What's better on hills: DWL or Pulse up, Glide down?

[monroe74]

duff: "It is amazing when talking to laypeople (I’m not calling anyone here that)"

I'm definitely a layperson, so feel free to call me that!

"there may be no appreciable air increase in the last 10 degrees and this means no pumping losses saved either"

I understand why the last 10 degrees would have a small effect. But I don't understand why the effect would be considered zero, or negative. In other words, I don't know why 80% would be considered preferable to 100%, if one is trying to minimize pumping losses and optimize BSFC.

I think the traditional answer is that we want to avoid inducing open-loop mode. But I have a feeling that a system with a wideband sensor doesn't suffer from that problem, or at least not as much.

[monroe74]

diff: "My Jeep makes all of it's torque at 3k and under, peaks at 3k and drops like a rock over 3200k. Redline at 5k."

That sounded fishy to me, so I looked it up:

http://www.edmunds.com/used/1995/jee...364/specs.html

Edmunds thinks your torque peaks at 4k, not 3k. And the redline is 4750.

"I can still engine brake once up to the right speed, saving even more energy."

Engine braking doesn't save energy, in this scenario. If you build up so much excess speed going downhill that you then have to use engine braking, you've wasted fuel.

"if I pedaled as hard as I could up the hill, and then coasted down"

It doesn't matter when you add momentum to the vehicle. Either way, you're converting chemical energy into other forms (kinetic energy, to the extent that you add speed, and potential energy, to the extent that you add elevation). The only difference between the two strategies is that in your strategy, you create a large differential between downhill speed and uphill speed. The result is that you find yourself going too fast, at the bottom of the hill. A sign of this is that you need to use engine braking. When you do this, you're converting valuable momentum into waste heat inside the engine. That momentum wasn't free; you paid for it. But instead of getting to use it, you're throwing it away.

You could avoid this need to slow down if you use a strategy that doesn't create excess speed in the first place. The result would be the same average speed, but with lower fuel consumption.

[SpartyBrutus]

Probably good to describe the boundary conditions and constraints when trying to max FE over hills: min/max speed at start base/ crest / and end base.

Seems to me that if the mins are say 50mph/50/70 versus 50/10/30 versus 50/30/50 then you may have different strategies to climb the hill. The 50/10/30 case may yield the best FE but of course only averages ~30 mph...

I tend to let my speed drop up to the amount that it will gain on the downslope. To me a straightforward way to determine proper throttle going up would be to run (repeated) trials at various rpms/shift points/throttle position (Xs) and see impact on whats important: ave speed, FE (Ys).

My 2 cent.

[Duffman]

Originally Posted by monroe74

diff: "My Jeep makes all of it's torque at 3k and under, peaks at 3k and drops like a rock over 3200k. Redline at 5k."

That sounded fishy to me, so I looked it up:

http://www.edmunds.com/used/1995/jee...364/specs.html

Edmunds thinks your torque peaks at 4k, not 3k. And the redline is 4750.

I think Edmunds is wrong, I dont think the suburban was over 4 either but your honda probably is. Most domestics (6 cyl and bigger at least) tend to be well below 4000, and trucks will be very much below that.

[Duffman]

Originally Posted by monroe74

I understand why the last 10 degrees would have a small effect. But I don't understand why the effect would be considered zero, or negative. In other words, I don't know why 80% would be considered preferable to 100%, if one is trying to minimize pumping losses and optimize BSFC.

Not zero and definately not negative but too small to matter. Pulling numbers out of my ass here but if you go from 100 CFM at 80 degrees and 102 at 90 degrees you have had a 2% increase in airflow but at the same time your A/F goes from 14.7 to 13.5 it has become 8.9% richer. 1.02x1.089= 11% more fuel. This jives with what you are telling us, but you have only maximised airflow by 2% which is really negligible and overall your FE will be negative becuase you are using 11% more fuel.

Thats why I put more value in the A/F ratio, the fuel injected may infact be very close to linear with pedal position but the airflow wont be throughout the whole range.

[monroe74]

"I think Edmunds is wrong"

If Edmunds is wrong, then Jeep is also wrong. They report the same data as Edmunds. Peak torque is at 4000 rpm. See here: http://www.jeep.com/en/2008/grand_cherokee/. (I'd provide a more direct link, but they don't make it easy to do that. Just click on Features & Specs, and then Technical Data.)

"I dont think the suburban was over 4 either"

The Suburban currently has 5 engine choices, all V-8. The torque peaks range from 4200-4400. See here: http://www.chevrolet.com/suburban/specifications/.

"Most domestics (6 cyl and bigger at least) tend to be well below 4000, and trucks will be very much below that."

I'll buy you a gallon of gas (or beer, your choice) if you can prove that claim.

To be fair, I do notice that Ford has a couple of examples tending in that direction. I see a V-8 with peak torque at 3750, and the V-10 peaks at 3250. See here: http://www.fordf150.net/2007/2007-fo...ifications.php

In the Chevy Silverado line (http://www.chevrolet.com/silverado/specifications/), there is a V-6 that peaks at 2800, but then there are 5 V-8s that peak from 4300-4800.

"Pulling numbers out of my ass here"

No comment.

[Duffman]

Seems to be significant RPM creep over the last few years. DifferentPointOfView's jeep it is a 95 model with likely an I6, I still firmly believe him when he says 3000. You are right on the suburban, I am very much out of touch with new vehicles and what I said applies better for vehicles from a decade ago:
http://media.gm.com/us/gm/en/news/events/pig_index.html

[JohnnyGrey]

What did you do to it?

I used the wideband's 0-5v output together with a custom circuit to alter the narrowband switchpoint anywhere from 12:1 to 17:1.

I also wonder if an engine with a wideband sensor would have more of tendency to behave this way (i.e., be more linear than one would expect).

That wouldn't make a difference. Ignoring fuel completely, your engine's torque curve is a product of how much AIR it can suck in at WOT at a given RPM. The reason torque falls off at high RPM is because at those speeds, your intake valves aren't open long enough to fully charge the cylinder. The reason you may have more torque at 3k than at 2k is because of the resonant effects of the intake and exhaust systems.

[Duffman]

Originally Posted by JohnnyGrey

I used the wideband's 0-5v output together with a custom circuit to alter the narrowband switchpoint anywhere from 12:1 to 17:1.

Did you build and design it all yourself or did you get a kit somewhere?

[JohnnyGrey]

Did you build and design it all yourself or did you get a kit somewhere?

I designed and built it myself. A basic version is essentially an NPN transistor surrounded by the right resistors, and will run you about $12 in radioshack parts. Mine cost about $25 because of the LEDs, and dual AFR switching.