jimmyblue

iTrader: (11)

So I've been chasing around after a spurious unlock
behavior following my recent converter install. This
is the third 10" converter but the first to give me
any such problem.

I logged a bunch of misfire stuff and the only one
that looks in any way correlated, is the "Misfire
Cylinder Mode Index Level". And it's not the level
that appears relevant, but a sharp negative slope
seems to immediately precede an unlock / slip.
Meanwhile there are -zero- current misfires shown.

This is all pretty baffling, I don't understand what the
name means and it seems that whatever it means,
the "active principle" is not the value but the change
in it. And I can't see any "precursors" to this "index
level" in what I've logged so far. Even more importantly,
the misfire tables I have seem related only to the
level and not the slope, so what to do? Besides the
good old, "kill 'em all, 32767 times" approach?

Any misfire gurus in the house?

ToplessTexan

I'm vaguely recollecting there some bit fields within the TSTATExx PIDs that indicate the cause of a release. You might look at there to get some hints as to whether or not the funky item you're looking at is potentially causal or just correlated. It looks to my eye like it's related to RPM, maybe it's first or second derivative. It seems less like that toward the end of the scan, so maybe I'm just seeing things.

In the scan you posted, it looks like it occurred just after lockup. Is it always like this? Just for grins, have you logged the adaptation cells?

In the short run, have you tried the heavy handed approach yet? Does that in fact allow it to remain locked?

jimmyblue

iTrader: (11)

The "kill 'em all" tune is in, this morning, but haven't
driven it yet.

I can't see TCC adaptation but I have the shift
adaptation set to zeroes, trying to hand-tune the
line and not have it drift around on me. I've always
suspected there is some TCC pressure / % adaptation
but have never seen any access to alter it. I have
the TCC min % set to all 98s and it stays put. During
the flare events I see the PWM staying put, but on
the TCC mode I see either --3-- or a series of
"Release", --3--, "Apply" zip on by. Unfortunately
the current scanner version will not plot or export
TCC_MODE (because it seems to log as "friendly"
text, not a number). I've gone and dug out my old
EFILiveV6 cable and have to see whether there's
any more visibility.

Bad30th

iTrader: (13)

I'm sure you already have played with it, but there is also a required TPS% for lockup, as well as lockup MPH. I see your first post says it locked reliably with the other converters and the same tune ?

Does it lock and stay locked when commanded to do so by VCM control ?

Rob (Bad30th)

jimmyblue

iTrader: (11)

I didn't play with the direct control but did see it
was "misbehaving under command" in last night's
logs. So I put all the misfire tables to 32000 (had
someone tell me it worked better than 32767, can't
say why) this morning and drove about 70 miles
today, street, highway, cruise, accelerate, even
"waffling", and nothing but sensible operation.

So it's pretty definitely the misfires. Just don't get
why the most basic ("current misfire") PIDs show
nada, but this obscure "index level" does, and
what the hell it all means. But I'll settle for dead.
Though I'd rather have functional misfire detection
for when things go wrong.

I had thought maybe it was a wiring intermittent,
connector, maybe damage because sometimes it
seemed like I'd go over a bump and it'd get better
(for a bit). But last night I read in the Helm book,
that misfire detection, hence unlock, is suppressed
when the ABS senses "rough pavement conditions".
Go figure.

I'm coming around to suspect that it may be the
headers, installed at the same time, maybe throwing
my tune off and elevating the misfires beyond what
they were to be, when "the others worked". I keep
on violating the first rule of troubleshooting, "One
change at a time."....

ToplessTexan

Now that you know you can influence it via the misfire threshold, start cranking that back down to a more useful level. As far as misfire counters go, I believe they are not incremented until it's actually decided that it's misfiring. The unlocking is part of that decision making process so you wouldn't see increase yet. Not sure about the "index level" thing either, but I am curious now.

I want to say the TSTATExx PIDs further describe some of the TCCMODE values such as the cause of the release. 3 is 'Apply Enabled', it looks like a normal sequence of transitions. Been a while since I looked at those and my memory is a little rusty.

If you dig out your EFI cable you might want to also grab the 7.3 software. You'd need a V7 license but that just an email.

jimmyblue

iTrader: (11)

I've determined that "Index Level" is in fact the
value in the misfire table (seeing it now at a
constant 32000). So I infer that Misfire Index
is probably a pointer to the table or cell in play.
I still do not know what the real meaning of the
cell value in the table is (time between events?
Yeah, OK, what event, what timescale?).

My current misfires are all now consistently zero,
so I'm figuring that the cell value is actually some
sort of misfire yes/no threshold.

BLK02WS6

iTrader: (2)

Had the same issue with a customers car - big cam was throwing misfires in the cruise rpm range in 4th and randomly locking and unlocking the converter during highway cruising. I raised and flattened the mid range tables (this was a pretty big cam, so it was causing a problem above the idle and low cells)and it got rid of the problem.

ToplessTexan

One other thought is that anything affecting the rate of rotation of the crank could show up in the misfire world. Is you new converter heavier/lighter than the others? Balancing? I'd expect any issues along those lines to have a very systematic impact on "misfire", much more so than normal sources of misfire. Just a thought.

ToplessTexan

More thoughts...

The tranny state PIDs I was thinking of were TSTATE09 and TSTATE10 which both have bits indicating why the TCC is off.

Also, I see in EFI's scanner a PID for not only cylinder mode index and index level but also revolution mode index and index level. That reminded me that there are two misfire detection modes. I think the same accumulator is used for both and you're in one mode or the other. Also, I noticed a misfire diagnostic delay in units of "engine cycles" (which I assume is either one or two revolutions) in the cal. If your misfires correspond to shifts, you might look at increasing this as an alternative to raising the misfire threshold (which I believe is a duration in microseconds.)

Despite my comments above, I'm still thinking I'd expect to see the misfire counters increment though...