PDA

View Full Version : LS2 VE Table values???



Acer
November 18th, 2006, 06:26 PM
Have just started adjusting the tune in my HSV VZ GTO (E40 ECM) using Flashscan V2 & noticed some strange values in the VE table. The values range from under 17% to ~ 59% which are much lower than what I would have expected. By comparison the LS1's typically have a range of over 40% to near 100%:confused:.

At first I thought there may have been an error in the units displayed via EFI Tune 7.3.3 (Build 520) but when I select other units in the properties such as gm*K/kPa (default) it still shows the table as a percentage. Did the same test after extracting the tune file using V2 on our LS1 VX SS & it works correctly, i.e, I can change the units & ranges are as expected.

Is this a bug, misnomer or different application by GM of the VE table?

In the mean time I suggest everyone be careful about SD tuning.

Cheers,

Michael

GMPX
November 18th, 2006, 09:39 PM
Try not to compare them to the LS1, in reality they are just numbers 0 - 4096 which in my mind (and suggested by others) is not very friendly to work with, so we chose to scale that to 0 - 100%.

Cheers,
Ross

Acer
November 18th, 2006, 11:27 PM
Ross,

I understand that they are ultimately just numbers but it is always I good idea to understand the logic behind it if you want to calculate the values instead of just guessing them. Also, you talk of 0 - 4096 yet I can only get % out of Flashscan V7.3.3:nixweiss: irrespective of the units I select. Suggest you check this table again as other programs are showing the usual range of VE values for LS2s.

Thanks,

Michael

GMPX
November 19th, 2006, 01:02 AM
Well, I guess the question should be, how would you prefer to see the VE numbers represented?

All we are doing right now is scaling 0 - 4096 into 0 - 100%, it's not really an indication of anything, but for most, 0 - 100% makes more sense. You cannot change the units to display something else because we have not implemented any others.
I guess what you are used to seeing is 100% on an LS1 is probably going to be at the upper end of the table given all other factors are correctly adjusted (Injector table etc). Yet, we do allow up to 500% to be entered in there because that is what mathematically it will calculate out to.

We are open to suggestions on what you feel comfortable with, if we were to simply show values 0 - 4096 then we would have it correct because that is how it is treated in the ECM, if tuner 'A' wants to show it as scaled to 0 - 100% then that is not wrong, if tuner 'B' wants to show it scaled to 0 to 7000 hogs breaths that is also not wrong, they are just a different representations of 0 - 4096. Though, I don't think that is an industry accepted measurement yet :-)

Cheers,
Ross

Redline Motorsports
November 19th, 2006, 04:02 AM
I see both sides but the reality is that the number doesn't really mean anything until you can relate it back to fueling. At that point its all relative. Bigger number more fuel and vice versa. It will be a bit tough for many moving into the LS2 as we so geared for the LS1. It might not be a bad idea to post a quick comparison of differences we should be looking for. This is apparently one of them!

Ross, if the stock PCM resolution goes to 4096 and you condensed it into 100%, then seeing percentages in the mid 300's in going to be normal...correct?

Howard

Acer
November 19th, 2006, 07:21 PM
Hi Ross,

It is not the end of the world but irrespective of what we are used to on the LS1, volumetric efficiency is a standard measure of an engine's efficiency & that is how many will interpret the current table. In fact what it is from your explanations is just a % scaling of the range of numbers that GM will allow in the table.

Personally I would much rather see the raw numbers in their units (say gm*K/kPa or however GM have done it this time) as then I can actually comprehend what they mean & compare with calculations.

I have had excellent results with SD tuning by using EFILive with the MAF connected to log data on the road & convert using a custom formula (similar to the methods described on this forum before but a slightly different equation) to convert to VE values in gm*K/kPa for the different map/rpm zones. Every time I have done this with enough logged data the LTFTs have been exactly where we wanted them (0 or slightly negative) when running in SD (without MAF) & the cars idle & run smoothly throughout the RPM range.

There are probably almost as many views as people here but from my perspective I would certainly appreciate going back to units or at least offer both display options as per the LS1s:) .

Many thanks,

Michael

redhardsupra
November 19th, 2006, 07:27 PM
i made a LS2 VE spreadsheet a while back, it works in both LS2 VE units, as well as the 'theoretical VE' so you work with it any way you'd like.

http://www.ls1gto.com/forums/showthread.php?t=106410

Acer
November 19th, 2006, 10:48 PM
Marcin,

Thanks very much for the links to your articles & tools. You have done an excellent job & generously share your insights with the rest of the tuning community:thankyou2: . I have generally stayed away from just relying on the LTFTs to empically tweak the VE tables, as the way you drive seems to have an impact, but it should work if you do it in sufficient steps & are consistent in the way you drive the car during each attempt.

Cheers,

Michael

redhardsupra
November 20th, 2006, 05:03 AM
it works with AFR%error too, or any other proportional error measure really.
it really doesn't matter whether you do ti in 'GMVE' as i call it, or the 0-100% 'theoretical VE', they're both directly proportional. however, the 0-100 is easier to remember and easier to understand, so i figured i'd give you a tool to do so

joecar
November 20th, 2006, 06:33 AM
The VE table is not really a volumetric efficiency table, but rather an airmass table.

redhardsupra
November 20th, 2006, 09:00 AM
The VE table is not really a volumetric efficiency table, but rather an airmass table.
correct, however airmass changes with things like pressure and temperature, thus airmass, while more expressive and descriptive, it is not as universal. VE (the theoretical one) is the purest form, as it is independent of atmospheric conditions. So you really have to look at VE, GMVE, airmass, and airflow forms to get a really good feel for how the car is running.

GMPX
November 21st, 2006, 01:39 AM
Oh no, not the airmass debate, lets get Paul and Ben Strader in on this one :muahaha: .

Hey, we just write the software, we don't tune cars all day, whatever works for you guys we will change it to, I guess I was thinking along the lines of that the native 0 - 4096 does not relate to anythng we know in the LS1, but those that use aftermarket ECM's as well would be familiar with a VE% table, even though it's not a true VE table......oh no, here we go.....

Cheers,
Ross

joecar
November 21st, 2006, 05:20 AM
I'll take raw 0-4095. :cheers:

redhardsupra
November 21st, 2006, 05:32 AM
ok, since i have admin's attention...
so is the 0-4095 an internal value, how it's stored in the pcm? i'm concerned for the LS2 guys, cause they got not that high to go with it, if stock you get 2.7, and peak is 4.1, that's not much, especially with FI.
same concern goes for the IFR table--63.95lb/hr? c'mon, my grandma could use bigger injectors and she's dead... ;)
are these things gonna be addressed in COS? this would make all FI GTO guys (and theres a lot of them!) convert to EFILive

Acer
November 21st, 2006, 07:43 AM
Ross,

I have no interest in protracted debates, just making the software as user friendly as possible & getting the car tuning right:) .

Tried to send you a PM yesterday but it bounced because your inbox is full. This was about the value of 4096 you quoted. I have never actually seen such a high value in an LS2 table before (see http://ls1-australia.com/forum/viewtopic.php?t=3904 as an example). Is this the maximum value GM will permit in the table but not actually used? The maximum I have seen using other tools on my LS2 is ~ 2530 (gm*K/kPa I think). This would explain the low VE% in your table & is a trap if people start thinking 4096 corresponds to 100% efficiency.

Cheers,

Michael

GMPX
November 23rd, 2006, 11:00 AM
O.K, see screen shot, this is the raw numbers, we will be happy to change them to show this instead of the theoretical VE%.

Yes PM box was full Acer, sorry about that.

Injector flow change is not an easy one, from what I can see it would involved a fair amount of code changing. 63.95lb/hr is the mathematical limit using the scaling GM have for the injectors.

Cheers,
Ross

GMPX
November 23rd, 2006, 12:04 PM
O.K, Paul got it figured out, how does this look now :) It was so obvious I feel rather silly :redface:

Cheers,
Ross

Black02SS
November 23rd, 2006, 12:22 PM
Ross, can you have settings in the tuning tool similar to how we have for LS1's for those that want multiple options? I don't care personally as long as there is something there to change. :lol:

Acer
November 23rd, 2006, 08:45 PM
Wonderful Ross - thanks very much for following it up:cheers: . Only other minor comment is that like all earlier versions of Flashscan I think you are a factor of 1000 out if the units are gm*K/kPa which I always used to compensate for in my calculations. I'll double check this & let you know if what I recollect is correct.

IMHO these units are enough in the VE table & I wouldn't worry about calcualting % VE.

Cheers,

Michael

Acer
November 23rd, 2006, 10:28 PM
Paul & Ross,

I have checked my earlier calcs & believe your VE values are a facor of 1000 out (should 1000 greater) to correspond with the units. Suggest you do your own check on the conversion of the values in the ECM & see what you come up with.

Thanks,

Michael

GMPX
November 24th, 2006, 12:21 AM
Ross, can you have settings in the tuning tool similar to how we have for LS1's for those that want multiple options? I don't care personally as long as there is something there to change. :lol:

Yes, Paul is adding that in for the next release.
Acer, g*k/kpa I think is correct (I'll happily be made a fool of again), if you were in mg*k/kpa then yes were are out.

Cheers,
Ross

oztracktuning
November 24th, 2006, 01:44 AM
With LS1s , I prefer to switch between g/s or g/cylinder (since it then can be setup to correlate with the timing tables). But i dont mind just changing a number by % errors collected in a chart.

Acer
November 26th, 2006, 09:20 AM
Ross,

Double checked everything again & your "raw" numbers are of the correct magnitude if in g*k/kPa:) . The reason why I was used to seeing numbers 1000 greater within the ECMs table is that GM appear to be working in mg*K/kPa to calculate the injector pulse width from the VE table. Are you intentionally converting the GM table to revised units?

Is the example of the table you provided (see link below) from a standard LS2 or a forced induction car as the values still seem on the high side?

http://forum.efilive.com/attachment.php?attachmentid=1098&stc=1&thumb=1&d=1164326640 (http://forum.efilive.com/attachment.php?attachmentid=1098&d=1164326640)

Have sent you another PM with adjusted factor to calculate the correct % VE.

Cheers,

Michael

redhardsupra
March 4th, 2007, 08:12 PM
some newer research done on how to work with GMVE, what does it really mean, how to tune it, etc...

LS2 airflow uncorked? (http://redhardsupra.blogspot.com/2007/03/ls2-airflow-uncorked.html)

Natural limits and bad weather (http://redhardsupra.blogspot.com/2007/03/natural-limits-and-bad-weather.html)

The Alchemist
May 8th, 2007, 07:32 PM
Hi there, has any one done the uncorking process and reported back to you yet with what happened?

regards,

Mike



some newer research done on how to work with GMVE, what does it really mean, how to tune it, etc...

LS2 airflow uncorked? (http://redhardsupra.blogspot.com/2007/03/ls2-airflow-uncorked.html)

Natural limits and bad weather (http://redhardsupra.blogspot.com/2007/03/natural-limits-and-bad-weather.html)

redhardsupra
May 9th, 2007, 04:09 AM
yes and no. i have the formula to calculate it, but it returns very noisy signal. basically the temperature model goes out the window because the BIAS table needs to be adjusted, and there's no way i know of to really arrive at the correct numbers in them. you solve the bias table, and we can get the full thing easily. any takers? i hear it's some heat transfer engineering thingy...

redhardsupra
May 12th, 2007, 05:16 AM
i got the BIAS formula from EFI's help files, 273.15+ECT+(ECT-IAT)*BIAS, but i cannot seem to match the scanned manifold temperature to the one calculated with the formula. what am i missing? where does the 'filter' value come into play?