wes65

iTrader: (7)

Okay, but we have decided that the backfire is not related to the priming pulse because i still experience it when i have the priming pulse set to 0.0ms.

Laur3ns

I currently have:

The above settings are MSPNP defaults.

So if I read this correctly I will need to disable the Priming Pulse. The priming pulse is something that is injected before cranking?

I will try:

Matt Cramer

Spooky, that should turn the priming pulse off.

Braineack

iTrader: (62)

both you have data sheets on your injectors that say they don't leak?

wes65

iTrader: (7)

Yeah try that. That's how mine is set up and when i do see the backfire, it is not nearly as severe as before.

Laur3ns

I have a datasheet of my 550cc's that they flow 549-551cc/min. I dont think the mention leaking specifically but they should not as they are new. Also, if I leave the car for 24hrs this problem does not happen.

@Brain: isn't it like this: Turn key to IGN, MSPNP boots, outputs are all high, then fires a Priming PW directly, twice or after 2s. Then, only when you move your key to START, it will use the Cranking PW? I mean Priming is not Cranking, is it?

wes65

iTrader: (7)

It is very possible that mine leak but i have gone from 460's to 550's and the problem still occurred. Now, im not saying that both sets werent leaky, it just isn't likely.

wes65

iTrader: (7)

I'm glad we are able to have a half-way intelligent thread about this. One that i tried to start a while back got locked.

https://www.miataturbo.net/forum/showthread.php?t=24827

Laur3ns

Ive disable the Prime PW for now. I am going to calibrate my LC-1 first, and will report later.

Braineack

iTrader: (62)

I dunno what to tell you then...maybe I'm wrong, but then again the (2) cars here with 550s that I setup don't have these issues, nor does mine with 460s. I don't believe I've heard any complaints from the MSes I built and loaded with basemaps for 460s and 550s in the past few months either.

that one was locked because there were two threads that were started like the day before you posted that one and i was in a bad mood

chucker

iTrader: (4)

Same here. Cleaned tan tops and clean red tops. Injector Rehab made no mention of leaks one way or the other. My **** popped with both sets, installed in two different motors; only common factor was the MS. I've tried everything listed in this thread as well as many other threads.

Next step is to get a fuel pressure gauge and CONFIRM whether or not they leak. Need gauge, need motivation.

Joe Perez

iTrader: (8)

I'm not buying the idea that this is something caused by / preventable with a software setting, be it priming or whatever.

We're talking about a pop that occurs immediately when you turn the key, right? Take a look at the circuit diagram for how we tend to set up our spark outputs:



Ok, I know not everybody understands how to read this. Vcc at the top, that's +5. Both of 'em. Where it says "Squirt-1" is where the pin from CPU goes to drive this thing, and where it says "to IGN" is the output that goes to the igniter. When that goes high (+5), the coil starts to charge. When it goes low (Gnd) the coil "fires" and you get a spark.

Q6 is an NPN transistor. When a positive voltage is applied to the "base" (pin 2) then current is allowed to flow from the collector (pin 3) to the emitter (pin 1) which as you can see, is connected to ground. When there is no voltage at the base, no current flows from collector to emitter.

Ignore the LED and R24. They're irrelevant here.

So the way this works is that the CPU is normally putting out +5 on the pin connected to the base of the transistor. This turns the transistor "on", which causes it to conduct. You've got +5 going through a resistor (to limit the current) and connected to the collector of the transistor. The output is also tied to this point.

So when the transistor is on, whatever limited amount of current can flow through that resistor and then to ground. (It's 5ma* in that particular example.) Because of this, the voltage potential at the output is roughly zero. All the current that can flow through the resistor is going to ground, there's none left to feed the output.

When the pin on the CPU wants to start the dwell cycle, it sets the output appropriate pin low, and the transistor turns off. Now, the voltage at the output is no longer being dragged down- it goes to 5v, the igniter turns the coil on and the primary starts charging.

At the calculated firing time, the CPU goes high again, turning on the transistor and setting the output back to 0v. The igniter turns off, the magnetic field in the coil collapses, and you get spark.

(yay!)

Let's think about what happens when you first turn the key on.

The CPU inside the MS takes time to boot up. Not a lot of time- it's much faster than you desktop PC, but still, it takes the tiny little man in there a few milliseconds to stabilize, load the program out of ROM, initialize the output pins, and start running the main loop.

During that time, the output pins are low. Well, technically they're indeterminate, but they tend to be low. This means that the spark outputs are high and the coils start conducting.

As soon as the CPU comes to life, it immediately sets the output pins correctly, so the coils turn off. This is when you get your "power up spark."

Now, think about this for a moment. The spark happens at basically the instant that the CPU first comes to life. Whatever priming pulses you have set aren't going to get sprayed until after the spark happens! The CPU has to be "alive" before it can command the priming pulses, and yet the very act of the CPU coming to life is the thing that causes the little spark to occur. There's no way that priming fuel could get sprayed from the injector all the way into the combustion chamber in the timeframe we're talking about.

Whatever fuel is getting ignited had to have already been there before you turned the key. How did it get there?

We're not talking about a huge injector leak here. Fuel does not have to be gushing out of them like it's a heavy day at the brothel. A couple of drops over the course of the evening, vaporized and distributed throughout the intake plenum, should be more than enough to give the spark plugs something interesting to do when you turn the key the following morning.



(*) It's been pointed out by DIY, and verified by me by scoping, that 5ma isn't quite enough drive current to make the Miata igniter completely happy. Be reducing the value of this resistor to, say, 470 ohms, one can hit the igniter a bit harder and get a proper rising edge on that line.

chucker

iTrader: (4)

This would explain why my pops are of greater magnitude after longer periods of non-op.

I'm measuring for a fuel pressure drop tonight. Goddamnit.

But, if you're suggesting a mere "couple of drops" could create such a scenario, then it's possible that there could be near-nil observable fuel pressure drop overnight and still get a pop day next, yes?

I guess what I'm saying is that it would seem unlikely that a couple drops would be measurable as a drop in rail pressure...

I'll report back when I get data.

Edit: And ****, since we're talking only a couple drops, then any slight drop in rail pressure over night could be attributed to other minor leaks, such as back through the pump's check, right? Or past the FPR? So it's reasonable to say that attempting to isolate the problem as a barely leaking injector via measuring overnight fuel rail pressure drop seems at least slightly invalid.

So I guess the question has been answered, Joe. Other than injector leakage, there is no other realistic/practical way for fuel to be introduced in the sequence needed to cause such an event.

wes65

iTrader: (7)

so what we're saying is that the fuel that is being ignited here is coming from leaky injectors. Correct?

chucker

iTrader: (4)

I can't speak for others, but it's what I'm saying, currently. And if this is indeed the cause, I ain't doin' **** about it.

And Joe said this: "Whatever fuel is getting ignited had to have already been there before you turned the key."

Don't know of any other way to interpret it.

I will maintain that it seems quite odd that TWO sets of injectors have done this in my car (and yours too, I believe?). But then again, if the amount of fuel is slight, then, well...

Laur3ns

Hmm Prime PW to zero has no effect and still pops/backfires. Interesting post Joe!

wes65

iTrader: (7)

My fuel pump does not prime when i turn the key on. I dont know if that has anything to do with it.

Joe Perez

iTrader: (8)

I doubt it. Again, this initial spark event takes place before the MS really has a chance to do much else in terms of priming events. If you look at the source code, you'll see that the relevant output pins get initialized way before the real meat of the main loop starts happening.



In other news, the following question was posited to me earlier today by PM, in response to my observations about the borderline inadequacy of the pullup current on the "standard" output driver mods:
Not at all. In fact, current availability to the igniter has little to do with the transistor chosen.

The problem here is that the instructions in the MSExtra documentation (which it appears were the foundation for the Miata HowTo article at DIY) instruct the builder to use a 1k resistor as the pullup. This resistor sets the limit for the maximum amount of current which can flow out to the igniter. Using the formula I=E/R (current = voltage / resistance) this means that the current in this circuit is limited to 5 milliamps.

I don't know if the Miata igniter is unusual in this regard, but it seems to be a highly capacitive load. Its drive current requirement is not a constant; it's quite large initially, tapering down over time. When you try to draw more current than a source can provide, the consequence is that the voltage on the circuit drops. Thus, in the case of the igniter, the rising-edge of the trigger signal winds up being really ugly- more of a rising slope. Whether this has any impact on the coil primary current I don't know, as I didn't have a DC current probe with me when I was scoping the triggers. My gut feeling is probably not, as MOSFETs are pretty non-linear in terms of source-drain current vs. gate voltage. At worst, you might lose a few tenths of a millisecond of productive dwell, but it still annoys me in general to see imperfect signals.

I discussed this with Jerry @ DIY, and discovered that they'd run into the same quandary when designing the MSPnP daughterboard, Instead of 1k, they decided to go with 100 ohms for this pullup. That gives you 50ma of potential trigger current to play with, which is more than enough to make the igniter happy. So I was actually being a bit overly conservative when I called out 470 earlier. A 100 ohm resistor will be burning 1/4 watt of power whenever the transistor in on, and I didn't want people re-using leftover 1/8 watt parts in that location and destroying them. But as long as you use the right values, by all means, go for the gusto.

Getting back to the transistor, it needs only to be rated sufficiently so as to survive whatever current we're allowing here. Remember- for the vast majority of the time (whenever the coils are not dwelling) these transistors have to take the whole pullup current and sink it directly to ground.

The 2N3904, small as it is, is entirely adequate for the task as it's rated for 200ma continuous collector current. Assuming VF = 2 for D14 & 16, then IF = 10ma through each LED. That leaves 190ma of surplus capacity available for the igniter drive part of the circuit, of which we're demanding only 50ma in the case of the 100 ohm pullup.

Ergo, there is nothing at all to be gained by installing a larger part here.

patsmx5

iTrader: (16)

Do this. Before you shut the car off, bring the motor up to like 3 or 4K. Turn it off and smash the throttle open so the engine is pumping lots of air through it. This will damn sure expel any fuel that's still inside the engine. See if it still pops now.

elesjuan

iTrader: (11)

If I've been switching the ignition on and off a lot without cranking / starting, once in a blue moon I'll get the 'pop.' Doesn't happen very often though unless I'm messing with something.

This might not be a real great idea. :P