MSgpio.com read only forums (Jan.15,2018)

Gurov,

I have attached 4.004c code that backs out virtually all of the changes made to the shift lever function made between Feb.18 and Feb.26 of this year (around the time 2.00Fx was released). So it *might* work for you, or at least help us know what isn't the problem.
This uses the same 4.004 INI as above.

Lance.

here are the other two videos

http://www.youtube.com/watch?v=eRvAqB6JdWo


this is 200Fx_test working through the same series of tests, same msq

http://www.youtube.com/watch?v=z-Fnha3l5HU

gurov,

I can't tell a lot from the videos, but it sounds like you are shifting continuously with the lever (less than a second between shifts). You probably shouldn't do that if your debounce value is high. Give the controller at least second or two to recognize the new gear.

But also try the code I just posted.

Lance.

Lance wrote:gurov,

I can't tell a lot from the videos, but it sounds like you are shifting continuously with the lever (less than a second between shifts). You probably shouldn't do that if your debounce value is high. Give the controller at least second or two to recognize the new gear.

But also try the code I just posted.

Lance.

will do tomorrow, possibly. i'm going away for a week to california, so i won't have the cars to play with. the 2nd video has more delays where i wait between shifting, and it does work, as soon as i shift into D (which is 5th) it stops recognizing the inputs.

3rd video has me doing the same pattern, and shifting into D as well, and you can see it count down from 5th to 1st.

i'll try the new code though.

exactly the same situation at 500 - 6000 debounce, with restarts between, and giving plenty of time between shifts.

back to 200Fx_test.

i have logs that i will upload and msq

i AM setting shift_mode to 0, dont know if that would make a difference.

it is also not picking up whether the up/down shift paddles are being triggered and stays in manual mode, with both up/down being reporting active.

MSQ and datalog(s)? BTW, have you tried 2.103? It has fewer changes than the 4 series code, and if it works for you this would help narrow down the problem to the changes made for 4.00x. If it doesn't work for you, then the problem must be in the transition from 2.00Fx to 2.103. Either way, knowing this will help get to the issue quicker.

Lance.

Lance wrote:MSQ and datalog(s)? BTW, have you tried 2.103? It has fewer changes than the 4 series code, and if it works for you this would help narrow down the problem to the changes made for 4.00x. If it doesn't work for you, then the problem must be in the transition from 2.00Fx to 2.103. Either way, knowing this will help get to the issue quicker.

Lance.

i will try to install 2.103

gurov,

If you could do that, it would be great. If 2.103 works for you, I will add 2-3 and 3-2 shift pressure limiting (as 2.104 code) so you have something to use while I sort out the 4.00x code.

Lance.

gurov,

I had done some thinking on this, and think I *might* know what is going on. I believe the issue might be related to setting the default shift mode to manual (normally it can be only one of the two auto-shift modes, but I believe you have yours set to manual mode on start-up). I will start testing this theory shortly, and fix the code where I can to make it work better.

In the meantime, if you want to try setting the default shift mode to one of the auto modes (1 or 2 in 'General Settings/Shift Factors/Auto Shift Mode'), this might allow you to use the 4.00X codes. You would have to select manual after engaging drive, though.

While I am doing this, I will also set the shift buttons so that the first press f either shift button while in auto mode changes to manual mode, but does not shift. The second and subsequent presses will do the manual shifts, of course.

Lance.

gurov,

What I think is going on isn't really a code problem at all, instead it's a conceptual design limitation.

When you shift from PNR to Drive, the function of the shift lever changes. In PNR, the lever sets the gear, with no ambiguity - the selected gear follows the lever position. However, when you shift to Drive, the lever doesn't set the gear at all, it just sets the highest allowed gear (5th in your case). The controller doesn't know if you want 5th, or 4th, or 3rd, or 2nd or 1st.

In the older codes, the first gear to come up was the highest gear, so the controller used that, then downshifted based on the current rpm and the uner-rev limiting algorithm into first (the clunking down the gears you described). However, in the new code, the logic was tightened up, exposing the confusion over what gear the controller should aim for. As a result, the code becomes very confused if you are in manual mode on a shift from PNR to Drive.

In auto mode, on a shift to Drive the target gear is then grabbed from the shift table, the appropriate gear is selected, and all is well.

However, if you are in manual mode when you shift from PNR to Drive, the code has to know what gear to shift to. The code used to assume you wanted first, but users (including you perhaps) didn't like that behavior (it was throwing the trans into 1st at speed if there was a reset). This could also happen if the user was coasting down a hill in neutral, then shifted into gear (not recommended, but some people do it). So the gear was left undefined. However, undefined is a bad thing for the control logic!

This was the reason I originally prevented the use of manual mode in start-up, but it took a long time for me to remember that (so I am sorry for advising you that you could try manual mode for start up simply by editing the INI - I believe no-one else did this, and that's why no-one else was seeing the problem you reported).

In theory, we could have an algorithm to select an appropriate gear when entering Drive, but that's exactly what the auto-mode is, and it seems inefficient to have a separate mode for just that.

So the codes I will release in the next few days will revert to auto_mode any time you shift to PNR, and will prevent the user from overriding the selection of sequential or skip shift auto_modes for start up. Users will have to select manual mode only once they have moved the shift lever to a forward gear position. This way there is no confusion for the controller over what the current gear should be.

However, if it helps at all, I have added the function you requested where the first press of a shift button changes the mode from auto to manual, and only the second and subsequent presses actually shift gears. That way if you are in auto mode and first gear (or any other forward gear) a press of either shift buitton will put you in manual mode, but leave the trans in the same gear (first in this case). The second press of either shift button will then change the gear.

The code is done, but I need to do a bit of testing before releasing them (they will be 2.105 and 4.004, announced in the "Latest Code Updates" thread: viewtopic.php?f=4&t=20 )

Lance.