4126b and earlier issues with voltage gear lever

[ashford]

it seems to work, only gotcha is that hysterisis has to be broken after physical shift to electically shift it

[Bernard Fife]

ashford,

Okay, that's some progress.

Yes, you need to exceed the hysteresis settings, because this is still an auto mode shift. The way to avoid that is to use manual mode (and the shift buttons).

Lance.

[Bernard Fife]

ashford,

It eventually occurred to me that we can clear the shift hysteresis flag whenever the new gear lever position doesn't match the previous position, so the shift will happen right away if the shift table calls for it. I will also set up a gear lever timer, so if a valid gear lever position isn't found after a specified length of time (say 2 seconds as a default), an error is reported.

These will be in the next version of test code, which I will post here in a day or two.

Lance.

[ashford]

ok ill give it a shot when it comes, im currently installing a 4l80 into a 70 nova i need to finish before i tear mine apart again. it appears i have a pinhole in the converter at a weld.

[Bernard Fife]

ashford,

Here is some 4.127k test code:

Monitor_4127k.abs.s19

Use new INI (attached)

(71.32 KiB) Downloaded 751 times

GPIO_MShift_4127.ini

For use with 4.127k test code.

(215.29 KiB) Downloaded 819 times

This code adds the capabilities mentioned above. The INI has an added parameter (under General Settings -> Error Behavior) that let you set a time during which lever errors are ignored (the default is 2.0 seconds, the values can range from 0.3 to 25.5 seconds). When a gear error occurs, a timer is started. Until the timer period expires the last known valid gear lever position is reported and used by the code.

Also, manual lever position changes will now cancel all hysteresis gear shift and TCC restrictions.

If these work well, they will become standard in the next release.

I haven't tested this code much (but have loaded it and verified some aspects), so use it with caution, and be ready to switch back to older code at any time if necessary (i.e. don't load this code and head out on a long trip without your laptop!).

Lance.

[ashford]

i tried out the k code. it act the same as the base 4127. the only time that a lever movement does not force 1st is a 4th to 2nd downshift but not always.
the earlier msq is a ignore time of 2. the second is an ignore time of 20

logs and msq attached

[Bernard Fife]

Okay, thanks. I'll have a look.

[Bernard Fife]

ashford,

I think I see what is happening.

The code checks if the lever position is invalid, if it is then the code starts a timer. Otherwise it uses the found gear lever position.

But think your lever is going through a state where the gear voltage is that for first gear (even if briefly). This is valid, so it uses it. But then the lever voltage enters an invalid state, so it continues to use first gear (and reports -99 when the timer expires).

So it is not just invalid lever positions (i.e. invalid voltages) that cause the problem. Instead it is any lever position that is either:
- invalid, OR
- more than one away from the current lever position,
that is a problem and should start the timer.

So I need to fix this in the code, and will post new code here when that is done.

Lance.

[Bernard Fife]

ashford,

Here is new test code. In addition to checking for lever position error conditions, this code will check if the current gear and newly selected gear are forward gears. If they are, they are only allowed to differ by a maximum of 1 gear (the default, this is a user setting under 'Shift Input Patterns -> Lever and Shift Buttons Configuration'). If the new gear is different by more than that (such as when you are in forth and the lever indicates first), it is considered to be an error, and ignored until either:

- a valid gear is selected, OR
- the gear error timer (as above) expires.

In most cases, I expect that a valid lever position would be found before the timer expires (especially if the error state is caused by being between lever positions).

Monitor_4127m.abs.s19

4.127m test code, use INI below.

(71.46 KiB) Downloaded 756 times

GPIO_MShift_4127.ini

For use with 4.127m code

(215.34 KiB) Downloaded 742 times

The heart of the problem is that we still need to be able to go from 4th to PNR, etc. so I am not sure if this will work. This code is untested. If you decide to give it a try, use great caution.

Lance.

[ashford]

i gave the m version a try, 1-2 works and 2-1 obviously works. 2-4 or does not and 4-2 sometimes works. i set max gear change to 2 and it is still the same. v2 and 4116 and earlier(very little testing though)of the code i tried did not have this problem or the "h" version. there are a few things i think i can try to see if there is arguments in the code. one i can try is to change solenoid states in neutral to something different, currently it is the same state as 1st. the other thing i can try is skip shift instead of sequential.