MPGuino release one workspace

[Njay]

Originally Posted by monroe74

On the systems that I know about, the fuel pressure regulator maintains a constant pressure. But I'm definitely no expert, and I'd like to understand this better. What car has a system that works the way you described?

I'm building this for my car, a 2000 Ford Fiesta (it's an European model). The 1st version of this car came out in 1995/96, and already worked like this at the time. See the pic below, it's a diagram of the fuel supply system. It's commented in my mother tongue but the translation is:

1) Fuel tank
2) Fuel pump
3) Fuel filter
4) Injector rail or "injector feeder chamber", I don't know the exact English name
5) Injector
6) Fuel pressure regulator valve <- note that it is connected through a tube into the "air intake" ("entrada de ar")

The fuel pump pushes the fuel through the fuel filter into the injector fuel chamber, creating a pressure in there. That pressure is regulated by the valve (6), by creating a "difficulty" for the fuel leaving back to the fuel tank. But the "difficulty" varies depending on the air pressure in the air intake. And that's all I know, I'm the electronics/software guy

[monroe74]

Njay, that's very helpful. Thanks for answering so promptly and clearly.

I wonder how prevalent that sort of system is, since it does seem to complicate the analysis.

And from an engineering perspective, I'm a little surprised that the system would be designed that way. It would seem to make more sense to tell the ECU what the intake pressure is, and let the ECU decide whether or not to let that information influence pulse width. In other words, if I were the designer I would address the issue by making the software a little more complicated, instead of making the hardware a little more complicated.

But maybe the system isn't very common, and maybe this is the reason.

[Yoshi]

Originally Posted by AndrewJ

Still working on calibration.

Just FYI...
I would like to explain how fuel parameter is set for Honda VX on my SuperMID.
http://ecomodder.com/forum/showthrea...6949#post26949

I use the fuel parameter 14220 for Honda VX.
The count tick is 5 usec, therefore the 14220 means 71100 usec per 1 milli-liter fuel usage.
Please note that I subtract 100 ticks (500 usec) from each injector on time for the mechanical delay (injectorSettleTime on your source which never been used).
so...
71100 usec/milli-liter -> 71100 X 1000 X 3.785 = 269113500 microSecondsPerGallon

Regards,
Yoshi

[MilesPerTank]

Originally Posted by dcb

Also, need more evidence that we need extra stuff to protect a $3.00 chip You are NOT going to like my extremely bare bones schematic

I've also got a fuse on mine as well as a transient single direction surge suppressor (Although mine is ~20ish volt clamp because its what I had lying around). A fuse should have been on the original board. A cars electrical system has tons of inductive loads all over the place that are high current compared to our little board. Depending on how they are clamped when disabled, you could see all kinds of wild ringing.

It's not significantly more money to add the two. I saw a blown up 7805 on a product just last week although it was probably a chinese clone to begin with . At least we have a diode on the input as a first stand against any negative swings .

Hopefully I'll wire mine into the car today. I think i may have the ugliest looking one this thread has seen yet

[cmags]

FYI, looks as if NKC is replacing the serial 'duino with a v2.0 which should have a more reliable serial communication, on par with the USB one. Figures that they put this out like 2 days after I ordered v1.0...

http://www.nkcelectronics.com/freedu...ompatib20.html

[Njay]

Monroe, I'm reading a few more things about this engine, and it may be constant pressure after all... need to do further investigation.

7805's maximum input voltage is around +30V. Any 12V automotive device is supposed to be able to withstand up to 42V big transients. And (big) negative transients also. If you're interested, I can post a link I have here somewhere to a commercial transient Test Equipment where you can see the types of transient normal auto electronics should withstand, and you'll be impressed. There are norms that define these transients.

Concerning the fuse, it is essential also for security. You risk a fire if you don't have one, specially taking in consideration the auto environment and the typical low reliability of home-built prototypes. The place where you're getting the power for your circuit is most probably already protected by a car fuse, by that won't save the device, because those fuses are for lots of Amps and your "thinner" circuit and the wires connecting to it will burn before the "big" fuse.

For the low price of both devices, a transient suppressor and a fuse, I think you unnecessarily risk too much.

I was unable to fully understand what electronics should I have in my circuit to protect against the standard auto transients. Therefore what I did was to add a fuse and a "powerful" voltage transient to each power input (I have 2 inputs). Furthermore, I added a series diode to further protect the voltage regulator against negative spikes and added a big and a small capacitor to the regulator's input. I want to have a rest sleep at night . My device will be permanently ON. To protect the signal inputs I used 2 common diodes and 2 resistors, plus a capacitor (adds considerable protection against transients) on the injector input. I didn't use zener diodes with resistors just because they are more expensive.

You can add a considerable protection to your microcontroller if you add an extra series resistor after the zener diode. Although a 5.1V zener diode looks perfect, the truth is that they have tolerances like any other devices, that can be up to 10%. This means 5.1 + 10% = 5.6V which is already more than the Vcc+0.5V (5.5V) the micro can handle; not to mention that, as input voltage raises, the regulated voltage will also raise, although slowly. Furthermore, the 7805 also has a tolerance, of +/-4% over manufacturing and temperature (you should take temperature into consideration in an auto circuit, because it will vary a lot), meaning that in some chips and some conditions your Vcc can in fact reach 5V - 4% = 4.8V, worsening the case, now it's 5.6V - (4.8 + 0.5) = 0.3V; with time, ATmega8 internal pin protection diodes will end-up degrading and burning, because current will not be limited on the pin. If you add the series resistor (10K or higher) you'll be protected against this.

Electronics design is fascinating

[monroe74]

Originally Posted by Njay

I'm reading a few more things about this engine, and it may be constant pressure after all

Interesting. In the drawing, the tube connecting the pressure regulator to the intake manifold definitely implies that the fuel pressure is not constant. But I'll be very interested in hearing whatever you figure out.

[Njay]

Originally Posted by monroe74

Interesting. In the drawing, the tube connecting the pressure regulator to the intake manifold definitely implies that the fuel pressure is not constant. But I'll be very interested in hearing whatever you figure out.

I hooked up at a sentence in an engine manual (in Italian...): "the valve makes the pressure difference constant". If you think about it, for the fuel to go through the injector opening at the same speed in all conditions, the pressure *difference* from one side of the injector hole (the injector's fuel chamber) and the other side (the cylinder entrance) must the constant... this is my suspicion. I'll have to take the car for revision soon, maybe I'll try to talk to someone there at Ford.

[dcb]

I don't know if that would be good news or bad news Might be good news for my saturn. Anyway, I'm declaring the MAP signal out of scope for release one (the title of this thread), at least until we have real data from the field that indicates we need to complicate the circuit and installation and code. maybe another sensor thread or something to sort that out would be good. And a way to generically read it on all cars would be extremely helpful too. I would like to eventually get a handle on the other sensors, but it has to be as generic as possible.

[monroe74]

Originally Posted by Njay

I hooked up at a sentence in an engine manual (in Italian...): "the valve makes the pressure difference constant". If you think about it, for the fuel to go through the injector opening at the same speed in all conditions, the pressure *difference* from one side of the injector hole (the injector's fuel chamber) and the other side (the cylinder entrance) must the constant... this is my suspicion. I'll have to take the car for revision soon, maybe I'll try to talk to someone there at Ford.

Very good point. I didn't think of that. So the term "constant pressure" can be misleading, unless we specify what we mean. Constant pressure relative to the atmosphere is not the same thing as constant pressure relative to the intake manifold.

I have a feeling that a random person at the dealer won't be a good source for a reliable, knowledgeable answer.

Now that I investigate some more, I think there actually isn't a problem. Looking at the service manual for my car ('95 Civic VX, p. 11-105), I see that the fuel pressure regulator has a hose connecting to the intake manifold. Fuel pressure is supposed to be 40-47 psi with the hose disconnected, and 31-38 psi with the hose connected (and engine idling). So the system is designed to drop fuel pressure when manifold pressure drops. So the system keeps fuel pressure more-or-less constant relative to manifold pressure, not relative to atmosphere. (I say "more-or-less" because I don't know for sure that the relationship is purely proportional. Although I bet it is.)

Anyway, I have a feeling that my car is just like your car, in this regard. And I have a feeling that this is the prevailing design, for EFI. Which should mean that pulse width is a good proxy for true fuel flow, after all.

It actually makes a lot of sense that engineers designing injectors would want to be able to assume a constant relative pressure. This would make is easier to predict, with a great deal of refinement, the exact mechanics of how the injector will behave.

So dcb, I have a feeling we've been spinning your wheels on this point, and this is something we actually don't need to worry about.