Multiple problems with version 3.4 Pleae help

[10_SS]

The actual knock retard one was the one I removed, but never could confirm whether it was something up with the graph itself or the parameter... The parameter did cause issues when I substituted it into the knock sensor graph, but worked fine on it's own?

well good you removed that one, because i removed the other one and it locked up again at 32m 15s. I'll run 3 again just to be sure, then I'll remove the other knock retard you had.

It runs noticably slower on this tablet (graph updates slow down when RPM and things change fast.. faster the change the slower the updates but doesnt seem like that is affecting the data freeze yet.

My specs on this tablet are worse than yours, I have 2GB RAM you have 4GB, I have 1.33ghz 4 core you have 2ghz 2 core? I'm getting 30mins out of it. Not sure if the computer specs even has anything to do with it yet but if something is eating up memory then yes it might.

After all this I'll add all my graphs to all your graphs and see what happens.

UPDATED POST #98 with freeze updates

[wasyl]

RAM, we need more RAM! - said russian scientist

best regards

[10_SS]

RAM, we need more RAM! - said russian scientist

best regards

could be, I have only 2GB on this Winbook.

I ran another repeat test, starting scanner and ECU from full off, then stop scanner once, then restart. Froze after 13min. UPdates in post 98.

Update: the 4th test froze exactly same time as 3rd test.. 13min 2 sec. One more freeze then I will delete the RETARD chart.. and see how it runs

[GHuggins]

Sounds like your on to something there - Mine always seemed to freeze right at 13 minutes...

[10_SS]

Sounds like your on to something there - Mine always seemed to freeze right at 13 minutes...

5th freeze: 34m 46 sec (full power off ECU, closed and opened scanner after 4th freeze)

I cant tell... not as consistent as I would like... removing KNOCK RETARD from the Charts next..


.. all info is in post #98

update on post 98:
6th freeze: 33min 22 sec - Removed the last RETARD Chart which uses KNOCK RETARD restarted scanner and ECU...

Now wondering if it's just this crappy Winbook.. restoring my standard layout that I dont have problems with on my Yoga Core I5..

[rolls]

Like I said before, that's the only way we'll be able to isolate this. Until you can tell us that it works with X, but not when I do Y, where Y is some very simple action or change, we won't know what the problem is... because the problem is Y. We can't just guess at what Y is. The software is too complex for that.

UI logging is pretty easy to hook up in Winforms. Could add something like that to the software in less than a day to help solve problems like this.

[Keith@HPTuners]

rolls, VCM Scanner is multi threaded, so, it's a bit more complicated.

10_SS, we don't have to have it crash exactly at X time.. just be repeatable to crash within 1 or 2 hours. We can set it up on a bench and let it go.

We need the crash to be repeatable on a bench though, with no changing inputs. That is a lot easier for us to duplicate than specific vehicle conditions.

[10_SS]

ok well we have you covered I think.. check post #98 for the running log of what freezes up.
With GHuggins problem LAYOUT, My bench setup is crashing regularly, it doesn't fully crash, just new data stops coming in and the log timer stops... the chart keeps rolling, just nothing is in it.. you can see a little white blip move across the bottom of the chart and the stop button and other features still work. All data digits freeze in the chart and they scroll off the screen then gone forever until you stop logging then they reappear and become the end of the log and it looks like you ended it normally. Have you tried GHuggins "problem" layout on a 2010 Camaro or 2012 ZL1 E38?

I am in the process of deleting his GRAPHS one by one until it stops locking up... he has less GRAPHS than me, I have maybe 3x, but some of the graphs he has refers to MATHS that have WAY more formula in them than mine. I feel this will work since I just put MY normal LS3 layout back in with no other changes and still logging after 1hr 10mins... big difference!

Question if you are still following.. lets say I include all of his MATHS.. if none of the GRAPHS reference his MATHS, do his MATHS still get processed like they do if a graph is using one? (by removing a problem graph is it also removing the associated MATH from being processed?)

My bench simulator has pots for MAP, coolant, RPM Crank and cam, IAT, pretty much everything except MAF and O2 sensor feedback (I have MAF capability, just dont have it connected right now because it doesnt seem to need it).
Pretty much I run it around 1800rpm, everything else normal, and constant and it will lock up. Once or twice I'll rev it up or increase MAP... but that doesn't seem linked to anything yet, though on this Winbook (slower Win10 tablet) while changing RPM/MAP you will see the chart updates slow down, the faster you change things the slower the updates are. On a fast AMD 6 core 8GB ram SSD desktop computer that is silky smooth (faster computer still locked up once or twice though).
I still have yet to try all this on my Intel based Yoga I5 SSD 8GB ram laptop... this one used to lock up on 2.24 though.. forget which channel was causing it but I removed it and it was fine every since.

[rolls]

rolls, VCM Scanner is multi threaded, so, it's a bit more complicated..

Of course. I'm suggesting you could set it up to log all mouse and keyboard presses. Eg every object with a clickhandler logs to a collection.

So you could then recover a list of exactly what buttons/menus were pressed at what time. Doesn't matter if its multithreaded as all the events are time stamped and logged to a threadsafe collection.

This way you could replicate the exact process the user followed.

Here is one that I used in a project before. It may look a bit "hack" like, however the simplicity means you can get it up and running in less than an hour. You could then send this with the RedGate crash report.

http://stackoverflow.com/questions/3...ption-handlers

https://github.com/MeaningOfLights/UserActionLog

Surprisingly helpful for replicating what a user did. Especially when it turns out it is the exact order they followed to create the crash. Eg add variable x, delete variable x, add variably y crashes but adding variably y only does not. You don't have to rely on user reports then, as you have exactly what they did.

If you couple it with something like Sentry.IO then you don't even need to wait for them to complain about the issue. It is already in your inbox with the stack trace, which user, the callstack, exact list of steps required to reproduce the error, their debug.dat, appdata directory.

MASSIVE time saver setting something like this up, pays for itself in a matter of months.

Sentry.io can cache reports offline and send them when the user next has internet as well.

[Keith@HPTuners]

10_ss, can you submit a support ticket with a vcm suite info log connected to the vehicle, your vcm scanner.cfg, channel config, layout config, and steps to reproduce the problem.

rolls, we have other things higher on the priority list right now. Last time we had a crash bug like this was about a year ago.. Time before that, I can't recall.

[10_SS]

10_ss, can you submit a support ticket with a vcm suite info log connected to the vehicle, your vcm scanner.cfg, channel config, layout config, and steps to reproduce the problem.

rolls, we have other things higher on the priority list right now. Last time we had a crash bug like this was about a year ago.. Time before that, I can't recall.

ok, but MY LS3 camaro layout just ran fine for 1hr 30min and I stopped it.. this is after having 99% repeatable problems with GHUGGINS ZL1 layout.. so do you want BOTH sets of files (vcm scanner.cfg, layout config)? channels are the same.

Last question, the vcm scanner.cfg is that the one in HP Tuners\Settings\VCM Scanner.cfg (the one that contains the MATHS), or the one in HP Tuners\VCM Scanner\Configs folder.. (confusing since there are two files with .cfg extension and I havent done this in awhile)

EDIT: [#68506]: HPTuners.com Other Contact Request - 20249. I sent everything, well labeled in a zip file let me know if you dont see the zip.

[rolls]

rolls, we have other things higher on the priority list right now. Last time we had a crash bug like this was about a year ago.. Time before that, I can't recall.

Fair call.

[Keith@HPTuners]

10_ss, submit both sets of files, but make sure they're labeled separately so it's very clear.. you could prefix each file with 1_ and 2_ so we know what group they are from.

We need the VCM Scanner.cfg that has the math parameters, it's in hp tuners\settings.

The more simple you make the ticket for us to read/process, the quicker we can get this bug fixed for you guys.

[10_SS]

10_ss, submit both sets of files, but make sure they're labeled separately so it's very clear.. you could prefix each file with 1_ and 2_ so we know what group they are from.

We need the VCM Scanner.cfg that has the math parameters, it's in hp tuners\settings.

The more simple you make the ticket for us to read/process, the quicker we can get this bug fixed for you guys.

yes, done, thanks [#68506]: HPTuners.com Other Contact Request - 20249. I sent everything, well labeled in a zip file let me know if you dont see the zip.

[10_SS]

yes, done, thanks [#68506]: HPTuners.com Other Contact Request - 20249. I sent everything, well labeled in a zip file let me know if you dont see the zip.

7th Freeze: 36min 1sec - GOOD NEWS for debugging I guess is this lockup occured on a totally different higher end laptop.. Yoga 11s Intel Core I5 8GB RAM, SSD, using the same GHUGGINS LOCKUP layout.

Since I've tested both a low and higher end laptop I'll probably lay off the continuous posts, I'm convinced it's just the scanner or Layout settings, but if I find the Graph that stops it from freezing I'll definitely post up the results..

[10_SS]

Well now I just ran into something different..
After the last lockup on the higher end yoga laptop I unplugged the USB cable and put into my desktop 6 core AMD 8GB RAM, SSD, and it started the scan the inital but instantly didnt have any data... RPM, etc showed 0rpm, but the same appearance was there with the white blips moving across the graph at the bottom..
I restarted the software 4 times, then unplugged the USB cable twice and opened/closed scanner two more times, same thing.
Only after cutting power to the ECU, after that it started scanning normally... almost like the ECU OBD communication was frozen... however, the MPVI Pro box was powered from the ECU every time I unplugged the USB cable.. I never reset it until restting the ECU... so.. hope that helps

The only thing that didnt seem right was the little box that pops up when you first hit "Start scanning"... the box that says "detecting vehicles, protocol, etc" it was very short like it wasnt getting any feedback... after the ECU/MPVI power reset, everything back to normal.

EDIT: There was just more weird stuff. when I said it was normal, everything seemed normal except the Graphs startup default "Main Spark" weren't seeing a change in Cyl AIrmass (in my case, MAP), but the MAP channels and Charts data was updating just fine.. so my RPM was updating but it was all staying in the 0.08 row for cyl airmass.. a quick reboot of the ECU and scanner fixed that. I guess I'll ignore that for now..

Edit again: Desktop with 6 core AMD 8GB Ram SSD locked up after 1hr 59min, quite a bit longer than the Winbook tablet

[GHuggins]

^^^^^ This is what mines been doing when it loses the wideband inputs or inputs to particular graphs - shows them being present in the channels as far as not blanking out the wording, but show absolutely no data... Maybe my math equations are too complex, which considering others are using the same equations I find hard to believe? My math equations do reference a lot of quickly and finely changing parameters - perhaps that's bogging it enough to kill it?

[10_SS]

shows them being present in the channels as far as not blanking out the wording, but show absolutely no data...

yes I think that is consistent with the original lockup freeze when the plots (charts) also stop showing lines.. however this Graphs startup default "Main Spark" thing was weird.. it was the only thing not working, and it was just the cyl airmass axis always showing 0, RPM was moving normally.

[GHuggins]

Keith, 10_SS pointed something out to me that I thought you should know - he said the scanner was freezing for him initially without any of my maths being present or used... I've currently made an entirely new layout for a jeep I'm working on and it's doing the exact same thing as far as things showing no data in the cells on startup - it's using about 1/10 the chart info and only has 3 graphs for lt+st, knock retard sensor and spark advance - haven't run it long enough to check for freeze up yet and won't until next week as I've got to go meet up with a customer right now, but anyway - thought that tid bit of info might possibly help?

[10_SS]

Im staring to see a trend... using the same Ghuggins problem layout, the better/higher spec'd computer, the longer the scanner will run. For example, 6 runs with a low grade tablet tends to last 13-30mins... that same setup on mid grade laptop lasts 36+mins.. on a desktop higher grade PC it lasts 1hr 20mins to 2hrs.

Now, if I take that 12min-30min low grade tablet and delete BOTH knock Graphs, now it will run ~2hrs then freeze.. so it still freezes.. but much farther out..

Im thinking maybe everyone's scanner will freeze if they run it long enough.. just depends what specs you have, and how many things (just Graphs tested at this point) you have going on. Lots of people usually run under 30mins... so.. they also likely wont see this problem.

All data is in the original #98 post.

Other bugs... TWO times now, I started scanning and the Graphs Y axis will not update... stays at the lowest reading.. stopping and restarting log doesnt fix it.. closing scanner and restarting does. (this is what Ghuggins is talking about in #119 above I believe).