Super MPGuino?

[bobski]

Simple points wouldn't be enough to meter fuel flow. It would respond to changes in engine RPM, but the duty cycle (ratio of on/off time) would be fixed, so the fuel flow would actually be constant. It would need some means of varying the duty cycle and/or the fuel pressure applied to the injectors.

I worked on a friend's '89 VW Cabriolet (rabbit) with a bosch (mechanical) constant injection system. It had a fuel fuel distributor setup with braided stainless lines running to nozzles on each intake runner. The distributor had some kind of pump or hydraulic sensor pickup attached to the engine block, as well as a well-balanced deflection plate placed in between the throttle body and air filter to gauge air flow. As the plate was pushed further and further by increased air flow, it would crank up the pressure to the injectors, forcing more fuel through.

[edit]
link: http://home.hiwaay.net/~langford/CIS.html
[/edit]

[dcb]

Volkswagen Type 3 - Wikipedia, the free encyclopedia,
"the Type 3 engine got a larger displacement (1.6 L) for 1966 and modified for 1968 to include electronic Bosch fuel injection as an option, making it the first mass production consumer cars with such a feature "

not CIS though.

If I recall, the ecu was basically a metal box full of transistors

[gtkid2002]

Well, it'll be a bit longer than expected. I have finals this week and next, so I won't get much done at all. I am also not driving the Fiat. The shift lever has a rubber rung on it, and over time these degrade. Guess what finally gave when I was shifting into fifth gear on my way home from getting the last parts for the license plate lights.

So now it will be sitting until I can get done with my finals and get the parts installed.

The fiat has an odd chunk of wood on the dash for the center, and it is warped, and doesn't really seem to fit right. So I think I'll be replacing that, so when I do that, I'll just install the Super MPGuino smack dab in the dash. I'll leave a USB port visible underneath the dash so I can just get a laptop in the car and play with it as I go.

It does have a MAP/MAF whatever sensor *braindead* smack dab on the intake. Only problem for me is that I can't install a hot air intake, or a cold air intake, because if I take that off, it won't even idle for more than thirty seconds. It's a massive air restriction as well, so I might try and consider updating it with a sensor from a Honda or something. It's literally a massive flap that gets pushed back by the incoming air. It takes quite a bit of force to move it, and I can hear it pop back into position when I punch and drop the throttle. It's very odd, but on this car, there isn't much that seems normal to me.

I'm still not sure if it's a points system, because it has the condenser, but I thought I read somewhere that in 1978 Fiat moved away from the points system. I'll find out AFTER I can get the 70 dollar manual, haha.

So much for Aeromodding this though, which kind of sucks. I could mod the headlight buckets with some clear acrylic or something, but it depends on how badly that kills it's looks. It is a classic Fiat, so it should have some class still.But at least it is a really light car, so it should get better city mpg that my Prelude could. That and it's a five speed, with no CEL indicator, which means I can do the temporary killswitch on it, and it won't go into limp mode *rich and retarded tuning to prevent damage* like the Prelude did as well.

I love this car, but at the same time I want to just sell it. For now I'm gonna keep it and fix it up. It's just going to be incredibly expensive until I get the rag top fixed. Then it should be cheap.

But yeah, just thought I'd update with a notice of no updates for a while.

[gtkid2002]

Well, the Fiat is kind of not working anymore, so it's back to the prelude, and back to the project. I should be able to start really digging in a couple weeks from now, but I'm at least going to be starting on it again.

On the Fiat though, there was a mechanical Bosch ?-type injection system, and a vane type MAF sensor. But it didn't measure density, and it would cause my car to sputter badly if I cornered fast enough. (Gotta love gravity and its ways) MPG wasn't pretty either, as the two times I was able to top off the tank after I got it (Fuel pump D/Ced and I lost around 10 gallons of gas. Was not happy) I only ended up with 19mpg around the city. So I tried to baby it a bit more, shift at 2.5krpm, keep to the slow lane (not a problem, thing was louder than hell anywhere over 50mph), and was easy on the gas. Got 18mpg. Not sure how that happened.

Then a few days ago I "magically" get five nails and five staples in my two rear tires. Not a single one in either of the front two tires. Seems like sabotage, but I didn't even park that day. Hell, it was only out of the garage for an hour to drop some people off then go home so I could sleep.

But I'm back to the Prelude now. Rod knock or not, I have no other vehicle to drive right now. And now I don't have to worry about someone slashing my rag top, having wet/moldy seatbelts, worrying about cornering in a RWD car (Yes, it makes a difference. I almost flipped it on the second day I really started driving it), worrying about getting siphoned, etc.

No clue on my mpg right now on the prelude. I can't fill up, so I'm going by my tried and true method of driving till the low fuel light first comes on, which means I have 2 gallons left, or roughly 50-60 miles, depending on how frequently I use the fuel cutoff.

On a side note, anybody know how much my Drag Coefficient goes up when I use my flip-up headlights?

[gtkid2002]

Life! I have a button!

After a long while of this thing sitting, and switching between a car with no VSS (My 1980 Fiat) and a car that's a year before OBD-II (the '95 Prelude that I hate), I've finally started playing with this again.

And this time I got somewhere.

After looking at the schematics, I tried swapping the standard button pin with a bunch of other pins. Digital pins in the low 90's, 60's and 50's. Nothing worked. But when I waited a while, the screen would dim. Then I accidentally touched the INJ IN wire, and it registered.

Huh.

So I tried putting the button pin on PWM pin 0. And I got it to move!

Now if I recall correctly, I can switch some of the higher digital pins to PWM with some added code. If that's the case, I'm just going to try that. Otherwise I might have to go to a serial screen and then use the rest of the PWM pins.

For the time being though, I'm just going to use the remaining 4 PWM pins, and see how that goes.

----------------LATER------------------

Okay, I've soldered in the pins. And I can only move the underscore to the right. None of the other buttons work. Doesn't matter what I tell which button should be controlled by 0, I can only scroll right.

I put it on pin TX3 or 14 on the COMM area, adjust the PCINT just below the button input, and even if it's the "middle" button, it still scrolls right. In fact when I do this, it only scrolls right untill I press the button. I've got charts for the Arduino mega, but not the Seeeduino mega. I'm not sure if it;s just that unpopular or if it's still too new. But I am still fighting this. I'm gonna try switching the INJ with the VSS and see where that gets me.

A futile attempt later...
Nowhere. Touching the wire only gets GPH to twitch, reguardless on whether it's set to be INJ or VSS.

I really hate to scrap this project, but something isn't working right, and I'm not talented enough to know what's going on. If I can find a use for the microcontroller besides this, I'll scrap it and stick it in my solar backpack. In which case the thread can be locked or left open if someone else wants to take over it. I'm just about done with this though. It's easier to just go out and buy a complete MPGuino and just install it that way.

So if anybody wants to trade me an MPGuino for a Seeeduino mega and screen with some solder on them and a bad wiring job that can easily be undone, send me a PM.

[gtkid2002]

Yaay. Old thread back from the dead.

Summary: Seeeduino Mega's do not work with stock MPGuino code. I've tried this, but you're welcome to try it again if you want.

I'm now rockin a '328, in a 2004 Honda Civic DX, standard screen.

Since the 328 still has more code room, I'm putting the temp sensor idea on hold, and putting an RPM meter on the front burner. A lot of people seem to be interested in getting RPM, so why the heck not. I'm also scrapping the accelerometer idea for now. I'm no longer in a performance handling car (the prelude had 4WS, that's why I wanted it), so it's moot for me.

Right now I'm attempting to duplicate the VSS code but have it renamed RPM and instantRPM. I don't know if that's how to do it or not, but it's what I'm trying.

If someone wants to give me some input on what I'm thinking, that would be -awesome- since I have NO CLUE what I'm doing:

A: Should I have the input pin on the Digital I/O side, or the PWM side with the VSS?

B: Would it be workable (albeit primitive) to have a "custom 2" screen that simply displays RPM? I mean have it just take the raw input from the RPM (resistored of course), and have it run through lcd.print command? *I know this probably makes little sense, but I'm trying *

If somebody already has this coded, they should post it in the Code Hacks section, as they would make probably a ton of people happy.

[bobski]

You already have an input you can use for RPM: The injector signal. Its only shortcoming being that it will read zero during deceleration fuel cut.
You just need to measure the time between the start of each injector pulse. If the MPGuino code used the up-to-date Arduino IDE and libraries, I would tell you to put something in the injector interrupt routine along the lines of:

routineCallTime = micros();
injectorPeriod = routineCallTime - lastInjectorOpen;
lastInjectorOpen = routineCallTime;

And in the main program:

if((injectorPeriod == 0) | (injectorPeriod > 1000000)){RPM = 0;}
else{RPM = 30000000 / injectorPeriod;}

Then do whatever you like with RPM.

But MPGuino is sort of off on its own development fork, so the above probably won't work.
If you want an RPM input that will work under all conditions, you could feed a conditioned tachometer signal to an interrupt input. Oh wait. Both interrupt inputs are being used for the single injector signal, thanks to the limitations of the older Arduino IDE. Huh. :/

[meelis11]

instant RPM function is commented out in mpguino code but is still there
maybe in newer versions it is removed, dont know...

I have working RPM in my mpguino
To make it show realistic rpm, I had to set injPulsesPer2Revolutions=1 in settings menu, before I think it was 2 so it showed me 2x more RPMs

function is this:

unsigned long instantrpm(){
init64(tmp1,0,instInjCount);
init64(tmp2,0,120000000ul);
mul64(tmp1,tmp2);
init64(tmp2,0,1000ul);
mul64(tmp1,tmp2);
init64(tmp2,0,parms[injPulsesPer2Revolutions]);
div64(tmp1,tmp2);
init64(tmp2,0,instInjEnd-instInjStart);
div64(tmp1,tmp2);
return tmp1[1];
}