jayrex

Thanks for posting those pictures. They look pretty knarly. ;D

If ClimberD at Hexmods is not making them anymore, I'm not sure where to get the custom bushings. I'm going to keep mine for now. If you can get the specs on the sizing, I bet you could get a local place to fab some up at a fair price. They are basically just brass cylinders, properly toleranced. When I get around to digging mine out, I could give you a rough idea about the size, but I don't have the proper tools to measure precisely. Not sure when that will be.

I really need to open that up to take a peek, and a pic, of the failure for you guys.

I was just anxious to get the car up and running after almost of year of down time so I didn't bother with the bushings. I didn't want to add more unknowns to the repair either.

Yeah, I'm up in the air about the necessity. Either way, I will not be launching my 5EAT. I didn't do it often, and I bet they help, but I certainly wouldn't call the brass bushings launch proof.

Theoretically the brass bushings should keep the gear mesh aligned correctly.

I definitely had that failure prior to the install. 3 teeth broken off the intermediate shaft. All the planetary gears were fine. (pictures earlier in this thread). I say failure, but the trans functioned like normal. I only saw the issue because I got eyes on it.

My bushings went in tight, but I didn't need to freeze them or anything. The bearing portion of the setup is the inside diameter. So everything spins freely, no reason to worry about them being tight. You really want them as tight as possible, while still being able to assemble. Just tap them in there. My tolerances were good, not too tight, not sloppy.

Also that isnt a high rpm area.

My opinion- the alignment under load is more important than the friction reduction of the needle bearings. I don't think they could bind or even get tight from 150-180 degree oil bath. Maybe bake the assembly in the oven at 300 deg and see if the tolerance changes.

So now my turn to thread jack you

 

 

Exactly which table(s) do you modify in order to improve your long term fuel trims across the board?

Lol you can have it, I'm done here.

You should be able to clean that up using the maf calibration in the real-time tables folder.

You can just multiply the effected cells with x0.95 for 5% less fuel or x1.05 for 5% more. You probably don't want to tinker with the values in open loop unless you have a wideband to read the afr accurately. So like Maf values 80 g/s and lower. Thats about where the fuel trims stop and open loop starts. Smooth out any changes to 1 or 2 adjacent cells.

I use the trace function in Cobb ATR so I can see where I'm idling right on the table. You'll need to log those values to see them at higher values.

I don't know a lot about opensource, if that's what you're using, but its going to be the same table.

I posted over at our forum-brother Nasioc to get some more insight.

I ended up figuring some stuff out. Its not epic, just an oversight on my part.

I started hijacking the post around #20

https://forums.nasioc.com/forums/showthread.php?t=2857248

I'll post my take-aways so if others stumble onto this thread.

From other thread:

--------------------------------------------------------------------------------

So my main realization is that Comm Fuel Final is definitely not the same as Closed Loop Fuel Target(CLFT). I think that's my biggest mistake here. It looks like you cannot really the dictate CLFT. You are kind of at the mercy of the ecu?

CLFT logs much closer to my actual AFR and is certainly driving the corrections and trims. Maybe others can learn from that mistake.

I bet the rest of the discrepancy has to do with the ecu and the magical rear 02 sensor even though I have them zero'd out. Or maybe its just poor injectors settings.

I did just try turning off closed loop completely and it made a huge difference. I feel like you can actually see the result of adjustments and control the AFR Targets. Tip In is even easier to see.

With open loop only, the AFRs are sitting pretty good, I just need a little better MAF calibration on the low end(a bit too lean). I feel like running open loop 24/7 because of the better control. It may not be the best idea longterm, but we'll see how it works out for a bit.

----------------------------------------------------------------------------------

so with the new trans....Did you do anything about the center diff bushings? Or did you just leave it stock?

No they are still sitting in the old trans. I haven't even seen them yet.

If I decide to later, I can pop them in pretty easily without removing the trans.

I will post a pic if/when I get in there. I think they have about 20k miles on them now. They might be tore up from the failure. I'm thinking that was the area that probably exploded. I highly doubt they were the cause thou.

I believe I zeroed those out to eliminate that, but I will double check. Thanks.

EDIT: Yes they are all zero, any other possibilities?

There are a few tables with some slight corrections possible, but nothing that would show .8 to 1.0 rich.

Like Closed Lop Fueling Target Compensation Tables and TPS fuel Adder Tables.

I believe all those calculations are taken into account to create Command Fuel Final.

Basically my Command Fuel Final value is not driving my afr corrections or learning values correctly. When its richer than target it keeps adding fuel.

Its almost being ignored or modified elsewhere.

Is Command Fuel Final the actual final value or are there further modifications to that value during corrections. What could I be missing?

I'm using Cobb ATR to tune. I've been doing it a while and I am familiar with just about every table. This happens with gas or E85, pretty much 100% of the time. I can't find anything that would cause this significant of a difference.

05 LGT - lots of mods, big injectors, namebrand fpr, all the fuels.

Example: Command Fuel Final is say 14.7 while cruising, but I can't hold anything leaner than 13.9. Idle is the same. open loop targets appears unaffected by the issue.

My Wideband matches my stock sensor.

Thanks for any help

https://cobratransmission.com/

Some of the gaskets you may want to get from a subie parts dealer. Like the gasket for the pump. Both times mine came off pretty mangled and corroded. That is prob the only one you'll need. Definitely plan on a valve body shift kit if you're adding power. I have a spare stock VB to sell if anyone wants. Just pm me.

TLDR: The new trans is in and working!

I did not open up the old trans yet, but I found bits of center diff gear teeth in the pan. I even had the special brass bushings in there to prevent this.

So the longer version with pics...

Removed the engine and trans in one shot! I saw this on an instagram post and decided to give it a go because aligning the trans to the engine or engine to trans in the car is a PITA either way. They popped right out with very little issue. Just make sure you have some good tilt possible and enough headroom. Separation and reconnecting works soo much better on the bench. Not to mention the dangers of trying to lift a transmission up into the car from the bottom. It can be sketchy solo or with help. I did this all solo.

Here is 4 pics of it on the way out:

Stealing the modified VB out of the busted trans:

VB out of the new (used) trans:

This is my ghetto DIY drain table. Shimmed the bench top at an angle and drilled a hole for a drain hose. Taped down some rope to guide the perpetual waves of trans fluid. This worked good, last time was a mess.

Clutch packs all taken out, and a pic of the new frictions:

This used trans was spotless inside. I didn't even need to change the frictions but I didn't know until getting it all apart. I changed everything out anyway because I had already bought everything. (~$200) I saved the old frictions and steels to use later if I burn these up. I seem to be good at that.

Here is everything buttoned up and rejoined together:

Back in the car and doing a little TRANS-fusion refill:

Everything is working great. Had an issue with one of the intercooler pipes popping off, making me thing the trans popped within 200 yards of my driveway. Then the car wouldn't start cuz the battery was dead. It was kind of confusing because I didn't find the intercooler pipe issue till 2 days later. I was sure I had killed the new trans. Was super bummed after all that work. Its all good now and every is working great! Even have her back on E85.

Bonus pics of valve cover resealing and turbo rebuild!

Pretty sure this video is what you need:

I've done it a few times myself, its not too bad. Little annoying from under the car cuz it never stops dripping on you.

Also some good info in my 5eat rebuild thread.

Trans exploded. Not sure exactly what broke yet, its too cold to take it out. I'm thinking center diff. Definitely not rear diff. Bought the exact serial number $450 + $150 shipping.

Ref trans serial number: TG5C7CWABA (05 Legacy GT Wagon) Not sure what other numbers would have worked. I'm assuming any 05 or 06 LGT 5EAT, not sure if wagon specific is important. 07-09 have different valve bodies, so I'd assume they would not work as a swap.

I will likely start another thread for the carnage pics and rebuild of the new trans. (will post a link if I do).

I might just switch the valve bodies and pop the trans in there, instead of changing the clutch packs. Haven't decided yet.

A2WC500C was not loading correctly. So I just moved onto A2WC521C as you suggested.

I found a greyed out 'Calculated Torque A' table, Edit, Edit Definitions.

http://legacygt.com/forums/attachment.php?attachmentid=261504&stc=1&d=1517457442

There is some weird box stuck over some of the text fields, but you can still make out what is in there.

Used these values:

Columns: C15DC

Rows: C1610

Data: C164C

This is the resulting table:

http://legacygt.com/forums/attachment.php?attachmentid=261503&stc=1&d=1517457442

Could not change some of the labels, maybe if I tinker with the 32BITBASE tab, more options for the table were in there.

Do these numbers look right and should I try playing with these values? The trans likes to let go around 3500+ so start increasing above that range?

I'm on 1000cc injectors and my injector scale is 2700 in RaceTuner, not sure if that translates to 270 here. I see 520 as the stock value in EcuFlash, Where I think Race uses 5200 stock scale value.

Ohhhh, I see the issue. These tables are defined for ECUflash, not Romraider proper. The definition format for Romradier is kind of a pain in the ass, so most people just use it for logging now.

Oh I see, I though you might have just misspoke with the 'ecu flash xml', figured you meant ecu_def.xml

Seemed like all the pieces were there to make the definitions work for Romraider. I'll take a stab at it with ECU Flash in the next couple days.

Thanks for all the help on this. I don't want to give up on my 5EAT just yet!

I'm hoping this fix plus the modified vb(installed), brass bushings(installed), and a fresh set of frictions should set it up to hold some decent power.

I tried various tweaks to get 521C or 500C to work, but I just can't get the table to populate into RomRaider ECU Editor.

I'm going to start digging in here over the next week or so:

http://www.romraider.com/forum/viewtopic.php?f=40&t=6303

I dove head first into romraider, but I'm a little stuck trying to get that table to show up, maybe I'm making too many assumptions. Any help would be greatly appreciated!

After connecting, my ECUID is:

A2WC500C

First assumption is that A2WC500C uses the same addresses as A2WC521C. I looked into ScoobyRom to confirm myself, but that was a struggle outside of my comfort zone.

So based on your post to the RomRaider Forums:

Columns: C15DC

Rows: C1610

Data: C164C

In case someone is searching later the above also applies to A2WC521R and A2WC521C.

You also mentioned:

"15 rows by 13 columns, rpm and base pulsewidth as axes. Thanks utc_pyro for pointing that out."

I've had no luck trying to figure out the exact syntax to edit the 'ecu_defs.xml' file to bring up the Calculated Torque table when I load A2WC500C.hex or A2WC521C.hex files (I guess they are interchangeable?)

It looks like I just need to add the table to 32BITBASE definitions. Not sure if I need to add something extra to the specific A2WC500C 'romid' section that calls on the 32BITBASE definition.

Here is what I tried adding to the 32BITBASE:

<table name="Calc Engine Torque" category="UTC_Pyro Beta" address="C164C" type="3D" level="1" scaling="RequestedTorque(rawecuvalue)">

<description>Calculated Crank Engine Torque by Pulsewidth.</description>

<table name="PulseWidth" address="C15DC" type="Y Axis" elements="13" scaling="PulseWidth"/>

<table name="Engine Speed" address="C1610" type="X Axis" elements="15" scaling="RPM"/>

</table>

Am I even close? I feel like there are many errors in my attempt here, lol.

Also, once I get this working, I see you tried multiplying the entire table by 1.2, is that still working out for you?

Ok it looks like they figured out how the scaling works over on the Romriader forums. dschultz and NSFW descuss it here: http://www.romraider.com/forum/viewtopic.php?f=40&t=8950&start=0

 

The upper row is base injector pulse width, and is calculated by multiplying the raw load and and the raw injector scaling. There are also a bunch of compensation tables that COULD be used but appear off in our roms. At least on dschultz outback the code actually went through them still so you could actually use them. Not sure if the comp tables are applied pre or post sending the torque value out the CAN bus though without IDA knoladge.

 

Also I seem to be seeing the shift retard tables, but I can't figure out which does what without knowing IDA. They look like the ones defined here:

http://www.romraider.com/forum/viewtopic.php?f=15&t=1158&start=60

I found where you were talking about it. Thanks again.

I've done quite a bit of reading on Subaru tuning, and this is news to me!

Thank you for taking the time, and I will look into migrating over to a different tuning platform. I like some of the opensource MAF tools anyway. I always just stuck with Racetune because I had the AP and figured it would show all editable tables. I never imagined things would be 'missing'.

Yes I did, but I'm not certain it is correct. (Requested Torque?)

I'm a little foggy in this area of tuning. Also not currently on E85, its a little wonky in the winter here.

Feel free to school me or shoot me a good link to read.

It looks like Requested Torque only applies to Target Throttle Angle

But I have read the 5EAT some how uses that value to adjust clutch pressure? And some say 410 is the max that changes anything?

throttlehappy has a bunch to say about it in this thread:

http://www.romraider.com/forum/viewtopic.php?f=14&t=7539&sid=95a8b76083c25fe25908e1f551f849c2&start=30

Here are my tables as they sit now. I just kinda took a stab at it and raised them.

Brief followup on this trans build.

I had been using the stock VB for about a year and a half, and it started slipping very bad again, even at about 25% throttle. (19lbs on E85)

I finally repaired the modified VB and installed it without changing the failing frictions. The difference was nothing short of amazing. The VB mod is just incredible, but 6 months later and the slipping is back. My frictions were on borrowed time anyway. I bet I wouldn't have burned up them up if I had the modified VB working from the start. Trans still drives okay, but I just can't punch it on the highway. Waiting for some warmer weather here in New York before I open 'er up again.

It works! It works! It works!

Little laggy on the shifting... as usual, but everything is good.

No codes.

All is well in the Universe.

Super mad and happy at the same time... that's a weird feeling.

Pulled out my modified VB. I'm thinking this valve needed to be held it by a little more than air... Retainer must have fallen out during re-assembly of the 'slabs'. I can't believe I missed that. I guess I need to work on my 'attention to detail'.

To get it up and running sooner, I put in the unmodded VB. I can test that one, and test the rest of the trans without the complication of the modded VB, and I need some parts that are missing anyway. I don't know where those pieces ended up... quite the mystery. ClimberD is hooking me up with some parts from his spare pile. - Thanks dude!

I have not fired it up yet. Writing this while I get the trans fluid in. Trying not to get too excited, but am really hopeful this will clear up my issues. I will update with the results soon.

Check out my new TRANS-fusion setup.

Big pan to catch the never ending rain of trans fluid.

2 Gallon bucket with paint strainer screen.

Dirty bits screened out.

TRANS-fusion time! This thing works great - little 1/4" NPT ball valve with brass nipple to some clear 3/8" tubing which fits right down the dipstick tube. Not one stray drip, and you get it done in one shot. (Home Depot plumbing isle)

I was looking at the Gear Train Table a little more.

Regardless of the logs... it feels like it shifts up to 3rd, then 4th is the same as 3rd, and then 5th feels like neutral.

See if my logic makes sense.

Perscitus, thank you very much for taking the time to graph and check my logs. I appreciate the time spent on that.

Trans was warmed up prior to log, about 150 - 170 deg F, but I will log Temp next time. I will also look into getting better resolution. I am using the KKL cable.

I'd also like to be more methodical with the gear shifts. I actually took two logs, one with normal D, and the second in manual mode. The weird thing is, I only found one log saved. So I don't know if it was added onto the first, or over written. Not sure how RR Logger handles the files. I just clicked 'stop logging' and then a few seconds later clicked 'start logging'. My guess is that they were chained together. I'll figure that out and get a good

1 - 2 - 3 - 4 - 5 - 4 - 3 - 2 - 1 in manual mode.

One other weird thing I just remembered. I think it only activates the manual downshift I requested after I hit the gas a little.

ClimberD, thanks for the input. I actually didn't fill the TC prior - honestly didn't know it was necessary. After extensive googling, it looks like its a 'best practice' type of thing, not necessarily a show stopper. Many people say it fills right up after the pump starts moving.

The first step is definitely changing out that VB. That will certainly get me closer to the cause. I am spending a little time logging before that because I want some good data to reference back to, after I make the swap.

Condition 1- If the OEM VB solves my problems, then I'll have data to hopefully zero in on the issue, and fix the Modified VB. Elevating it from 'paper weight' status

Condition 2- If the issue remains, maybe I can compare some logs to further reaffirm the problem location in the mechanical bits of the trans. I really don't want to open up the trans again, just to go looking around for a problem. That seems like too much uncertainty for that amount of work.

I think I need to get more familiar with this table to understand where the problem specifically is.

Even when it does seem to be working a little, its not shifting well... very sloppy and 'slam shifty'. Once its settled in gear it does hold very firm, where my previous frictions would let loose under load. Where I am having problems now, I've never had slipping issues, so I don't think its the few lightly worn frictions I left in there.

Anyway, I'll get a few more logs and swap the VB in the next day or two.

Thanks for you help, this would be much harder without any input from others.

Attached a TCU log. I'm not sure if anyone has experience with these. Not really sure what I should be looking for. You can see with the gear changes, that the rpm does not change, just keeps climbing, which is what it feels like in the car. Feels like a CVT trans

FYI - still my modified VB, not the new to me - used stock one.

5EAT.csv