Results 1 to 18 of 18

Thread: ZF8 shift pressures and timing - for Steven@HP

  1. #1
    Advanced Tuner
    Join Date
    Nov 2017
    Posts
    383

    ZF8 shift pressures and timing - for Steven@HP

    Under Trans>shift pressures, there are 3 tabs.

    General, upshift, and downshift.

    These seem to be related to fluid pressures in the various clutches. There are also some time based parameters.

    I've been doing my best to apply what I know about tranmissions from the past to these ZF8s.

    As I understand it, in order for a shift to take place one clutch must release and/or another must apply and they must be coordinated to avoid overlap.

    First off, will logging Clutch Shift Type reliably identify the tables (normal, perf, sport, max) that are being used? In logging, I only ever see normal clutch shift type displayed no matter what shift mode, shift ID, driver type or aggression is present.

    Second, can you give a very basic run-down of what is happening as the trans completes a shift based on commanded pressures, fill times, and slip times?

  2. #2
    Logging Clutch Shift Type should be reliable, assuming we've mapped it correctly. It pulls from the internal value that is used to decide which set of tables the TCM actually looks at.

    So as you mentioned, its releasing a clutch and applying a clutch.
    This is true of every shift the transmission does.

    Every gear change the ZF8 commands is one of these "clutch" to "clutch" shifts. The ZF8 has 5 of these, and has 3 active at a given time for the gears 1-8 and reverse. Park/Neutral has 2 active, so shifting from Neutral to Reverse/Drive has a single engagement.
    So it releases one clutch and applies another to shift.


    The releasing clutch is called the offgoing clutch, and the engaging one the oncoming clutch.
    Depending on whether its an upshift or downshift, under power or coasting, the actual "stages" of shift are different, but those are mostly unimportant, and in the case of "under power" or "coasting" is mapped in tables as positive or negative torque.

    First, understand that a clutch pack can hold a certain amount of torque depending on how much pressure its held together with. This pressure is transmitted by various means, but its usually a small chamber (a clutch chamber/volume) that has a piston that pushes the clutch packs together, and a spring that returns it back to open (so that with no pressure the chamber empties). By default, when you're driving down the road, the pressure doing this is the "line pressure". Higher pressure has higher torque capacity, but if you look at the line pressure table, you'll see they modify it as torque moves. Higher pressure requires the oil pump in the transmission to work harder, and causes more drag, so its controlled to have enough torque capacity but not excessively so, in order to keep efficiency high.

    Now when a shift happens, we want to remove the torque capacity from one clutch, and transfer it to another. To reduce drag and improve efficiency again, we typically have clutches sitting empty to help minimize any movement they have when not active. So we have to fill our oncoming clutch, and empty our offgoing clutch, but do so in an organized manner so that we don't break anything inside the transmission and hold as much torque as we can during the shift, and go as quickly as we can.

    So I'll divide the shift into a few sections:
    Fill->Torque Transfer->Inertial Phase->Maintenance

    Fill
    During the fill phase, we fill the oncoming clutch with oil again, up until the point it starts actually building pressure in the chamber. Essentially the clutch solenoid valve targets a set "pressure" (really its just an opening of the valve, look at the characteristic curves for the solenoid)
    There are actually three sub-phases: A prefill, where the ATF is pumped into the chamber just until it starts to hit resistance from the piston/spring, a fast fill to move the piston to move it so the clutches are just nearly touching, and a stable fill (we call just fill) that moves the clutch packs until they just "kiss" and have 0 torque capacity.

    Tuning this stage is complex, and if you want to learn more, I'd suggest this paper: Dynamic Analysis and Control of the Clutch Filling Process in Clutch-to-Clutch Transmissions
    There's a lot of math in here, but the ideas are also expressed in somewhat plain English if you'd like to learn more.
    The time this takes we want to be as quick as possible, as its responsible for a big amount of delay we feel in a shift.

    Torque Transfer
    During this stage, we transfer the torque capacity from the offgoing clutch to the oncoming clutch. We do this using a targeted oncoming rate and pressure, and a targeted offgoing rate and pressure. Now, when calibrating we put in the oncoming target pressure and offgoing target pressure, as well as ramps.
    These are however only guidelines, the transmission internally is trying to hit a targeted slip while doing this transfer (for the ZF8 I have not identified what this is or how its controlled, but I know it happens). So the transmission itself attempts to adapt these pressures while shifting to hit these goals.
    At the finish of this stage, our oncoming clutch is now holding all the internal torque of the transmission, and the offgoing is holding none.

    Really the only tuning parameter we have for this step is the Oncoming and Offgoing rates. And we only have the offgoing control during downshifts, during an upshift, we don't have control on the ZF, as its internally managing this to manage the RPM drop and internal slip.
    I should mention, the oncoming clutch is still "slipping" at the end of this stage, but its now transferring full torque. This slipping is required as the output of the clutch isn't necessarily at the same speed as the input. Reducing this slip to zero is the goal of our NEXT phase.

    Inertial/Slip Phase
    This is the stage we normally think of as the "shift". Its when the engine speed starts to change (and in fact the speeds of all the stuff inside the transmission as well!) and we start to feel like we're shifting. During this phase, we're basically waiting for the oncoming clutches slip to reduce to zero. We do this by controlling its pressure.
    Now that torque has been transferred internally, everything has to "catch up". These internal torques now act on new planetary gear sets, the input shaft, and turbine. During this step is when RPM actually drops or rises as the turbine drags the engine down or up. If you were wondering why our oncoming pressure is adjustable by gear, its because these internal torques change per gear.

    So, we have a target slip time. This is how long we expect this ratio change to take up. If we want to speed up our perceived shift, we target a shorter time here. We likewise will need MORE oncoming pressure to deal with these higher internal torques. Likewise, this is where torque management comes into play in a very important regard. During an upshift, any torque the engine is producing has to be handled by the transmission in addition to what its already handling as internal components are speeding up or slowing down. By having increased engine torque, we have increased pressure desired, and we have increased the amount of energy that the oncoming clutch bears. To make this job easier, we have upshift torque management, to reduce the strain on the transmission, and in this case it will actually SPEED UP the process of the shift as we'll have less slip and energy to control. So we can actually complete our shift faster with the correct amount of torque management. Too little and we're harming things and slowing it down, too much and we're slowing the car down when it comes out of the shift unnecessarily.

    On a downshift, we reverse this process a little, we actually want torque from the engine to help us speed up the shift (raise the engine RPM and rev match!). Too little and we're again just wasting time, and too much and we're overreving and causing excess wear as the clutch has to bring it back down.

    But typically, faster slip times and more engine torque mean we need more clutch torque capacity, and so you see as we target more torque or faster shifts, our pressures increase.

    As well, the offgoing clutch really shouldn't be apart of this phase. Once the oncoming clutch is handling torque, if the offgoing one was as well, we'd start to have parts of the transmission fight each other. This is called "tie-up" and slows our shift down (and puts a lot of strain on the gears and clutches!). Hopefully we picked good values while ramping it off previously!

    Now if we don't apply enough oncoming pressure to handle everything, we might actually have the engine kind of overshoot its target RPM on shifts. We call this "flare" on a downshift (the engine overspeeds the target input speed before its brought down) and "slip" on an upshift (it feels like the engine kind of slowwwwly slides into gear). So, our goal here is to set pressures knowing our targeted shift time. Ultimately, we're limited by how much pressure we can generate, and the torque capacity of our clutches. The faster we want this phase to happen, the more internal torque we need to do so, and the higher the pressure and more slip we potentially face. OEMs balance this with comfort in mind as well. If this phase isn't smooth, the user feels it. A fast shift that's a little sloppy might have a burst of torque or a drop in it on the output shaft, and you feel this (the shift "chirps the tires" or "bogs down").

    Most of your time tuning a transmission, your goal is to speed up shifts or change the quality of them, so you play a careful balancing act between shortening the slip time and increasing the pressure.

    As well, internally there is adaption happening where the transmission is trying to adjust the oncoming pressure to hit the target slip time. If it has to adjust too far, you actually will get DTCs indicating clutch failures (and it'll even know which clutches are possibly failing because of the shifts it notices the problem on!

    My best advice for tuning this is shortening the shift or reducing the torque management means more pressure is needed to handle the higher internal torques. If you allow flare or bogging to happen, all that energy is being dumped into the clutches wearing them out prematurely. There is an eventual limit where you simply can't go any faster before it happens that you will hit, and that's a limitation of the clutch materials and pressures you can apply. A beefier trans that can better handle those stresses is necessary, but then you may also have to retune all the low power shifts so they still feel soft and gentle Keep in mind raising the target pressure also extends the torque transfer phase, as you need more time to get the pressure and torque capacity to that level!

    Maintenance
    This phase is boring. Oncoming pressure is set to line pressure to make sure the clutch doesn't slip. During this step, its just making sure the oncoming clutch pressure allows zero slip. The offgoing clutch has no pressure applied, and is slowly draining away (there's an empty time that is mapped to indicate how long this takes).
    Tuning this means tuning the line pressure tables.


    I know thats a lot to digest, and if you want to learn more, I suggest taking a read of Virtual Clutch Controller for Clutch-to-Clutch Shifts in Planetary-Type Automatic Transmission.
    While reverse engineering the ZF8, I found this to be very helpful in trying to figure out how they were managing the shifts and plotting them out.

    Also as a quick aside to what "phase" you should target while tuning:
    If you want the "delay" from when a shift is commanded to when you feel it "start", you need to tune the fill and torque transfer (ramp) phases. You don't feel anything happen until this completes and inertial transfer starts! An aside here, a lot of people perceive lots of transmissions to have long delays from the triggering a paddle to the "shift" happening. Most of the time, its not a hardcoded delay, but is actually these two phases executing. Part of what makes the ZF8 a great transmission is that it can accomplish both these phases very quickly, making it feel like when you press the paddle, it shifts right away!

    If you want the "delay" as the ratio is changing to be different, you need to tune the inertial/slip phase. Keep in mind, the faster you want to do this, the more torque you see internally, the more energy the clutch sees, and the faster it can wear! This is balanced by OEMs to be comfortable and provide good service life. We may have different goals in mind in the aftermarket (or in fact beefed up the clutch packs so they can handle more energy!).

    Hopefully that's as thorough as you need, or more than that. If you want to learn more, check out the links I provided.

    Good luck!
    If its not broke, just give it time.

  3. #3
    Tuner
    Join Date
    Feb 2019
    Location
    Owensboro, KY
    Posts
    178
    That's a lot of information, but as i read it a few more times i believe this is very beneficial in learning how the ZF8 transmission works. I haven't got to the tuning side of the trans yet but this explains a lot of what we need to be looking at when it comes to applying changes to the trans to get the shift preferences we want and to also keep the trans from just eating itself in a short time.. Thanks You for this information i will be bookmarking this!! Keep this type of stuff coming i love learning the why rather than just being told to change this and that.
    Last edited by FstdodgeKY; 11-26-2019 at 03:09 AM.

  4. #4
    Advanced Tuner
    Join Date
    Nov 2017
    Posts
    383
    Thank you again Steven! This is way more than I expected and by far the best write up I've seen on the zf8.

    A small additional question related to how the trans is mapped - specifically for torque management - the tables show turbine torque as the row axis. Is that correct? I asked for turbine torque to be added as a loggable parameter and have been logging it. To me It seems like input torque is actually what is referenced since turbine torque is very small any time the TCC is locked, but torque management is still dropping torque on shifts. Or am I missing something?
    Last edited by spoolboy; 11-26-2019 at 04:53 AM.

  5. #5
    Quote Originally Posted by spoolboy View Post
    Thank you again Steven! This is way more than I expected and by far the best write up I've seen on the zf8.

    A small additional question related to how the trans is mapped - specifically for torque management - the tables show turbine torque as the row axis. Is that correct? I asked for turbine torque to be added as a loggable parameter and have been logging it. To me It seems like input torque is actually what is referenced since turbine torque is very small any time the TCC is locked, but torque management is still dropping torque on shifts. Or am I missing something?
    We may need to update some axis descriptions. Turbine torque isn't the reference for those, I will have them updated.
    "Turbine Torque" as seen by the ZF is more like the torque going through the converter's fluid section. It will reduce when locked as now the turbine is ideally locked (though only truly 1 to 1 locked when in Overlock) to the impeller and there's no torque difference across the converter.

    My recommended one to log is 4403 - Trans Engine Torque. This If it's not available on yours, send a ticket in. It's available on the ZF8HP2 TCM for sure as I mapped it for that.
    If its not broke, just give it time.

  6. #6
    Advanced Tuner
    Join Date
    Nov 2017
    Posts
    383
    Channel 4403 was not mapped for me - 2016 6.4 Challenger 8hp70.

    I sent a ticket into support. Ticket ID: 196347

    Is there a way to log:
    shift time
    slip time
    flare rpm during shifts
    commanded shift oss


    I don't see a channel for these.

    From infolog:

    TCM: 7E9, CAN, TCM -TransmisCtrl - ZF8HP, Dodge
    Serial: T05UN216693450
    OS: 05035474AD
    Base OS: PRQ17000_00000
    Calibration IDs: 05035474AD
    Calibration VNs: 406863CC
    Basic PIDs: 12
    Trouble Codes:
    Controller Type ID By OS: 21060
    Diagnostic Requirements: Init CAN 0x10 0x03, Unlock 0x27 0x01 0xAABBCCDD, FCA SGW Bypass
    Scanning Methods: ISO 15031-5 Mode 0x01 Functional, ISO 15031-5 Mode 0x01 Physical, ISO 14229-1 Mode 0x22, ISO 14229-1 Mode 0x23 with FCA SGW Bypass, CAN Listener
    Last edited by spoolboy; 11-27-2019 at 09:02 AM.

  7. #7
    Advanced Tuner
    Join Date
    Nov 2017
    Posts
    383
    Actually, I'd like to have a lot of these mapped for this application if possible

    transchannels.jpg:

    Maybe these aren't options?

    I have shift ID, shift situation, shift type.
    Last edited by spoolboy; 11-27-2019 at 09:05 AM.

  8. #8
    Senior Tuner
    Join Date
    May 2016
    Posts
    1,036
    Thanks for the detailed post. I've done a lot of experimenting with making 8HP70 survive bring raced a lot at high torque in a heavy car and your explorations go right along with what I've found. It really is a balancing act to get everything right but makes such a big difference and keeps the transmission healthy.

  9. #9
    Quote Originally Posted by spoolboy View Post
    Channel 4403 was not mapped for me - 2016 6.4 Challenger 8hp70.

    I sent a ticket into support. Ticket ID: 196347

    Is there a way to log:
    shift time
    slip time
    flare rpm during shifts
    commanded shift oss


    I don't see a channel for these.

    From infolog:

    TCM: 7E9, CAN, TCM -TransmisCtrl - ZF8HP, Dodge
    Serial: T05UN216693450
    OS: 05035474AD
    Base OS: PRQ17000_00000
    Calibration IDs: 05035474AD
    Calibration VNs: 406863CC
    Basic PIDs: 12
    Trouble Codes:
    Controller Type ID By OS: 21060
    Diagnostic Requirements: Init CAN 0x10 0x03, Unlock 0x27 0x01 0xAABBCCDD, FCA SGW Bypass
    Scanning Methods: ISO 15031-5 Mode 0x01 Functional, ISO 15031-5 Mode 0x01 Physical, ISO 14229-1 Mode 0x22, ISO 14229-1 Mode 0x23 with FCA SGW Bypass, CAN Listener
    I'll try to take a look this week and see if I can't find some parameters that would work for this.
    If its not broke, just give it time.

  10. #10
    Advanced Tuner
    Join Date
    Nov 2017
    Posts
    383
    Thanks for all the hard work Steven.

  11. #11
    Tuner in Training Speedy!'s Avatar
    Join Date
    Jun 2018
    Location
    interwebs
    Posts
    26
    This is great information and was timely for me as I've been digging in to the trans calibration the past week to better understand it on my 2016 Hellcat A8.

    It seems when I log transmission PIDs my VCM Scanner gets sluggish and it jumps a lot in displaying the data. Anyone else see that issue? I can literally have only 10 or 12 pids logged for engine and it's all good, I add "ShiftID" and everything gets sluggish like the RPMs aren't logged smoothly, etc.

    I also reviewed the other thread about driver type mapping to aggression tables and had a question. Is there a reason you couldn't just change the driver type table to be all aggressive to force it to use the aggressive tables and limit the number of tables you have to go modify?

    So in these tables just fill the driver counter to all "600" or even "100" to make it happen sooner and the upper and lower driver type and aggression "14" and "3" respectively?

    Capture.JPG
    Last edited by Speedy!; 12-05-2019 at 10:37 AM.

  12. #12
    Advanced Tuner
    Join Date
    Nov 2017
    Posts
    383
    I think you can eliminate or at least minimize your scanner issues by prioritizing polling speeds for the things that happen quickly. It works for me.

    As far as changing the tables you have shown, I don't think you can. But you can certainly copy/past multiple tables so that they are mapped the same way.

  13. #13
    Tuner in Training Speedy!'s Avatar
    Join Date
    Jun 2018
    Location
    interwebs
    Posts
    26
    Thanks, I didn't realize you could prioritize PIDs, I'll see if I can figure that out!

  14. #14
    Tuner in Training Speedy!'s Avatar
    Join Date
    Jun 2018
    Location
    interwebs
    Posts
    26
    Ok, seems the pid poling times were already set to 0ms for quickly changing items and slower for others so that's not the issue for me it seems. I can try lowering the slower ones further but some are 200ms, 500ms + already.

  15. #15
    Tuner
    Join Date
    May 2016
    Location
    Denver, North Carolina
    Posts
    116
    One of the biggest issues I have with these transmissions (specifically the 8HP70) is hitting torque limits after doing heavy engine modifications. I can get around it by disabling a lot of torque management but I don't like doing that. On these if we change the torque limits (Trans---General--->Max Torque Limit and Trans----General-----Gear Based ----Normal) it causes the trans to go into limp mode with 2 codes, P1DD2 and U0402. More specifically this seems to really only be an issue when adjusting the Gear Based----Normal setting. This would very much help to allow us to push up the power levels without completley disabling torque management

  16. #16
    Senior Tuner
    Join Date
    May 2016
    Posts
    1,036
    Quote Originally Posted by mxatunerjg View Post
    One of the biggest issues I have with these transmissions (specifically the 8HP70) is hitting torque limits after doing heavy engine modifications. I can get around it by disabling a lot of torque management but I don't like doing that. On these if we change the torque limits (Trans---General--->Max Torque Limit and Trans----General-----Gear Based ----Normal) it causes the trans to go into limp mode with 2 codes, P1DD2 and U0402. More specifically this seems to really only be an issue when adjusting the Gear Based----Normal setting. This would very much help to allow us to push up the power levels without completley disabling torque management
    Send me a PM and tune. U0402 I'm not sure about but I turn off torque management as far as power limits go but use torque management on shifts on 8HP70 without issue. I do disable some trans tests though.

  17. #17
    Advanced Tuner
    Join Date
    Feb 2016
    Posts
    319
    Quote Originally Posted by mxatunerjg View Post
    One of the biggest issues I have with these transmissions (specifically the 8HP70) is hitting torque limits after doing heavy engine modifications. I can get around it by disabling a lot of torque management but I don't like doing that. On these if we change the torque limits (Trans---General--->Max Torque Limit and Trans----General-----Gear Based ----Normal) it causes the trans to go into limp mode with 2 codes, P1DD2 and U0402. More specifically this seems to really only be an issue when adjusting the Gear Based----Normal setting. This would very much help to allow us to push up the power levels without completley disabling torque management
    You maybe need a different approach.......rather than broadcasting to the world (and the trans) all your newfound power, consider a more stealth method.
    Get the PCM to under-estimate the tq you are making. Remember these are just numbers...no correlation to real tq.
    It will take some rescaling of airflows etc but you can feed it whatever numbers needed to keep you under limits.
    As the PCM is under reporting tq, you will now have to rescale pressures in TCM(as it will see lower tq numbers than actual) but it will remain cool and unflustered.
    You can the crank up pressures/shift times till your neck hurts or it breaks the trans (obviously not desirable) but not hit any limits.
    The tq management can be kept in and tuned to suit the new numbers also.
    Obviously it is a bit of work and PCM/TCM must be tuned together, but if you have the patience the results can be amazing.
    They are a gooood trans.

  18. #18
    Tuner
    Join Date
    May 2016
    Location
    Denver, North Carolina
    Posts
    116
    Quote Originally Posted by Hemituna View Post
    You maybe need a different approach.......rather than broadcasting to the world (and the trans) all your newfound power, consider a more stealth method.
    Get the PCM to under-estimate the tq you are making. Remember these are just numbers...no correlation to real tq.
    It will take some rescaling of airflows etc but you can feed it whatever numbers needed to keep you under limits.
    As the PCM is under reporting tq, you will now have to rescale pressures in TCM(as it will see lower tq numbers than actual) but it will remain cool and unflustered.
    You can the crank up pressures/shift times till your neck hurts or it breaks the trans (obviously not desirable) but not hit any limits.
    The tq management can be kept in and tuned to suit the new numbers also.
    Obviously it is a bit of work and PCM/TCM must be tuned together, but if you have the patience the results can be amazing.
    They are a gooood trans.
    I actually took an approach basically identical to what you just described on a test mule last weekend. I had been thinking about it and remembered we had the ability to control the estimated tq values. I figured if I could get the PCM to think the tq value was lower the TCM wouldn't step in and reduce tq. Once I figured that side out on the dyno, then I worked on the tcm shifting pressures as well as how much tq reduction was used for shifts. When finished it drove like stock under normal conditions, and then when hard in the throttle shifting became very firm and quick. I had some others drive it as well and they loved it so I'm going to implement it into one I'm working on now to see how if I can replicate the results on a different setup. Thanks for the advice, good to hear someone else came up with the same approach and has had success with it as well.