Memory dump of ECU

[b3lha]

Originally Posted by Trevor

As a result of this post I will no doubt again be accused making an attack, but so be it. Incorrect information is not in the interests of members and requires correction. This particularly applies in respect of the arduous work covered by this thread.

Thanks Trevor. I do not read your post as an attack, merely a differing opinion. I asked Harvey to post here because I value his opinions. I also value yours and those of anybody else you wants to chip in.

Originally Posted by Trevor

Solenoid valves 1, 2, & 3 are normally open devices and close and shut off fluid when energised (electrically turned on). In this state they do not bleed off pressure and close the bleed so that pressure is applied.

These normally open solenoid valves are turned ON in order stop pressure bleed off and as a result then provide control pressure in respect of the selection of ratios 2, 3, & 4.

Solenoid valves A, B, & C, in the case of transmissions incorporating a centre clutch, are all normally closed devices. Solenoid valve C is a normally open device in the case of transmissions incorporating a centre differential.

In regard to this project, I don't see that how the solenoids operate is important. That subject has been done to death a few times already. The information I am looking for is why the solenoids operate. What is the TCU trying to achieve when it increases or decreases a given duty cycle. What is the nature of relationship between, for example, the solenoid C duty cycle and the torque split.

Originally Posted by Trevor

Thus in both systems the respective C Solenoid valve is near fully energised for virtual lock up of front rear, without requiring a different configuration of electrical or hydraulic systems. Both C valves bleed off fluid as a means of control.

That's the sort of thing I'm looking for. On my car, with the JDM box, when the stick is in N, solenoid C is at 5% duty. On Nomake's car, with the USDM box it is 95%.

I don't know how the duty cycle relates to torque split on the JDM box.

On the USDM box, Nomake is assuming that 95% is FWD because it starts out at 95% and drops on hard acceleration when the TCU is presumably transferring power to the back. You seem to be suggesting the opposite. This discussion is certainly worth having.

[b3lha]

Originally Posted by Nomake Wan

Anyway, some observations from this run.

1.) Those weird secondary percentage indicators appear on all three of the Duty Solenoids. Bug in the GUI.

2.) Duty Solenoid B remains at 5% when unlocked. When the torque converter locks, the solenoid jumps to 50% and then transitions from 50% to 95% gradually. If the transmission is in "Normal" mode when you disengage the cruise control, it drops to 55%, but does not go down to 5%. So when driving around without Power Mode, the torque converter appears to be in a semi-locked state. Odd. It'll go down to 5% if you start to brake for a stop, though. Presumably to prepare for sudden acceleration.

3.) Duty Solenoid C is at 65% about 95% of the time when driving around normally. So whatever 65% represents torque-split wise, that's what the USDM cars maintain. Power or Normal mode doesn't make a difference. Also, as before, the only time that Duty Solenoid C went to 95% was when the shifter was in P or N. The most I've ever seen it at while driving was 70%. And the lowest of course was 5% when launching from a stop.

4.) Duty Solenoid A remains at about 58% most of the time when just cruising, and alters based on a direct relation to throttle input.

I also took a better screenshot... I'm a sucker for clarity.



Hope this helps some!

Thanks Nomake,

A superb write up as always. This is getting interesting. I hope we can make some sense of it all. I'll try and fix up the software a bit later this week.

[Trevor]

Phil,

You are unable to access what is going on if there is confusion between what happens as a result of a control signal energising or alternatively de-energising a device. If your expectations are back to front you will be looking for a paddle.

Sincerely, Trevor.

[svxistentialist]

Originally Posted by b3lha

I don't know how the duty cycle relates to torque split on the JDM box.

On the USDM box, Nomake is assuming that 95% is FWD because it starts out at 95% and drops on hard acceleration when the TCU is presumably transferring power to the back. You seem to be suggesting the opposite. This discussion is certainly worth having.

I'm a layman when it comes to automatics, but this fundamental difference is what I would expect, given that we are dealing with fundamentally different machines.

The US gearbox is essentially a FWD box with rear assist, Subaru calls it the Active Torque Split AWD box, although a very clever and effective rear assist, unlike the rubbish from Honda and others.

The JDM/EU gearbox is the fulltime AWD box, they call this the VTD-AWD box, and is biased [around 65% if memory serves] to rear drive before the TCU starts sharing torque around.

Now if you take both these vehicles and give them a hard launch, you should get exactly what you are describing here; Trevor's jdm being rear bias, if it loses traction at all it will likely be at the rear, so the computer will shift the torque to the front to eliminate slip.

In the US car, it starts off front biased, if you launch it hard the front wheels will lose grip, and the computer shifts torque to the rears to avoid slip.

Two differently biased systems, they have to work in opposite directions where slippage occurs.

Joe

[Trevor]

Phil,

Re the question i overlooked ---

On the JDM box a signal providing maximum energised time and current in respect of normally open solenoid C, will increase front rear drive to virtual lock up. You can prove this by inserting the control fuse. Useful if you get stuck in the mud.

The US arrangement with a normally closed C solenoid, reduces pressure by opening the solenoid and consequently the clutch is not operated and there is no front wheel drive. Proven because a faulty C solenoid results in lock up.

This arrangement prevents the need for mechanical alterations in respect of the hydraulics used in the two systems, as was proposed some time ago by Harvey.

I trust that you now are aware just why solenoid operation must be correctly understood. Bringing up trivia was certainly not in my mind.

[b3lha]

I'm having trouble getting my head around this, so I'm going to quote something from the USDM manual.

Originally Posted by USDM 1992 Factory Service Manual Section 3-2 Page 177

Transmission Control Unit (TCU) I/O Signal
Duty Solenoid C, connector B68, pin 3.
Fuse on FWD switch: 8-14 volts
Fuse removed from FWD switch, with throttle fully open and select lever in 1st gear: Less than 0.5 volts.

I know you can't measure a duty cycle with a voltmeter, but I interpret this as follows:
When the car is in FWD mode. No power to the rear, solenoid C is fully energised. A high duty cycle (100%).
When the car is in AWD mode and the stick is in 1 and the throttle is fully open, going for max acceleration, solenoid C is almost completely de-energised. A low duty cycle.

This seems to agree with Nomake's observations that a lower number indicates lockup (F50:R50) and a high number indicates power towards the front (F90:R10).

[Nomake Wan]

Originally Posted by Trevor

On the JDM box a signal providing maximum energised time and current in respect of normally open solenoid C, will increase front rear drive to virtual lock up. You can prove this by inserting the control fuse. Useful if you get stuck in the mud.

The US arrangement with a normally closed C solenoid, reduces pressure by opening the solenoid and consequently the clutch is not operated and there is no front wheel drive. Proven because a faulty C solenoid results in lock up.

This part really got my attention. So basically, all they did was reverse the solenoid to adjust for the different transfer clutch mechanisms. Certainly a cost-effective method, and fully explains why energizing (95% duty cycle) Solenoid C on a USDM car results in FWD.

[b3lha]

Here's something else interesting from the service manual.

The control unit stores optimum transfer clutch torque data for a variety of driving conditions. When actual driving conditions (vehicle speed, throttle opening, gear range, wheel slip etc.) are detected by various sensors, the control unit selects a duty ratio most suitable to the given condition from memory.

It would be interesting to see if I can find these maps.

[b3lha]

Originally Posted by Nomake Wan

This part really got my attention. So basically, all they did was reverse the solenoid to adjust for the different transfer clutch mechanisms. Certainly a cost-effective method, and fully explains why energizing (95% duty cycle) Solenoid C on a USDM car results in FWD.

I don't know about hardware differences in the trans, but there are certainly software differences in the TCU.

I don't really understand the part where Trevor says,

Originally Posted by Trevor

The US arrangement with a normally closed C solenoid, reduces pressure by opening the solenoid and consequently the clutch is not operated and there is no front wheel drive. Proven because a faulty C solenoid results in lock up.

If the solenoid is normally closed, then a high duty cycle will open it. I believe that a high duty cycle gives FWD but Trevor seems to be suggesting that opening the solenoid results in "no front wheel drive".

[Trevor]

Originally Posted by b3lha

Here's something else interesting from the service manual.

It would be interesting to see if I can find these maps.

Phil, I find this BS originating from the sales dept. I could be wrong but keep that in mind and don't get screwed up on it.

Why would the system not simply adjust to the prevailing conditions, while include what amounts to a time delay and an upper and lower threshold, so that a mean is established. This is normal within industrial process control, but the automotive world glorifies the process as a sort of learning thing.

Both you and Nomake in your last posts, show that you have cottoned on to what is correctly involved outside of your black box. Well satisfied, you can relax a little, withdraw within and shut the lid again. Hooray!

[Trevor]

Originally Posted by b3lha

I don't know about hardware differences in the trans, but there are certainly software differences in the TCU.

I don't really understand the part where Trevor says,


If the solenoid is normally closed, then a high duty cycle will open it. I believe that a high duty cycle gives FWD but Trevor seems to be suggesting that opening the solenoid results in "no front wheel drive".

When the normally closed solenoid is energised via a high duty cycle the valve will open and control pressure will be bled off. Reduced pressure will result in the clutch opening and therefore no FWD. N.B. Edit --- by this I was indicating, no Four Wheel Drive, i.e. only Front Wheel Drive. Sincere apologies. I now realise that "FWD" is totally ambiguous.

[b3lha]

Originally Posted by Trevor

When the normally closed solenoid is energised via a high duty cycle the valve will open and control pressure will be bled off. Reduced pressure will result in the clutch opening and therefore no FWD.

That's what doesn't make sense.

The quote from the service manual and the results of Nomake's experiment suggest to me that a high duty cycle is FWD and a low duty cycle is 50:50. Furthermore I was under the impression that "no FWD" (ie. RWD only) is an impossibility on a USDM tranmission.

Please explain.

[Nomake Wan]

I think it's a mistake in wording. To make things more simple, I believe it is this:

JDM: Normally Open. Closing [Energizing] the solenoid will activate the transfer clutch and result in 50/50 split. A failed solenoid will default to the 35/65 split.

USDM: Normally Closed. Opening [Energizing] the solenoid will deactivate the transfer clutch and result in FWD. A failed solenoid will default to 50/50 split.

This also makes sense since opening a solenoid decreases line pressure (deactivates clutch) whereas closing it increases line pressure (activates clutch).

[Calum]

Originally Posted by b3lha

I know you can't measure a duty cycle with a voltmeter,

Some multi-meters can actually. Fluke has made models that can for years, and some of the cheap chinese meters picked up this feature (lucky for us cheapskates).

http://www.mouser.com/search/Product...U97kdFdw%3d%3d

For $30 it might be worth picking one up just to test your theories.

[Calum]

dang it, double post. Sorry!