Grinding synchros...

[vrg3]

It's a specification from the American Petroleum Institute. The GL stands for "gear lubricant." I discussed it a little in this thread:

http://bbs.legacycentral.org/viewtopic.php?t=18054

In retrospect I should have just posted in this thread and linked to this thread there.

GM Synchromesh fluid, from what I can tell, is similar to a lightweight GL-4 fluid. In fact, Redline recommends using their MTL (a synthetic 75W80 GL-4 oil) as a replacement for Synchromesh. Synchromesh isn't actually specified to meet any API requirements as far as I know. They call it a GL-S oil, with the S standing for "special."

If not using GL-5, I'd start with something like Redline MT-90 (which at least has the correct 75W90 viscosity for our cars) before experimenting with something lighter like Synchromesh. I know MT-90 worked amazingly in my transmission.

[evolutionmovement]

Josh - it may be your transmission in particular as the ones I've had are notchy, but very easy to get in gear, sometimes almost guiding themselves. It's not a Honda, but I prefer it to most cars I've driven. I drove an '01 Maxima today that was far sloppier, notchier, and with a longer throw.

Steve

[90LegAWD]

well, being that it gets so damn cold here in MI winters, i'll be running 2qts syncromesh and 2qts gear oil then. my car will see a lot of winter driving, including ice racing.

[Legacy777]

Josh - it may be your transmission in particular as the ones I've had are notchy, but very easy to get in gear, sometimes almost guiding themselves. It's not a Honda, but I prefer it to most cars I've driven. I drove an '01 Maxima today that was far sloppier, notchier, and with a longer throw.

Steve

Yeah I don't know....but I really don't like it. The other thing too....without having two cars sitting side by side and having two different people drive them and give their impressions....you really can't compare them.

My opinion of not going into gear easily may be totally different to you. Personally I think the throws in the sti short shifter could be shorter, but that's just me.....which may account for me not being happy with the shifting.

I've got a set of kartboy bushings I'm going to install this weekend, and hopefully the neo gear oil will dramatically improve things. I'll be sure to post my opinions when I get it in.

[fishbone79]

I don’t speak up much, mainly because I like to know exactly what I am talking about when I do. I have pretty extensive experience with quirky old transmissions, particularly extremely old ones with square cut gears and no synchros. I live in upstate NY (quite close to vrg), and with all that works against me, I am usually able to get >225k out of any car I’ve owned (the last sube I dumped at 309k, original tranny). The trick I’ve devised for shifting these transmissions is somewhat involved, but extremely effective.

That being said; I’ll take the sube out of a drive, running through all the gears, and meticulously write down what gear I am in, and the start and finish RPM of the engine at my shift. I’ll shift each gear at a bunch of different RPMs to increase my sample size. Then I throw all this data into a spreadsheet, and graph engine RPM against gear (I know this sounds like a long process, but it goes quickly and you will absolutely thank yourself). If you look at these graphs, you’ll notice a trend in your engine’s preferable start and finish shift RPMs, as well as start and finish RPMs for different goals in any drive (i.e. takin’ it easy, goin’ balls out, and romping a buzz box)… the rest is simple. Knowing where to shift and when completely eliminates the need for double clutching (I first tried this on my 46 ford truck, no synchros and square cut gears), as well as silly shifting mistakes. The entire process can be done in your head, but I prefer to keep track so I can make predictions on where and when I should shift, as I drive …. When you get good enough, you will forget you have a clutch unless you are starting or stopping.

Sorry to bore you all with a longwinded response, I just hope someone else gives it a try. The tranny in my current 92 is almost completely devoid of synchros (I know because the majority of their remains drained out in my lube)… and I drive it everyday.

As an aside, there are 9 AWD subes (90-94) in my local junkyard, which has transmissions for 29.99 every other weekend. I picked up one and am rebuilding it for my inevitable tranny explosion. Hope all this helps, and I hope you all don’t think I am an arrogant asshole for my novel of a response… I have much to learn from all of you.

[vrg3]

Where exactly are you located, fishbone79? Which junkyard is this? The one in Horseheads?

I don't fully understand what you are describing with the preferable start and finish shift RPMs... Are you basically saying you collect data on how much the transmission layshaft slows down just due to friction during a shift, and selecting a shift RPM for each gear that makes the engine speed difference before and after the shift match the layshaft speed difference?

Wouldn't that mean you'd lose your say in when to shift? It seems like it would certainly rule out shifting at redline like when driving hard.

[fishbone79]

Junkyard:

http://www.garysupullit.com/

I guess maybe I should be more detailed… I’m digging for the info on my car now.
Vrg, the basics you’ve pinned… when you do the math on the data, you come up with a few different explanations, but the best way to verbalize it is this: (anyone who has ever driven an old tranny from the 40s or below will know exactly what I am saying) you have a particular “window” of engine RPM in which a shift is precise... it helps to think backwards. For instance, if I was in first gear, wound my motor up to 3000 RPM, the shift into second gear will place my engine at a particular RPM, say 2200 (which is not exact, but im looking for the data …. also given no other external influence, hill, rough road, etc.) The “shift window” would be 800 RPM. However, this window is not fixed up and down the RPM scale; it broadens or narrows depending on the gear. If I wind first gear up to 4500, my second gear shift may put my engine at 3100RPM (again, for instance). The only way to achieve complete success with the method is to get a lot of data, I use a tape recorder. Shift all gears at any random combination of RPMs, low, high, in between. Graphing the data points will show you an exact set of lines for each gear.

I suppose this also depends on your method of shift, I will generally double clutch all the shifts to be sure I am finding the exact RPM to match the engine/lagshaft/axle speeds (I guess I should have noted that above, apologies). Many freight trucks have gages to measure all of this in real time because truckers never use the clutch. Correcting for a non-double clutch shift is simple… initially, you must be exact in your time lag during a shift (for instance, keep the time period from when you leave first gear in the tranny to when you are into 2nd around half a second). This keeps the loss of lag shaft speed due to friction at a relative constant (it is by no means a true constant, but is close enough for our first order approximation).

Given that you already know where it should be shifted, you should be able to calculate the loss of lag shaft speed during a shift by recording the engines RPM right after a non-double clutch shift. Does this make sense?? I have a hard time putting it all into words… its all about matching the speeds of the 3 different spinning things in your car. Skip barber’s book goes into great detail, and does a much better job explaining it, but I’d have to go hunt it down. If you think about it though, it makes perfect sense.

That said, in practice, you can shift at whatever RPM you’d like, at redline… wherever, but the engine RPM and transmission speeds follow these rules strictly… there is a specific difference between the Initial and final engine RPMs of a shift that can be graphed and calculated. Although this “window” does not necessarily remain constant, different road/terrain conditions effect engine RPM dramatically… and all that could be graphed with hill pitch percentage and road friction coefficients if you were really ambitious… someone should defiantly let me know if they are .

I fear I’ve gotten way to theoretical for this post… there is a quite bit of literature on the subject, some of which I may dig up and put in a new thread if anyone expresses an interest. Let me know

Cheers

[fishbone79]

A little update: These are technically called shifting curves or RPM curves, and the shift window is termed “Engine RPM shift Compensation.” I’ve found all the data for my 92 legacy; its on my old lapper which currently has no method of data transfer (haha, no burner, no floppy, no LAN card, and no working USB port). I will use its modem later tonight to transfer the data and make it available if someone seems interested. I find it fascinating.

I’m fairly sure you can request all this data from subaru’s initial dino tests, complete with rating curves… how long it would take to get, I have no idea. Also, and I am serious about this, if anyone is interested in collecting data for me, particularly concerning hill grade percentage, road roughness coefficients, tire inflation differences and possibly even engine/tranny lube weights for some statistical analysis and modeling concerning RPM variations, we could be looking at some publication possibilities. Vrg, you can get me on my gAIM handle, fishbone790, later… please no pranking.

Cheers,

Morgan

[vrg3]

Oh, Gary's! Cool... you're near Binghamton?

I think I kinda understand what you're talking about... but not fully.

How can you calculate the loss of layshaft speed by recording the engine speed after a non-double-clutched shift? When you're in gear engine speed is rigidly locked to road speed, no?

The size of the window you describe is determined solely by three independent variables: the gear you're in, the gear you're shifting into, and the speed (pick whichever speed you want -- road speed, pre-shift RPM, or post-shift RPM). Right? But I don't see the significance of the window.

I see what you mean about it not necessarily limiting when you can shift... It just places constraints between the speed of a shift and the time you can make the shift. Right? Like, a shift at 2000 RPM can be done much more quickly than a shift at redline, because you don't have to wait as long for the layshaft to slow down due to friction.

I'd really like to understand what you're trying to explain... if you happen to find some of the literature about it please do post it in another thread.

[evolutionmovement]

I may need to hit that place in the future for an AWD tranny. At that price I can have it cryo-treated and built up. But I wonder what a new Legacy GT 5 spd would cost ...

Isn't this graphing idea basically what you learn as you drive the car?

Steve

[fishbone79]

The loss of layshaft speed is only while shifting… when the shifter is in the N position between 2 gears; the shaft loses it’s RPMs. As we all know, the point of the double clutch is to compensate for this loss. When the shifter is in a gear, the layshaft RPM is dependant on the road speed. The math behind gearing is complex, but simple in that all changes that take place can be considered relatively linear (for our purposes).

The RPM window is not terribly significant in and of itself. The pre-shift and post-shift engine RPMs are the important factors, but people seem to have trouble thinking that way for some reason, so the window idea is simply for idealization. The size of this window, as you stated, is certainly dependant on three variables… but the road speed variable can be thought of as almost irrelevant if we assume the testing to be done in flat conditions on normal tarmac where all shifts are timely. Road speed can almost be considered increasing linearly while in gear, (although it is certainly not nearly this simple…) and for our purposes, with the above assumptions, holds relatively constant throughout the duration of the shorter shift, although it does decrease slightly.

You are absolutely correct, on your third point, if you look at the trends in pre/post shift RPM, you’ll see that a shift at a higher RPM is generally slower because the engine speed has a larger ‘window” by which it much be reduced before the next gear can be advanced. AND the layshaft has a larger RPM gap to make up as well.

All this leads into theory on the most efficient methods in racing… essentially, it can tell you where your shifts will be quickest and give you the most power. Considering the road speed as relatively constant during a swift shift is fine, but road speed decreases can not be ignored during a much slower shift. Think about this situation, you want to pass someone and are in fourth gear… at 2800RPM, when during the pass should you shift into 5th to make the pass the most expedient? If you give it WOT, then you’ll certainly pull worlds ahead, but the lag time between your shift from WOT in 4th gear to 5th gear is much larger than if you had shifted at 4500 RPM in 4th.

I don’t think people have any idea how much more goes into driving a car truly without sychros… And yes, all this is learned when you drive the car, all quantifying it does is allow you to make subtle changes, keep records, and increase the lifespan of your transmission exponentially. If you think you know how to shift, try driving to work or wherever without touching the clutch other than to stop or start… if you know all this information, you can do it in a breeze without any ill to your transmission. If you don’t, you will mess things up and put a lot of pressure on the synchros.

I guess I went on a rant… sorry all, I just find this stuff fascinating. And I’m sure my convoluted explanation of it all is not helping a bit.

Cheers,
Morgan

[fishbone79]

Sorry, maybe I should clarify that engine RPM in the first couple of posts is being used as an indicator rather than a dictator of transmission/road speed. Since it’s the only thing we can gather empirical numeric data from, it must be this way. Once the data has been assembled, we can turn the situation around and predict what will happen with engine speed changes in terms of their effect on road speed, tranny speed and shift lag.

[vrg3]

Well, the variable of road speed could just as easily be replaced with pre-shift RPM or post-shift RPM. Or you could use pre-shift RPM, post-shift RPM, and the starting gear. My point was that there are three independent variables, and no more.

I think I understand now... You're just describing "matching speeds" without double clutching or otherwise rev matching. When driving a certain car that I don't especially care for I can occasionally get "into the zone" and do just that... I drive it all around town using the clutch only to keep the engine from stalling when the car stops.

[fishbone79]

Precisely, it’s as simple as that.

[fishbone79]

Another note, vrg, attaching a data logger and temporal positioning indicator to your BC/BF scan tool would be an incredible advantage. I know this has been mentioned in the past, but I can only imagine the amount of data that could be collected with it… just need some way to distinguish data points by gear…

And yes, Binghamton... Gary's is an awsome resource. They are always getting subes in.

[vrg3]

Yeah, the scan tool could use a lot more working over... at some point it'll get that attention.

The few times I've been to Gary's I saw a few Legacies but they'd been already stripped of a lot of parts... maybe I need to go on the day they put the new cars out in the lot.

[-K-]

Let me put in one vote against thicker oil.
I got a little grind 1-2 and 3-4 no matter how easy I would shift. When cold it would not shift for anything, it would even grind going OUT of 1st!

So I drain the old stuff and it's way thicker than any 70-90. I guess the previous owner thought thicker oil would help the grind, it DIDN'T and I think it made it worse.

I decided to use MT-90 to replace it, I'm less worried about breaking the tranny because it's only GL-4 than the grinding. So far I like the MT-90 and just being able to shift in the morning is nice. I could still grind 'em if I shifted real hard but I'm easy on the tranny and don't get any grind now.

EDIT for some confusion

[Legacy777]

Wait....so you drained out the MT90, or you used the MT90 and didn't like it? But then you say you do like it......

[legacy92ej22t]

OK, so ever since I did my clutch (I used the cheapest gear oil we could find) I've steadily gotten crunchy gears in pretty much every shift. So, I want to get some quality gear oil. I'm probably going to go with Redline. I was wondering what the difference is between the 75w-90 and the 75w-90ns? And is the concensus that the Redline 75w-90 is the best to get? TIA.

[Tleg93]

http://racerpartswholesale.com/redtech4.htm

It seems like good information but first hand info is probably better. Hope it helps.

[vrg3]

Redline 75W90 has friction modifier additives in it which are needed in mechanical LSDs. Redline 75W90NS doesn't have those modifiers.

They're both GL-5 gear oils, so either one should be safe to use. I don't know anything about how friction modifiers work so I have no idea whether they could affect shift quality.

Redline recommends their 75W90NS for our gearboxes and rear differentials.

I don't think there's any kind of consensus among us though.

[legacy92ej22t]

Scott- Thanks, that is pretty good info even though it mainly about MT-90.

Vikash- Thanks, 75w-90ns it is!

I thought maybe the NS stood for Non-synthetic but it was still listed as a synthetic oil, so I wasn't sure.

[entirelyturbo]

I should add, I might need to start trying different gear oils again.

I dunno why, but my gearbox has gotten so grumpy, I almost would rather drive the wagon just so I don't have to deal with it. Upshifting is fine, no complaints. But downshifting? What a chore. And I don't mean dropping it into 2nd at 50mph so I can be bouncing off redline through a corner, I mean easing it into 2nd at 20mph so I can make a 90º into a parking lot. It doesn't grind or anything, but I really have to have patience, or I have to force it in and feel guilty about it.

Believe it or not, Ginger is waay more cooperative than Patti is. I think the reason was my very poor driving habits when I was younger. When I was 16, I would downshift and use engine braking, as in put it in 2nd at 30 and slowly take up the clutch to slow the car down, and I would do this almost every single time I slowed down. I'm fairly certain that's why my clutch conked out at only 110k, and I think that's taken its toll on the synchros. I've never done that with Ginger, because I learned better by then.

What would be the ill effects of running straight GL4 Redline MT90? Would I screw stuff up worse?

[legacy92ej22t]

Well, I ordered 4 qts of Redline 75w-90ns and it was only like $40 shipped. I thought it was going to be a lot more expensive then that. I'll let everyone know how it works out.

[vrg3]

I thought I had written up this explanation already... I guess I forgot to. Here's what I've been able to piece together over time:

Differentials use bevel gears to connect shafts that are perpendicular to one another.

Back in the day, they used helical bevel gears in rear differentials. These gears are pretty much bevel versions of what we normally think of when we imagine helical gears. The driveshaft ran down the middle of the car and had to be in the same plane as the rear halfshafts. So you had this huge hump along the middle of the interior of the car.

Then they invented the hypoid bevel gear. The clever design of these gears allows the input and output of the bevel gearset to be in different planes. That means the driveshaft could sit lower (greatly reducing the size of the hump) while still allowing the axles to remain at the same height. Brilliant!

The thing about hypoid gears, though, is that they involve a lot of hardcore metal-against-metal sliding. Normal oil would just get sheared apart. To deal with this, oils for hypoid bevel gearsets include extreme pressure additives. EP additives are sulfur compounds of some type that bind to the surface of the metal and remain between the two metal faces even when subjected to extreme load. EP additives are essential to the longevity of hypoid bevel gears.

These sulfur compounds do have a downside, however. At least in certain formulations, they can be corrosive to yellow metals such as bronze and brass. Most transmissions' synchronizers are made of yellow metals. Apparently it is possible to formulate EP additives that are inert with respect to yellow metals, but it is either harder or more expensive than simply using the common formulations.

From what I can tell, the primary difference between the GL-4 specification and the GL-5 specification is the quantity of EP additives. GL-5 oil has a lot more EP additives than GL-4 oil. Most of the time, GL-4 is specified for transmissions with helical gears, and GL-5 is specified for differentials with hypoid gears.

So the synchro-eating sulfur in GL-5 is not usually a problem. I mean, transmissions and differentials are separate things, right? So that's fine -- the EP stuff goes in the rear end, the synchro-friendly stuff goes in the transmission, and we're happy.

But what about front-wheel-drive cars? It turns out they're okay, too, because with a transverse layout you end up with things laid out in such a way that helical gears work fine. So we use GL-4 in FWD transaxles (transmissions with inbuilt differentials) with no problems.

But then... along comes Fuji Heavy Industries, with a really odd design. A traditional longitudinal layout, but front wheel drive (for the purposes of this discussion we don't need to worry about AWD). The layout is pretty clever, but it does require a hypoid differential in order to fit everything in a reasonable amount of space. So we have a transaxle with a hypoid bevel gearset.

This hypoid transaxle makes us nearly unique in the automotive world, from what I can tell. We need EP protection from the same oil that we need to protect our synchronizers. Subaru wasn't just being dumb when they specified GL-5 for our gearboxes even pretty much every other Japanese car uses GL-4; there was a good reason. And this is why it is ill-advised to use oils like Redline MT-90, GM Synchromesh, or BG SynchroShift, which work like magic in most synchronized transmissions, in ours -- you risk damage to the differential.

I'm running MT-90, a GL-4 oil, anyway, but I can't in good conscience suggest to anybody that it's a smart thing to do. I sleep at night because of four thoughts:

- GL-4 does still specify some EP additives, just not as much as GL-5.
- Brand new quality synthetic GL-4 may well have nearly as much EP protection ability as 100,000-mile-old dino GL-5.
- I've never heard of a Subaru front differential grenading under reasonable use.
- My car shifts soooooooo well, whereas just before the change I couldn't get into 2nd gear without matching very very carefully.

Whew. Sorry for being so longwinded. For hardcover copies write to...