Hi all. I have a 1991 Accord LX Wagon. The speedometer and odometer died recently, so I had to replace the gauge assembly. (The pulse-counting circuitry is all on the main print board.)

Since then, the automatic transmission is not shifting into 4th gear. Easy to tell, no RPM difference when popping back and forth between D3 and D4 while moving at a decent speed.
I can drive in 3rd gear just fine, but with the higher RPM I may get worse gas mileage. And I'm not sure the lock-up hydraulics are engaging either. It would seem that if the ATU doesn't think we're in D4, probably not.

I have high confidence in the speed sensor, have tested the solenoids, and have two different A/T controller units (ATU) to try that both were working before the change. Both are acting the same way now, no D4.

Here are two notes about the new (used) speedometer.
1. The D4 bulb in the shift selection indicator is burned out. I should have disassembled the gauges and replaced it but I didn't.
2. This gauge assembly does not have the "S" indicator like the old one did.

I've checked integrity of the D4 wire from the ATU to the shifter. Both the line from the gauge assembly bulb and from the ATU get grounded when one shifts to a particular position.

So now I'm wondering if the burned out bulb (that does not light up D4 on the gauges when shifted to that selection) could actually be affecting the ATU behavior.
Has anybody encountered this kind of effect? I'm considering pulling the gauges back out to fix this. But this seems like a long shot, it's just a light bulb.

Similarly, any reason the missing "S" indicator would throw off operation of D4? Again, I think it's just a light bulb, why should it care?

Thanks

 

I doubt the bulb is causing the shifting issue. It does sound like the replacement cluster is not the same as your original and might cause some problems. I'm trying to figure out if the instrument cluster sends any voltage/signal to the TCU from the VSS, tachometer, etc.

The 91 accords have a window on the TCU and ECU to look at codes. You may want to check what codes are flashing on the TCU to help give you direction on testing.

 

Thanks PAhonda for your attention and response.

I did try replacing the D4 bulb and got this behavior:
* D4 indicator is lit all the time
* Turning on the ignition acts like it does when there is a TCU code: check engine light on, fuel pump does not initialize the pressure, can't start the car
* Putting a jumper in the diagnostic connector does not cause the D4 indicator to flash codes, even though I see in other threads that some other models do use that indicator
* The original TCU does have a window (visible through a hole in the passenger kick plate). However, no codes there either. Maybe there's a bulb in there I need to check.

So, I didn't like how things changed, and pulled the good bulb back out of the gauge assembly. And now I'm able to start the car again. Haven't driven it again but expect I'm still in the same state as before.

You are correct, this gauge assembly isn't a perfect match for the original. I could not find a match out there on the internet but maybe I didn't try hard enough.
I had several to choose from and decided I wanted the cruise control indicator more than the S indicator.

 

I thought I had seen a window through the kickplate, but it turned out to just be the multi-colored metal housing of the TCU. Neither the ECU nor the TCU in this car has a window to see codes. I may rig up something to emulate the S indicator if I can be sure of the way the lines into the gauge assembly are used. I do have the main service manual and the electrical service manual for this car.

 

I did an experiment today. I put in the original cluster (with dead speedometer and odometer), and took a drive. Fourth gear worked just fine. I thought this experiment would be useful if the TCU really is throwing a code; it should have reported it if in its memory, when I turned on the ignition. But apparently its conscience is clear.

Even with the replacement cluster, the check-engine light is not on. I think if the TCU has something to say, it would tell the ECU to turn that light on.

Looking at the circuit diagrams, it appears that the TCU wants to control the S indicator through a transistor. If the replacement unit doesn't have that circuit, I bet that output of the TCU is staying high (5V) and throwing off some other logic. So the TCU knows something isn't right and stays in the problem state, no fourth gear and a bit wimpier acceleration starting up from a stoplight. (Is this what others have been calling "limp mode"?)

 

Reading on the web, I guess what I'm experiencing is not as severe as "limp mode".