Adding a generic PWM output for increased trans support

[gross polluter]

It appears that a few of us are either running or will soon be running an Aisin Warner transmission on Mshift. I've been running an A340LE from a MKIV Supra Turbo for close to 10k miles on Mshift. The transmissions control functions are close enough to the 4L60e that out of the box with a few circuit mods Mshift works just fine. However, in addition to Line Pressure and TCC, many of the AW transmissions have a solenoid that controls the backfill on the accumulators (SLN in the wiring diagrams.) With more people running transmissions other than what's been documented on this site, I think a good feature of Mshift would be an AUX PWM table for transmissions that have more complex functions, such as the accumulator backfill found on the AW transmissions. I think similar functions in other transmissions refer to this as accumulator pressure control.

After probing a stock MKIV supra, here's what I found on the operation of the accumulator backfill:

-In any shifter position the PWM duty on the SLN solenoid is 20% at 300hz
-The PWM duty on the SLN only changes when there is a shift event. On the MKIV supra the output changes about a half second before the shift through a half second after the shift, about two seconds total
-When the PWM duty on the SLN changes during a shift the duty cycle directly reflects the load on the engine: Higher engine loads have lower duty cycles during a shift.
-According to both the factory service manual and my findings, Higher PWM duty cycles equate to more accumulation and softer shifts, lower accumulator pressure.

It seems that such functions could be accommodated with a simple generic PWM output table. The PWM frequency in the case of the AW transmissions is the same as the line pressure control solenoid so I'd imagine you could select an unused output pin that's common with the timer on the line pressure output, presumably Output 3 or Speedo output for those not using the speedo output. The table would have the same structure as the line pressure table, load vs. vss. If the table could be as large as the line pressure table that would probably work best, but minimally it should be as large as the shift table. I think to broaden the functionality of this output there would be a few things that dictate its behavior: Does the output follow the table all of the time or does the output only occur during a shift. If the output occurs during a shift, how long before a shift does the output occur and how long does that event occur for? If the user selects the function that the output only occurs during a shift then it would be nice to have an option to define a single value that the PWM duty is at when the output is not affected by the table.

I think such a feature would help broaden the support of more transmissions on Mshift. Thoughts? Questions? Possibility?

[Bernard Fife]

gross polluter,

It's certainly possible, but I will have to think about it a bit. We aren't necessarily tied to a timer output, as we can always bit-bang an output to PWM it (Outputs 1,2,4,5, & 6 are done this way). Using the speedo output allows us to potentially use a wider frequency range, but I suspect lots of people are using the speedo output. Using another, less commonly used output might work better, but bit-banging is limited to around 2000 Hz max. (with better resolution at lower frequencies). However 300 Hz is no problem when bit banging, and gives us adequate resolution.

Here are my initial thoughts: Would it be better to have a PWM ability using the 'clutch' function on the spare output (http://www.msgpio.com/manuals/mshift/V2tune.html#sc)? The clutch outputs (on either or both of the spare outputs at the moment) act only during the duration of a shift (as defined by the user-settable shift delay factors - for line pressure adjustment and shift completion - http://www.msgpio.com/manuals/mshift/V2tune.html#gs). At the moment they are ON/OFF. I could easily add PWM to these.

- Do we need a 12x12 table (which since it it based on speed and load doesn't really reflect shift behaviour unless the user compares it to the shift table and aligns it properly), or
- Would it be better to have a fixed PWM percentage for all shifts (along with user defined on and off PWM% like Output3 when not shifting), or
- PWM% for each up and down shift, or does this need to depend on load?
- If it does need to be load dependent, is it sufficient to have it be PWM only if the load is below (or above) a user set value?
- Or should there be a, say, 10x1 table of 'output PWM during a shift' versus load, applied to all shifts? Or one such table for upshifts and one for downshifts? Or do we need an individual table for each possible shift (with up to 8 forward gears, this could be up to 14 tables)?

But all this depends on which outputs are available in the configurations most likely to use such an output, so I need more feedback.

Until I get some feedback, my inclination is to have one 10x1 upshift table of shift PWM versus load for each of upshift and downshifts (i.e. two tables in total), and have this applied to SP1 (since SP2 can be used for an addition shift solenoid output). I will wait for people to chime in, though!

Lance.

[gross polluter]

Here are my initial thoughts: Would it be better to have a PWM ability using the 'clutch' function on the spare output (http://www.msgpio.com/manuals/mshift/V2tune.html#sc)? The clutch outputs (on either or both of the spare outputs at the moment) act only during the duration of a shift (as defined by the user-settable shift delay factors - for line pressure adjustment and shift completion - http://www.msgpio.com/manuals/mshift/V2tune.html#gs). At the moment they are ON/OFF. I could easily add PWM to these.

That sounds like it would work just fine. In the case of the AW30-40 transmissions the PWM output is only active during the duration of a shift.

- Do we need a 12x12 table (which since it it based on speed and load doesn't really reflect shift behaviour unless the user compares it to the shift table and aligns it properly), or
- Would it be better to have a fixed PWM percentage for all shifts (along with user defined on and off PWM% like Output3 when not shifting), or
- PWM% for each up and down shift, or does this need to depend on load?
- If it does need to be load dependent, is it sufficient to have it be PWM only if the load is below (or above) a user set value?
- Or should there be a, say, 10x1 table of 'output PWM during a shift' versus load, applied to all shifts? Or one such table for upshifts and one for downshifts? Or do we need an individual table for each possible shift (with up to 8 forward gears, this could be up to 14 tables)?

Thinking about it, the table doesn't HAVE to be 12x12. The PWM output controls the pressure that acts on the accumulator so the PWM duty will vary with load, a fixed value won't work well. A simple 1x10 table for setting a PWM duty based on TPS or MAP should work just fine. A single table is sufficient for both upshifts and downshifts and all gears. I honestly don't see the need to have individual tables for each gear and up/down shifts.

[Token]

will wait for people to chime in, though!

Just cruising in behind gross polluter's wake. Very excited to see the GPIO be able to control this tranny.

[Bernard Fife]

Okay, I will look at adding this for the next release (it will likely be a few weeks off, though). Lance.

[Token]

Awesome.

[Jim Blackwood]

That all sounds really good to me. I'd have to dive into this stuff to refresh my memory enough to contribute so I'm sorry to not be much help right now. I do remember the AA80E had some really cool adaptive logic going on that is probably too advanced for the hardware we are working with, but the documentation may suggest possible paths for us to consider.

Jim

[Jim Blackwood]

I just checked and the AA80E does not have a SLN solenoid.However I do not yet know the function of SLT. This could do the same thing and just be labeled differently. There are still some unknowns with this transmission.

Jim

[gross polluter]

SLT is generally line pressure.

[Jim Blackwood]

Thanks. I saved that in my copy of the spreadsheet on the AA80E.

Jim