The D4 light

Hold on now...

I've spent some time with the shop manual here trying to make some sense of
this.

You mentioned flaring on up and downshifts, specifically 3-4 and 4-3. This
is critical to diagnosis, so I hope you've been complete in your
description.

One major difference between the shift from 3 to 4 is Shift Solenoid "A"
turns off going up, then turns back on again going down. Solenoid "A" is
off when the car is at rest in Drive. It turns on again for Reverse, so if
you have no delay when shifting to Reverse, Solenoid "A" is surely fine.

However, the shop manual lists these things as possibly being wrong:
1) Throttle cable adjusted too long (probably not the case here, but worth
checking because it's cheap)
2) Vacuum modulator valve stuck (not the case here)
3) Orifice Control Valve stuck (possible; also accounts for harsh upshift)
4) Feedpipe O-ring broken (not the case here)

If the Orifice Control valve is stuck, this is due to sludge and neglect.
It is also possible the 3-4 shift valve is sticking. This is inside the
main hydraulic valve body.

With my copy of the manual, I can't tell if you bypass the Orifice Control
Valves when shifting manually.

Also, there are line pressure inspection bolts on the side of the
transmission. These can be removed and a gauge plumbed in so line pressures
can be monitored as-it-happens. If you've got a slow pressure buildup and
release to 3rd/4th parts (which is what appears to be happening here), this
will become apparent when measured.

Honda automatics haven't changed substantially since 1990, so if you bring
the car to your local dealer, the techs there will have a pretty good idea
what to do and will have the necessary testing equipment.

Good luck.

 

No, it flares into third under load, whether on upshift (2-3) or
downshift (4-3). If I ease up on the throttle when the shift is about
to happen it does not flare. I haven't noticed any other transmission
anomalies besides what feels like late shifts in general.

Yeah, reverse seems fine.

Eh, it's possible. Remember I had this thing ripped to shreds not too
long ago. But I don't know what the throttle cable could have to do
with it, the ECM sees the TPS position and the transmission sees
manifold vacuum, so how does the cable adjustment change that?

Actually, it does kind of act like an open modulator in that the
shifts are late, I couldn't find any vacuum leaks by a quick check
though.

Oh, that's interesting. I assume those specs are in the manual. What
is the title/number of the manual you are looking at?

 

TPS reading is entirely dependent on throttle cable movement. If the
cable is out, the TPS will receive incorrect readings (example: pedal
hits floor, throttle still not completely open). Seems to me this would
be manifest in any gear though, the reason I don't think it's the issue
here.

The modulator is only ONE factor. Are ALL shifts late, including
Reverse? If not, the modulator is fine.

I'm looking at a copy of the Helm manual for '89-'91 Accord.

Problem is, I have to pair that with my original '91 Integra factory
manual because the generous dummy that copied the Accord manual for me
neglected to include ANY indices, most of the wiring diagrams, and a
bunch of other pages he figured weren't really that important. I guess
he did this to save time. He's not a mechanic, of course...

Another problem is that the hydraulic diagrams are small, busy and
dense, plus containing a generous salting of legal-notice-size 6pt type,
not easily photocopyable. And the 'Teg's hydraulics are a little bit
different from the Accord's.

The Integra's AT is very similar to the Accord's, differing primarily in
manual gear selection choice, and in the inclusion on the Accord of a
"linear solenoid", something the 'Teg doesn't have.

The one thing that overwhelms me while attempting to comprehend this, is
the crushing, brain-busting complexity of the whole electronically-
controlled automatic transmission idea. The hydraulics are bad enough,
and then you have to layer the electrical control on top of that, plus
all the sensors the TCM and ECU depend on.

Well, the basic concept of the tranny is pretty simple, but how Honda
actually goes about making the beast work is frankly boggling. How long
did it take them to come up with this? How many times did the manual's
artists and writers have to go back to the engineers before their
diagrams and text were correct? Do I really wanna know?

A manual tranny is almost caveman-like in its simplicity and crudity:
you have a man in the driver's seat stirring a big stick around. By
contrast, in an auto tranny you have a computer stirring the big stick.
But...the computer, lacking eyes, ears, fear, wife in the passenger seat
or mother-in-law in the back, has to be able to figure out on its own
exactly what needs to be done in the very next instance, every time.
It's a hard job.

A final note I forgot before: The TCM (Transmission Control Module).
These are known to go bad and this can account for all your symptoms,
including the lack of trouble codes. You can get replacements for about
ten or twenty dollars at the wreckers. Might be worth a shot.

Do you wonder now why good techs are paid what they're paid?

 

Oh right, I get what you meant now.

No, seems to go into reverse at the right point.

Yeah, the layman tends to assume that the engine of a car is the most
complex system but modern A/T's seem to have claimed that title. It's
amazing how reliable they are, on the whole.

Yeah, I was going to check into that. Looks like leaking caps are
pretty common on this one...