PDA

View Full Version : Need a couple of BETA testers...


JayPSC
07-13-2005, 03:39 PM
...That are tuning MS with WBO2, I've basically compiled Matt's/the poi's WB tuning spreadsheet into a VB program so I need a few people to beta test it for me.
I need people who are actively tuning their car this week, ideally using Innovate 0-5v output although I can write some code to accommodate other WBs (which I'll have to do eventually).
It's now very fast & can crunch a 41k line log in less than a minute.

http://www.jaypsc.co.uk/vexme/ve1.jpg

http://www.jaypsc.co.uk/vexme/ve2.jpg

Send me a PM or reply here if you wanna beta test it.

Jay

kyle242gt
07-13-2005, 04:15 PM
system requirements? (IE Excel, WinXP, etc; I have a bare-bones 400Mhz Celeron Win2K machine)

Tech Edge WB 0-5 volt support, and I'm down :rockon: The TE volt/afr is different AFAIK.

JayPSC
07-13-2005, 04:24 PM
Requires Excel as the AFR targets are stored in there, should work fine on win2k although I have no way of testing on anything other tha XPpro or Server 2003.

400mhz should be enough although it will obviously take a bit longer, running on a phasechange cooled A64 here :-D hehe.

Get me the TE volt/AFR equation & I'll add it asap - as long as it isn't some stupid lookup table, then it may take a bit longer lol.

edit: Just got it running on my laptop on powersaving mode at 1.2ghz & it isn't all that bad, still only a couple minutes to crunch a 4mb log :)

the poi
07-13-2005, 08:33 PM
Hey jay. I'm leaving to go for a tuning run in volvorod85's car right now. We'll see how magical this bastard is :-D And awesome frickin work!

Two 87s
07-13-2005, 09:09 PM
Brilliant.

I am using an Innovate, and I have it set to output the AFR as a voltage. In other words, it outputs 14.7 AFR as 1.47. I'm not sure if that's similar to any other setup, as I had to tell the Innovate to do that. It made it easier to convert to AFRs for the Poi's spreadsheet.

If your program will work with those outputs, e-mail it to me at aaronreedbaker at gmail dooooot com.

(Even if it doesn't, e-mail it to me. But tell me what I ought to make the Innovate output look like, as I can change that to anything I want, I think.)

I will definitely do some BETA testing. Post up a walk-through of how to use your program, if you would. From the screenshots, it looks like you input an ideal AFR table, load a VEX file of your current map, load a log and it outputs a VEX of your desired map. Am I getting anything wrong? It basically does what you would want to do with Poi's spreadsheet by hand, only it automates it, yes?

Thanks for the hard work!

Aaron

ovlov760
07-13-2005, 10:29 PM
sent a PM

linuxman51
07-13-2005, 11:53 PM
i'd love to but i dont have (gasp) excel (or office for that matter....)

the poi
07-14-2005, 03:30 AM
damnit man! i tried it tonight, but i dont knwo what format the datalog should be in!

JayPSC
07-14-2005, 04:39 AM
I've been using normal full logs from MSnS-extra, is there any different formats I need to be aware of? The rest of you will all have PM's

Jay

JayPSC
07-14-2005, 05:18 AM
Right now basically you want to get into Megatune & open the VE table you're using & output the vex file, then open my prog & click the 'Open VEX' button, this will load the VEX file into the top table & also the AFR targets into the bottom table. If you need to change these they are kept in the 'AFRTargets.xls' spreadsheet in the program directory.

Now you can select the Minimum hit counts by dragging the sliderbar below the tables, currently goes from 1-50, a tooltip should appear while you are dragging it to tell you what number you're on. Also select which WB you are using by clicking the radio buttons in the top right. Innovate 0-5v is 10-20 AFR, Techedge 0-5v is 9-19 AFR, both linear outputs ie AFR=Vout x2+10 or AFR=Vout x2+9 in the case of the TE unit, I'm just assuming this is correct for the TE as I've never used it.

Then simply click the 'RUN' button & browse to the datalog you want to import, a sample datalog is now included in the .zip file. I've been using the standard extra datalog so it's just jumping to the cells that i specify & ripping the data out, if needs be I can make it more robust.

When this has finished processing it'll write a new VE table in the bottom box with all the changed bins highlighted in pink. You can now click the 'Save New VE' button to export it as a .vex file.

This is basically exactly the same as Matt's/the poi's spreadsheet but the calcs are slightly different as there was overlap between bins which I have tried to keep to a minimum to make it more accurate. If anyone has any problems/errors/questions/wishlist then just Pm me about it.

Chris can you send me the top 10lines of the datalog you were trying to use, jaypsc at pornstarr.co.uk ;-)

benflynn
07-14-2005, 07:52 AM
i'd love to but i dont have (gasp) excel (or office for that matter....)
are you still in school at uab we can get most of the microsoft stuff for under 10$

swedefiend
07-14-2005, 07:55 AM
Are you still in school? At UAB we can get most of the Microsoft stuff for under 10$

Punctuation my young Jedi.

swedefiend
07-14-2005, 07:56 AM
i'd love to but i dont have (gasp) excel (or office for that matter....)

No Open Office there Linuxman???

WTF?

If you don't have it, holla at ya boy.

Matt Dupuis
07-14-2005, 08:56 AM
Dammit, I want to be part of the testing procedure, but I won't be putting that many miles on my car in the next little while, and it's still in break-in mode so I'm not doing the full hardcore tuning just yet.

However, THANK YOU for doing this! I was using my spreadsheet last week and not only is it intensely clunky and slow, but it's so intensely clunky and slow that I can't use it on my laptop! I'm sure you've streamlined it to refine the hundreds of kb needed just to store the calculations!

I can't wait for the opportunity to try it in the next couple of weeks.

By the way - Techedge NBlin AFR = 9+(V*2), 0-5 volts = 9-19:1 AFR

JayPSC
07-14-2005, 09:18 AM
Dammit, I want to be part of the testing procedure, but I won't be putting that many miles on my car in the next little while, and it's still in break-in mode so I'm not doing the full hardcore tuning just yet.

However, THANK YOU for doing this! I was using my spreadsheet last week and not only is it intensely clunky and slow, but it's so intensely clunky and slow that I can't use it on my laptop! I'm sure you've streamlined it to refine the hundreds of kb needed just to store the calculations!

I can't wait for the opportunity to try it in the next couple of weeks.

By the way - Techedge NBlin AFR = 9+(V*2), 0-5 volts = 9-19:1 AFR

Cheers Matt, thought that was the calc. I'll send you the file to have a look at being as it was mostly yours/Chris' calcs that got this working. Orginally I had it accessing your spreadsheet but as you used more & more lines it just got slower & slower, now it just runs the logic on each line of log. It won't give you exactly the same results as your spreadsheet (even using exactly the same code in both I can't get them to tally up lol) but I've checked everything manually & it should work 100%.
There's no reason why there should be any issues with the calcs as you guys have been using them for a while now, just wanna make sure everything else is fine, I've loaded in & exported both 8x8 & 12x12 tables & all seems fine.
Can't actually go out & do any tuning myself this week as I got so stupidly sunburnt on tuesday that I'm complete lobster hahah.

Matt Dupuis
07-14-2005, 09:44 AM
Things look good. Don't have a VEX at work so I'll play with it some more at home.

Question: How far around the column target/row center does the program grab samples from the datalog? i.e., if my columns are at 300, 900, 1500, etc., does it grab info from 200-400, 800-1000, 1400-1600, or is it only looking for 300, 900, 1500? Are there any provisions for adjustment in this area?

Question: Does the datalog have to be named anything special? Does the datalog, VEX, and AFR have to be in the same directory as the EXE file? Do they need to be in any special directory? Can you grab info from multiple datalogs, or must one combine all datalogs to be sampled from into one uberlog?

JayPSC
07-14-2005, 10:42 AM
It'll grab in equal spaces ie if your rows are at 500, 1000, 1500 it'll grab data from 0-750, 750-1250, 1250-1750 if you want the facility to narrow it down more then I can, but being as MS interpolates between bins I dunno if it's neccesary.

AFRtargets file has to be in same dir, everything else can be called whatever & be located wherever you want as there is an open file dialog to browse to them. One datalog at a time at the moment, Iíll look into opening multiple logs soon.

www.jaypsc.co.uk/pics/12x12.vex
www.jaypsc.co.uk/pics/8x8.vex

Couple of vex files if you wanna have a play at work, should be a 100line log included in the .zip

Jay

Matt Dupuis
07-14-2005, 10:53 AM
Holy hell, that is slick~!

kyle242gt
07-14-2005, 01:18 PM
:please: I didn't get a PM </whine mode>

I'll install Excel if I hafta :)

JayPSC
07-14-2005, 01:45 PM
:please: I didn't get a PM </whine mode>

I'll install Excel if I hafta :)

Lol sorry, looks like I forgot you. Working on an excel-free version at the mo, but may take me a while to code. You have a PM in the meantime to test this version. I've also just added a tooltip that appears over the new VE table to show you how many hits you got in that bin :cool:

kyle242gt
07-14-2005, 01:50 PM
thanks, got the PM. That latest tooltip is a great idea. :cheers:

ovlov760
07-14-2005, 02:15 PM
Very nice so far. It runs great on my win2k machine but my old laptop(p133,win98,80mb ram,4 gig hd) complains of a "automation error" after loading a .vex file.

JayPSC
07-14-2005, 02:51 PM
Ahh I have no way of testing on a win98 machine so we'll just have to say it's win2k upwards only. Just sorted the poi's log problems, you can now use custom logs as long as you keep the column headings the same ie: 'RPM', 'MAP' & 'O2' this should help you keep log size down to a minimum.

Jay

linuxman51
07-14-2005, 02:51 PM
took care of some routine maintenence (oil leaks... damnit) this morning, hooked the wideband up to megasquirt, and am now off to play with this tool, field report coming shortly

linuxman51
07-14-2005, 03:09 PM
"unexpected error with vex file: 462: the remote server machine does not exist or is unreachable"


....? I'm going to download your 12x12 vex and look in it and compare, will post any differences in a few

benflynn
07-14-2005, 03:14 PM
Punctuation my young Jedi.
:booty:

linuxman51
07-14-2005, 03:14 PM
bombs out with your vex file as well (office 2k2 if that makes a difference), any services that should be runng that one might have turned off?

JayPSC
07-14-2005, 03:17 PM
bombs out with your vex file as well (office 2k2 if that makes a difference), any services that should be runng that one might have turned off?

Hm that is weird, office shouldn't make any diff as the vex is just a txt file. Can you send me your vex please jaypsc@pornstarr.co.uk, what os are you using???

linuxman51
07-14-2005, 03:19 PM
Ok it appears that with com+ support turned off it won't work (fyi for anyone else)

kyle242gt
07-14-2005, 04:16 PM
OH NOES! It won't fit on my 800x600 laptop screen... that could prove interesting.

Moving on, opening my vex, I get Unexpected Error with .vex file
429: ActiveX component can't create object
I just installed openoffice calc 1.9.113

Trying it on my desktop with the shipped files, it loads the afrtargets.xls on load of VEX. It doesn't do that on my laptop.

MSnS-E doesn't use a 12x12 AFR target, it uses 8x8. Does VE interpolate from an 8x8 table? Similarly, I have different RPM bins (kinda/sorta every other one) and MAP bins in the the target table. Can VE straighten that out, or should I create a new AFRtargets.xls file with 12x12 at the same RPM/MAP bins?

Off to try MSExcel... :x:

//edit - yahoo, that did it! for some reason, OO didn't cut it; probably missing some critical piece of bloatware MS includes.

Runs well, and is totally bad ass! Huge props!

JayPSC
07-14-2005, 04:38 PM
OH NOES! It won't fit on my 800x600 laptop screen... that could prove interesting.

Moving on, opening my vex, I get Unexpected Error with .vex file
429: ActiveX component can't create object:

Do you have Excel on ya laptop??

MSnS-E doesn't use a 12x12 AFR target, it uses 8x8. Does VE interpolate from an 8x8 table?
It uses VE target values from the AFRtargets.xls file, similar to the spreadsheet, you input what AFR you want it to correct to for each bin. Nothing to do with AFR targets in extra.

Similarly, I have different RPM bins (kinda/sorta every other one) and MAP bins in the the target table. Can VE straighten that out, or should I create a new AFRtargets.xls file with 12x12 at the same RPM/MAP bins?
You've lost me here lol

kyle242gt
07-14-2005, 04:52 PM
You've lost me here lol

See my edited post above... with excel loaded, it's working :party:

As it stands now, my AFRtargets.xls was manually (mentally) modified to use the same 12 rpm and 12 map bins. In MT, there's only 8 rpm and 8 map, and they're different than the main VE table.

I'm wondering what would happen if I simply created a new 8x8 AFRtargets.xls using the 8x8 map output from MT in vex format. Will VE interpolate AFR targets even though the map is a different size and has different bins?

This thing is soooo killer. One of my beefs with MST3K is you can't really tell before and after what it's suggesting. The two-up view of the tables is excellent. :party:x10!

JayPSC
07-14-2005, 05:01 PM
See my edited post above... with excel loaded, it's working :party:

As it stands now, my AFRtargets.xls was manually (mentally) modified to use the same 12 rpm and 12 map bins. In MT, there's only 8 rpm and 8 map, and they're different than the main VE table.

I'm wondering what would happen if I simply created a new 8x8 AFRtargets.xls using the 8x8 map output from MT in vex format. Will VE interpolate AFR targets even though the map is a different size and has different bins?

This thing is soooo killer. One of my beefs with MST3K is you can't really tell before and after what it's suggesting. The two-up view of the tables is excellent. :party:x10!


Ahh gotya, if you just use 8x8 table in the xls it wont interpolate, best of doing it manually otherwise it'll just have blank AFR bins

linuxman51
07-14-2005, 05:14 PM
It seems to be a bit slow to react.. should one have ego correction turned on or off? (I've got accel enrichment off, decel fuel cut and over run turned off to "pureify" the logs)

Two 87s
07-14-2005, 05:22 PM
I did a little datalog and exported my VE table. The program seems to work fine for me.

Only problem is that I don't know that I have my Innovate outputing the right voltages right now. If you have the Innovate settings the same as the factory default, there are two possible types of output, on output 1 and output 2. I cannot remember, for the life of me, which output wire I'm hooked into. (It's two different wires on the same stereo out connector...)

So my question is, for the Innovate 0-5v setting, what should your raw datalog look like? It's set up so that V*2+10 = AFR? Is that the standard setting for one of the two Innovate outputs?

Thanks,
Aaron

Edit: For clarification, Innovate says:

"As shipped, one output simulates a typical narrow band oxygen sensor, the other one is programmed to show AFR/10 for gasoline on a digital voltmeter (for example AFR of 14.7 outputs 1.47V)."

I'm tapped into the second output, I think. Which means that my O2 table formula should be AFR = V*10. Is there a way to add an extra option for that?

kyle242gt
07-14-2005, 05:32 PM
Suggestion: De-pink (unhighlight) the changed cells on a re-run; I can never remember to change to TE before running... :oops: but it'd also be nice when increasing the hitcount bar and re-running.

Speaking of which, I'm not seeing the tooltip for number of hits in the VE bins, only on the slider.

//edit - this is a piddly whiny bitchy naggy request, but here goes :-D could the progress bar be moved to the right hand margin, between the radio buttons for WB type and the open/run/export buttons? And could said buttons be moved up a smidge? I can't see the progress bar or the export button at 800x600, which is all I can run.

//edit 2 - I'm seeing some cells highlighted (highlitten? highlander?) that aren't changing; does that just mean that there were enough hits to generate that VE?

//edit 3 - Kenny raises a good point. Does this thing work off Gve or some permutation of Gve*Gwarm*Gego*Gammae*...ad nauseum? If that's the case, as I suspect it is, then all Gamma adjustments (accel/decel/Target AFR/etc) should be off. With those on, the O2 voltage will not be a function of Gve...

JayPSC
07-14-2005, 06:14 PM
It seems to be a bit slow to react.. should one have ego correction turned on or off? (I've got accel enrichment off, decel fuel cut and over run turned off to "pureify" the logs)

What do you mean by slow to react? It's best to run with all enrichments off so you have nothing else affecting your AFRs

I'm tapped into the second output, I think. Which means that my O2 table formula should be AFR = V*10. Is there a way to add an extra option for that?

Ahh so that's the Innovate 0-2v setting, I'll add that tomorrow

Suggestion: De-pink (unhighlight) the changed cells on a re-run; I can never remember to change to TE before running... :oops: but it'd also be nice when increasing the hitcount bar and re-running.

Will add it to my list for tomorrow

Speaking of which, I'm not seeing the tooltip for number of hits in the VE bins, only on the slider.
Not in your version yet, will have it in next release

//edit - this is a piddly whiny bitchy naggy request, but here goes :-D could the progress bar be moved to the right hand margin, between the radio buttons for WB type and the open/run/export buttons? And could said buttons be moved up a smidge? I can't see the progress bar or the export button at 800x600, which is all I can run.
I'll sort something for you 800x600 users (using 2560x1024 here lol)

//edit 2 - I'm seeing some cells highlighted (highlitten? highlander?) that aren't changing; does that just mean that there were enough hits to generate that VE?
They're probably pretty much spot on then & changing so little that it rounds back to the same number, I'll put some logic in to avoid this tomorrow

This is great guys keep em coming ;)

coldfusion21
07-14-2005, 11:33 PM
hit me up, im not going to be using it right away but would like to keep it around just in case...

linuxman51
07-14-2005, 11:38 PM
What do you mean by slow to react? It's best to run with all enrichments off so you have nothing else affecting your AFRs



Ahh so that's the Innovate 0-2v setting, I'll add that tomorrow



Will add it to my list for tomorrow


Not in your version yet, will have it in next release


I'll sort something for you 800x600 users (using 2560x1024 here lol)


They're probably pretty much spot on then & changing so little that it rounds back to the same number, I'll put some logic in to avoid this tomorrow

This is great guys keep em coming ;)


Well on the plus side of things (no complaints but I do have several observations, pending a few questions...) it makes things a lot more consistent (my map now looks like total peaky ****, but transitions are fairly smooth), and the lower rpm medium load (call it 60-70kpa) is just about spot on, as are the idle afr's.

however the low load (which i set up to be the same afr as medium load) above say 2200 rpms and medium load upper rpm's is about a point off and gets worse higher in the rpm band (low load high rpms, while generally thought theoretical but often encountered trying to entice people to race) are just way off, but it seems to be "learning".

I sat in the drive way and datalogged me reving the car up keeping the vaccum gauge a little below and above 15 mmhg (~40's kpa) from idle to rev limit, the first run had 23:1 afrs, the second run had 20:1, and the third run was closer to 18:1, but it never got any better (for reference, low load i'm shooting for 16-16.5:1)

It also has a nasty rich spike on tip in at low rpms (no tps and mapdot turned off, i've got no accel enrichment), and it tends to skew towards richer below 100kpa than leaner (by richer i'm seeing in the 12's, and have 13.8 across the board at 100, and 13.4 at 120. at wot it goes to ~95kpa).

I've had a few problems aquiring larger datalogs (it seems to do a lot better the bigger the file), those have largly been my fault, and at one point it clobbered a datalog with a vex file with .xls extension hehe.

All in all its good stuff, i've refrained from opening the fuel map and "fine tuning" anything myself while i've played around with the software and to get a more thourogh(sp? alky) review for ya since you do want feedback.

One thing of note... on the techedge, its reading close to 5 volts before it gets to 20:1 (it was like 4.8-4.9 in MT at 18.5ish, not sure if thats normal or not) and 2.5 volts isnt 15:1, its more like 14.5:1, dunno if anyone else has noticed that (here comes the confession... I've never hooked my wideband up to the megasquirt, i've been more of a "on the fly" tuner, but would rather abandon that in lieu of something safer, easier, and quicker, msTweak just never quite worked for me tho :( ). Before the obvious chime in, yes, its setup in the settings for wideband, and in the ini file its setup for the TE wideband.

linuxman51
07-14-2005, 11:40 PM
Or maybe not. lemme charge the laptop and go back and datalog some more with it setup right... I grabbed the innovate line.


<--- noob.

ovlov760
07-14-2005, 11:53 PM
So far i'm pretty pleased with your little prog. I've done about 2-3 datalogging runs with it. I've been comparing my logs to my map and the map it generates as I go. IMO it seems to be pretty accurate and in most cases dead on. My wish is for the ability to load multiple logs and a 800x600 mode for the old laptop. I'm having way more success with this than mstweak...

linuxman51
07-15-2005, 02:02 AM
Ok back from another road trip with the right voltages this time, datalogged a trip of about 8 miles, varrying speeds (hitting all the nasty spots, wot, full decel, etc). chewed it, vex'd it, rode out, markedly better. went another ~5 miles just hitting some part throttle stuff that didnt wanna go right the first time, better still.

Datalogged the long trip home, hit all the rough spots again (let the car hang out there for a little bit to max. the effort) couple wot pulls in lower gears (and one or two in higher gears, dont tell the cops), this took a little longer to process, so i checked things under the hood, etc. loaded the vex, started another log (you all get the idea by now) car is driving superb now, just some odd spots here and there (ones you wouldn't normally hit anyway, but we're shooting for perfection right?), low rpm full throttle (yeeeek 10:1).


Its still struggling with narrowing down part throttle above 2750 it seems, but its zeroing in fairly well. All in all I like it (and i suspect if i just datalogged the car driving around for a day it'd be all good, just let it munch on a big log for a while... sounds perverted)

JayPSC
07-15-2005, 08:30 AM
Right everyone 1.2Beta is Here (http://www.jaypsc.co.uk/ve.zip)

All suggestions have been taken care of apart from the multiple log loading, which I'll start working on, for now you'll have to cut/paste them together.

It now pulls the AFR targets out of a .txt file (you still need excel for the logs though) & you can also change the bins within the prog, just doubleclick on any AFRtarget bin & type your new AFR & press enter to set it. It also saves any changes you make to the AFRtargets when you close the prog, only prob you might have with this is if you go from 8x8 to 12x12 as your tables will be a bit messed up & bits missing, shouldn't be a prob for most though.

Odd why some parts of the table are proving more difficult for you Kenny as the calcs should be the same for all parts of the VE table, all I can suggest is turning the hit count up to filter it a bit more. You should also check to see if you have any offset between what your TE is reading & what it is outputting to MT, I know I have a slight deviation between what my LM1 displays & what MT picks up.

Kyle I've made the prog as small dimension wise as I can & it seems to fit on 800x600 ok

Jay

sberry
07-15-2005, 09:07 AM
The logs are store with an .xls extension but they are actually tab delimited text files
Why not just parse the datalog and nix the need for Excel?

split(/\t/,$row);

steve

Matt Dupuis
07-15-2005, 09:25 AM
This is my drum, and I'm gonna bang it:

One reason for certain bins going wonky might be the injector opening time. Again, depending on how many squirts you're using and how far off your IOT figure is from reality, the "simple" calculation required to get from one AFR to another could be pretty far off. Since Kenny's getting close eventually, and they generally seem to be low pulsewidth bands (high RPM low load), I suspect his IOT is playing havoc on the program the first time 'round. Because he'll eventually get there, the program is proven to work, but it might take a few tries for a few users...

Point of interest - with my new engine, I just played with the IOT figure and observed the results when switching back and forth from 1 squirt to 2 squirts. I know it's not 100% accurate, but I did find that IOT needed (at idle) was between .1 and .2 milliseconds, compared to the suggested .9 to 1.2 ms. At higher RPM the IOT requirement might change for me, since my closing time depends on how un-saturated my injectors are during the PWM phase of the injection cycle, and at idle that's only about .6 ms. I'm not sure of how saturated injectors behave, since I've been using peak-hold since MS' birth.

JayPSC
07-15-2005, 09:41 AM
The logs are store with an .xls extension but they are actually tab delimited text files
Why not just parse the datalog and nix the need for Excel?

split(/\t/,$row);

steve

Yup that will be the next step, was just easier to get it working off the bat with excel

linuxman51
07-15-2005, 11:04 AM
Sorry, shoulda stated, i generally use a minimum of 15 hits, it keeps the program from tweaking things i dont want tweaked :)

kyle242gt
07-15-2005, 01:17 PM
DL'd last night, 408K. Giddily made my changes, working from 45 hits down to 3 just for kicks. Changed my map, and drove again today.

The most heavily logged areas, 20 and 50Kpa, left me running far richer than I wanted, 14.5-15 versus 16-16.5.

Logged again, and lo and behold, the areas that were too rich are leaned by the VE program 1-3 points, which is roughly 4%, or .6 AFR. We shall see, but I think it's getting close without getting all the way to my target.

Matt, you have a knack for making my head hurt :omg: are you suggesting that our IOT is incorrect? I have 1.0 on greentops (I know I know you're a badass with a V8 and everyone's setup is different therefore you can't help, only confound :-D)

Jay, that tooltip on each cell is the sh!t. Seriously. :urgod: And thanks for the 800x600 accomodation.

//edit - the de-highlighting is nice!!

//edit 2 - how about having two tooltips for cells that don't make the hitcount - number of hits and suggested VE. That'd be great for holes in the new VE table, maybe even a way to override and enter it, like on the afr table?

Matt Dupuis
07-15-2005, 01:28 PM
If your IOT is correct, your AFR should stay the same regardless of how many squirts/cycle you choose. If you go rich when you double your squirts/cycle, your IOT could be too high. You could also be running into rounding errors, but that's another story for another day.

Megasquirt calculates the pulsewidth from Reqfuel, VE, and all the enrichments, divides it by how many squirts/cycle, and then adds on the IOT. If your IOT is too high, you'll be squirting extra fuel in for each injection. Doubling the injections will double the error, which is a quick & easy test to see how far out you are. Warning: it could be different at high load compared to low load...

And therefore, when this program calculates how much richer the VE table must be, it's affecting the calculation BEFORE the IOT is added on. If your IOT is incorrect, so will this program be (to a lesser degree).

The closer your VE map is to correct, the closer this program will get you. You can start out with a really poor one, and this program will get you closer each time you run a new datalog thru it.

It's not a fault with the program - I noticed the same thing with my early Excel version. It also wouldn't matter if you told this program what your IOT was in Megatune - you'd need to know how far out it is to make it work properly, and if you know that you might as well put that info into Megatune!

Edit - I'm using greentops in my V8 as well (yeah, gotta have SOME Volvo parts!) and using the flyback circuit my IOT is 0.2ms. I even think that's too high for idle, but it might be wrong for higher RPMs - I haven't tested it yet.

linuxman51
07-15-2005, 02:30 PM
here's an idea to make it a little more agressive, have the program "compare" two vex's and logs, for instance, you go out, log it for a bit, crunch the numbers, spit out a new vex, load that, datalog that, take the first unchanged vex, the changed vex, the first datalog and second datalog, and be able to see what kind of effect its having and ramp up aggressiveness or tone it down perhaps.

just an idea :)

Matt Dupuis
07-15-2005, 02:41 PM
Or you could just work on your IOT, so your CLT, IAT, and baro corrections work the way they're supposed to... :-P

kyle242gt
07-15-2005, 02:49 PM
Or you could just work on your IOT, so your CLT, IAT, and baro corrections work the way they're supposed to... :-P

.....so! Now you've got my attention. Are you suggesting that fixing the IOT will correct the CLT/IAT/Baro problems many of us have experienced? (IE too rich when it's cold out, etc).

:-D

JayPSC
07-15-2005, 02:52 PM
here's an idea to make it a little more agressive, have the program "compare" two vex's and logs, for instance, you go out, log it for a bit, crunch the numbers, spit out a new vex, load that, datalog that, take the first unchanged vex, the changed vex, the first datalog and second datalog, and be able to see what kind of effect its having and ramp up aggressiveness or tone it down perhaps.

just an idea :)


As awesome as that would be, I have no idea where to start coding that hahaha. That would definitely be the way to go, could tune the complete VE table in a few runs. Anyone care to gimme a hand with this project? Don't care if you don't know VB as I can do all that, just need a hand with the calcs as my brain is pretty fried from coding everyday this week lol.

As it stands I think this s/w even with it's basic calcs is far more useful/accurate to WB users than MST3000 as it seems to work very well, even if it takes a few runs.

Someone think of something decent to call it while you're at it hahah ;)

Kyle I'm not gonna add an override box on the new VE as that can just as easily be done in MT, I'll see what I can do RE: the suggested VE on the bins below the hitcount threshold though ;)

kyle242gt
07-15-2005, 02:58 PM
Name:
Vexme
InstoVE
Magic
MSTweak_is_teh_ghey

JayPSC
07-15-2005, 03:05 PM
Name:
Vexme
InstoVE
Magic
MSTweak_is_teh_ghey

LMFAO VexMe sounds good, will do for time being hahahaa

Matt Dupuis
07-15-2005, 03:28 PM
.....so! Now you've got my attention. Are you suggesting that fixing the IOT will correct the CLT/IAT/Baro problems many of us have experienced? (IE too rich when it's cold out, etc).

:-D

Nope. Too rich when it's cold out comes from the Airdenfac.inc file being too literal of the ideal gas law. It's one of my other bitches...

You could fix your Airdenfac.inc file and recompile the S19 file, like I've done.

But if your IOT is too high, the CLT/IAT/Baro problems won't be LINEAR and predictable. Let's use an example, shall we? Let's say your IOT is 0.6ms too high (you've got it set to 1.0 but it's really 0.4), and you're driving around on a cold day. Your IAT tells the computer to richen it up by 10%. At idle your total pulsewidth is normally 2.0ms and cruising down the highway it's 4.0ms.

Because your computer THINKS it's only squirting fuel for 1.0ms, the IAT enrichment will bump the calculated squirt up 10% to 1.1ms at idle, which will bring the total time to 2.1ms. But in reality your squirt time is 1.6ms, and the 0.1ms the IAT enriches things by is only 6.25% when it's trying to be 10%.

Now you hit the highway - your normal cruise pw is 4.0ms, the computer has calculated 3.0ms and it richens it up by .3ms because of the air temp. Because of the IOT error, the actual squirt time was 3.6, is now 3.9, and the IAT has only enriched things by 8.33%.

If your IOT was set correctly, the enrichment would have been 10% in both situations, rather than 6.25 at idle and 8.33 at cruise.

Now, to be completely accurate, IOT is USUALLY around 1.0ms. However, there's also an injector CLOSING time that's based on how quickly the current breaks down in the injector and the pintle closes. Think of this as "bonus" fuel that happens after every injection. IOT is actually the amount of time it takes for the injector to start fueling, minus the amount of time it takes to stop fueling. That's why I'm getting numbers that are so "small".

Saturated injectors flow less current so they tend to take longer to open, and the closing time is relatively small. Peak/hold injectors open quickly and then the current ramps down to about the same as saturated injectors, so they close in about the same time as saturated ones do. If you're using resistor packs on your greentops (or what have you), consider them saturated.

Now, go give it a try and see what you come up with!

Edit - Also keep in mind that the way you've wired your injectors - the sequence, the gauge of wire, the length of wire, any fuses or relays - plus the alternator, battery, and whatever switches you're going thru, will all affect the current available at your injectors, and therefore your IOT. This is why I say that every car's different!

linuxman51
07-15-2005, 04:03 PM
Or you could just work on your IOT, so your CLT, IAT, and baro corrections work the way they're supposed to... :-P

.... I got the opening time for the injectors i'm using from rc engineering

linuxman51
07-15-2005, 04:13 PM
As awesome as that would be, I have no idea where to start coding that hahaha. That would definitely be the way to go, could tune the complete VE table in a few runs. Anyone care to gimme a hand with this project? Don't care if you don't know VB as I can do all that, just need a hand with the calcs as my brain is pretty fried from coding everyday this week lol.

As it stands I think this s/w even with it's basic calcs is far more useful/accurate to WB users than MST3000 as it seems to work very well, even if it takes a few runs.

Someone think of something decent to call it while you're at it hahah ;)

Kyle I'm not gonna add an override box on the new VE as that can just as easily be done in MT, I'll see what I can do RE: the suggested VE on the bins below the hitcount threshold though ;)

I would think (might have to test this theory somewhat) that by having the start and end point and the data used to get there that one could come up with a local scaler partial to each ve bin, and then if say it took rasing bin X by 5 to illicit a 2 afr change and its still 1 afr off target you could bump it by 2 or 3 to get that last little bit, or perhaps an easier way to do this would be generate a scalar table for the first datalog, an actual afrs table from the before datalog, an actual afrs from the after datalog, look at what it took to get from point a to point b, and apply that forwards to datalog C (not sure if i'm being clear here, lemme pseduo diagram it):

Initial run ->
Datalog A
Vex A
Parse this with vexme, and have it generate an initial actual afr table (based on what the logs say)
Datalog B
Vex B
Feed this stuff back into vexme along with the orriginal actual afr table, and have it generate a current afr table and look at what the value of change was between vex A to vex B, what afr change that illicted, and then generate a scaler table based on this information, apply this forward to the next vex file with respect to the scalar (might want to have a table of pair data for faster crunching, the scale value and the change in afr, csv works well for this).

kyle242gt
07-15-2005, 05:33 PM
Just ran a 1.4M DL, man I love this program. I now have hordes of cells whose values aren't changing, even with 100+ hits. This makes the third time I've run it since turning off the enrichments.

the poi
07-15-2005, 05:35 PM
Just ran a 1.4M DL, man I love this program. I now have hordes of cells whose values aren't changing, even with 100+ hits. This makes the third time I've run it since turning off the enrichments.
thats what i was getting back on the old excel spreadsheet. its a pretyt awesome feeling, especially when you start driving around watching your AFR and its ****ign spot-on everywhere. Matt rocks for getting the idea down and Jay now rocks for making it so damn easy to use. Kick. Ass.

JayPSC
07-15-2005, 06:56 PM
Another minor change just for you Kyle lol

http://www.jaypsc.co.uk/vexme/ve3.jpg

Blue= Cells with no change
Red= Cells that need to go richer
Green= Cells that need to go leaner
White= No hits

Look ok? Much more visual IMO - Full blue table & you're finished lol

edit: Moved the files btw, now available HERE (http://www.jaypsc.co.uk/vexme/)

ovlov760
07-15-2005, 07:46 PM
The new download location isn't working for me. I get a 404 file not found error when I click on the file link? If you want a mirror to host your file for tbricks members let me know.

JayPSC
07-15-2005, 07:55 PM
Fixed, was me being a dick lol. Should be fine for hosting, have unlimited bandwidth/space on my host :cool:

linuxman51
07-15-2005, 10:20 PM
the change in afr targets was most distressing. only complaint with the new stuff tho, bouta go out for some evening fun and datalogging

kyle242gt
07-15-2005, 10:55 PM
Colors didn't work?

another 1M log, and I've got a wide swath of no-change.

Serious props on this thing, can't forget Matt Dr. Braniac.

linuxman51
07-16-2005, 12:45 AM
Colors didn't work?

another 1M log, and I've got a wide swath of no-change.

Serious props on this thing, can't forget Matt Dr. Braniac.


no, how it stores and reads the target air fuels changed. took me a minute to figure out what was up (i.e. its not reading the txt file anymore)

part throttle is doing quite well, it keeps overshooting and overshooting the targets at wide open throttle, everything else is great (i think this has something to do with the "lag" between the map sensor change and the actual afr change)

ovlov760
07-16-2005, 03:58 AM
is there any way to make it so it can use both excel & notepad AFRtargets? I personally find it more difficult entering the #'s in notepad.

JayPSC
07-16-2005, 06:51 AM
is there any way to make it so it can use both excel & notepad AFRtargets? I personally find it more difficult entering the #'s in notepad.

Yup same, unreliable too so gone back to Excel, it saves the tables out if you change them within VexME now

kyle242gt
07-16-2005, 01:07 PM
Got it working (didn't notice the file name change, so shortcut didn't work).

The AFR tables were distressing :) but such is the life of a beta user.

Colors are slick! I'd still like to see a popup of VE values that didn't make the cutoff, and a way to enter/override the calcs.

coldfusion21
07-17-2005, 10:43 PM
can someone lend me a AFR table? im not running extra code and so i dont have a table of my own, and when i try to make one in megatune and export it, it wont let me import or export anything, so if someone could lend me one so i could see the format, i would apreciate it.

linuxman51
07-18-2005, 12:31 AM
well I guess it doesnt like the 16v-ness of the car, but i keep getting "ballpark" maps regardless of how large or small the datalog, i raised my IOT back to what rc eng. says it is (i've been using 1.2, i was told its in the neighborhood of 1.5-1.6, so I set it to 1.6 and let the program work its magic)

I'll post my more recent vex and datalog, along with the target afrs txt so you can see what i'm shooting for (nothing oddball here, i promise), maybe someone can shed some light on why exactly it hits above and below so much.

(bundled in a zip file since vex isnt a valid extention)


*note that sun-3-vex is the resultant of the datalog and sun-2-vex (generated with 50 hits)

JayPSC
07-18-2005, 05:51 AM
I'll have a look at the logs later when I get a chance, for now I've been working on the comparison mode, how does this look to everyone?? -

Load Vex1, Log1 & Targets to output Vex2

Say for a given cell VE=43, ActualAFR=15, TargetAFR=12.5

VE x ActualAFR / TargetAFR = VE2
43x15/15=51.6

Load Vex2, Log2 & Targets

On this run the AFR on that cell has actually only gotten to 13 when we were aiming for 12.5

So we calc the difference between what we were shooting for & what we actually got
ActualAFR2/TargetAFR=Scale factor

13/12.5=1.04 or +4%

Next time we calc the vex we use

(VE2 x ActualAFR2 / TargetAFR) x Scale factor = VE3

(51.6x13/12.5)x1.04=55.8

Does this look the right way of going about doing this, or does anyone have any better ideas?
I wanna try to get some kind of comparison mode coded this week so we can get it tested, if we can nail it then theres no reason we cant get the accuracy bang on...

kyle242gt
07-19-2005, 12:57 PM
I'd say that sounds about right, but this is kinda Kenny's baby.

As far as I'm concerned, if you can run four or five ten minute logs, crunch the numbers, and spit out a solid map, I'm happy.

Now, how about something for spark? :rofl: :jk:

linuxman51
07-19-2005, 02:18 PM
I'd say that sounds about right, but this is kinda Kenny's baby.

As far as I'm concerned, if you can run four or five ten minute logs, crunch the numbers, and spit out a solid map, I'm happy.

Now, how about something for spark? :rofl: :jk:


sure thing, just datalog an ion sensor and map it for peak cylinder pressure at 14* atdc

kyle242gt
07-21-2005, 12:47 PM
Back to reality, after daydreaming about ion sensing and overcoming the shock that something so complex a) is of interest to me and b) is at least vaguely understood...

Cut/n/pasted two logs together, crunched away, hit save new VE, then thought, ya know, I wanna run this one more time..so I hit cancel and got a "Run Time Error: Cancel was selected"

Any love on the popup of suggested values for bins below the hitcount?

JayPSC
07-21-2005, 02:47 PM
Yeah I noticed the error on cancelling hahhaha, afraid I forgot to turn the error catching on in that subroutine when I compiled that version. Couldn't recompile as the code is radically different now with the comparison mode & I didn't save the source for the earlier version (which I really should start doing). You'll get your suggested values, don't worry jim-lad, I'll do it when I get bored of doing comparison coding.

I'll get as much done as I can this week & next as after that I'm back doing contract work for a month or so so won't have as much time to dedicate to it.

Jay

kyle242gt
07-21-2005, 03:01 PM
All good man, the comparison mode gives me a headache anyway :-P This software is completely usable and, frankly, indispensable as is...

JayPSC
07-21-2005, 07:38 PM
Comparison mode is done, at least the base code anyways. Whether the calcs are right & actually work or not is a different matter. Aint had chance to test as my car is in bits again, so I have Kenny out giving it a go. I'll tidy it all up, go through it, make a few changes that I wanna do & release new beta tomorrow.
Kyle your tooltip code still won't be done, I've had a look & it requires a surprising amount of coding, but I'll at least make a start on it.

Jay

linuxman51
07-22-2005, 02:23 AM
Ok some nitt-picking and casual notes regarding the comparision mode stuff.. its better but not quite the 1-2-done (I didnt expect it to be either, so no worries there).

It won't write out a new VME after the second pass (first pass is no comparision, heres the datalog and the vex, just like the old software), it had multiple colors for things, and here's what I noticed.. (i'm changing my fuel map entirely tomorrow to accomodate wierdness in tuning i've noticed with the 16v, so much of these issues might clear themselves up anyway, without the software i was having some strange tuneability problems, so I dont expect the software to jump up and go "oh hey i can tune this" on the first pass).

It does better, its still a bit overzealous (again though that might be my map layout), but *MUCH* more consistent, whereas before with the software, i would have this magical barrier at 3k where everything would go to hell regardless, this thing cleaned up stuff clear to where i reved it, leaned out my WOT where it needed to be and added a touch to give me a fairly smooth 13.5-13.9 (target is 13.9) wot range.
It should also be noted that this trial was done with two short datalogs, and like I said, i'm going to re-do the whole fuel map tomorrow, i'll let it cut its teeth on that and report more in the afternoon.
All in all, its excellent, you've outdone yourself, and this should serve to prove to those who doubt, that stateful software is always mo betta than stateless software ;)

the poi
07-22-2005, 03:35 AM
Just came back from tuning volvorod85's car with it. I had used my spreadsheet for most of the map in the last few weeks, but tonight, I used the latest version on the Vexme site (so i have no diea what this compariosn nonsense is :-P ) and it works fantastically well. Tuned some highway and in town crusing numbers, and only had to come up and down on a handful of bins. went out driving and all AFRs are spot on. Works ****ign great, so much faster than crunching lines through the spreadsheet! awesome work jay.

JayPSC
07-22-2005, 09:52 AM
Nice to hear it's working well for you guys, 1.22 beta is Here (www.jaypsc.co.uk/vexme) along with full instructions, this contains all the comparison mode & a few other changes.

Let me know how you get on & if it seems to be working fine I'll release it on MSEFI too.

Jay

blkaplan
07-22-2005, 10:47 AM
i got to test this out,

maybe ill give it a go tonight

linuxman51
07-23-2005, 02:52 AM
Breakthrough (well maybe). MS is only seeing 3.96v max from the TE (this obviously hampers how the program would work, esp with the afrs i had in the file).
I'm gonna hook the laptop up to it and see whats up, might take it appart and see if i've bridged something or maybe grabbed the wrong pin? (anyone got any suggestions here? :-D)

the poi
07-23-2005, 05:41 AM
Breakthrough (well maybe). MS is only seeing 3.96v max from the TE (this obviously hampers how the program would work, esp with the afrs i had in the file).
I'm gonna hook the laptop up to it and see whats up, might take it appart and see if i've bridged something or maybe grabbed the wrong pin? (anyone got any suggestions here? :-D)
BWAHAHAAHAHA. if i had a nickel for every time......


cut diode D11. (i think its d11...) its a 5.1v zener protection diode for the O2 circuit but it starts clamping quite a bit earlier (around 3.8v) :-P

JayPSC
07-23-2005, 07:05 AM
How's that Kyle?

http://www.jaypsc.co.uk/vexme/ve4.jpg

I'm not writing the facility to be able to change the VE bins within VexME because that'll completly screw stuff up if people start changing VE bins & then run the comparison mode, you can change them within MT anyways.

edit: This has had an added benefit actually, in comp mode when you hover over bins that have had a scale factor applied to them, the suggested VE that pops up has no scale applied to it so you can see how much difference the scale factor is actually making to that bin.

Has been uploaded now, still v1.22

Jay

linuxman51
07-23-2005, 09:22 AM
BWAHAHAAHAHA. if i had a nickel for every time......

Thats it, i'm retroactivly taking my toys and going home!


cut diode D11. (i think its d11...) its a 5.1v zener protection diode for the O2 circuit but it starts clamping quite a bit earlier (around 3.8v) :-P


cool, yea i thought it was funny that it would stop *perfectly* at 3.96 every time (I measured out like 4.5 with the volt meter before installing, :sigh: easy fix)

linuxman51
07-24-2005, 01:01 PM
Ok found a potential bug in the software, results from running comparision mode a second time (with the second vme), clobbered my afrtargets excel file several times until i deteled everything and rebooted (first time i tried running it a second time, so it could have been a fluke too)

JayPSC
07-24-2005, 01:26 PM
Fixed ;-)

linuxman51
07-24-2005, 02:13 PM
*bling*

JayPSC
07-24-2005, 04:18 PM
DIY-WB non linear support also added :twisted:

linuxman51
07-25-2005, 02:26 AM
now it asks if i want to save "copy of afrtargets" when i exit. other than that this thing is working like a champ dude!

JayPSC
07-25-2005, 04:44 AM
Cool, weird how it's asking you to save that though, the prompt should be turned off, can't recreate it on my laptop either :???:

linuxman51
07-25-2005, 04:54 AM
Cool, weird how it's asking you to save that though, the prompt should be turned off, can't recreate it on my laptop either :???:

could just be XP telling me its time...

JayPSC
07-25-2005, 05:47 AM
Lol, 1.24 now ready for download, added Zeitronix WB support ;-)

kyle242gt
07-25-2005, 01:14 PM
I'm using 1.22, thanks for the popups!

I liked the old colors better; the new ones somehow look a little washed out.

Two suggestions, one reasonable, one not:
Reasonable: Some way to bypass the "normal/comparison" selection
Unreasonable: I had a fab idea that you will not want to implement, but it's so glorious I just have to bring it up... and yes, I'm beating a dead horse :-D here we go:

Have the popup be either the calc'd VE or the original, depending on whether or not it made the hitcount. Then, make it possible to doubleclick to toggle calc'd or original.

That would be sweet.

Told ya it was unreasonable. :-D

Kenny, you tried comparison yet?

K

JayPSC
07-25-2005, 02:32 PM
I'm using 1.22, thanks for the popups!

I liked the old colors better; the new ones somehow look a little washed out.

Two suggestions, one reasonable, one not:
Reasonable: Some way to bypass the "normal/comparison" selection
Unreasonable: I had a fab idea that you will not want to implement, but it's so glorious I just have to bring it up... and yes, I'm beating a dead horse :-D here we go:

Have the popup be either the calc'd VE or the original, depending on whether or not it made the hitcount. Then, make it possible to doubleclick to toggle calc'd or original.

That would be sweet.

Told ya it was unreasonable. :-D

Kenny, you tried comparison yet?

K

Yeah colours are washed out in normal mode, but there's gotta be a way of easily seeing difference between normal & scaled bins in comp mode, that's why I changed them.
Bypassing the selection screen is a pain unless I add a settings file which I really don't wanna do right now unless it gets more complex.
As for the switching bins between calc'd & original, could probably add it & switch off the flag for that cell so in comparison mode it isnt affected, will have a look at it if get a chance.

Jay

linuxman51
07-25-2005, 03:34 PM
Kenny, you tried comparison yet?

K


yes, i discovered the bug in it foo :-P

linuxman51
07-25-2005, 07:44 PM
So I was talking to jay early this morning (aboot 3:30am or so) and mentioned something i'd noticed, and its happened again.

In comparison mode it will modify the map as per usual (well the criteria present), but the suggested values dont always match what its generated..

for instance, at 1200rpms 85 kpa, the old value was 83
the suggested is 99, but the generated is 118... (dark pink). It should be noted i've noticed this more when its used in VME mode..

linuxman51
07-25-2005, 07:48 PM
i might be on to something here, more in a few...

linuxman51
07-25-2005, 07:58 PM
Ok the maps in the VME are different from the VEX near as i can tell, the VEX has my 85/1200 bin as 50, the VME has it as 83. i'll zip the two up and post for jay's benefit, it seems to only happen after a cycle or two through comparison mode (kinda like maybe the scalars are added to the map or something to that effect?)

linuxman51
07-26-2005, 08:27 AM
man yall other folks must suck at the beta testing, i'm racking up on the bug count here :-D (probably much to jay's chagrin hahahaha)

version 1.26 forthcoming...

JayPSC
07-26-2005, 08:38 AM
man yall other folks must suck at the beta testing, i'm racking up on the bug count here :-D (probably much to jay's chagrin hahahaha)

version 1.26 forthcoming...

Quite lucky you found that actually, quite a large error in comp mode - fixed now in 1.26.
MSII support too if any of you are that lucky yet :-P

linuxman51
07-26-2005, 08:47 PM
ok new issue (you still love me, right baby?)

"Unexpected error 30009: Invalid row value" when saving the vme.

JayPSC
07-27-2005, 04:25 AM
Fixed Biiiiiiatch

JayPSC
08-03-2005, 04:19 PM
I've also just had a look at a mode using the Gego in the log for use with AFR targets with extra mode, grabs an average Gego for each bin then multiplies the old VE by that number. Anyone interested? If so I'll release tomorrow or next day, very busy at the mo, don't have much spare time.

Jay

DeathWagon
08-03-2005, 07:21 PM
I've also just had a look at a mode using the Gego in the log for use with AFR targets with extra mode, grabs an average Gego for each bin then multiplies the old VE by that number. Anyone interested? If so I'll release tomorrow or next day, very busy at the mo, don't have much spare time.

Jay


I'll give it a whirl, hit me up on aim or forward it to me.

-linux

JayPSC
08-04-2005, 03:12 PM
v1.28 available now

Matt Dupuis
09-09-2005, 05:26 PM
** Poof ** Back from the nearly dead.

About 2 days after the early releases of this program, my new engine took a ****. 6 weeks later, I'm back on the road, so I'm ready to test s'more. Kinda surprised nothing's happened in the past few weeks on this topic, but W/E.

So this weekend I'm going to do some driving 'round and tuning, though I'm still in the baby-forth mode 'cause I'm worried about lunching another cam and/or thrust bearing.

I was also talking with another local MS'r, and he's been running MSII. One thing lead to another and I'm wondering how much effort it would be to produce an MSII version? I'll get you a VEX and a few datalogs in a few days so you can play...

Edit - has this been introduced to the MS board yet? I'm sure they'd be crazy for it...

kyle242gt
09-09-2005, 05:39 PM
Hey Matt - killed an engine eh? Reminds me of my painfully expensive foray into the world of BBC... I think it took two cams and three rebuilds to get it right :roll: what happened?

Vexme has MSII support, or at least it appears to. Nobody on MSII here that I know of.

I used this enough to get my fuel map spot on, then I turned accel/overrun/EGOadj back on and ignored the laptop completely... only got 17MPG :grrr: even though I kept my foot out. Revising spark map and DLing again in preparation for further Vexmeing.

I think the lack of activity in this topic is just because it works so nicely, no need to talk about it. I never saw a release of this on MSEFI, but I don't get over there too often.

benflynn
09-09-2005, 06:07 PM
it's there

Matt Dupuis
09-09-2005, 08:56 PM
What's where? MSII support or VexMe on MSEFI.com?

The gear on the cam that drives the distributor/oil pump failed. I dunno what the problem was. Crane billet roller cam (used) and steel Explorer cam position sensor gear (also used), so I think the used-ness of each piece contributed to the failure. I had plenty of float on the CPS shaft, plenty of float on the oil pump drive shaft, only ran 10w30, and though the pump is a HV, it's only a standard pressure and the loose bearing clearances keep the pressure below 60 psi. I added an oil spray on the cam to help alleviate the problem, and I hope my brand new Crane cam / MSD gear combo lasts.

The steel shavings in the oil for 500 km wiped out the thrust bearing, and possibly the proximity of the clutch cable to the headers caused a constant load on the TO bearing... I've got an oiling mod done for this too, suggested by Clevite, plus I've loosened the cable quite a bit compared to the last time.

I really wish I knew EXACTLY what happened, but I don't.

benflynn
09-09-2005, 09:24 PM
...
VexMe on MSEFI.com

kyle242gt
09-10-2005, 12:16 PM
:yikes: mystery failures are the worst. My BBC's woes turned out to be oversize cam bearing journals - the bearings would rotate, chew up the cam, send junk into the crankcase, etc. Only discovered that when I took the entire engine to the machinist and said "HELP!"" Many $$$$ later it was done. :x: here's hoping that your problems are at an end, my overpowered friend.

JayPSC
09-11-2005, 08:33 AM
What's where? MSII support or VexMe on MSEFI.com?

http://www.msefi.com/viewforum.php?f=86&sid=c582c5e22554f5e7eed0adcb289df0bb

Have my own sub forum on MSEFI now, not had a lot of time to work on it recently, far too busy with work, but tbh it works pretty well as it it. Full MSII support now & most WBO2s too.

Jay

benflynn
09-11-2005, 11:30 AM
now it is my turn, what are you guys running for target afr's?

Matt Dupuis
09-15-2005, 11:52 AM
http://www.msefi.com/viewforum.php?f=86&sid=c582c5e22554f5e7eed0adcb289df0bb

Have my own sub forum on MSEFI now, not had a lot of time to work on it recently, far too busy with work, but tbh it works pretty well as it it. Full MSII support now & most WBO2s too.

Jay

After it was confirmed that it's on MSEFI, I dropped in and found your forum. Good thing too, 'cause I had troubles with the .dll file that needed to be deleted and replaced...

The Compare mode is wacky, right? I've tried it and everything it wanted to change, it tried changing in the opposite direction: i.e., original bin was 73, compare suggested 68, normal suggested 77.

Also when saving the VME file and loading it back up in the compare mode, my AFR table went all screwy in some of the lower bins - is that normal? In any event, I wasn't allowed to change it in Compare, but was allowed to correct it in Normal mode.

So far so good, though - car runs much smoother, much quicker than my old spreadsheet. I'll use this to get the map close to 14.7 across the board, switch to a narrowband and run MST3K to get the peaks & valleys lined up, and then switch back to get the VE's & AFR's finalized.

kyle242gt
09-15-2005, 12:11 PM
now it is my turn, what are you guys running for target afr's?
25 - 17
50-90 - 16
100-110 - 14 (except 2800/3200 - 14.5)
120 - 13.5
140 - 13
180 - 12.93
200 - 12.50

I'll use this to get the map close to 14.7 across the board, switch to a narrowband and run MST3K to get the peaks & valleys lined up, and then switch back to get the VE's & AFR's finalized.
Choomean "get the peaks and valleys lined up"?

Matt Dupuis
09-15-2005, 12:26 PM
In the VE table.

The calculations in Megasquirt take into account manifold pressure in the fueling calculations. If your engine were equally efficient at all RPMs and manifold pressures, you'd enter the same VE into all 144 bins. In fact, you'd only need one bin...

But your engine wants different amounts of fuel at different RPMs and manifold pressures, which is why we've got a 12x12 table. The reason for this is because every component in your engine becomes more or less efficient at different RPMs, and they overlap and combine to make a bunch of "peaks and valleys" across your VE table. Your intake manifold goes through different harmonics as the RPM rises, and those harmonics aren't the same at all manifold pressures, for example.

Megasquirt interpolates between VE bins - if you've got a bin at {1000 RPM / 70 kPa} and the entry is 60 VE, and your next bin is {2000 RPM / 70 kPa} and it's entry is 68 VE, MS will smoothly interpolate at an 8 kPa/100 RPM slope. Therefore if you've got a relatively even "ridge" in your VE table - a peak with a more-or-less constant slope from low to high load - you should set one RPM column at the low point of that ridge and one at the high point of that ridge, so MS can accurately predict all points in between.

What's more important, however, is that a peak or valley doesn't occur IN BETWEEN rpm columns. If that happens, MS is going to try and interpolate between the columns, but the engine actually wants more or less fuel than MS thinks it wants. Sure, the O2 sensor will compensate to a degree, but only when it's activated... If the engine is in warmup, accel, decel, or if the VE table is greater than your O2 sensor range, it's gonna run poorly.

It's one more step in getting an engine tuned properly.

I've gone thru this all in the MST3K thread, in the Topics forum.

boosted
09-23-2005, 12:25 PM
ok after perfectly tuning a bmw 325, i can really say.

VEXME IS THE ****!

ovlov760
09-26-2005, 05:12 PM
any word on whether or not this will be integrated into future versions of MT? :badboy:

coldfusion21
10-17-2005, 10:51 PM
ok, so i ran this fine with all my 8x8 maps but when i try to run a 12x12 it gets all pissy and gives me an error code 6, map overflow. the problem is that this is a vex file spat out from megatune and should be just fine. i looked at the two in notepad (a regular 8x8 .vex and the 12x12 that was being pissy) and the only difference, asside from have 4 more of all the bins, is that the 12x12 has
"Scale 0.352000
Translate -28.400000" before it starts describing the tables.

http://countofnowhere.homeip.net/~jared/10-15-05-extra.vex

ive also been playing around with my .vex in this program

http://countofnowhere.homeip.net/~jared/VEXer.xls

could that have something to do with it?

JayPSC
10-18-2005, 05:42 PM
Yeah that might cause an issue, also are you using MSII because your RPM bins are not defined in the same way as a normal MS1 .vex ie just the 1st 2 digits.

Jay

coldfusion21
10-18-2005, 05:47 PM
nope, ms1, s9 code...

coldfusion21
03-10-2006, 12:31 AM
sorry top bump the old thread, but im having an issue, i pm'ed jay about, but im assuming he's to busy to respond right now and was hoping someone else could chime in.

when i try to load a log, i get an error 13: type mismatch

im using MT2.25 and code version c13. it loads my VEX just fine, but no matter the data log, it refuses to open it.

ovlov760
03-10-2006, 12:36 AM
sorry top bump the old thread, but im having an issue, i pm'ed jay about, but im assuming he's to busy to respond right now and was hoping someone else could chime in.

when i try to load a log, i get an error 13: type mismatch

im using MT2.25 and code version c13. it loads my VEX just fine, but no matter the data log, it refuses to open it.
Does the first line of your datalog contain some information about the code version your using? If so, delete that line, save and retry vexme.

coldfusion21
03-10-2006, 01:34 AM
Perfect!

:-d