TCC behaviour

[gui67]

I found 2 weird behaviour of the TCC:
The minimum TCC lock-up speed is 50km/h. When my speed is just on 50, the TCC locks and unlocks very quickly:

TCC50kmh.png

Then the second is that the TCC stays locked:
Max TCC load(kPa) is set on 95 and Lock TCC above Temperature (°C) is set on 100.
even with hig load (MAP=112kPa) and a temperature of 82°C, the TCC stays locked.

TCC 112kPa.png

Is it possible that the temperature seen on Tunerstudio and on the log is wrong or does the soft uses a temperature of 80°C to lock the TCC? (With 78°C Trans temp, the TCC unlocks)
I think what I see in the log and in Tunerstudio is right: When I start the car, the temperature shows 23°C which was approximately the temperature of the garage.

[Bernard Fife]

gui67,

There have been a few fixes in the TCC logic in the last few weeks. One of these fixes a temperature offset issue in the INI, and this is likely the cause of one of the issues you raise. So try upgrading from your 4.125 code to 4.127: http://www.msgpio.com/manuals/mshift/V41code.html to see if that helps.

You have your TCC hysteresis time set to just 1.3 seconds, so the TCC can switch on/off every 1.3 seconds. If this was mine, I would set it to at least 5 seconds, and probably 10 seconds would be even better (you can set it as high as 25.5 seconds). That should help the TCC cycling issue.

Also, I don't think you have this problem, but others that stumble upon this thread might also check their full time TCC mode settings to make sure it isn't set to be on when it's not wanted (full time TCC is intended to allow the TCC to lock at high speeds to allow a little more top end speed). See: http://www.msgpio.com/manuals/mshift/V41tune.html#nt

Lance.

[gui67]

Thanks Lance,
I will try that new code for the temperature offset.

but for the On-Off switching, I first tried a longer time but prefered shorter and decrease it to 1,3s. If I set it to 10s, theneverytime I brake I have to wait this time until it goes on again... I felt it was just on a few seconds before I need to brake again.
by the way, on the log, the TCC switches 6 times during 3.5 seconds: it did not used the TCC hysteresis time.
Would it be possible to add a TCC speed hysteresis?

[Bernard Fife]

Would it be possible to add a TCC speed hysteresis?

gui67,

Yes, this should be possible.

Lance.

[gui67]

Lance,
Do you have an idea why it switched 6 times in 3 seconds, the time hysteresis was not used.
Guillaume

[Bernard Fife]

Guillaume,

I would need to see a log to try to figure it out. In general, if the value is set to 1.3, that is the time after it turns ON before it can turn ON again. Turning OFF is not limited (so that things like brakes will unlock it immediately).

So it could turn off any time, then turn on again after 1.3 seconds, then off right away, then on again at 2.6 seconds, then off and on at 3.9 seconds, and so on. The serial logs don't keep the time perfectly synchronized with the parameters either, so you have to allow a margin for that too.

Lance.

[gui67]

I saved the portion of the log where it appened.
There is definitively less then 1.3 sec between 2 turning ON. I also felt it while driving

I though the hyseresis was the minimum Off time, not between 2 turning ON. So I could try to make it a bit longer then.

[Bernard Fife]

Yes, something is odd as shown in the log, because while TCChyst counts down initially, it is never reset when the TCC is turned on again. I will have a look in the code.

[Bernard Fife]

Guillaume,

Here is some test code. It uses the standard 4.127 INI. This code should:
- start a counter to count down the TCC hysteresis (the user value, it is in the datalog) anytime that TCC is called to be locked.
- TCC will only lock again if and when the counter reaches zero.

Note that the TCC can be called to be locked even when it is already locked. So the TCC won't lock again until the TCC is not requested to be locked for the hysteresis time, and then it is called to be unlocked. If it is called to be locked again any time before the timer expires, the timer starts over again.

I think this should work better (surprisingly, the TCC logic is one of the most complex parts of the code, and the logic is both wide-ranging and subtle). But we will get this sorted out!

Lance.

[gui67]

Hi, lance,
I tried it this morning, the TCC did never went on.
look at the log in the link below:
http://g.scheidt.free.fr/logs/2013-07-29_07.15.22.msl

Guillaume