TCC behaviour

[Bernard Fife]

Guillaume,

Okay, thanks for testing that, it was worth a shot. I'll have a much closer look , and post something here when I have a likely solution.

Lance.

[Bernard Fife]

Guillaume,

Here's a better effort:

Monitor_4127b.abs.s19

(71.1 KiB) Downloaded 771 times

The INI has a new gauge for the TCC hysteresis counter ("TCC Hyst") so you can see what the code is 'thinking'.

GPIO_MShift_4127.ini

Adds TCC hysteresis gauge

(215.25 KiB) Downloaded 769 times

The new code appears to work as designed on my bench.

This code:
- starts a counter every time the TCC unlock function is called,
- does not let the TCC lock until that counter has counted down to zero.

That means the TCC will not be applied until after the conditions to lock it have been sustained for a while. If those conditions are then not met, the TCC unlocks immediately, the counter is reset, and the TCC will not lock again until the counter expires. This means the TCC can lock and then unlock quickly in some circumstances, but then it will not lock again for a while (the default is 7.5 seconds).

There were a couple of problems causing the problems you noted, one was that the TCC PWM was updated in the clock interrupt, but this didn't take the TCC hysteresis into account (it does now), and the TCC hysteresis counter was not initialized properly (so the first on/off cycle could happen rapidly).

I remembered why we don't have speed hysteresis on the TCC. Suppose you have the cruise control on, but hit a hill. The throttle increases (to maintain a constant cruise speed), and the TCC might reach conditions should unlock, despite the constant speed. In fact, there's exactly that circumstance on my daily drive, so it can't be too uncommon. If we had speed hysteresis, the TCC wouldn't unlock then at all, which isn't good.

Lance.

[ashford]

Guillaume,

I remembered why we don't have speed hysteresis on the TCC. Suppose you have the cruise control on, but hit a hill. The throttle increases (to maintain a constant cruise speed), and the TCC might reach conditions should unlock, despite the constant speed. In fact, there's exactly that circumstance on my daily drive, so it can't be too uncommon. If we had speed hysteresis, the TCC wouldn't unlock then at all, which isn't good.

Lance.

i had an idea for this a while ago without an actual hysteris. have the normal min speed to activate tcc then a speed setting to deactivate tcc a few mph lower than activating. i found it quite annoying whenever the speed limit was the same as tcc min speed, so i modified the ini toallow small numbers and set it to 20 mph and minimum gear to 3 and worked around it but the torque converter is always locked as soon at it gets into 3rd and below kpa limit.

[gui67]

Hello Lance,
I tried the new code yesterday evening.
The TCC now works again, no more temperature problem or quick on-off.

However, by looking at the logs, it seems that sometimes, it goes on despite a high load. Max load is set at 95kPa.
I have one example on a log where the TCC counter counts down to 0 and locks the converter while the boost is at 115kPa.
and a second example where boost increases above 95kPa during the countdown, without resetting the counter.

But on some other times, the converter unlocks during boost or the counter is reset. It might be due to communication between MS3 and MSGPIO? If the GPIO does not get the MAP information often enough or something like that.
Would logs help you?
Guillaume

[gui67]

Lance,
you can see at 504s what I said: the boost increases and the TCC stays on.
(I now realize that it only stayed less then 1 second.)
I first thought it was a longer time.

[gui67]

In this one, you can see that the counter counts despite the fact that boost is higher then 95kPa.

TCCboost.png

[Bernard Fife]

Guillaume,

Yes, the counter counts down regardless of the level of MAP, *unless* the TCC is called to be unlocked, in which case the counter is reset to the user value to start counting down again. In your case, the trans is shifting, and the TCC MAP isn't checked while shifting. So the counter counts down during the shift and is reset, then starts counting down again after the shift because the MAP level is much lower. It counts down to 1.9 before the MAP exceeds 100 kPa, then it is reset. That's the expected behavior.

The important thing is that the TCC is never engaged when the MAP is > 100 kPa (except during a shift; or if full-time TCC parameters are met, or the temperature is high, etc., etc.).

Lance.

[Bernard Fife]

Guillaume,

Also, while the TCC always unlocks on downshifts, upshift TCC behavior is user configurable. See: http://www.msgpio.com/manuals/mshift/V41tune.html#nu

Lance.