4L80E code

[Bernard Fife]

Guys,

Here is some better code. It has a number of small changes that *should* improve things on the shift to neutral issue and possibly the manual mode selection issue as well:

Monitor_4141.abs.s19

(72.71 KiB) Downloaded 478 times

GPIO_MShift_4141.ini

(218.44 KiB) Downloaded 484 times

Lance.

[ashford]

tested this code out on a 4r70, it seems to make the problem worse for me.
what i have observed is
1. when error happens cgear rapidly switches from 1 to 2. tgear reports 0 but physically in 2nd
2. any amount of speed being read when switced to n or in n triggers it.
3 will only come out of it at a stop when switched to n then back into gear, it corrects random.


i have 2 logs, one of driving around and one in my driveway rolling down the slope in neutral then into gear

[Bernard Fife]

Ashford,

Okay, this is very helpful info, and should narrow the problem down quite a bit. I will post again when I have a solution - hopefully soon.

Lance.

[Bernard Fife]

Ashford,

I did find something in the code, and this attached code might fix it. It uses the same INI as before:

Monitor_4141d.abs.s19

Test code.

(72.71 KiB) Downloaded 484 times

Let us know if it is any better (or worse). At very least a datalog of this code in action will narrow things down even more...

I apologize that this has not been straightforward to sort out. The code is surprisingly complex in this area due to the interaction between the various parameters and possible usages, as well as the recent addition of user-adjustable 'fault tolerance' in the shift lever position.

Lance.

[ashford]

my 1 minute drive to work showed no error, i tried 2 times and nothing. it looks good ill test further after work

[ashford]

ok i took a bit of a drive with it and it is almost there. when shifting to n while moving than back into drive makes it shift to 1st. ther are 2 things that can bring it back
1. breaking the load hysteresis 20% tps for me
2. going into a spot on the shift table where there is 1st then leaving it will shift.

[Bernard Fife]

Ashford,

Okay, that's very helpful info. I will look into the code and put some fixes in place. I will likely post improved code tomorrow. Thanks for both your help and your patience with this!

Lance.

[Bernard Fife]

Ashford,

This code might help:

Monitor_4141f.abs.s19

(72.82 KiB) Downloaded 463 times

GPIO_MShift_4141.ini

(218.44 KiB) Downloaded 488 times

This code clears the hysteresis flags when a lever error is encountered (to allow shifting immediately). It also has some changes to the target gear code when the lever has an error condition.

Lance.

[ashford]

it seems its a one step forward and 2 step back deal. the d version seemed to work well with the exception of needing to break load hysteresis. the f version has the 1-2 oscilation but only ocurred the first time it saw neutral, not since, i wasnt even moving that time. the need to break load hysteresis to shift is still present, the oly difference i see between this code and the d version is that the d version had an immediate shift on a manual 1-2, and f version needs a throttle blip. 2-4 needs a throttle blip as well on both versions but goes from 2nd to desired gear instead of through 1st.

speed hysteresis seems to have no effect on resetting a stuck gear only load. another point i can bring up is that in the log where i stand on it, it shifts from 3rd directly to 1st and i doublechecked that i have it set to sequential the log confirms what i felt it went straight to first i normally feel 2nd get bumped in that situation.

not sure if this is related but slip seems to report garbage for the last ferw versions i checed in memory. i also noticed on a 1-2 shift only pc is not reduced.

something unrelated is i see about a 5mph jump in speed while the tcc is pwmed, if memory servs me when i set up the controller i put an inline resistor to get a total of 10 ohms and also a flyback diode. are there any other suggestions for this.

[Bernard Fife]

Ashford,

It appears that the problem is caused by the target gear being set to -1; that's when the oscillation begins (in conjunction with a gear lever error).

You might try the 4.142 code from this thread: viewtopic.php?f=4&t=829&p=4337#p4337

It has a number of improvements in this area.

Also, does it help if you set the code to ignore the lever errors with the latest code(s)? ( http://www.msgpio.com/manuals/mshift/V41tune.html#ge )

Let's deal with one issue at a time (or at least one per thread - you can start another thread for other issues), otherwise the conversation gets very confusing.

Lance.