An experiment using ScanGuage

[paulbates]

I've reached the point where I've chased my in-dash instant MPG gauge as far as I can and its time for a scanguage. I will be getting one in the next few weeks. I've read the manuals and get the basic operation of it. After I get used to using it, I have a first job for it, determining the most efficient way to accelerate from stop to 43 MPH.

Two things at play here: Most of my starts are of this type and 43 is where lock settles in and MPG is highest, and more than anything else, these starts crater my MPG. I learned with my I-4 Milan that revving only to 2000 rpm actually made things worse than revving to 2500 - 2600 RPM. By ~2 MPG average. Meaning that there is a similar sweet spot for every vehicle and it has to be found. But without measurements it was hard find without lots of driving.

Here is my idea for an experiment and I'm interested to hear from those that have tried something similar or have some experience that would help me.

1. I have a an area I can use for testing that's 1/2 mile long, flat and straight
2. I'll perform 0 - 43mph starts over the same area at different accelerations and RPMs
3. Measure fuel used over each of 5 reps each of each type of acceleration

For #2, I'd like to use a gauge to try and keep each run true and repeatable. Any ideas on a gauge and how to use it? While I know not all gauges for a particular interface (VPW for me) work in all cars, I listed the set of stock scanguage + VPW gauges to try on my Montana. I picked and ordered these from what I think might be best

Throttle Position Probably best for keeping my foot consistent
Engine Load Popular for DWL
Engine Torque 3.4L V6 has a broad flat torque curve
HorsePower Will go up with rev (?)
Fuel Rate What I'm trying to fix but probably moves
A/F Ratio ?
Manifold Pressure ?
Ignition Timing will change with rev
??

[SentraSE-R]

Paul,

The problem with ATs is torque converter slippage. You'll need to watch parameters such as engine load and/or throttle position sensor position while also making sure your TC doesn't unload and shoot your rpms from 1800 to 2800 during a test. Otherwise, you'll introduce a bunch of other variables besides acceleration rate.

Should you choose to test acceleration with constant LOD or constant TPS, letting the TC do its own thing, then you should also note rpm ramp-ups during each test. It's possible, but unlikely, that TC slippage to 3000 rpm in second or third gear uses less fuel than steady acceleration below 2000 rpm in third gear.

I like your plan of doing five replications of each parameter change. If you have a tachometer in your dash, I'd monitor TPS, LOD, AVG, and MPG. At the end of each test, I'd write down the maximum RPM recorded in the current trip data. If you buy an UltraGauga instead, I'd use the UG equivalents, plus average G/H

[PaleMelanesian]

I'll throw in my experience with the Odyssey 3.5V6 5AT. I've not scientifically tested this, but I've driven it enough to be confident that this is the best for this vehicle.

I try to keep it under 2,000 rpm, or possibly touching 2100 if it's an uphill takeoff. Ideally I hold it at 1800 or so for accelerating and then 1500 at cruise, but the trans will run up to 2000 before shifting. I use the highest load I can while still keeping within the rpm parameters.

[paulbates]

Thank you both. I'm open to any and all ideas to make this work.

Should you choose to test acceleration with constant LOD or constant Tps , letting the TC do its own thing, then you should also note rpm ramp-ups during each test. It's possible, but unlikely, that TC slippage to 3000 rpm in second or third gear uses less fuel than steady acceleration below 2000 rpm in third gear.

That's very interesting. I was going to try going up as high as 3000 but didn't have a good reason or understanding of why. And to your point nothing says that a flat TP is the right thing..I understand the really basic principles of AT but not at this level. SG should give me targets at different points to keep it consistent and fuel usage at the end to identify progress. Having said that, it took two weeks of DWL practice before my right leg quit cramping

I try to keep it under 2,000 rpm, or possibly touching 2100 if it's an uphill takeoff. Ideally I hold it at 1800 or so for accelerating and then 1500 at cruise, but the trans will run up to 2000 before shifting. I use the highest load I can while still keeping within the rpm parameters.

I have a 4 auto and 5400lbs so it tends to be higher rpms then that without the acceleration being glacial, but but this is easy enough to test. With the SG LOD can now be measured

Another question, what is being targeted for LOD? Like instant MPG keep it as high as possible?

In other news the SG showed up on the doorstep today. But in the spirit of Christmas gifts it will have chill under the tree..

Paul

[MaxxMPG]

Converter slippage is minimal above 2000rpm in today's transmissions. The real zone you're looking for would be found on a BSFC chart for the engine. And I'm not sure if there are any available online for the 3400 engine.

The transaxle does have a "Electronic Capacity Controlled Clutch" - GM-speak for a converter that can partially lock. So you will find that the converter clutch isn't fully locked until a few mph above that minimum engagement speed. I use the tach to see when the converter is fully locked and when it is allowing slippage. On my car, I can get down to 37mph in overdrive with the converter locked, but if I increase speed to 40-41mph, the tach goes down slightly as speed goes up. My 3500V6 is an evolutionary update of your 3400, and I find that I get better fuel economy on the highway at 55-60mph (1600-1700rpm) than I do at 45mph (1300rpm) or 40mph (1100rpm). Can't know for sure about 2000rpm, because air drag would cancel out any increase in FE, but it appears that the engine is more efficient at speeds closer to 2000rpm than 1000-1200rpm.

[PaleMelanesian]

I would focus on LOD rather than tps. LOD (or rather, vacuum which is directly proportional) is what drives the shifting logic on my van.

Also do a test with an average lod, but letting up at certain points to encourage an upshift. I expect that will give better results than a constant lod.

[paulbates]

Ok, thanks and thanks again. I think based on what i am hearing I will have to set up several different ways to approach it and see which one gives the most consistent measurements and results.

Chris, couldn't find the BSFC as you predicted, but I can capture the information with the scan guage and make one. Just not in the units typically used. And it would be based on scanguage readouts but should be accurate relative to each other. (?)

I'm wondering what that will get me? make a graph and chart the peaks?

And I found this, the GM power curve for the LA1. Not exactly my version but close enough. The question is, given tourqe maxing out closer to 3000, how much might that play in, as mentioned by Darrell above?

http://www.v6z24.com/specs/3400specs




This guy put a turbo on the LA1. I don't want to do that, but he provides a number of details about how the LA1 is set up, valve behavior ect. I've captured the important bits here:

• GM LA1 engine is already pretty hefty as far as bottom-end components are concerned. It comes factory with a forged, internally-balanced crank and forged connecting rods. With the stock cr at 9.5:1
• LA1 is the stock camshaft. I do not have the exact factory specs, but for all practical purposes, it should have overlap to benefit the NA engine
• The stock 3400 injectors are set at 22.5 lb/hr

I found this info on Chevy and Pontiac Vans from 2000 - 2004,

First Gear Ratio (:1) 2.92
Second Gear Ratio (:1) 1.57
Third Gear Ratio (:1) 1.00
Fourth Gear Ratio (:1) 0.71

Final Drive Axle Ratio (:1) 3.29

[MaxxMPG]

In simplest terms, the torque peak is where the engine is most "efficient" at turning gasoline energy into kinetic energy. For the LA1, that appears to be 4000rpm, but you can see it's within 10 lbft from 2200 to 5000, and within 10% of peak torque from about 1300rpm up to near redline. I see a slight rise in the torque curve from 1800 to 2300, so I suspect that the optimal shift point will be somewhere in that range.

When evaluating torque curve, keep in mind that the 'best' shift point may vary based on which gear selection is involved.
1st-2nd - is 2.92:1 down to 1.57:1 - so tach will drop to 53.76% of what it was before the shift. So forcing a 1->2 at 1800rpm drops the engine to 970rpm. In the real world, it doesn't drop that low because the converter isn't tight at that engine speed, but it shows that a shift at 2200rpm, dropping to 1200rpm in 2nd, puts you on the low end of peak torque. This also reveals why the transaxle is so stubborn about completing the 1-2 upshift below about 1600rpm. It's hard coded in there to keep the transmission from shifting at a speed so low that the engine is idling when 2nd applies.
2nd-3rd - is 1.57:1 down to 1.00:1 - so tach will drop to 63.69% of what it was before the shift. Now, that 1800rpm shift is no longer a liability because it drops the engine down to 1150rpm (neglecting converter slippage).

Your strategy of testing on a lonely road with different shift points is a sound idea, but that "State of Nebraska" torque curve makes me think the variation is going to be slight. So while recording the numbers, you may want to keep all those decimal places to track any trends or conclusive evidence.

The gear ratios listed are for the 4T65E. It's a pretty sound design, and it has less than average internal hp loss relative to competing 4ATs.
Here's a pdf file with more information than you ever wanted to know about your transmission - www.transtec.com/tech_insert/IS-4T65E-5.pdf
The bottom of page 1 shows which clutches/bands are applied in different driving ranges. The 4T65E shows its age with its design using three bands. Look carefully and you will see that shifting from P to R does nothing more than apply the reverse band. Shifting from R to D releases the reverse band and applies the forward band - no clutches apply or release.

One area to explore with the SGii is fuel cut. Your van should have it, but you usually have to downshift from (D) to D and then lower to get to it and stay in it.
I use this strategy to slow down on highway exits, feeling for the fuel cut as variations in engine braking. From a highway speed of 45, I will skip 3rd and go straight to 2nd. Within a second or two, I feel much more engine drag and then it seems to let up a bit when the engine speed drops to about 1500. Then, I drop to 1st and the cycle repeats, with the engine braking starting to let up at 1600-1700rpm, and then a sort of 'bump' at around 1200-1400rpm (seems to vary) as the injectors return to always-on. Because of the loss of speed, I only use fuel cut when loss of speed is necessary, as in steep downhill grades or shorter highway exit ramps that require heading into them 'hot' but ending with a red light, stop sign, or very sharp curve at the top.

[paulbates]

Thanks Chris, those are all great thoughts and help. I can say confidently that I have learned more about my car from you here and when you welcomed me to the board. You introduced me to the fuel cut (an non-flat towable limitation) then. I use the instant mpg to tell when I am in fuel cut, I am not able to hear it.

I am going to peek at that pdf now.

While possible to set up random tests to see which works best, it helps me to understand some the complexities involve.

I'm going to think about it for a few days and then put my test down on a collection spreadsheet. I'll put it back here for comments.

I confess that this whole thing is dripping with geekyness and proves that I need more to do. But I appreciate. I will keep this van maybe 80 - 100K more miles and getting this right will be worth it.


Paul

[CarlD]

This is the best technical article I have found on fuel economy, and your tax dollars paid for it:

www.epa.gov/oms/models/ngm/420p05001.pdf

This link shows the coastdown coefficients for all vehicles for any given year to use in the equations:

http://www.epa.gov/otaq/crttst.htm

As an example, the Prius is well documented and its minimum BSFC point is about 18kW @ 2100RPM. This is a plot of accelerations assuming constant engine operating points of 10kW @1600RPM and 18kW @2200RPM (70% of engine power makes it to the wheels.) In this case it would burn less fuel to accelerate at the higher power level. For the higher power level, the BSFC is 190 g/kWh or 0.95grams/s. For the lower power, the BSFC is 238 g/kWh or 0.66 grams/s. Higher power takes about 19.2s to reach 40MPH, the lower power takes about 43.2s. So we have 18.2 grams of fuel vs. 28.6 grams. In reality constant power is not best but rather a gradually increasing power level gives the best FE, if you can keep the engine near min BSFC as you accelerate. Most vehicles won't let you do this as they want some reserve torque in case you decide to mash the go pedal.