Finer points of P&G

[daqcivic]

So I am new to employing P&G as a dedicated FE technique, and I want to work out the finer points. I've read up quite a bit on what people have to say about it, so I think I've done enough research to get into a discussion about it.

As I understand it, the goal of P&G is to maximize FE by maximizing coasting distance/time (not sure which is better to think in terms of) and minimizing BSFC while DWL. (I understand min BSFC to be the RPM and load level at which the engine uses the least amount of fuel for the power being produced.) To put it another way, to get the best aggregate FE between coasting (EOC if possible) and pulsing (accelerating at the load and in the RPM range that will minimize BSFC, given the speed, power output, and gearing of the vehicle). It seems to have been very well demonstrated that this achieves significantly better FE than "egg-under-foot" acceleration and steady coasting for a given average speed and RPM, though the benefits are far greater for lower speed ranges.

The general wisdom seems to be that min BSFC (for the power output and speed/gearing of the vehicle) is achieved in low RPM ranges and with higher loads (however this is quantified, usually vacuum pressure, SCII readout, or pedal/throttle opening percentage). For my vehicle, a 92 Civic DX 5MT, since I cannot take advantage of a SCII and do not currently have a vacuum gauge installed, I have found PaleMelanesian's advice description of his P&G shifting and load scheme a helpful starting point, since he has a similar car, has refined his technique using a SCII, and describes his load in terms of throttle %. I have also found a helpful online speed/RPM in gear calculator for Hondas (Fatboy Raceworks » Graphical Transmission Gear Ratio Calculator), esp. since I have a different tire size than OEM. Here is my basic scheme (numbers are rounded) for for shift points and pulsing:

Gear beginning MPH/RPM ending MPH/RPM
1 0/idle 10/2000
2 10/1050 21/2150
3 21/1450 32/2200
4 32/1700 44/2300
5 44/1800 55/2200

This will no doubt be refined over time, but I think it's pretty close to ideal given real-world driving conditions. PaleMelanesian has also said he uses about a 50-60% load in lowest gears up to about 80% in higher gears (80% seems to be the consensus on the best general load to achieve min BSFC), so I am trying to emulate this as a starting point. (Forgive my lack of citation links, as I haven't figured out how to embed links yet.)

Aside from the way P&G goes against the grain of my intuitive sense, I have some questions and concerns about how to do it properly:

1. Should you keep the load steady through the whole speed/RPM range in a given gear, or should you gradually increase load through the range?

2. Should you peg the throttle at the given load level the split second the gear is engaged, or should you take a second or 2 to engage the gear smoothly and get up to the desired load? In the first case, the jerk of the engine would seem to waste fuel, but it would seem to be necessary to attain the desired load for the greatest percentage of the pulse time. In the second case, I figure smoother is usually better, but especially in shorter gears taking a second or more to smoothly get up to load means that ideal load is achieved for only a few seconds due to the rate of acceleration. In addition, in top gear there is very little acceleration at the low end of the RPM range regardless of the load level, so it would make sense not to use a higher load at such RPMs if no additional acceleration is achieved. My Civic takes forever to go from 45 to 50, but above that it gets much better.

3. Should you clutch-in and throttle-off the instant the high end of the speed range is achieved, or should you take 1/2 to 1 second to ease off the throttle before disengaging the clutch entirely? This is especially relevant to at least early to late 90s vehicles since lower emissions are achieved by having the computer let the revs hover a second after the throttle is let off suddenly. That extra second of revs could add up to a lot of wasted fuel, but if the throttle is decreased slowly in gear the effect would be reduced.

It is going to be a frustrating couple weeks before I can get through a full tank with P&G to see how well I'm doing with it, but I'd love some more discussion on the finer points so I can get the best start possible!

[mobilerik]

If you wanna get right to the "finer points" and aren't afraid of some high school math, take a look at this party-killer I posted a few months back.

The gist is that if you have a gauge that'll give you your instantaneous GPH, you can easily calculate your break-even pulse-to-glide ratio for different scenarios.

For those who want to see how to compute this:

* P&G Gas Used = PulseRate x PulseTime + GlideRate x GlideTime
* DWL Gas Used = DWLrate x DWLtime


Plug in your averages:
You'll need to choose a cycle time to compare, ex. 5 second pulse and 10 second glide = 15 total seconds.

* P&G gas used = 4.0 gph x 5 sec + .4 gph x 10 sec = 20 + 4 = 24
* DWL Gas Used = 1.3 gph x 15 sec = 19.5

So you can see that using a 5 to 10 pulse-to-glide ratio (or 1 to 2) is not fuel-efficient for me.

In the final analysis, all you're really aiming for is to use less gas in a certain amount of time than you would cruising. Since every car is different, the best approach IMO is to get a gauge and do the observation/calculation. Good luck!

[daqcivic]

Thanks, akashic, I had actually read that post earlier and found it helpful in recognizing how simple it is in principle to compare a P&G scheme to DWL.

Some way of measuring instantaneous fuel consumption would be wonderful, but I am not aware of a way to do that with an OBD-I vehicle. The next best thing seems to be a vacuum gauge, and I'm definitely considering getting one, but there isn't necessarily a precise relationship between manifold pressure fuel consumption.

The "finer points" I am most interested at this point are the numbered items in my post, that is, whether it is better to gradually build load during acceleration through a gear or to keep load steady the whole time, and whether it is better to give priority to smoothness or quickness during shifting, P>G, and G>P transitions.

What I'm trying to get at is, given that you've learned the ideal acceleration and P&G schemes for best FE, how do you refine your technique further?

Anyone out there with a 92-95 Civic DX/LX who's got P&G nailed?

[mobilerik]

Originally Posted by daqcivic

Some way of measuring instantaneous fuel consumption would be wonderful, but I am not aware of a way to do that with an OBD-I vehicle. The next best thing seems to be a vacuum gauge, and I'm definitely considering getting one, but there isn't necessarily a precise relationship between manifold pressure fuel consumption.

The "finer points" I am most interested at this point are the numbered items in my post, that is, whether it is better to gradually build load during acceleration through a gear or to keep load steady the whole time, and whether it is better to give priority to smoothness or quickness during shifting, P>G, and G>P transitions.

What I'm trying to get at is, given that you've learned the ideal acceleration and P&G schemes for best FE, how do you refine your technique further?

Anyone out there with a 92-95 Civic DX/LX who's got P&G nailed?

If you're interested splitting hairs, you need to be able to see them. The mpguino is what you want. It's for pre-OBDii cars, and it's all over this site. Without one, you're like a competitive sprinter without a stopwatch or a standardized track. You need the instantaneous feedback. Right now, even if your tank miles are going up since you started P&G, you can't conclusively know that it was the P&G that helped. One reputable ultra-hypermiling member with a Civic, after installing his mpguino, reported that he got WORSE mpgs with P&G. You really need the feedback to sort any of this out.

[daqcivic]

I get what you're saying, I'm not sure I can figure out the mpguino install and programming. I don't even know how to read electronic schematics or tap into a wire.

[MetroMPG]

daqcivic: if I'm not mistaken, you can now buy a pre-assembled MPGuino programmed with the latest code. You'd still need to install it, but I suspect you could also get help with that from the members in the MPGuino subforum.

And to split hairs as precisely as you want, it seems like you'll need a BSFC map for your engine.

[PaleMelanesian]

You rang?

What I'm talking about is lower THROTTLE in the low gears and at low rpm, and higher throttle at higher rpm. At low rpm, it takes less to fill the cylinders, since the valves are open longer. (lower rpm = longer time for each cycle) Watching the MAP and LOD gauges on the SG, I aim for ~50% throttle, building up to 80% as the rpms climb. If you're in the gear for more than a few seconds, definitely taper it (heavier) as you gain speed.

Gears:
Your shifting scheme looks pretty good. You can go a little lower on the rpm, shifting a little sooner. If it growls/rumbles a bit, that's fine. As long as it can gain speed it's ok.

I hardly use 1st gear. On a downhill start, I don't - I go straight to 2nd. On flat/uphill starts, I only use it to get rolling, and shift to 2nd about as soon as it's fully engaged. (3mph or so?)

The transitions between pulse and glide are a small part of the whole, but when you do them so much, I guess it can make a difference. If you can minimize the number of P&G cycles, they matter less. I finish my pulse, clutch in, and 1/2 sec later key off, right as the revs reach the lowest point. Even the Scangauge can't measure at a fine enough resolution to show the best method here. My reasoning is that this way I have the shortest time in the key-off position, since the low-rpm engine dies that much sooner.

For the pulse, I do recommend getting a vacuum gauge. I've seen that ~80% load on the Scangauge is about 12psi MAP, so you can aim for that.

If your rpm is low enough, shifting with both quickness and smoothness is really really easy.

As for BSFC charts, the colorful one in this article seems to be good enough for my purposes. Just aim for the red zone. Article: Browser Warning

Overall, I find that getting the very best pulse less important than getting the best glide. You can do a 1:1 pulse to glide ratio, with a light pulse, or you can do a heavy pulse with a glide 3x as long. Both will give you similar results. Gauge the terrain, traffic, and other conditions to get the best glide you can. If that means pulsing on an uphill, do that. You can then coast up and over the peak and down the other side. You have to fight the urge to "do something", and just wait it out.


Ok, rambling done. Hopefully some is useful.

[daqcivic]

What I'm talking about is lower THROTTLE in the low gears and at low rpm, and higher throttle at higher rpm. At low rpm, it takes less to fill the cylinders, since the valves are open longer.

That makes perfect sense. I was a little confused from an earlier post where you listed a throttle % according to gear, not RPM; but of course, the engine doesn't care what gear it's in or even really what speed the vehicle is at, just what RPM it's at and what load is being demanded based on the throttle. So if there is enough time spend in gear I should begin initially with about 50% and increase up to about 80%. However, for my car's gearing, since RPMs at initial engagement speed in 2nd and 3rd are below the rule-of-thumb optimal BSFC range of about 1500-2200RPM, (around 800-1000 in 2nd and 1200-1400 in 3rd), it would seem best to begin in these gears with less than 50% throttle, until the RPM gets up to at least 1500.

Your shifting scheme looks pretty good. You can go a little lower on the rpm, shifting a little sooner. If it growls/rumbles a bit, that's fine. As long as it can gain speed it's ok.

This is something I definitely need a gauge for. Going lower seems like it wastes fuel because I end up spending a couple more seconds at those lower RPMs with almost no acceleration. Then again, Lower shift points mean fewer overall cycles, so it would take some pretty precise testing to with good instrumentation to figure out where the optimal point is. On the upside, a few hundred RPMs difference in shifting points is probably going to have a minuscule effect on FE if the load is managed smartly.

I hardly use 1st gear. On a downhill start, I don't - I go straight to 2nd. On flat/uphill starts, I only use it to get rolling, and shift to 2nd about as soon as it's fully engaged. (3mph or so?)

I hear most people saying this, and I don't necessarily disagree (as if I could provide any solid evidence to the contrary), but it does seem to contradict the BSFC logic of using the engine in its most efficient range because 2nd gear is used for several hundred RPMs before 1500. I'm going to accept the prevailing wisdom of this for now, though I hate lugging the engine.

Overall, I find that getting the very best pulse less important than getting the best glide. You can do a 1:1 pulse to glide ratio, with a light pulse, or you can do a heavy pulse with a glide 3x as long. Both will give you similar results. . . . get the best glide you can.

This is really helpful! I had gotten good at maximizing coasting using very slow, light load acceleration, but it is quite a bit different and more difficult to time things (esp. in rush hour city traffic with hills and such) using heavy load acceleration. But if I know I can freely adapt the P&G ratio to fit the circumstances to get the most gliding I will probably do a lot better than trying to stick to a rigid pulse scheme.

It's definitely starting to sink in that I can't really refine things much and know if it's helping unless I get some instrumentation. Thanks MetroMPG for the advice on mpguino, I imagine I could find out where those few leads are by looking it up in a manual.

And thanks to all for your advice and willingness to help a clueless beginner!

[PaleMelanesian]

The uphill pulse / downhill coast thing helps in traffic. You keep a fairly constant speed, building potential energy instead of kinetic during the pulse.

Short of getting some instrumentation, you need a reliable/consistent fillup routine. Same station, same pump, same time of day, etc.

[MetroMPG]

Also, since you willingly label yourself a "clueless beginner", please take the time to develop your technique in a quiet environment before unleashing it in traffic!

(And I'll assume you're aware of what happens to your power steering and vacuum assisted brakes when driving with the engine off.)

Driving safely takes enough effort on its own. Throwing a P&G learning curve into the mix can divert a lot of attention from other tasks. Later, when it becomes an "automatic" habit, that's less of an issue.