legacycentral bbs

The N/A trannies DO have a higher final drive, which does reduce the load on the gears (at the cost of shitty gear ratios).

I ran my stock tranny for about 2 years. I beat it up real bad though.

phil have you seen anyone run a cryo treated stock gearset?

First, remember that an RA gearset is an OE gearset, so yes I have. But they are all idiots.

I've spoken to a lot of vendors about cryo-treated and micropolished RA gears.

I was told, "of all of the broken RA sets we get back per month, about 50% of them are cryo-treated or polished." That does not instill confidence.

From an engineering standpoint, cryo-treating a gearset has the potential to actually reduce the strength of the gearset. If it's done incorrectly, you end up trading fragility for hardness. That is, the gearset might be HARDER, but it will be more britle. Gears don't wear out. That's never a cause for failure.

Anyone who's running cryo-treated gearsets has wasted their money.

If cryo treatment isn't a part of the production process, it's pretty much snake oil.

Look at it this way, PPG gears aren't cryo-treated, and they are the strongest gears in the world.

Micro-polishing is pretty cool. I'm skeptical that it makes any notable difference in performance, but at least it makes the gears shiny.

Cryo-treat brake rotors, not gearsets.

Phil is definately correct on the cryo treating. You don't want to make gears more brittle. They may show a higher hardness on a rockwell tester, but its the fatige strength that I worry about. The more narrow you make the linearly elastic region in stress/strain, the worse off you'd be.

Micropolishing helps in theory. If you reduce the amount of crack tips on the surface, the less likely they are to propogate into failure. Think of it like a potato chip bag. You get a little tear, and it easily runs down the whole bag. If you round that tear out before it runs, it'll hold.

Micropolishing with a little ectching might help more, but we're still avoiding the fact that the gears are weak to begin with.

Plus, depending on the system, micropolishing can be macropolishing. Some places will sand blast the shit out of a gear or part and call it micro polished. Take away too much material, and you start changing the profiles of the teeth, which thus throws off pitch angles, and provides gross inconsistency.

"Micropolishing" is really only done well by sonic polishing, and again done very carefully!

Got two 1.7xx 60' times and two 112 traps today on my NA tranny. Nah Nah Nah..... Damn clutch holding me back Grrrrrrrrr.

as far as cryo treating, it must depend on the gears because it is popular for R&P's in the rock crawling world where it is widely considered to be effective. different manufacturers, scenarios, etc

this is a little outside the box but... what about making some braces for the case out of 1x 3/16 to 1/4 or so flat strap that can be tightened up, since isnt case flex a factor for subaru transmission gears breaking? or is that more a function of the shafts flexing. just thinking out loud is all. also what about homebrew reinforcing of the shift forks?

Case flex is a myth.

People have argued with me about this, and they were all wrong.

It's not an issue with shaft flex or case flex or anything like that. There's no "stock gears are strong, it's the case's fault that they break!"

PPG's use the stock case.

If you look at the transmission case, it's a 1" thick aluminum cradle with 10mm bolts running all through it. Plus, it's split vertically. If the case DID open, the gears would be forced downwards and upwards, and that would damage the front ring-and-pinion too.

Because of the case design, the shafts cannot flex either. The main cradle has a section width of 6". 1st through 4th gear are crammed into a highly reinforced, 6" wide, 1" thick box. PLUS, the bearings are held in place by some massive bearings (the rearward output shaft bearing is actually bolted to each half of the case). AND, if the shafts flexed, the output shaft would seize. There is an inner and outer shaft on the output side. Flex that (which would be impossible), and the two shafts would contact eachother.

The shift forks are pretty low-quality cast aluminum. They are plenty strong to handle stock-ish power and they don't really break with a synchronized transmission.

That's the biggest load of horse shit I've ever heard. I'm pretty sure most of that information is from nasioc.

Gravity pushing the gears together, are you kidding me? The main shaft spins up to 6500 rpms max on a stock legacy, gravity is not going to be pulling that down anywhere if the case flexes. Even if you are talking about failures in the lower rpm range like 2000 rpms, that's still too fast for gravity to be doing anything productive. Whoever wrote that doesn't realize just how weak of a force gravity is compared to other forces.

The case doesn't have 10mm bolts running all around it, they're 8mm bolts around the majority of it, the 10mm bolts are around the differential and at the top. And so what if the case has 8mm bolts running all around it? The bolts only go around the case. There is nothing holding the case together in the middle. Between the main and drive/driven shafts is where there needs to be bolts. If there is enough force or heat applied to anything it will deform.

I don't think that whoever said that understands just how minute of a distortion is needed to cause a major problem. The case doesn't need to distort an inch for damage to occur, just a few thousandths of a mm can cause nasty problems if the conditions are right.

Now, do I believe that case flex is the problem? Don't know, nor do I care. I just know that whoever originally wrote that crap doesn't know much of anything about stress, froces, or pretty much any engineering. I'm not saying that the 5mt is a bad transmission, I'm just saying that the design of it wasn't made for the kind of torque these engines produce. There is a reason the 6mt was produced. Part of it was to stay competitive, but there were other reasons also, they definately could've modified the 5mt for another gear, and they didn't.

I happen to think case flex is real. I have seen the inside of the transmission and do not think the case is as strong as you might make it out to be.

PPGs being a solution is not an example of why case flex cant exist. In fact, its a perfect explanation for why case flex is the culprit.
For instance, if gears did break because of case flex, stronger gears such as ppgs would certainly help. The larger pitch and the increase in tooth size would mean that the teeth could walk on eacah other a much greater distance. PPGs have HUGE teeth - probably 3 to 4 times the size of the stock teeth.
Has anyone hardness tested the stock gears vs ppgs?

The newer (99+) cases are supposedly also less prone to flex - why they break less?

You have to remember, this transmission was originally designed to handle 130hp.

I actually just read on nasioc that rallispec tested the newer gear boxes and said that the main problem was input shaft flex. Apparently they measured wayyyy to much tooth separation at only 275 ft lbs.

Also, the stock sti gears are only marginally larger than our stock 5 spd gears. Their case however is much stronger... This is more evidence to support the case flex theory.

Regardless, until someone actually does their own ACTUAL experiments we can only theorize...

Danny

gto7419 wrote:You have to remember, this transmission was originally designed to handle 130hp.

More like 70-80

Case flex isn't real. Trust me. Over 800 wheel horse power, and we've never, ever, ever, ever, ever, ever seen a failure of the case.

Are you telling me that I don't know anything about engineering? Hah. Good one bucko.

First gear on the output shaft rides on a film of tranny fluid a few hundredths of a mm thick. If the gears are pushing apart, first gear on the output shaft would contact the output shaft. Yet, that doesn't happen, does it?

The 99+ transmission cases are identical internally than the pre 99 cases. I know, I've had them next to eachother. Actually, the pre-99 cases have more gussetts externally than the newer cases. The only thing that changed in 99 was the synchro engagement hub design. A WRX gearset will drop into the Legacy Turbo case without case modification.

The pre99 case is design is exactly the same as the STi RA design, which was running 276hp. The post 99 case design is the same as the 5-speed STi's from MY00 in japan, again running 276hp.

Rallispec's 276ft-lbs rating was against a stationary output shaft. They locked the input shaft and output shaft, and applied a load. To assemble a gearset, the locknut on the input shaft is torqued to something around 250ft-lbs. Either Subaru is stupid, or Rallispec is wrong.

The 6-speed gears are MASSIVE compared to the WRX gears. Easily twice the size. And the case is different because of an entirely different internal design. The 6-speed and the 5-speed, are like apples and oranges. They are two totally different beasts.

The rotational speed of the shafts doesn't impact the effects of gravity. Someone who says that clearly doesn't understand physics. Gravity is a constant force, there can be forces opposing it, but gravity is a given, and the acceleration due to gravity is constant unless other forces are introduced to oppose it.

Think of it like this. A straight cut gear is stronger than a helical cut get. This has been well proven, despite what some vendors would like to tell you to boost RA gear sales. A straight cut gears produce only seperative forces. The gears spread apart as loads increase. Helical gears produce lateral loads as load increases. It would stand to reason that if case flex was really a problem, a straight cut gear would be weaker than a helical cut gear, because the case would flex more with a straight cut gear.

The reason Subaru gearsets break, is because the gears aren't strong enough. Plain and simple, the gears are just weak. That's all. There's no bandaid fix. The case is bolted to the motor, and bolted together. Unless you stretch one of the bolts, or crack the case, it's not going to flex apart. I've seen more blown transmissions than your average person. I'm speaking from actual experience.

I assure you that strapping something around the transmission isn't going to help prevent gear breakage. The only way to prevent gear failure is to upgrade, or make less power.

I am going to assume that you know what the difference is between plastic and elastic deformation, and as such I am going to ask you how you can honestly say that the case does not flex. Don't get it twisted, I'm not saying that it's a problem, or that it even is a concern. I'm just saying it's silly to think that there isn't some change in the physical characteristics of the tranny with enough force applied.

I don't know what kind of transmissions you're putting together, but the stock ones have all the driven gears pressed onto the driven shaft, none of them are riding on GL-4/5. If the gears were riding on oil ontop of the driven shaft the car wouldn't go anywhere, as the gears would just spin.

If your point is about how the driven shaft sits on needle bearings on top of the pinion shaft, okay, whatever, don't care.

Just reread the part about how I said that I don't give two fucks if case flex is a problem or not, just that it exists. Everything with enough force will deform. You can argue all you want about whether or not it causes the gears to break or not. I agree with you that it isn't what the problem is anyway. Like I said, they went to an entirely different transmission for a reason.

The point of this discussion is whether or not case flex is a problem. My point is that case flex does not exist as a problem for gear failure. I'm not a child, nor do I live in some fairy tale world where objects behave in perfect motion.

Does the case flex? Given the right circumstances, yes. Atoms are constantly vibrating and moving, so technically, the only possible situation where any matter will not move is absolute zero. But that's not what we are talking about. Your picking a fight about one thing when the conversation is about something completely different.

As for transmission assembly, clearly you have NO idea what you are talking about.

First, reverse, and second gear are all machined onto the input shaft. First gear on the driven shaft rides on a film of gear oil. The 1-2 shift collar is fixed to the shaft. When you engage first gear, you are locking first gear to the output shaft. Second gear is exactly the same way, but it does ride on needle bearings. Third and fourth gear are the opposite, they are actually one machined piece that is pressed onto the output shaft and locked in place with a woodruff key, and the matching gears are set on the input shaft and supported by needle bearings.

If every gear on the driven shaft was pressed on, the car would be in first and second gear at the same time, all of the time. You'd have no neutral. Talk about not going anywhere! I'm sure you could see the problems with that situation.

The 5-speed is still the same as it was in 1989 on the Legacy. Minor changes, but the basic structure is identical.

This conversation has been a little over my head, and I don't really know why I bothered to read it all, but at least I got a new sig quote out of it.
It was just too good to pass up!8)

I think Phil has his info generally right...


It has always been my understanding that most Subie 5 spds break due to shock loads placed in the input (1st,2nd,Rev) shaft, and cause the shaft to distort and momentarily lose mesh with its complimentary gears...most all of the broken 5 spds I have seen the guts of support this theory...I don't honestly see how the internals of a Subaru Transaxle could place such a torsion load on the case that it would cause it to flex...the only way I see this happening is if the internals were to seize, and transmit the power from the engine to the case, in which instance, I am sure the case would flex, as that is just the tendency of materials when forces are exerted upon them.

I've spent a LOT of time with subaru transmissions.

The shafts in the transmission are not transmitting very much torque to the case. The axes of rotation of the shafts are perpindicular to the axes of support. The shaft is also supported on two ends by journal bearings. IF a load is placed on the shaft (from the shafts being forced apart), there are two points of support that lie in the plane of the case half helping to oppose that load. If the case was split horizontally instead of vertically, I could see the potential for seperation, but since it's split vertically, you have a better chance of STRETCHING the case than actually flexing it.

The shafts simply cannot move apart when they are bolted inside of the transmission. If that were the case, first gear would break far more often, because it is the only gear that doesn't ride on a needle bearing.

The only thing that I can visualize for the causes of gear breakage is the following:

Shockloads that exceed the shear strength of the gear tooth.

That's it.

It's not really possible for the teeth to lose mesh with it's complimentary gears, because that would mean that the shaft would need to flex and back off more than a centimeter (to totally disconnect). I don't know if that's what your saying or not, Nick. I think you mean back off enough to no longer be in it's designated place.

The problem with that, is that helical gears don't LIKE to back off. Helical gears like to walk apart from one another laterally, rather than pushing apart and backing off of one another. That's why the thrust surfaces are so critical to the 5-speeds operation (which, again, is a film of oil, not an actual bearing).

So really, I think the fact of the matter is that the stock gears are just not strong enough to handle the shock loads placed by higher output.

But does it explain how transmissions blow up while "just driving along"? Well, when you aren't launching and you aren't shifting like an idiot, the impulse loads aren't as high as if you were driving like every stoplight was a drag strip. So it does stand to reason that at certain power levels, the standard loads created by the engine are enough to match or exceed the shock loads created by launching hard or shifting fast. At that point, your fucked, and the trans is going to blow up.

I ran some arbitrary numbers through some calculations for gear strength between straight cut and helical, and if the stock gears were straight cut, they'd be somewhat stronger. But is it enough? I don't know. Plus it places different loads on the transmission that it might not have been designed for. Through experience, however, I've found that the straight cut gears in the Subaru 5-speed case are plenty capable of holding up to power.

I'm glad after all of these years, I've made someone's signature

What I meant was that the amount of runout between the gear teeth (which there is some, as most all machined parts have a wee bit) could cause the breakage of teeth when a shock torqueload is placed on them.

It could be, but I'd think the effects of runout would be marginal at best.

Again, first gear on the output shaft rides on a film of oil. Any significant increase in runout would shove it against the output shaft and weld the two together. Even when disengaged, first gear on the output shaft is still spinning, and any flex would surely manifest itself there first. At least with some scoring or something. But that never happens, at least I've never seen it.

I HAVE seen a first gear FUSE to the output sleeve, though. It was the result of using some bobo-brand gear oil in a high horsepower application. The fluid sheared, and first gear contacted the sleeve. It welded the gear to the sleeve. Don't use Some ghetto gear oil. This also happens when you have no gear oil in the tranny at all.

Not to change, but to add to this well-informed thread:

Built 4eat.

i blew 2nd gear again last night on my jdm 4.11 tranny

jake15 wrote:i blew 2nd gear again last night on my jdm 4.11 tranny

Ahhh man, that sucks!

Bummer Jake. Whats your next box going to be?

Hey, my NA 5mt MAY have broken today!

But I am suspecting something in the clutch (crossing fingers).....

It will only go into 1st, but I did get it to go into reverse for a second to back into the garage.....

It makes bad bad bad noise WHILE rolling, like something is spinning around and around and around inside. When you are stopped, NO noise.

I can put the clutch in, no difference in noise.

while going slow, it will sometimes bind up with a "bang!"

I think the pressure plate or 6 puck or is damaged and not the tranny because there isnt any "clatter" that sound like teeth, and the fact that the weird/bad noise only happens while moving....

Hmmmmmmm

If there's no noise at idle, clutch in or out, then that's odd. If it was the clutch, it'd make noise when you let the clutch out in neutral.

Tranny has to come out anyway to fix it. Inspect axles first, to make sure you didn't grenade a CV joint, and go from there. If the clutch is fine (which it probably is), then pull the tranny apart. It ONLY goes into 1st and that one time reverse?

I wonder if you broke a selector fork. That'll produce all of the symptoms your experiencing. It's likely the 3-4 fork that broke.