Yet another theory on how to tune...
06-19-2007, 02:35 PM
#42
TECH Addict
Thread Starter
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by NicD
Ahh, I didn't see your ninja edit. Just wanted to make sure we are on the same page here.
06-19-2007, 02:38 PM
#43
TECH Enthusiast
Join Date: May 2007
Posts: 538
Likes: 0
Received 0 Likes
on
0 Posts
Waste O2 and Fuel add up quickly. The maf is fine the way the engine is using fuel is whats different. Look at your BSFC numbers. They add a bit of enlightenment to the subject manner.
Originally Posted by SSpdDmon
With my SLP MAF and SLP Lid, the AFR is off significantly (-9%~11%) at different points along the curve. Currently, it's set as follows:
Idle Frequencies: Had to add ~3% to keep AFR at 14.5~15:1.
Low TPS (3250~4250Hz): Had to pull 6~8% out of the curve to keep AFR at 14.5~15:1 or it would run too rich.
Moderate TPS (4500~6500Hz): Varies from -4%~4% to maintain stoich AFR.
Heavy/WOT TPS (>6500Hz): Varies from 4%~11% to get commanded AFR to equal WBO2 AFR.
That's a 20% variance! Any reason why?
Idle Frequencies: Had to add ~3% to keep AFR at 14.5~15:1.
Low TPS (3250~4250Hz): Had to pull 6~8% out of the curve to keep AFR at 14.5~15:1 or it would run too rich.
Moderate TPS (4500~6500Hz): Varies from -4%~4% to maintain stoich AFR.
Heavy/WOT TPS (>6500Hz): Varies from 4%~11% to get commanded AFR to equal WBO2 AFR.
That's a 20% variance! Any reason why?
06-19-2007, 02:45 PM
#44
TECH Addict
Thread Starter
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by LS1curious
Waste O2 and Fuel add up quickly. The maf is fine the way the engine is using fuel is whats different. Look at your BSFC numbers. They add a bit of enlightenment to the subject manner.
06-19-2007, 04:51 PM
#47
TECH Addict
iTrader: (3)
Join Date: Aug 2003
Location: IL
Posts: 2,561
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by SSpdDmon
Yeah...that's new to (beyond) me.
It's a measure of efficiency in an engine. This information routinely provided to you if you have your engine tested on an engine dyno.
It is simply the amount of fuel your engine uses in lbs for every "horsepower" it makes.
Cut and pastes from various sites:
To calculate BSFC, use the formula BSFC = Fuel_rate / Power
Where:
Fuel_rate is the fuel consumption in grams per hour (g/hr)
Power is the power produced in Kilowatts where kW = w * Tq / 9549.27
w is the engine speed in rpm
Tq is the engine torque in newton meters (N·m)
Note: The Power in the BSFC calculation is not weather corrected.
BSFC
This is the ratio of the engine fuel consumption to the engine power output as measured at the flywheel. BSFC has units of grams of fuel per kilowatt-hour (g/kWh) or pounds mass of fuel per brake horsepower-hour (lb/bhp-hr). BSFC is a measure of engine efficiency.
06-20-2007, 12:28 PM
#48
I thought this thread will turn ugly 
I fully understand that the IDC will be off when you change the IFR table. Also PE and VE with change IDC.
For the end result of how a car runs I've seen no proof of any kind why you can not use the IFR to tune in the fuel trims. Now ussing the MAF should be limited and I will only go 10% max.Over that the drivability can suffer from what I've found.
Believe me,been tuning these since the first software come out and tried every new Fad that came out.I have 3 different ways I tune depending on vehicle setup. IFR table is still one of them and has been dead reliable.
I fully understand that the IDC will be off when you change the IFR table. Also PE and VE with change IDC.
For the end result of how a car runs I've seen no proof of any kind why you can not use the IFR to tune in the fuel trims. Now ussing the MAF should be limited and I will only go 10% max.Over that the drivability can suffer from what I've found.
Believe me,been tuning these since the first software come out and tried every new Fad that came out.I have 3 different ways I tune depending on vehicle setup. IFR table is still one of them and has been dead reliable.
06-20-2007, 08:02 PM
#49
TECH Addict
Thread Starter
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
For ***** and giggles, I set my MAF back to stock. So now, I have a stock MAF table, a stock VE table, and a relocated IAT sensor. I went and did an IFR tune this afternoon with my MAF threshold set to 400rpm still (MAF only full time). After a few flashes, I'd say it worked well enough to convince me that this way is still valid. I think the keys to what worked for me are threefold:
1) I dropped the threshold to eliminate the VE's influence. When I tried this before (other thread that was around a week or so back), I didn't do this and the results were rather shitty.
2) My IAT has been relocated to a much more accurate spot. No more heatsoak retained by the lid skewing the IAT readings.
3) I applied 50% of my correction factor to avoid overshoot. Example: If commanded was 10:1 and the WB read 9:1, 9/10=90% or reduce fueling by 10%. I would only reduce it by 50% of the correction (10*0.5=5.0).
After only a few flashes, I had part throttle and WOT dialed in. My BENs are fairly close to 1.0 (-3%~1.5% variance). It definitely was a lot easier to work with one column instead of the 20 or so in the VE table.
Take it for what it's worth...not saying this is any more right or wrong than VE tuning. But, it sure seemed to be a little faster for me. I'll keep an eye the rest of the week to see how it does day to day and report back.
1) I dropped the threshold to eliminate the VE's influence. When I tried this before (other thread that was around a week or so back), I didn't do this and the results were rather shitty.
2) My IAT has been relocated to a much more accurate spot. No more heatsoak retained by the lid skewing the IAT readings.
3) I applied 50% of my correction factor to avoid overshoot. Example: If commanded was 10:1 and the WB read 9:1, 9/10=90% or reduce fueling by 10%. I would only reduce it by 50% of the correction (10*0.5=5.0).
After only a few flashes, I had part throttle and WOT dialed in. My BENs are fairly close to 1.0 (-3%~1.5% variance). It definitely was a lot easier to work with one column instead of the 20 or so in the VE table.
Take it for what it's worth...not saying this is any more right or wrong than VE tuning. But, it sure seemed to be a little faster for me. I'll keep an eye the rest of the week to see how it does day to day and report back.
Last edited by SSpdDmon; 06-21-2007 at 08:32 AM.
06-21-2007, 11:24 AM
#51
Originally Posted by Slowhawk
I thought this thread will turn ugly
well, if everyone's whipping out their mad ideas, slowhawk, SSpdDmon, and whoever else that gets it, have you seen the 'complex math' thread in the advanced section, and my Bias Table adjustment adventures? any comments?
06-21-2007, 12:32 PM
#54
TECH Addict
Thread Starter
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by RedHardSupra
06-21-2007, 12:36 PM
#55
my last post in that thread demos how i did it for a small range of cells. i'm working on automating it but right now it's kinda manual.
just IM me or something, i'll tell you what to log and we'll do an experiement on yours, you're a prime candidate since you moved your IAT.
this is what i've been trying to say the whole time, i understand you notice inconsistencies and trying to adjust for them with IFR, but i don think that's the culprit, BIAS is.
just IM me or something, i'll tell you what to log and we'll do an experiement on yours, you're a prime candidate since you moved your IAT.
this is what i've been trying to say the whole time, i understand you notice inconsistencies and trying to adjust for them with IFR, but i don think that's the culprit, BIAS is.
06-21-2007, 01:22 PM
#56
TECH Addict
Thread Starter
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by RedHardSupra
my last post in that thread demos how i did it for a small range of cells. i'm working on automating it but right now it's kinda manual.
just IM me or something, i'll tell you what to log and we'll do an experiement on yours, you're a prime candidate since you moved your IAT.
this is what i've been trying to say the whole time, i understand you notice inconsistencies and trying to adjust for them with IFR, but i don think that's the culprit, BIAS is.
just IM me or something, i'll tell you what to log and we'll do an experiement on yours, you're a prime candidate since you moved your IAT.
this is what i've been trying to say the whole time, i understand you notice inconsistencies and trying to adjust for them with IFR, but i don think that's the culprit, BIAS is.
If you look at it from a general perspective (without all the mathematical equations), the purpose of B4901 is to 'guess-timate' charge temp. So, what are the factors we're dealing with? Engine bay temps and heat radiation (if that's even a correct term) from heat soak along the intake tract (lid, MAF, TB, intake, and heads). The only two souces of measurement in the LS1 PCM's are ECT and IAT. If one or both readings are inaccurate (e.g. IAT - lid subject to heatsoak), then you're fudging things. You can relocate the IAT to a more accurate location like I did (within a range of maybe ~5*F error). But, you can't do that with the ECT.
On a side note - Are the heads really only 180*F like the coolant or is the coolant really a lower temp when entering the head and receiving only a portion of heat transfer by the time it hits the temp sensor? Intake runner/head temps could very well be more than the 180*F the coolant is.
Anyway, back on topic. Assuming my IAT is accurate, the airmass starting at that temperature has to travel a certain distance over a period of time. We measure that at a fixed point (the MAF) in grams per second, right? During the course of it's travel through the intake tract, the temperature of that air is going to be influeced by the warm~hot surfaces directing it to the cylinders.
The 'simple' questions are:
(1) How long does it take the air to make it through to the cylinders? --OR maybe better stated as-- How long does it take for the volume of air in the intake tract to be cleared through to the cylinders at certain airflow rates (like when a full bath tub drains to empty)? Volume is fixed here...should be an easy calculation.
AND
(2) How much post IAT heat is really transfered to that air at each of the airflow rates given in the charge temp blending table? In other words, isn't there already an equation for temperature transfer over time? If so and we can figure out #1, we should be able to figure out #2.
Probably over simplified there.....but, if we can determine that, then we can figure out the true charge temp, use an accurate IAT sensor temp, and back fill the factor based off an accurate ECT sensor temp.
Last edited by SSpdDmon; 06-21-2007 at 01:36 PM.
06-21-2007, 01:58 PM
#57
you totally missed the point....
1. i'm creating a general model for fueling that can work with our sensors and tables we got. bias is just step 1, it's kinda hard so i spent some time on it, i'm not attributing complete responsibility for all errors we see to it, that'd be ridiculous
2. there's a bunch of other things we gotta model before it will be all done. IFR is a hugely important one, but at we dont even know how short pulse adders and offset really work. a lot of work to be done there, ultimately we're getting the model better, piece by piece.
3. i know we're dealing with imprecisions, that's why we got a general model for it, and that's why i'm optimizing it in such a way that it fits all cases, not just some. the fact that the temp of aircharge wont be ECT or IAT by the time it gets to the cylinder is irrelevant, that's why i compare my estimate to the temp it would've been to create the fueling observed.
1. i'm creating a general model for fueling that can work with our sensors and tables we got. bias is just step 1, it's kinda hard so i spent some time on it, i'm not attributing complete responsibility for all errors we see to it, that'd be ridiculous
2. there's a bunch of other things we gotta model before it will be all done. IFR is a hugely important one, but at we dont even know how short pulse adders and offset really work. a lot of work to be done there, ultimately we're getting the model better, piece by piece.
3. i know we're dealing with imprecisions, that's why we got a general model for it, and that's why i'm optimizing it in such a way that it fits all cases, not just some. the fact that the temp of aircharge wont be ECT or IAT by the time it gets to the cylinder is irrelevant, that's why i compare my estimate to the temp it would've been to create the fueling observed.
06-21-2007, 02:15 PM
#58
TECH Addict
Thread Starter
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by RedHardSupra
you totally missed the point....
1. i'm creating a general model for fueling that can work with our sensors and tables we got. bias is just step 1, it's kinda hard so i spent some time on it, i'm not attributing complete responsibility for all errors we see to it, that'd be ridiculous
2. there's a bunch of other things we gotta model before it will be all done. IFR is a hugely important one, but at we dont even know how short pulse adders and offset really work. a lot of work to be done there, ultimately we're getting the model better, piece by piece.
3. i know we're dealing with imprecisions, that's why we got a general model for it, and that's why i'm optimizing it in such a way that it fits all cases, not just some. the fact that the temp of aircharge wont be ECT or IAT by the time it gets to the cylinder is irrelevant, that's why i compare my estimate to the temp it would've been to create the fueling observed.
1. i'm creating a general model for fueling that can work with our sensors and tables we got. bias is just step 1, it's kinda hard so i spent some time on it, i'm not attributing complete responsibility for all errors we see to it, that'd be ridiculous
2. there's a bunch of other things we gotta model before it will be all done. IFR is a hugely important one, but at we dont even know how short pulse adders and offset really work. a lot of work to be done there, ultimately we're getting the model better, piece by piece.
3. i know we're dealing with imprecisions, that's why we got a general model for it, and that's why i'm optimizing it in such a way that it fits all cases, not just some. the fact that the temp of aircharge wont be ECT or IAT by the time it gets to the cylinder is irrelevant, that's why i compare my estimate to the temp it would've been to create the fueling observed.
Last edited by SSpdDmon; 06-21-2007 at 02:31 PM.
06-21-2007, 05:32 PM
#59
well in that case congratulations on joining ranks of all the 'pro' tuners out there who decided that the most important metric measuring the quality of tuning methodologies is the time needed to cash in.
i still see no model, no evaluation methods, no proof that what you're actually doing is a cause, and not a sideeffect of something else.
is there an icon for 'disappointed?'
i still see no model, no evaluation methods, no proof that what you're actually doing is a cause, and not a sideeffect of something else.
is there an icon for 'disappointed?'
06-21-2007, 06:22 PM
#60
TECH Resident
iTrader: (17)
Join Date: Sep 2004
Location: NC - Charlotte area
Posts: 3,747
Likes: 0
Received 0 Likes
on
0 Posts
.....there needs to be a dissapointed icon. i'm a fan of Williams performance..he knows his ****. bottom line. i don't pay attention to a LOT on here...i read and do my thing...what works for me works.
there are just toooo many ways to skin a cat...what way is right?
there are just toooo many ways to skin a cat...what way is right?