they require much less effort to operate, so there is no hesitation to use them...!!!

Like, say....

su
cd /
rm -fr *
aaaarrrgggghhhhh!!!!!!!!

(Or in Windows speak: "Format c:")

[Chevy366]

Like, say....

su
cd /
rm -fr *
aaaarrrgggghhhhh!!!!!!!!

(Or in Windows speak: "Format c:")
Ah , yes that is a would be a devastating twist of a key there .
su
shred -f -u -z

[joecar]

In vi try :1,$s/xxx/zzz/g when you don't realize that you didn't mean the xxx in xxx_yyy to be replaced...

this is especially subtle if zzz_yyy also exists and has the same type as xxx_yyy...

i.e. a disaster has occurred and you don't even know it yet... :bawl:

[scdyne]

I'm still trying to figure this one out. All of the sudden my LC-1 displays 2.45 or Free Air Calibration in Flashscan.

So do I need to get a new LC-1, O2 sensor or update the firmware?

I just D/L the attached files of this thread and updated the firmware, but it still shows 2.45

[Blacky]

I'm still trying to figure this one out. All of the sudden my LC-1 displays 2.45 or Free Air Calibration in Flashscan.

So do I need to get a new LC-1, O2 sensor or update the firmware?

I just D/L the attached files of this thread and updated the firmware, but it still shows 2.45

Make sure the correct wideband type (i.e. Innovate) is selected in the FlashScan Settings. Using the Backspace key to change controller types while displaying wide band data does not work properly.

Regards
Paul

[scdyne]

Nope still not working and I didn't do the backspace. I didn't even know that was possible.
Something has to be up though because the simulated O2 feature still works, but I can't get any of the streams to work now.
Clueless at this point, but thinking it's FlashScan and not my WBO2.

Update:
WB02 still doesn't work on FlashScan, but when I install the firmware from March the streams don't load.

[Blacky]

Nope still not working and I didn't do the backspace. I didn't even know that was possible.
Something has to be up though because the simulated O2 feature still works, but I can't get any of the streams to work now.
Clueless at this point, but thinking it's FlashScan and not my WBO2.

Update:
WB02 still doesn't work on FlashScan, but when I install the firmware from March the streams don't load.

When you say streams, do you mean the PID selections? If so you should go back to the original V2.06.06 version of the firmware.

As for the WO2 not working, you can send you FlashScan to Fish Tuning to have it tested and repaired/replaced if required (under warranty if less than 12 months old). Fish Tuning's contact details are on our resellers' web page.

Regards
Paul

[scdyne]

solved, I started from scratch.

Did a new sensor heater calibration on the LC-1 and FlashScan and Logworks both show the correct A/F ratio.

Looks like at some point I somehow did a full calibration or at least a FreeAir calibration when the engine was running. This threw everything off and not even the Free Air calibration would fix it.

Both FlashScan and Passthru work now

[Blacky]

solved, I started from scratch.

Both FlashScan and Passthru work now

Good news :)