Computer Problems with Manual Trans Swap

This site may earn a commission from merchant affiliate
links, including eBay, Amazon, Skimlinks, and others.

I evaluated the EWD in relation to the auto trans. It does not have very many connections.

I can see that wiring the 'N' indicator to light up won't do anything because the ECU doesn't even detect this signal.

I do believe I could design a circuit to emulate the transmission and make the ECU happy in all conditions. I think it is possible to fool the ECU into thinking the transmission is there. It needs the following:
  1. ECUA-10 (S1) - Needs a 15 ohm, 20W resistor to ground. Might get warm
  2. ECUA-09 (S2) - Needs a 15 ohm, 20W resistor to ground. Might get warm
  3. ECUA-08 (SL) - Needs a 15 ohm, 20W resistor to ground. Might get warm
  4. ECUB-08 (VSS-) - MAY need a Vehicle speed sensor simulator. Requirement will need to be determined by o-scope tracing
  5. ECUB-08 (VSS+) - MAY need a Vehicle speed sensor simulator. Requirement will need to be determined by o-scope tracing
  6. ECUA-03 (HOLD) - De-pin from ECU
  7. ECUD-20 (PWR) - De-pin from ECU
  8. ECUD-19 (HOLD) - De-pin from ECU
  9. ECUD-21 (TFN) - Leave connected as is
  10. ECUD-15 (R) - De-pin from ECU
  11. ECUD-16 (2) - De-pin from ECU
  12. ECUD-17 (L) - De-pin from ECU
  13. ECUD-18 (OD1) - Leave connected as is
  14. ECUD-8 (SPD) - Is output from ECU to drive speedometer. Leave connected
  15. ECUC-12 (OIL) - Input from temp sensor in trans. I would use the actual temp sensor and clamp it to the engine so it gets readings.
With the items connected as above and not worrying about the VSS- and VSS+ signals, I believe you should be able to drive around all day with no CEL issues. The problem will be that your speedometer won't work, but that can be solved other ways.

The logic here is that the ECU has no idea if the trans is in D or N. It only knows if it is in R,2, or L. The absence of these signals indicate that it is in D or N.

I believe the ECU attempts to drive the trans solenoids just the same whether in D or N.

The resistors effectively simulate the solenoids.

With no VSS+/VSS- signals, it won't know the vehicle is moving and won't do much with the transmissions solenoids and should not throw any errors about it not shifting since it shouldn't be cross referencing the vehicle speed with gear selection/RPM changes. It may be wanting some specific signal on these pins even when the vehicle is stopped though, probably nothing more than a resistor, but I'd have to see a scope trace to know what it wants.

Don't try this at home, you might burn up your ECU. I have not tried any of this, it's all theory.

If this worked, the speedometer could be solved with a dakota digital hooked up to the SPD pin (depin from ECU) and using your own sensor.
 
So I guess to summarize, I theorize that three resistors and some de-pinning can make the ECU stop throwing transmission related codes. If you are game to try it, I would love to hear the results!
 
There's no way to modify the old ECU without having schematics and detailed documentation. There are two circuit boards - see image below of the processor board.

When you clear ECU codes then the various emission's related tests the ECU performs are marked as "incomplete" and you can't pass inspection with incomplete tests. Once you drive around some, and for at least one test it takes multiple driving cycles, the tests are marked as "completed." Of course by then the MIL light is back on.

89661-60652.jpg
 
sbman said:
I evaluated the EWD in relation to the auto trans. It does not have very many connections.

I can see that wiring the 'N' indicator to light up won't do anything because the ECU doesn't even detect this signal.

I do believe I could design a circuit to emulate the transmission and make the ECU happy in all conditions. I think it is possible to fool the ECU into thinking the transmission is there. It needs the following:
  1. ECUA-10 (S1) - Needs a 15 ohm, 20W resistor to ground. Might get warm
  2. ECUA-09 (S2) - Needs a 15 ohm, 20W resistor to ground. Might get warm
  3. ECUA-08 (SL) - Needs a 15 ohm, 20W resistor to ground. Might get warm
  4. ECUB-08 (VSS-) - MAY need a Vehicle speed sensor simulator. Requirement will need to be determined by o-scope tracing
  5. ECUB-08 (VSS+) - MAY need a Vehicle speed sensor simulator. Requirement will need to be determined by o-scope tracing
  6. ECUA-03 (HOLD) - De-pin from ECU
  7. ECUD-20 (PWR) - De-pin from ECU
  8. ECUD-19 (HOLD) - De-pin from ECU
  9. ECUD-21 (TFN) - Leave connected as is
  10. ECUD-15 (R) - De-pin from ECU
  11. ECUD-16 (2) - De-pin from ECU
  12. ECUD-17 (L) - De-pin from ECU
  13. ECUD-18 (OD1) - Leave connected as is
  14. ECUD-8 (SPD) - Is output from ECU to drive speedometer. Leave connected
  15. ECUC-12 (OIL) - Input from temp sensor in trans. I would use the actual temp sensor and clamp it to the engine so it gets readings.
With the items connected as above and not worrying about the VSS- and VSS+ signals, I believe you should be able to drive around all day with no CEL issues. The problem will be that your speedometer won't work, but that can be solved other ways.

The logic here is that the ECU has no idea if the trans is in D or N. It only knows if it is in R,2, or L. The absence of these signals indicate that it is in D or N.

I believe the ECU attempts to drive the trans solenoids just the same whether in D or N.

The resistors effectively simulate the solenoids.

With no VSS+/VSS- signals, it won't know the vehicle is moving and won't do much with the transmissions solenoids and should not throw any errors about it not shifting since it shouldn't be cross referencing the vehicle speed with gear selection/RPM changes. It may be wanting some specific signal on these pins even when the vehicle is stopped though, probably nothing more than a resistor, but I'd have to see a scope trace to know what it wants.

Don't try this at home, you might burn up your ECU. I have not tried any of this, it's all theory.

If this worked, the speedometer could be solved with a dakota digital hooked up to the SPD pin (depin from ECU) and using your own sensor.
Click to expand...

The VSS in the transmission is not used for the speedometer. The VSS in the output section of the transfer case is used for that purpose.

What do you mean "de-pin from the ECM"? Can you not simply disconnect the related connectors?

This is promising!
 
GLTHFJ60 said:
The VSS in the transmission is not used for the speedometer. The VSS in the output section of the transfer case is used for that purpose.

What do you mean "de-pin from the ECM"? Can you not simply disconnect the related connectors?

This is promising!
Click to expand...

The speedometer appears to be driven by the ECU from the SPD signal (pin ECUD-8). Refer to page 132 of the '96 EWD. Pin 3A of the combination meter appears to be the drive input for it, and the signal is generated from the ECU. I had assumed the VSS was in the transmission, but if it is in the transfer case, it will need to be disconnected as well. VSS- and VSS+ will definetly need to be disconnected.

By de-pinning I mean disconnected. theoretically all this could be done with a plug-in connector at the transmission connectors.
 
TrickyT said:
Your approach will not work. This has been discussed in this thread: H151 5spd conversion in parts for sale classifieds and this thread: A750F 5-speed auto tranny w/ paddle shifters in an 80? Yup.
Click to expand...

I can't tell from your posts how far you went with the testing to prove out your theory that it can't be done. Can you expand on your findings? As long as the ECU doesn't know you are moving, it shouldn't do any transmission stuff or throw codes. That's my theory at least from looking at the EWD and the associated notes.

There is another more complex (far more complex) potential solution, and it would actually have far reaching implications for tuners of 80 series in general. the idea would be to use a powerful microcontroller (STM32 or equivalent) to create a 'fake' OBD2 signal for the connector under the dash. It would get it's data from the real ECU (or any ECU you might happen to be using) in real time, and present it over to the diagnostic port. It would be hooked up to the CEL light and turn it on/off at the appropriate times. This would be a complex project and the legality of it is questionable. I'd never do something like that if it wasn't legal of course.... but there are open source libraries for OBD2 protocols and I do happen to be a firmware developer. just sayin.
 
With regards to the speed sensor, what if the transmission VSS is fed signals from the transfer case VSS via something as primitive as a vampire connector?

Nope, not if the point is to fool the ECU into thinking that the truck isn't moving.
 
GLTHFJ60 said:
With regards to the speed sensor, what if the transmission VSS is fed signals from the transfer case VSS via something as primitive as a vampire connector?

Nope, not if the point is to fool the ECU into thinking that the truck isn't moving.
Click to expand...

Exactly. you want the ECU to see 'solenoids physically connected' but no speed, ever. It should never try to match any shifting outputs with any throttle/map/etc... inputs because it isn't going anywhere. Will it ever do tests and go into closed/open loop and all that OK? Very likely that will be an issue and might the nail in the coffin that @TrickyT has already found, it sounds like he went pretty far with it.
 
sbman said:
I can't tell from your posts how far you went with the testing to prove out your theory that it can't be done. Can you expand on your findings? As long as the ECU doesn't know you are moving, it shouldn't do any transmission stuff or throw codes...
Click to expand...

I don't have an H151 transmission and didn't do any testing. All of my conjecture is from careful reading of the FSM and EWD. You may be able to avoid throwing transmission related codes by tricking the ECU into thinking it's standing still, but I would think it then couldn't do a very good job of controlling the engine either. At this point Johnny has nothing to lose by swapping back his original ECU and trying the 3 resistors.
 
From reading the rest of the posts as linked by @TrickyT the problem would be that the ECU won't do any of it's magic as far as readiness tests and open/closed loop without the VSS signal. With the VSS signal, it expects to see certain things happen when it tells the transmission to shift which would be impossible to simulate because of the real world (engine RPM can't be simulated, the ECU knows the real value). So the above posting I made could potentially keep the light off, but you still wouldn't pass smog because none of the readiness tests would be passed.

Man-in-the-middle approach to the ECU is the only reasonable solution then.
 
To expand on my man-in-the-middle approach, what I'm proposing is that you disconnect the OBD-II port from the ECU. Disconnect the CEL light from the ECU. Connect these to a custom processor that can be programmed the way you want it to be. Connect the Toyota ECU's OBD line to the custom processor. The custom processor is then programmed to let all of the OBD data go through as normal (pass through), EXCEPT transmission related codes. The custom processor would only turn on the CEL if the Toyota ECU was throwing an emissions related code.

This would be a large amount of work and quite complex. It would also work if you happened to have, say... a non toyota ECU controlling the engine. You'd just need to be able to read the data you need (RPM, temperature, timing, 02 Voltage, etc....) so that you could put out all the OBD data the smog computer wants to see.

I doubt that would be legal. no one should do that.
 
The only A343 tranny trucks that this swap will work on is the 1995 , at least here in MA as they scan 1996 and on.

I posted somewhere that my solution for this problem was to use a ECU out of a Supra . The only difference I could find was those use two forward O2 sensors .Otherwise it looks the same right down to a MAF sensor and similarly sized injec
 
landtank said:
The only A343 tranny trucks that this swap will work on is the 1995 , at least here in MA as they scan 1996 and on.

I posted somewhere that my solution for this problem was to use a ECU out of a Supra . The only difference I could find was those use two forward O2 sensors .Otherwise it looks the same right down to a MAF sensor and similarly sized injec
Click to expand...

I don't *think* they scan a 96 here in NC, but I'm working on a way to verify.

I was unaware that Supras used the 1FZ, or something close enough to emulate it. Which model Supra were you thinking?
 
It's the NA 95-97 Supra .Same achitectute with the exception of having two front O2.

I have the service manual and EWD for the Supra and have checked it out .Some of the sensors have slightly different ranges specd but nothing crazy.

Our injectors are rated at 305 cc/min while the Supras are 308cc/ min. That difference is a non issue as far as I can see .

As a side note, there wee three versions of the ECUs just as there are for our trucks . The production dates for both platforms of The ECUs are the same .

It looks like both systems were designed together.
 
No they don't. It's the newer style. All the connectors at the sensors and such are the same though. You would need to splice the appropriate pig tale onto the lc harness and then add the additional wiring for the extra O2 sensor.

I have the ECU, pig tale and good used harness .Just need the 93/94 exhaust manifolds to mount the front O2 sensors in
 
You must log in or register to reply here.

Similar threads

S
Problem - H150F manual gearbox
Replies
1
Views
384
Sondchara
S
R
Fuel issues after manual swap
2
Replies
22
Views
2K
Stuckinthe80s
Stuckinthe80s
fjbj40
  • Question Question
Manual flywheel
Replies
0
Views
310
fjbj40
fjbj40
Catsxb
  • Question Question
DEQ problems
Replies
14
Views
906
Catsxb
Catsxb

Users who are viewing this thread

Total: 1 (members: 0, guests: 1)

Share this page

Back
Top Bottom