How to Tune 5EAT Line Pressure

[ClimberDHexMods]

One way to find out...

 

Wait is the 409 hard ceiling for all 05 5eats?

[seabass07]

Yes, unless that table NSFW posted changes the ceiling. I don't know if the tcu has the ceiling, but the ecu certainly does. 100% pedal angle was giving me around 85% throttle plate when I upped it too far. 85% throttle plate was where 408 rt was on the map.

[bmx045]

i'd be curious to see if the 07+ tcu's respond differently to high rt

[seabass07]

There's a table for that... Have you upgraded to 521R yet?

 

   <table name="Maximum Requested Torque" category="Drive-by-Wire Throttle (DBW)" address="CA8BC" type="2D" level="1" scaling="rawecuvalue">
     <description>Requested torque values will be limited to this maximum.</description>
     <table name="Maximum Requested Torque" type="Static Y Axis" elements="1">
        <data>Maximum</data>
     </table>
  </table>

 

Is there more code to go along with this or does it just get pasted into the ecu defs file in the a2wc521r section?

 

*nvm dumb question. I didn't even think of using ecuflash for some reason.

[ClimberDHexMods]

As BMX045 alluded to, you can tune the two throttle maps such a way that, for example:

50% accelerator pedal position (APP) would request 350 requested torque (RT) which is a higher RT value for this APP than normal, but then scale down the corresponding throttle table so it's still outputting 50% throttle position (TP) or whatever it really is.

Second example:

75% APP, gets you 380 RT, but still the same old corresponding TP as before, whatever it was, call it 75% if going with a throttle cable-type mapping.

 

End result: You still get X APP to produce Y TP, so the gas pedal experience is unchanged. But the incidental RTs would now be a lot higher than they use to be for the same pedal and throttle positions. Thus, TCU would see the same APP, TPS as before, but also see a higher RT. We don't know the logic the TCU uses, but the FSM states RT is of primary importance.

 

This theory here should theorhetically tighten up mid-throttle line pressure. Whether scaling peak RT to 408+ will help over 300, either way seat of pants testing and logging RPM/time will tell the difference, if there is one.

[bmx045]

Well after my tuner utilized an increased RT in the mid range cells, also, of course correlating to early boost onset(easier with an a/m downpipe for being able to spool quicker). increased RT in the mid range cells secondarily benefits our line pressure.

[ClimberDHexMods]

Well after my tuner utilized an increased RT in the mid range cells, also, of course correlating to early boost onset(easier with an a/m downpipe for being able to spool quicker). increased RT in the mid range cells secondarily benefits our line pressure.

 

By the way your review of this is proof enough for me that it can work considerably well.

[bmx045]

One of the biggest seat of the pants improvement I've felt was midrange shifting in D, VERY firm, not so much at WOT though, me thinks line pressure isn't enough at wot. This can fluctuate depending on how the car was tuned. Why? Because if the TCU and ECU sees an RT of lets say 380, and your engine is not making much power, the tcu will shift hard. So keep that in mind. This confirms, atleast to me, that line pressure DOES increase a LOT depending on RT. This can also be noted when laying on the throttle and letting off a split second before you know the gear will change in D, you may get whiplash since you cut the throttle BEFORE the TCU had the chance to detect the new RT. I believe the TCU has a set time/rpm to shift in D, it knows when the shift is coming up, it takes a snap shot of the environment and then shifts. If you let off the throttle inside this calculation window, you will get a shift firmness equivalent of whatever RT you had depressed just a second ago, but when it shifts, you're off the gas and get jerked.

 

The TCU, and this is why I believe there is a paddle shift delay, does a quick calculation of certain variables. When using the paddles, you determine the time to shift no matter the variables, so I believe the TCU has to analyze the situation when you press the paddle, it performs calculations for a second, then shifts at the corresponding RT line pressure. Since I said these shift points are set in D mode, we really don't know of a shift delay since we aren't changing the gears, paddles are completely random shift inputs, and the tcu needs to do a quick read and analysis before it shifts

[ClimberDHexMods]

I call them "aftershocks", an easy term we should all use for when you were WOT, but then you let off and for the next couple seconds the TCU still calls for full line pressure.

Aftershocks are tuned into the TCU as a delay counter following WOT. I do not know why the factory tuners added that feature the way they did, but it is the biggest complaint I have from my customers about the F1 5eat, that the aftershocks become more harsh (because the F1 operates with higher line pressure). Would be nice to tune that out and have all line pressure calculations be instantaneous (like what ECU tuners do with zeroing out CL-to-OL-delay, very similar. Otherwise I don't see any way of curing the issue. Like I said before, there are hundreds of tables in ECU mapping, and it is most likely the same immensely complicated mess in the TCU. There are lots of compensation tables and probably a complicated line pressure master formula on par with the level of complication found in tuning ECU ignition timing. There are decay rates and varying slope rates for the engagement and disengagement of each individual clutch pack, and we haven't even gotten to the center diff or torque converter yet. And like how every ECU ends up being tuned a little different to make up for mechanical variances in each engine, I am sure now that every single transmission is a little different from the next. Each of my 5eat F1 customers reports a slightly different outcome, with the exact same hardware. This makes sense, as TCU learning is largely maxed out with the F1, so all the slight variations in each trans can no longer be hidden with stock learning. It's a mess, and I'm about to write a disclaimer for everything I sell that says no two results will be the same; you get what you get, and if you don't like it then I'll refund your money and help you get back to stock. It's quite frustrating. This must be how tuners feel when every customer has a slightly different mechanical powerplant, despite everything being identical on paper.

 

/Rant

[bmx045]

I'm guessing the response time in ms needs to be upped in order for it to make appropriate line pressure actions

[seabass07]

The aftershocks were definitely worse after upping the RT tables to 408.

 

I upped my max rt to 500 and scaled the RT table to 500 to see what happens. Basically a 56% increase over the stock settings. I hope to get some logs in today or tomorrow.

[ClimberDHexMods]

I'm guessing the response time in ms needs to be upped in order for it to make appropriate line pressure actions

 

I don't know why there is a delay, because the TCU should be able to calculate near instantly, and also change the line pressure solenoid pulsewidth almost instantly. Since fluid does not compress well, the pressure variation should be as quick as the rerouted fluid can move the relevant valves, which can and should all happen in a fraction of a second. All RE5R05A variants have the delay, so I suspect the contract for the Subaru RE5s aka 5EATs came with all the TCU architecture and map values ready to go. A few tweeks to dumb down the firmness for the soccer mom Subaru market, and it's off into production.

[ClimberDHexMods]

The aftershocks were definitely worse after upping the RT tables to 408.

 

I upped my max rt to 500 and scaled the RT table to 500 to see what happens. Basically a 56% increase over the stock settings. I hope to get some logs in today or tomorrow.

 

Logically this makes sense, since you're increasing line pressure. This suggests though that the aftershocks are using medium level line pressure.

If they were using 100% firmness (RT of 300) to begin with, then:

A) going from 300 to 409 really helped, or

B) going from 200 medium level average aftershock pressure to 300 is what you are feeling, or

C) a combination of both.

 

Would be nice if this develops into something my customers can use to tune their line pressure and weed out the aftershocks in exchange for somewhat lower clutch pack life, though in reality it would benefit everyone.

[bmx045]

What i think would be just as satisfying in terms of importance would be tc locking. From what i understanding is the tc is mechanical. Or does it get a signal from the tcu? Wondering if there was a way to override with a manual signal (like in a certain hertz) with a switch

[ClimberDHexMods]

What i think would be just as satisfying in terms of importance would be tc locking. From what i understanding is the tc is mechanical. Or does it get a signal from the tcu? Wondering if there was a way to override with a manual signal (like in a certain hertz) with a switch

 

It's activated by a PWM solenoid. It too operates on a ramp up of pressure. If the solenoid went from 0% to 100% with no ramp up over time, you would get a loud nasty bang from the clutches biting instantly. Perhaps diffuse with a capacitor, you could wire in diodes to keep the two inputs to the TCC solenoid isolated from eachother, and then yes you would have an easy way to lock on TCC for things like tuning/dyno'ing, fuel economy (though it already does fuel economy lockup by itself).

[bmx045]

Imagine if we locked up the tc and RT? Lol we would chirp gears all day. With a lot of tranny heat build up of course

[ClimberDHexMods]

Imagine if we locked up the tc and RT? Lol we would chirp gears all day. With a lot of tranny heat build up of course

 

Opposite: lockup is inversely proportional to heat. So locked up TCC = less heat. Higher line pressure = less slip = less heat.

 

After writing all this, I am again fairly sure I'm going to one day do a "manual swap" and control the 5eat with an aftermarket TCU, referencing boost for line pressure, and throttle angle for shift points, TCC locked up a lot more than it is now, etc. It's the only way to really get everything perfect, and all you need in order to accommodate transmission discrepancies and wear is to pull out your laptop.

[bmx045]

What i mean is a locked tc would secondarily transfer more forces on the clutch packs since theres no slip from the tc. I'd forsee the clutches getting real hot if you were to wot in D with a locked tc

[seabass07]

Logically this makes sense, since you're increasing line pressure. This suggests though that the aftershocks are using medium level line pressure.

If they were using 100% firmness (RT of 300) to begin with, then:

A) going from 300 to 409 really helped, or

B) going from 200 medium level average aftershock pressure to 300 is what you are feeling, or

C) a combination of both.

 

Would be nice if this develops into something my customers can use to tune their line pressure and weed out the aftershocks in exchange for somewhat lower clutch pack life, though in reality it would benefit everyone.

 

I noticed it with part throttle most. I'm not sure whether it was really increased at WOT. It was always a harsh jolt. I'm not convinced whether or not the max line pressure at WOT will increase. I don't see why it wouldn't.

[seabass07]

What i mean is a locked tc would secondarily transfer more forces on the clutch packs since theres no slip from the tc. I'd forsee the clutches getting real hot if you were to wot in D with a locked tc

 

But the majority of the heat in the trans is from the torque converter. Locking it up will only help temps. I think the little bit of slip in the clutches is just a fraction of the heat that the unlocked TC would make.

[ClimberDHexMods]

I noticed it with part throttle most. I'm not sure whether it was really increased at WOT. It was always a harsh jolt. I'm not convinced whether or not the max line pressure at WOT will increase. I don't see why it wouldn't.

 

It wouldn't if the TCU's highest RT referenced is 300. So we're all hoping it's scaled up to 409 in preparation for us...

 

But the majority of the heat in the trans is from the torque converter. Locking it up will only help temps. I think the little bit of slip in the clutches is just a fraction of the heat that the unlocked TC would make.

 

Assuming a WOT environment, clutch packs are still plenty capable of getting burned, and a locked up TCC will definitely not help things, but I do not know to what extent it would hurt either, but probably to a good amount. Of course we're talking about a stock VB here, since F1 should shorten up the period of dynamic friction vs stock.

 

What i mean is a locked tc would secondarily transfer more forces on the clutch packs since theres no slip from the tc. I'd forsee the clutches getting real hot if you were to wot in D with a locked tc

 

Yup, depends on how long they're slipping before they go static lock. 2 seconds vs 1/4 second is a massive difference.

[bmx045]

and this vb mod you speak of.....

 

......

 

 

...................................

[seabass07]

Im sure the clutches will burn up, but could it heat up the atf more than an unlocked torque converter?

[ClimberDHexMods]

Im sure the clutches will burn up, but could it heat up the atf more than an unlocked torque converter?

 

no

 

plus I really believe people are blowing this whole heat thing out of proportion. On the track, I get it. Rallying up a mountain pass, again I get it. But for daily street, I just don't see the problem, especially if running Amsoil, which supposedly can handle a good amount of heat just fine.

[seabass07]

Ive only had issues with back to back 3rd gear pulls in hot weather and excessive towing uphill.