Minimal Best MPG Throttle Display

[GasSavers_maximilian]

I was thinking about how to display the throttle position for optimal fuel economy in the simplest, most compact manner. That way your mounting options could be enlarged, maybe allowing locations where you don't need to take your eyes off the road. Anyhow, what I came up with was to use a 10 segment LED bar graph that is dual color (red & green). Of course you could use a different number of segments. If you are going too slow, it displays a green bar, starting with your current throttle position and ending at the optimal one. If you're going too fast, it displays a red bar, again starting at your current position and ending at the best one. I guess you could pulse back and forth fast (or just use a tricolor LED display) for when you're on target. Have to test it. You'd just ignore the indicator when other factors (like highway minimum speed limits) contradicted it. It would still give you the direction to aim for when possible. Cutesy stuff like having it "walk" the led from current position to goal would just be annoying I suspect. I know the concept's darned easy, but I drew up a picture and it gave me an excuse to try out Photobucket. Anyone have any thoughts on improvement/problems? Thanks.

[R.I.D.E.]

Best throttle position for acceleration is the lowest amount that gives you the lowest possible vacuum reading without delaying your transmission shift points (automatic).

If manual, its about 70% throttle from 1500 to 2500 RPM as a general rule.

regards
gary

[theholycow]

What data do you intend to use? You mentioned throttle position and speed.

Based on wanting to keep centered on a target and avoiding either extreme, I would say to use a pattern like this:
IIIIIIIIII

[GasSavers_maximilian]

Originally Posted by theholycow

What data do you intend to use? You mentioned throttle position and speed.

Based on wanting to keep centered on a target and avoiding either extreme, I would say to use a pattern like this:
IIIIIIIIII

First, let me say this is just design groundwork. I don't have time to build one right now. It'd hook into the OBDII port, so whatever info that provides. I have a rough algorithm for the control system sketched out, but I'd love input. I just want to boil down the control loop I went through when monitoring the Scangauge II. I realize there's a lot more going on when a person drives (like anticipating upcoming stops and accelerating right before a hill, never mind pulse and glide), but the Scangauge doesn't monitor those things either. You can always ignore it. I'm not sure I quite get your centering suggestion. Do you mean that the pattern would move or something, giving you an idea how far off you were? Could you elaborate a bit? I can envision a couple things you might've meant.

I thought that the size of the LEDs that were showing might serve that same function (longer line means further off). It also wouldn't necessarily have to be scaled exactly to the actual throttle range, although that might get confusing. Maybe it could give more resolution in the areas you tend to drive in and less further up (much like telephone A/D converters...think that's called "companding" if my memory serves me).

[theholycow]

I was talking about centering because of this:

Originally Posted by maximilian

If you are going too slow, it displays a green bar, starting with your current throttle position and ending at the optimal one. If you're going too fast, it displays a red bar, again starting at your current position and ending at the best one.

"Too slow" sounds as bad as "too fast", so I figured it should get the same color. "optimal one" and "best one" sound like the center.

[GasSavers_maximilian]

I'm just unclear how you know whether to speed up or slow down. Is it that the green bit should always be centered on the display? Sorry if I'm being dense.

I was just following the stoplight color paradigm: green means "go" (speed up) and red means "stop" (slow down).

[theholycow]

Originally Posted by maximilian

It'd hook into the OBDII port, so whatever info that provides.

Are you aware of the work required to get data from OBDII? Millions (literally) of people try to DIY a cheaper OBDII interface, none ever get a reasonable amount of success, and I've only ever been able to find one or two that had a slight amount of success. The ELM327 is the cheapest way to get access to OBDII data, and is probably what's inside the ScanGauge II.

[GasSavers_maximilian]

I've checked out some of the projects out there (a while ago, now). I have a little bit of microcontroller experience, although I've never read OBDII. Like I said, just brainstorming at this point. Display design is very intriguing as people can have very different ideas of what's intuitive.

[theholycow]

Originally Posted by maximilian

I'm just unclear how you know whether to speed up or slow down. Is it that the green bit should always be centered on the display? Sorry if I'm being dense.

I was just following the stoplight color paradigm: green means "go" (speed up) and red means "stop" (slow down).

OIC. I was thinking there's a slow end and a fast end; if it's reading on the slow end (which you might position on the left or bottom) you know to speed up, and if it's reading on the fast end (right or top) you know to back off.

I'm not sure it's all that necessary to provide a gauge for optimal speed, though...a willing driver can do a much better job of choosing speed than the bar graph, and an unwilling driver won't obey it anyway.

[GasSavers_maximilian]

Originally Posted by theholycow

OIC. I was thinking there's a slow end and a fast end; if it's reading on the slow end (which you might position on the left or bottom) you know to speed up, and if it's reading on the fast end (right or top) you know to back off.

I'm not sure it's all that necessary to provide a gauge for optimal speed, though...a willing driver can do a much better job of choosing speed than the bar graph, and an unwilling driver won't obey it anyway.

Yep, your scheme is certainly a valid way to approach it. Be interesting to see if anybody has any others. Unwilling drivers wouldn't install the unit, presumably. All I can attest to is it's what I'd want (well, I'd really want an actual mpg cruise control, but that's a lot harder and potentially dangerous). The algorithm I follow in my head isn't all that complicated, but it'd be nice not to have to bother.