You know if the injectors are running very low pulsewidth, they can become erratic just because they aren't opening and closing properly. I wonder if we are pushing the limits on the injectors too much and because they never open its causing some kind of electronic feedback thats messing with the injector triggerings...

Yeah I think it is worth a shot. At the very least it would be a good LRU shotgun maintenance test (Air Force aircraft term) to see if it is a hardware issue. Swap units, keep everything the same (same firmware, same ecu file) and see if the problem goes away. If it does not, then it either has to be a software issue, something with the adapter harness, or something with the car.

Not to debunk the therory, but I do know if I start it cold it will idle as long as I want it to. If I drive a mile and stop at the stop sign, it will idle as long as I want it to. If I start it hot, it won't idle for crap. If I drive it for a while to eliminate any possibility of heat soak (if this is actually the problem) after a hot start, it will idle semi-crappy, lean and surging.


That said, it did idle less crappy on stock injectors and I didn't have the dieing problem.

You know I considered that. But with the Megasquirt and the hi-res code I was able to idle the car at 12.5-13:1 with no problem, though I had to retard the timing in the idle cells a bit to keep it stable. I also had no problems with cruising. So I think even the 550cc injectors are capable, as further evidenced by the number of people successfully running them with the MS.

But if there is a significant difference in the control circuitry between the Adaptronic and MS, or the software, then like you suggest it could cause a problem I think. That is why I was wondering about using peak-and-hold versus 'regular' injector pulsing, and if Adaptronic had another customer running a similar engine with 550cc RX7 (high impedance) injectors.

The other thing too with my NB is the returnless fuel supply system. The rail is running at about 60 psi, so the injectors are actually flowing more than their rated 550cc. So, for comparison, what we would really need to know is if anyone is running high-impedance 550cc injectors (or larger) with 60 psi rail pressure.

Here is someone running 550cc injectors on a MX-5, though they are talking in this thread about tuning tip-in enrichment.
http://www.adaptronic.com.au/forum/i...hp?topic=464.0

We could try to change the maps over to saturation base but we'll have to increase the pulsewidth across the board. I also noticed the VVT controller was in full swing on your maps. I just removed all the VVT settings since you guys arn't running VVT. It kept showing very very strange readings on the VVT* column. Maybe it was confusing the ECU?

How did you do this, just zero out the VVT settings? I can try this too. I have no use for it since I do not have a VVT head.

Yes, and change it in the engine tab at the top and all the way to the right. The second trigger is marked VVT.

Travis

OK will do, thanks.

did a couple searches on the adap forum regarding high imp injectors. they say to still run them at .9A peak hold. who knows.

I read that one the other day with great anticipation, only to find out that he had it wired wrong all that time. I'm only ASSUMING that our PnP units were wired correctly.

I asked Mark at Deatschwerks about this peak and hold vs saturation thing and never heard anything back. Rob and Andy at adaptronic say that peak and hold will only make your injectors operate faster, which really doesn't matter. We can cancel it out with dead time no problem, so I could try to switch today and see what is necessary to make the change...

I say it is worth a try if it is potentially the cause of the problem. But if it is not, then I think the peak-and-hold scheme is a better one to use if it improves the opening time of even high impedance injectors.

I am going to go out on a limb here with some subjective reasoning. If the peak-and-hold scheme ends up running more average power through the injectors, then they could be getting warmer than normal due to the additional current flow. At low fuel flow levels, like idle and cruise, where there is not as much fuel flow through the injector to cool it, then they could overheat and cause vapor lock in the injector. Either that or their electrical characteristics change enough with heat that they cross over some threshold with the tuning or circuitry, and thus operate poorly. This would correlate with the heat soak condition.

This makes sense in a way to me for this reason alone:

It appears to be heat soak related, but not IAT heat soak. Here is a cold start and a hot start from yesterday. No changes to the ECU - just shut down and restarted five minutes later. First one IAT is 18*C and water is 18* C. Second restart is WT80*C, IAT is still 18*C so ZERO IAT heat soak.

But, look at the injectors and especially the fuel trim. On the hot restart, the trim is flat lined at 4. It doesn't react at all with the fluctuations of the injectors. This is not normal.

I don't know of anyone who has ever had a problem like this with stock injectors. That just occured to me. We are all running larger then factory injectors, and now that I think about it that wasn't 8 months ago that I had a similar problem. That was just after my larger injector install. I probably haven't experienced it because I haven't been running the car much. I've been testing the COP system, and just doing basic tuning for setting up the 600cc's...

Two more shots to look at. They are all from the same hot start log in the post above.

The hot start in the post above died. Here is the restart. You can see where the flatlined trim started to work and you can see how the AFR's started to settle down.

The second one is after arriving home and idling in the garage. Trim is working and idle stabilizing. So, I think this is tied to trim somehow.

Interesting Stein. I am going to have to go back and look at what trim is doing in my latest logs.

I have a question. I have been thinking about going through the idle tuning procedure in Sect 5.6 of the Installer's guide. In the NB base maps, Idle tab, 'Overall control' section:
- Minimum value is 60%
- Maximum value is 100%
- Step period is 11ms
- PID timeslice is 44ms

In the Adaptronic Installation manual it mentions the following:
"On some engines (such as the Mazda B-series DOHC) the idle valve starts to behave
non-monotonically below a certain duty cycle, so a minimum is specified to keep the
duty cycle outside this range. "
Since we have B-series engines, I am assuming this is why the minimum value is set at 60%. But is the 60% value a swag, or based on someone trying it out with a NB IAC valve?
According to this thread, should we not be using 22ms (instead of 11ms) as the PWM frequency for the NB IAC valve?

Sure looks that way. Can't hurt anything by trying it, right?

Stein,

Did you try to run the car after the changes?

Travis

You mean what you sent me this morning? If so, no. I'm going to try it tonight.

What do you think of my flatline trim logs above?

OK I changed it to 22ms, went for a drive, parked it, then waited... same problem, eventually it started running rough, then died. I restarted it with some difficulty, and let it idle, roughly, then after maybe 30 secs or so of rough idling it cleared up and idled normally.

Here is the log plot for this instance. At the beginning it is idling smoothly, then it begins running rough. Just before the center of the plot you can see where the RPM goes to zero twice (engine dies) and the two corresponding restarts. After restart #2 it idles, albeit roughly, then clears up just after half way. Then I blipped the throttle a few times (you can see the spikes in TPS) and then let it settle. It idles smoothly until the end of the log.