10R80 Clutch speed/ slip math

[murfie]

Clutch B Speed.MathParameter.xml
Clutch C Speed.MathParameter.xml
Clutch D speed.MathParameter.xml
Clutch E speed.MathParameter.xml

A's speed is "input shaft"
F's speed is "input shaft B".
Direct from the speed sensors.

Here are the maths to get B,C,D, and E clutch speeds.

A and B slip are PIDs.
When solenoid applies their speed should be 0. Any difference from 0 is slip.

When solenoid E is applied clutch E connects the TSS to ring3 / sun 4. When on, E's speed should be equal to TSS. Any difference is slip.

C,D, and F are all attached to the nod. When any one of them are on the nod's speed equals that clutch's speed. When two are on, their speed difference is slip, but from which you can't tell. If all 3 are on, in 4th,7,and 10th you can isolate which clutch is slipping. Either way you can see generally what speed the nod is going, and a general idea of slip.

You can see offgoing pressures, oncoming pressures, boost pressures, ramps of there pressures, and their resulting clutch slip. Engine brake torque and transmission fluid temp and you can follow along where you are in the tables. Make changes and see what they did. Its minimize slip, but subjective to how harsh you are willing to accept for a shift.

Shifts, clutch speeds and slip.jpg

Shifts, clutch speeds and slip 2.jpg

[mejohn50]

Thanks for posting this. Huge help. Regarding minimizing slip, is that oncoming only, or off going and oncoming? The way I understand how a shift occurs means there has to be *some* slip just by nature of the events taking some non-zero amount of time. I’ve read some resources that only focus on oncoming clutch and ramp times for the most part, but that seems like half the story or less.

[murfie]

The key to a good, well timed, well executed shift is the prep for it.
Just like planning a project, painting, or many other things in life.

The prep phase of a shift consist of two things:
-Boost pressure for the oncoming clutch (Boost pressure)
-filling the oncoming clutch (Stroke pressure)

With a clutch solenoid off(0 duty cycle) the clutch can be considered "empty" off fluid. An empty clutch would react very delayed and unpredictable. Low duty cycle is used to fill the clutch with low pressure fluid, so that you get a instant and predictable clutch application from solenoid duty cycle when the time comes. This fill takes time, but to much time builds unnecessary pressure, so it's a balancing act.

Now if you were to just do duty cycle for stroke pressure, it would be much slower, or you could get clutch drag/ flare on down shifts, as it slips the clutchs of the current gear being used if given too much time. This is where boost pressure comes in. It is applied in effort to not interfere with the current gear while oncoming clutch fluid is being filled, think in terms of time on this not pressure, you're not trying to build pressure at this point, just fill in fluid. More pressure can also be added to other clutch's responsible for the gear, 6R80 had a constant clutch, offgoing, and oncoming(2 applied clutches making up a gear). 10R80 has three constant clutches, offgoing, and oncoming(4 clutches making up a gear). With 10R any one of the clutch's could be the one slipping, but you can see that with these speed maths and comparing them. The goal is to not have to have excessive pressures, as that slows everything down.

There's actually a lot to get into on just the timing, verification, and the adaptiveness/ decision making of whether to shift or not with recent pedal changes during the prepatory phase.

Boost pressures in the log. This is not actual pressure, just what the DC of the solenoid is doing. Again, goal is quickly filling clutch with fluid not pressure. You can see how the steady clutches go to a high pressure for these events.

Boost pressures.jpg

Stroke pressure, and ramp. If the clutch is actually filled with fluid, these can actually be pretty close to representing the fluids pressure.

Stroke pressure and Ramp.jpg

I try to increase boost pressure, but reduce its time, therefore preparing faster, and reducing the shifts over all time. Its sort of guessing in the 6R80 just watching gear ratio, but with the 10R you can watch clutch slip directly and the effect on the current gear.

Comparing the timing of the off going clutch and starting the oncoming clutch's start of its ramp. It adapts to that based on OSS speed vs commanded OSS shift point, correcting to meet the targeted shift point. The drop of the off going clutch will create a "hole" in torque applied to wheels. The ramp rate determines how hard or soft this hole is filled.
That starts the torque transfer phase....

[mejohn50]

Looks like I have some data logging to do. I just don't have that Hawaii weather so it might be a while before I get any meaningful data to start making sense of this.

[CoreyMS]

Just talking out loud...tell me if I'm off...(looking at Murfie's stock file)

Shifting from 2-3 E is coming on
Wot shift, 340 KPA in 220 MS (boost time) while ramp time 0 is 200 MS

Shifting from 3-4 E is going off, F is coming on.
Wot shift, 540 KPA in 140 MS (boost time) while the ramp time 0 is 120 MS

Would you not want to either keep this separation of 20MS or at least keep them proportional trying to tighten them up?

Have you tried to just lower the time and see if you can actually tell in the log keeping the pressure the same?


For reference my 10R140 in my Superduty (the pressures are much higher)

2-3 E 700 KPA in 162 MS with Ramp Time 0 200 MS
3-4 F 625 KPA in 151 MS with Ramp Time 0 240 MS

IF they are close to equal I think you could bring the time way down on the 10r80

[Justinjor]

Subbing to this. I love tweaking and logging the 10r.

Between myself and my good friend Branden in GA - we've just about got this 10r absolutely perfect. Lightning fast shifts at wot with 1000+whp and grandma/grandpa cozy cruising at part throttle.
Best of both worlds

[mejohn50]

Subbing to this. I love tweaking and logging the 10r.

Between myself and my good friend Branden in GA - we've just about got this 10r absolutely perfect. Lightning fast shifts at wot with 1000+whp and grandma/grandpa cozy cruising at part throttle.
Best of both worlds

Meant to ask this a while back…are you willing to share some guidance on how you got there? Kind of looking for the path you took to get where you are with the A10.

[aksubiedubie]

Also interested in this, just started to poke around my F150 5.0 10r80.

[ghostofnyc]

Subbing this

[engineermike]

@murfie, is it correct to say that the usermath you posted is the speed of the clutch, relative to static, but only if that clutch is engaged?

Next, is it correct to say that the torque applied to that clutch is a function of the engine speed and engine torque, multiplied by the engine speed divided by the resulting clutch speed?

For instance, if brake torque is 100 ftlb, engine speed (ISS) is 3000 rpm, and clutch C speed is 5000 rpm, then the torque that clutch C needs to hold is 100 ftlb x 3000/5000?

[murfie]

@murfie, is it correct to say that the usermath you posted is the speed of the clutch, relative to static, but only if that clutch is engaged?

Next, is it correct to say that the torque applied to that clutch is a function of the engine speed and engine torque, multiplied by the engine speed divided by the resulting clutch speed?

For instance, if brake torque is 100 ftlb, engine speed (ISS) is 3000 rpm, and clutch C speed is 5000 rpm, then the torque that clutch C needs to hold is 100 ftlb x 3000/5000?

With 4 directional speed sensors all four of the gear set element speeds (ring, sun and carrier) can be calculated all the time. Heres a stick diagram. I know planetary gear sets and how you get different ratios out of one of them is hard for people to get their heads around, never mind working out 4, so maybe this will help.

10R Stick diagram .png

A is the ring 1 gear speed, from a sensor.

B is the speed of sun 1 or sun 2 gear.

C is the speed of ring 2 gear.

D is the the speed of the carrier 3 gear.

E is the speed of the ring 3 gear.

F is the speed of ring 4 gear, from a sensor.
IIRC thats what I did, but I may have did the speed of the linked element which should be the same, it depends on the gear ratios of each planetary I was able to find.

Clutch A is a brake clutch, when on it grounds R1
Clutch B is a brake clutch, when on it grounds S1 and S2
Clutch C is a rotating clutch, connects R2 to the node
Clutch D is a rotating clutch, connects C3 to the node
Clutch E is a rotating clutch, connects the input shaft (C2) to R3 and S4
Clutch F is a rotating clutch, connects C1 and R4 to the node
OWC is connected to S1 / S2 , allowing forward rotation, prevents backward rotation of S1 / S2

The torque applied to an individual clutch isn't really helpful to think about as 4 clutches are engaged in each gear creating the appropriate output ratio. It would get complicated figuring out the torque based on the gear ratio changes to each of the 4 planetary gear sets and the respective power flow through them.
What clutch is slipping is what you want to know and that is what can be calculated/ seen from watching different clutch speeds. Summarized in my original post and what I hope the stick diagram helps clarify.

example:
If C and D are on then R2 speed = C3 speed. If the speeds match we know both C and D are on. If the
speeds do not match we know either Clutch C or D is slipping, but cannot determine which clutch is the
cause unless all 3 node clutches are on.

[engineermike]

Thanks, @murfie. After I posed the question I realized the torque transmitted by each clutch was a bit more complicated.

I was basically trying to determine how the between-shift clutch pressures are determined by the logic. I've so far figured out that the clutch pressure applied is directly proportional to engine brake torque (actually, probably ISS torque after TC multiplication) but the slope is different for each gear. However, the slope is the same for 2nd and 4th, for instance, and a different slope for 5th and 7th. My bet is that the logic calculates the holding torque required for that particular gear and clutch, and modulates pressure to achieve it. The holding torque would be directly proportional to the hydraulic pressure applied. But I would need to know the torque transmitted by each clutch to prove that hypothesis.

[veeefour]

Thanks, @murfie. After I posed the question I realized the torque transmitted by each clutch was a bit more complicated.

I was basically trying to determine how the between-shift clutch pressures are determined by the logic. I've so far figured out that the clutch pressure applied is directly proportional to engine brake torque (actually, probably ISS torque after TC multiplication) but the slope is different for each gear. However, the slope is the same for 2nd and 4th, for instance, and a different slope for 5th and 7th. My bet is that the logic calculates the holding torque required for that particular gear and clutch, and modulates pressure to achieve it. The holding torque would be directly proportional to the hydraulic pressure applied. But I would need to know the torque transmitted by each clutch to prove that hypothesis.

Try min clip 30622 - this determines how little is left

[murfie]

Thanks, @murfie. After I posed the question I realized the torque transmitted by each clutch was a bit more complicated.

I was basically trying to determine how the between-shift clutch pressures are determined by the logic. I've so far figured out that the clutch pressure applied is directly proportional to engine brake torque (actually, probably ISS torque after TC multiplication) but the slope is different for each gear. However, the slope is the same for 2nd and 4th, for instance, and a different slope for 5th and 7th. My bet is that the logic calculates the holding torque required for that particular gear and clutch, and modulates pressure to achieve it. The holding torque would be directly proportional to the hydraulic pressure applied. But I would need to know the torque transmitted by each clutch to prove that hypothesis.

Just know there are no pressure sensors.
That all you are logging is solenoid DC to expected pressure based off fluid temperature.

Each clutch A-F has a directly proportional (0 current = 0 pressure, max current = max pressure) shift
solenoid that controls pressure to that clutch. The solenoids are controlled by “smart driver” chips – when
the main micro sends out the commanded state for each solenoid the smart drivers return fault information
(open, short to power, short to ground).
Line pressure is controlled by an inversely proportional solenoid (0 current = max pressure, max current =
min line pressure). The line pressure control solenoid is controlled by a “smart driver” capable of detecting
open, short to ground or short to power faults.

Aside from that it basic timing control involving the prep for the shift, the ramp rate of the shift, ect. Mostly to make the shifts smoother for passenger comfort or wear and tear. You can only get so much out of a stock 10R before you need physical modifications to the clutches or possibly the transmission oil pump/ seals to make it survive behind high power and torque.

[engineermike]

I got all that covered, murfie. I was trying to determine how the logic and math works to calculate the hydraulic pressure applied to the clutches between shifts. I might have to settle for "it's proportional to input shaft torque", with the primary learning being that your engine brake torque needs to be accurate in order to ensure the correct clutch pressure is applied. And if you are experiencing clutch slip then modify the calibration to increase the engine brake torque using the TTL, LTT, or some torque modifier and see if that improves it.

I wonder how many procharger or edelbrock supercharged cars using the supplied calibration are having clutch slip issues as a result.

[murfie]

I got all that covered, murfie. I was trying to determine how the logic and math works to calculate the hydraulic pressure applied to the clutches between shifts. I might have to settle for "it's proportional to input shaft torque", with the primary learning being that your engine brake torque needs to be accurate in order to ensure the correct clutch pressure is applied. And if you are experiencing clutch slip then modify the calibration to increase the engine brake torque using the TTL, LTT, or some torque modifier and see if that improves it.

I wonder how many procharger or edelbrock supercharged cars using the supplied calibration are having clutch slip issues as a result.

It's the same as reading someone saying they are trying to calculate a MAF transfer.
Ford gets their relationship for the transmission valve body through a bench test. The primary logic that keeps shifts operating smooth scales with engine brake torque. All the adaptive logic to pick up the error in that primary logic is based on calculated slip. It would be the only feed back to correct things to your liking. If they are not confusing the harshness of the TCC, People usually reset these and find it smooths smooths shifts out. The adaptive logic is probably to aggressive at minimizing slip.