BACK ON TOPIC.


I've been logging my AFRs in boost vs. my AFR Targets. VE Live is getting that shit pretty damn spot on without any effort.


also, I dunno why, but my wastegate cannot hold 12psi anymore and I'm now constantly boosting to 16psi....it's fun.

1. It truly is the best thing since sliced bread.

2. I sense a new motor in your future :giggle:.

1. Yeah it's neat.

2. Nah, i figured out the issue: I'm a retard.

<threadjack>
I'm sure it wasn't as bad as my friend's screw-up. He kept overboosting, when I looked at his car the first thing I noticed was a vac line running strait from his BOV to his wastegate actuator...no boost reference in sight.

Oh, and were you serious about getting rid of that vband wastegate housing? You never responded to my PM.
</threadjack>

well, I put my vacuum line on the wrong wastegate port when I was fooling around with it over the weekend.

yeah, you gotta pester me. I actually have it in hand now at least, I'm getting closer.


Yeah, VE Live is doing a damn good job I'm going to have to say:

(look at targets vs actual on the bottom)

http://www.boostedmiata.com/gallery2...serialNumber=1

How does the map look though? I find aesthetically it looks horrible!

I think I prefer using MLV as I can sanity check any changes against a datalog log. just me though, the live analyze is very impressive non-theless.

My map looks like a monkey bashed his fists against the number pad. Now, with that being said....Does it matter? If the logs show a dead on fuel curve, why not just roll with it?

note: that second part there was a legit question, anyone have an opinion?

I dunno. I think I trust the monkey fist numbers since its smart enough to interpolate between the cells. It kinda seems to know what its changing and how it smoothly relates to the cell before it or after it. Thats just my fuzzy logic on the whole thing.

What's the big deal with aesthetically pleasing fuel maps anyways? We don't live in a perfect world, and in a perfect world our engines would run as smoothly as your pretty fucking rainbow fuel map shit. If the AFR is fairly smooth, and it revs to red smoothly in all throttle positions, it's doing EXACTLY what I'd want it to do. And that's exactly what it is doing, and exactly opposite of what the MSPNP map ever did.

All very fair points, but it just grates, know what I mean. A smooth map you can take pride in.

Meh I shouldn't worry about that shit I have other more important stuff to mess with on the car!

I understand, I'm a little OCD too. But coming from a car that's been running like shit for months if not years with the bandaids, it's a huge huge huge improvement and I don't care at all what the fuel map looks like.

I mean, it works....

I went from this:

http://www.boostedmiata.com/gallery2...serialNumber=1

to this:

http://www.boostedmiata.com/gallery2...serialNumber=1

this was over the course of a week. I'll more than likely go back in and smooth out some of the outlander values.

ha...I love this 16x16 business...


my old map:
http://www.boostedmiata.com/gallery2...serialNumber=1

(the column 4K where the numbers are low were fixed...that was MLV fail)

Man I wish i would have taken a shot of my old "smooth" map. 3 runs of the new autotune and it looks choppy at first glance, but hovering my mouse over each cell it starts to make sense when you can see your targets and how they were interpolated from my 8x8 target table.

Needless to say the drive is butter smooth now and I think its partly because I started tuning with AE left on.

Does yous peoples tunes withs overruns ons?

custom filter [tps] > 1 or something like that.

Cant wait to use this.

I've always had an issue with MLV mainly offering me a tune that suited the way I drove for that particular tuning session. I know auto tune works in the same way, but live. If I drive through all the gears quite slowly (stuck in traffic or queueing around town) I go through parts of the map that I've spent most time cruising through the streets at and because I've been sat in these cells for 90% of the time i.e. 2300rpm/45kpa, if I then drive through that zone in lower gears, I go quite lean and power hesistates slightly (18:1). My way around it was to manually go through logs and dial fuel back in to those areas so that rolling through gears slowly at part-throttle wouldn't run me lean but then rely on EGO to take it back out in these areas when Im cruising in those cells again in a higher gear.

I've always logged with EGO on, AE and decel on.

Maybe I should let auto tune make my map over the course of a week or two and go from there?

i auto tune to and from work each day until it really just changes numbers back and forth.

then ill tune on the highway, once or twice and let the EGO handle the minute little variances.

I'll probably stop using this feature after next week.

How many of you are using this to tune boosted cells?

I was skeptical, but I tested it on a long straight road. I made sure the throttle came on nice and steady. I could feel the car make more power as it tuned the fuel.

My map kinda looks like ass as well. There is a really deep trough through it in 3d mode. I think it is due to the fact that it can dial in the cruise rpm so well. Lets face it, the cruise cells are only a small part of the map. Under hard accel you probably don't touch the cruise cells.

I'm going to tune for a week and then let it be. I tried to hand smooth my map and my car started to run a little rough.

This just goes to show how unique everyones setup is. I think it would be interesting for a few of us to start with the same map and see what we end up with over the course of a week.

Anyone offer some hand holding to transfer from MLV and MegaTune to Tunerstudio?

I am a green horn newb to MS. Cut my teeth with a DOS version TEC2. I have been making it ok changing somethings with MLV and MT. I'm a bit nervous to go to TS.

I have gone in and changed some of the INI's to show my AEM-Linear AFR's and such. When I imported an MSQ and tried to upload to the ECM this morning it gave me some error and asked if I wanted to continue and I chickened out.

I also have AEM AFR issues with MLV and MT not reading the same AFR's. I have read the posts and have not changed the 9.?? value yet. I think I might have a ground issue to take care of first.

I think I need to start a new post on questions about a car that is new to me..... ;)

BTW I can't find the AFR Target 1 on my MLV graph? Thanks

I do, it gives a damn near perfect afr during WOT and its pretty smooth on the ramp up to full boost.

TS reads the AEM WB perfectly without having to fool around with the .ini files. Functionallt the programs are identical, just load in your map and you should be fine. Any errors will show up on the log and will be stupid things like volts mismatch or something that wont effect your tune/settings.


AFR Targets only get logged in MS-II; I'm pro like that.

You can add a custom field in MLV to do similar with a MS1 though...

Would you be so kind as to offer the instructions on how to get this custom field? I'd find it handy to go over what autotune has been doing over the course of say an hours drive and see if the traces start to sync up more and more.

Sounds awesome and I want in.

Ok I won't mess with them. It seems that my MLV and MT AFR reading is 0.6 AFR leaner than what my AEM gauge reads. I take that this is not normal? Just wondered if this was a wiring/voltage difference issue. Which one do you believe?

I also would be interested in the custom gauge suggested richyvrlimited!

Sorry I'm such a :noob:

Thanks!

mine is also doing the same, the support for the aem gauge wideband is poor at best in both megatune and tunerstudio. i imagine it's gonna take fucking around with it to get it close like i had to do with megatune. i don't really understand the problem since it's output is 0v=10:1 afr and 5v-20:1 linearly.

For the gauge type in tuner studio did you select that gauge under project properties AEM gauge 30-42xx and then calibrate afr using aem linear 30-42xx?

Next step will be to email AEM for their AFR vs voltage table and plot it against what tunerstudio uses in the inc file folder using excel. If they match, then I would trust tunerstudio.

After looking into this a bit further I've found that it is only the non-linear aem that has a reference table. Go into the lambda.ini folder and look at the equation for converting the voltage to afr. The use the voltages that you receive from the AEM folks and see hat afr spits out. This is very easy to to do in excel.

I cant remember what I used, but in MT we could never get Saint_foo's AFRs to match well, in TS the were spot on. Yes the AFR targets were in volts, but whatever thats an easy conversion.

page 9: http://www.schnitzracing.com/manuals/AEMWBK.pdf

it outputs 10-18.50 AFR at 0-4.25 volts

I got the car out for a little while yesterday afternoon and played around with the autotune. Well, sort of. I unchecked the "update controller" box so I could see what changes it wanted to make before I allowed it to make them. I ran it on "very hard" since my table is already pretty well dialed in. There was traffic so I couldn't hit as many cells as I normally would. Previously I had been letting MLV do its thing but I would do some of my own smoothing. Like other people here, I found that the numbers autotune wants to run make my table look intuitively less smooth, but I'm just gonna let it do its thing for another few sessions and see how it ends up. It left most cells alone but it wanted to pull 1-2 points from most of my 2400 rpm column, and a few points here and there in the 20 kPa row.

I also took some datalogs at the maximum rate it will do, and TS still appears to be totally free of the occasional comm resets I was getting in Megatune. I would think a noisy CAS should be logged by both programs, but TS seems to be immune. Does TS run a filtering scheme to ignore the spikes?

Another feature I'm diggin is the compare tune function, where it will give you a difference report between two msqs. So if your junk happens to get corrupted you can compare to an older known-good msq without having to hunt for the problem or sift through every single window with Megatune's report function hoping to spot the problem.

a little confused here, if you look at the charts on pages 5 and 6 it shows 0v=10 and 4.99v=19.98
and for whatever reason ts shows mine in afr when in ms i could never it to do that

Scott:

I have had tons of problems with MT corrupting tables and files due to the numerous comm issues.
TS does not seem to do that :2cents:

:yippee: This is awesome can't wait to use it. This is why I'm getting all my buddies on ms.

I got TS up and running yesterday and drove the car some today with it. I have noticed that the AFR's still don't match on the gauge displays, but the AFR's on the AEM gauge seem to be very close to where I have the AFR programmed in TS. So not sure if that is just a gauge display problem in TS and MT.

So far I like it but I think it eats my laptop battery for lunch. It doesn't seem to last very long.

I have my settings on easy and have let it tune in boost. I have not studied my tables though because I am fighting a virus on my home computer which is taking all my free time. :vash:

My understanding is that the live tune will adjust your VE table to try to get the AFR as close to the target values as possible. If it is interpreting them incorrectly and displaying the wrong AFR on screen, then it should be tuning incorrectly as well..

I had a similar issue, ended up changing the formula used to convert volts to AFR in a couple of places (ini files). I can't remember how i came up with my formula but it involved data points, excel, and a trend line.

HTH

I'm going to work on relocating, rewiring and grounding my AEM and go from there. Thanks for the help. I may need more later. ;)

Let me know what your results are from regrounding. I had mine grounded to the head (brown wire) and chassis (black wire). Seemed ok. Then after reading the instructions on FM's website I put the black wire on the head and the brown wire to an MS ground. Seems about the same.

What do you guys use for Lambda Delay in ms ?

Not sure I understand that ( ? )

Also, Is it better to set your filter using low KPA rather than RPM ? I figure that way you can tune high kpa low rpm such as when the motor is under high load.


Thanks :confused:

Poobs!

I think you'll get some insight at msextra forum, thake a look at Megasquirt MSEXTRA and MS3EFI &bull; View topic - VE Analyze Live - A few questions.

Andreas

This may seem like dumb question, but where does VE Live get its target data from?

-Rouphis

Thanks !!! :bigtu:

From the AFR table

The 6x6 one in megasquirt? Or the 12X12 one in MLV. In ether case, it is nowhere near correct. In both tables my 101KPa target is 13.0 yet VE Live is trying to make 14.7.

-Rouphis

It will target whichever table YOU tell it to target. By the way the AFR target table in TS or MT is an 8x8 table.
Do yourself a favour and stick with AFR target table 1 (the 8x8) It's all interpolated so you'll have no gains by having more cells to enter "14.7" in or "12:1" for boost etc.

MLV can be told to base it's analysis on whichever table you tell it to look at.You have an option of three tables IIRC.

I know about MLV, MLV is fine. But in TS VE Live, its not interpolating correctly.

Strange, you would be the first to have this problem.

Is your wideband sending the correct values? Are you seeing the same values in TS and on your WB gauge?

My wideband is fine. And VE analyze in MVL works fine. I will post my .msq in a little while.

-Rouphis

Ok, I think I understand what the delay does but how do I determine what to set it to ?

Maybe I'll just leave it alone but again if something seems wrong how do I know if it is these settings that are at fault ? :noob:

Ok guru's... If I load my current msq in megatune like I have been for the past year and a half, the car runs as good as it ever has.
BUT... If I load the same msq using TS it runs like shit. Completely lean idle around 17 to 1, leans out to 18/19 to 1 while free revving the car in neutral. Didn't even try to drive the car as it was running horribly. The ve map is the same in both TS and Megatune, as are all other settings that I can see.
I can't figure this out, any advice would be appreciated. Why the difference?

have you setup a project in ts properly?

Did you choose the right firmware and set your wideband for your project in TS? Also make sure airden, matfactor, and thermfactor are in the project inc folder.

Followed the directions in the other thread to the best of my knowledge, hires firmware set, all three inc files are in project inc folder, innovate wideband default selected... I'll look again. Keep the tips coming, this is really frustrating me.

Is your afr the same in TS? I didn't know people ran the default LC1 settings.

yes they are the same, I run them on purpose after having a long talk with diy.

Is there a case/condition when you don't have to copy those files to your project folder ????

I'm using TS and MT interchangeably and I'm not having the above problems. !!!

super noob question...

whenever I use VE live it is targeting WAAAAAAY too rich, cruising in 5th I'm getting in the mid 9's.

I realize this is probably all my fault, I just need to know how to correct it. should I reset-up my LC-1, maybe I have the settings incorrect???

The VE live uses your AFR table as its target, check that.

go to ve analyze live > advanced options > AFR targets?

http://i6.photobucket.com/albums/y24...screenshot.jpg

well that explains why it's hitting 9s... rofl

try a table like this:

http://www.boostedmiata.com/gallery2...rgets_MSII.jpg

@jrw - the AFR table in VEanalyze should be the same as your normal AFRtable for VEtable1. It uses the same table.

JRW, that table looks like a classic case of table corruption. I had a few of those when starting out with MS and TS.

Use Brains and see what happens ;)