realtime monitoring tools for pre-obd2 cars

[theholycow]

Originally Posted by slogfilet

There's a lot of information about a DIY computer over on ecomodder.com. Parts look to be about $80... worth looking into!

http://forum.forum.ecomodder.comdisp...mputer-26.html

Broken link. I think it's supposed to be this link:
http://forum.ecomodder.com/showthrea...mputer-26.html
but that one requires you to be logged in.

[monroe74]

gary: "I found out when it [shift indicator light] comes on you can put it out by giving the car more gas, and that is exactly what I am learning to do, apply the minimum amount of throttle that keeps the light off, which should keep me in lean burn."

Interesting technique, but I have a hard time understanding how that's useful, for the following reason. Once you're in top gear, the SIL is off 100% of the time. And aren't you usually in top gear? I find that I am rarely in any gear other than top gear, unless I'm traveling under 25 mph. And when I'm traveling under 25 mph, it's usually only for a very short interval. I accelerate to 25, and then I'm in top gear.

It sounds like you're cruising in 4th, and that surprises me. And if you're in 4th, and applying enough throttle to turn the SIL off, I think that would produce a fair amount of acceleration, unless you're heading up a grade. So I'm a little confused about what's really going on.

"I dont think it would be possible to get the kind of mileage I reported without staying in lean burn a significant portion of the time"

There are some interesting differences between your driving environment and mine (although we have very similar cars and certain similar driving techniques). It sounds like you're putting on a lot of miles, in a very consistent environment (the same route over and over again). This means you're getting relatively frequent feedback at the pump, and this is enough to help you evaluate how you're doing.

I'm putting on relatively few miles, and my driving routine is erratic and unpredictable. So the interval between fills is long, and it's hard to make the connection with particular driving events and styles (because one tank, or half tank, includes a very diverse range of driving environments).

Also, it sounds like you're often in heavy traffic. I'm usually in light-to-moderate traffic. I think we're both concerned about being courteous to the person following us. But in my moderate traffic conditions, I still have room for fairly radical P&G, without him noticing. So I can do lots of WOT (at high gear/low rpm), alternating with neutral-coasting or even EOC, without bothering anyone.

On the other hand, your environment of heavy traffic calls for something much closer to a steady cruise. My hunch is that maximizing lean burn is the optimal strategy in your environment, while minimizing pumping losses is the optimal strategy in my environment.

I think the VX is an exceptionally flexible tool, with an unusual ability to adapt to a variety of environments, in the hands of a clever operator. There is not just one best way to drive it, under all conditions. At least that's my working hypothesis. I'm trying to develop techniques to test that hypothesis.

"I dont think knowing what my instantaneous mileage is at any point in time would change that much"

I think the interesting thing about mpg calculations is that they all involve choosing a span of time for the purpose of averaging. Even the term "instantaneous mileage" doesn't really mean "instant;" it just means a span of time that's relatively short (perhaps a second, or less). At the other extreme, the interval might be a tankful, or 90 days, or the lifetime of the car. Or the lifetime of the driver, or the planet.

I think we need better tools for zooming in and out, i.e., choosing a variety of intervals for mpg analysis. This is relatively unimportant in your situation, where your driving routine involves a lot of consistency and repetition. But I think you could see how it would be helpful in my situation.

[monroe74]

fume: "What you would need to monitor injector pulse width would be a speaker output transformer"

You're right, that's simpler than I thought.

"The pitch of the sound output would change with the rpm of the engine and the tone or timbre of the sound with pulse width."

That makes sense. But I'm not interested in monitoring rpm this way. After all, I already have the tach for that purpose, and I'm also hearing a sound just like that directly from the engine (a noise that increases in pitch as rpm goes up). I think what's more interesting is to modify the sound in order to simultaneously monitor another sensor's output (e.g., O2). I wouldn't know how to do that, but you probably do.

When I listen to a band, I can hear several instruments at once. Similarly, I should be able to use sound to monitor several sensors at once.

[monroe74]

slog: "a DIY computer over on ecomodder.com"

Yes, that project looks interesting. I'm going to keep an eye on it. One limit I don't like is that I want something that can capture data and send it to my computer. Oh well, it's hard to please everyone!

holy: "but that one requires you to be logged in."

I think this link might work:

http://forum.ecomodder.com/forumdisp...mputer-26.html

[Dalez0r]

To me, the simplest approach (simpler and cheaper than what the ecomodder guys are doing) would be to use a PIC microcontroller (under $10) and read the injector pulsewidth and VSS right off the ECU wiring harness, and send the data raw to a PC via serial/USB. On the computer end, select from preprogrammed vehicles or enter in your own vehicle info and bam! Fuel economy tool. :P

When I get back from Burning Flipside, I may need to work on this....

[GasSavers_RoadWarrior]

Well since all the data is there, and only needs signal level conversion and protocol translation, a PIC is a bit of a heavy handed way of going about getting a laptop PC involved. (FWIW anything pentium class and up should handle this kinda thing so we're talking $50 laptop, or scrounge one for free.)

Was pondering a way to kick it oldskool and use MAP signal, injector dwell, and a potentiometer+pendulum accelerometer data processed by linear integrators to give a unified econometer reading.

BTW I can "cheat" on marvin and just get a "traveller" computer out of another 80s-90s mopar and hook it up. Most domestic produced chrysler products from late 80s to '95 can use these. This may also be an option for other makes, for instance did Acuras have trip/mileage computers that could be retrofitted to hondas?

[Dalez0r]

I'd say a PIC is a whole lot better and lighter than using a Arduino board

And if you don't want a laptop constantly in the car, a PIC can quite easily output to a LCD for reading realtime...

[GasSavers_RoadWarrior]

True if you're gonna do it that way you can make it tidier than a laptop. If you find a free/cheap laptop with a dead screen however, and an SVHS output, you can get a portable DVD player free/cheap with a busted drive, build that 5 or 7 inch screen into the dash, hookup through the inputs (you buy one with inputs) to the lappys SVHS out... hide the laptop guts somewhere else... and do it that way.... your benefit doing it that way, is that you're well set up for taking advantage of any "tune on the fly" PC software that will work with an EPROM emulator and your ECU, so you can make yourself a custom lean calibration.

[monroe74]

dalezor: "On the computer end, select from preprogrammed vehicles or enter in your own vehicle info"

I'm interested in these various ideas that are bouncing around, so I'd like to ask you to explain more about what you're thinking, if that's OK with you.

What's the software you're picturing on the PC? I know there are various programs that are designed to work with OBD2 data, but that's not what we're sending to the PC, right? There are also programs like Crome that are specifically designed for hacking certain kinds of ECUs, but that's also a different story from accepting an incoming signal that represents realtime injector and VSS activity.

Or is there some kind of custom program running on the PC? I'm just trying to grasp what you're talking about.

[GasSavers_RoadWarrior]

Okay been looking at a couple of tach/dwell meter designs I've got in books, appears the simplest possible thing to do is put a 250 microfarad capacitor across a 15V meter, and connect that to an (or the if TBI) injector through a diode. Possibly a 12V meter would do if you only expect to see 80% duty cycle, and the diode will drop .6V if you use a general purpose silicon one.