Nistune 0.10.21 Release

[Matt]

Version 0.10.21

Enhancements:
Check when ECU images loaded that part number matches address file
Add IAT to consult view window
Add trial period as push button instead of entry
Format key differently in registration window and auto upper case
Make ignition scalings same as fuel scalings (right click map)
Auto open previous windows on startup

Fixed bugs:
Add crash dump reporting on shutdown (requires VS2010 for latest version)
Crash on Type 1 frame handle errors (fix threading issues)
Update users manual for knock reporting, input tracing, speed trial, active test, AFR auto tuners
Resolve stream mode issues causing lockup from 0.10.20 consult changes
Resolve freeze up COM port when checking USB consult connection
Crash on remote login and under Ubuntu Linux (WINE) from 0.10.14 due to Menu Setup failure in Windows
Add single time error box when sync failed
Add VTC documentation
Cannot rotate graph on extended monitor
Consult RAW data formatting check
Modify S14A/S15 boost control maps for AT and MT models in split address files
Verify RD28 Zexel ECU sensor updates
Zexel ECU still selected when switching from Zexel.adr file
Expired software reports negative days
Moates: Hardware maptracing dropout when changing values
Z31 fuel temp reporting issue. Change register address
AFR auto tuner direct entry fixes for colour correction
Update Z32 300ZX Type 2 document for knock board capacitor positioning
S13 CA18DET timing table scales dont match graphs
Fix TTPmin/max tables and tracing based on hardware tracing verifcation
Define correct U13 KA24E speed limiter
74F05 Nissan ROM image does not have ignition parameter in consult table. Fix original table
ER34 register update list between tuner/stream mode causes connection dropout
Register flags not updating in tuner mode (ER34/WC34 and others). Stream read all registers
Log recorder clears fields after loading address file
J30 VG30DE TTPmax works but TTPmin incorrect. Found and fixed address
Emulator improvements. Queue change events and better recovery when Moates returns no trace data
Verify warm/cold start maps. Tested HCR32, BNR32 and J30 VG30E. Correct
Log player opens wrong file in record mode out of standby/hybernation
Log player has no parmaeters on auto connect on startup
Constant window auto applys change when window first opened
Debug logging crash on close
Read CUID from ECU ID process parse for varying ECU register translations (ER34 alhpa display%)

Other changes not in release notes:

Added adjustment to injector sizes for some SR20 address files
Y33 adjustments (same as ER34/WC34 injection time scaling)
Fan control address updates added (available in separate thread)
Added PLX IMFD support but not tested. Dont expect this to work. I need to work out how they interpret the data
Fixed Constants table list crash
Added RPM and TP safety limits to RB30 and VG30

Additional info

This version is my stable release candidate. I've identified crash issues and added a crash dump reporter (only works for XP however). Any immediate issues will be rectified and then put into a 0.11 release with 6 month release life on it

Next plans are to move the current Nistune project to Visual Studio 2010 so I can integrate a Win Vista/7 reporter and re-test using that environment before continuing on

Attached is the current bug list. Getting a lot smaller but I need some dump files from any logging crashes since they are not repeating here.

Post any issues in this thread. Please limit discussion to major bugs only. Any enhancements go in the enhancements section

[Matt]

Just put the release up then to replace the earlier version for a customer today (Build Date Aug 1 2012 20:12)

[Croustibat]

hi, i just tried that version on my work computer and entered trial/trial to test. It expired within maybe 10 minutes then closed. Using winXP 32 bits.

Is that normal behavior ? I was just looking at some maps found in the type1 section.

[Croustibat]

Also the vertical scrollers dont work on the left menu. If i grab one and go down, it does not move and the menu does not scroll. The screen is rotated though so maybe it is related ? To see the rest of the menu i have to enlarge the main window.

[Matt]

hi, i just tried that version on my work computer and entered trial/trial to test. It expired within maybe 10 minutes then closed. Using winXP 32 bits. Is that normal behavior ? I was just looking at some maps found in the type1 section.

You have an earlier build of that version. New one has a 'trial' button. Anyway 10 minutes is the maximum trial time allowed (5 x sessions). Otherwise people use it for remapping EPROMs etc without purchasing the software (also had to disable sync for that reason in the trial version)

Also the vertical scrollers dont work on the left menu. If i grab one and go down, it does not move and the menu does not scroll. The screen is rotated though so maybe it is related ? To see the rest of the menu i have to enlarge the main window

Try putting the screen back to normal. I've never tested the scrollers on a rotated screen. I'll put that as a bug report item to look into. Could be a problem with the grid controller used if it does it with all grids

[Croustibat]

the thing is that i did not do anything else than watching a tune. No changes at all, just browsing. Maybe you could allow "unlimited" browsing and no mods on the trial version, or start the timer when the first mod is done ?

I had no connexion to nistune hardware obviously and made no changes at all.

I do understand why you do that, but that is a little bit extreme. You may as well not do a trial version. I cant see anyone buying a tuning software that just automatically close after 10 mins browing a tune.

[Matt]

The trial is just that. A chance to look through the software, connect to the vehicle and see how it works. Also it doesnt close when its not a trial version

Anyway as initially posted this thread is to discuss any serious problems with the version before putting out the 0.11 release

[Torque]

Looks much better than expected

Just my quick five cents:

Installer is using my custom install directory, and does not insists on going to C:\PROGRAMS which is nice.

Installer is ignoring prior Nistune settings for ADDRESS/LOG/WB .. (old issue) so these go
to the personal USER FILES, not where Nistune had it configure (I am talking upgrade here)

Unable to deselect driver install

I like the Delta Compare colours (finally!)
Font sizing in the grid is not implemented yet?

Large icons are great to have.

VTC CUT/RECOVER is mixed up (wrong way round) on Z32 Type 2


All tested on W7/64 17'' laptop


Well done Matt, it looks good !

[Matt]

Installer is ignoring prior Nistune settings for ADDRESS/LOG/WB .. (old issue) so these go
to the personal USER FILES, not where Nistune had it configure (I am talking upgrade here)

Thats on the enhancement list still (customise install settings)

Unable to deselect driver install

Start of installation it wont let you deselect and components. At the end of the installation there are three tickboxes for starting installation and Nistune. You cant unselect those?!

Font sizing in the grid is not implemented yet?

On enhancement list still

VTC CUT/RECOVER is mixed up (wrong way round) on Z32 Type 2

Makes more sense VTC on until RPM cut @ 6100rpm and then recover as revs drop @ 5900rpm. This is the activity I see on the bench

[Matt]

Got a crash dump report for the log player closing now which is good. I can see where it is shutting down (after playback not closing properly)

[RomChip200]

I'm sorry but the file name generation behavior in the log player/recorder is not better in this version !
Log recorder has still old file name on the next reconnection

Do this test:
0) Connect with F4&F5 .... Initiate logging .... file is Nistune_2012-08-01_0803_10.csv
1) 10 minutes of logging ....
2) Stop recording with 2 presses on STOP button, do you want to save .... yes .....
3) file name displayed becomes Nistune_2012-08-01_0813_43.csv
4) disconnect with F4 & F5
5) put PC in standby mode (so Nistune and log player/recorder are kept open)
6) 14 hours later, resume PC ....
7) reconnect with F4 & F5
8 ) press record button, the file name used is still Nistune_2012-08-01_0813_43.csv !!! whereas clock time is 6:00pm .....

Note: the only way for me to force the generation of a new file name in steps 7&8 is to click on Name button and Cancel, then file name displayed reflects the current date and time. But this is PAINFUL b/c it creates systematically a file on the disk with 0 bytes size (Nistune_2012-08-01_0813_43.csv file btw)

[RomChip200]

VTC CUT/RECOVER is mixed up (wrong way round) on Z32 Type 2

Makes more sense VTC on until RPM cut @ 6100rpm and then recover as revs drop @ 5900rpm. This is the activity I see on the bench

I agree with Torque, you called them the wrong way for Z32
0x79F5 is really a recover point (what you call Enable, I would call it "potential re-enable")
0x79F6 is a trigger point to switch off VTC (if previously on off course) (what you call Cut)

ECU tests the off point before the recovery point:

Code: Select all

CODE:D1B2 cv_6:                                             ; ...
CODE:D1B2                     cmpb vtc_release              ; 118*50=5900 rpm   CUT POINT
CODE:D1B5                     bcc cv_10
CODE:D1B5
CODE:D1B7                     cmpa vtc_const4              ; POTENTIAL RECOVER
CODE:D1BA                     bcs cv_7

Be very careful, A & B accumulators are used here, and some hysteresis is introduced on A and B just above (refer to full routine below)

The fact NISSAN put 6100 in 79F5 (i.e. 6100 > 5900 cut point) is just a trick from them, the way the code is written just means, once VTC is switched off, it will never be on again in high RPM range. Recover point would be useful on Z32 if 79F5 was lower than 79F6 (e.g. if 79F5= 4000rpm, ECU switches off VTC @5900 and would put it on again if RPM drops below 4000)

Remember my post:
viewtopic.php?f=5&t=2326&p=16989&hilit=VTC#p16987

Code: Select all

Moved code into other post to shorted this thread...

Matt, take it easy, I got the feeling you got sometimes wrong assumptions when testing on the bench, maybe this is the way you handle the code, you just have disassembly listing, right ? YOu don't use IDAPro, do you ? Maybe you should because IDAPro allows to follow the program flow through the branches very easily and so, following more easily when ECU checks many conditions to enable or disable something.

[Torque]

Thanks RomChip,

I was meaning to get back to Matt regarding this matter.
(I have my own address file anyhow)

It's good to know that the logs I took are valid and VTC is triggered the way I
understood it.

[Matt]

I got the feeling you got sometimes wrong assumptions when testing on the bench, maybe this is the way you handle the code, you just have disassembly listing, right ? YOu don't use IDAPro, do you ?

I still use DASMX since its quick to generate a listing and easy to use. However it is buggy and I have written PERL scripts to fix some of the output

I should get my head around using IDA Pro since it is a brilliant disassembler, however I tend to move from ECU to ECU very often so I dont spend long on one particular code base to construct a proper project for any particular unit. Constructing a symbol table for DASMX is quite a bit quicker but doesnt have support for showing individual bit fields or the graphical jump arrows which is nice.

In addition to this the 7790 processors are not supported so I have to use my own disassembler for this (swapping between 8/16 bit instructions is very annoying)

With the original VTC points I derived for the ECU from memory the cut/recover was based from following the program code. However looking at it again on the bench the other day it just didnt seem right having it in this order, when testing on the bench and most other parameters in the ECU have cut higher than recover.

This then got me thinking that I have got the current ones backwards. I'll put them back to how they were before

[Matt]

Re Logging: I'll have another crack at the file naming. I was using index mode to test, so wasnt really looking at the actual time stamp

However the logging is better than before. In the last version it was using the previous file names and writing over them with the new log after standby. I'll look into the 0kb file also since I did notice some of these during my testing but wasnt sure on the steps to repeat them