How to Determine Replacing Ring and Pinion Gears UPDATED

How to Determine Replacing Ring and Pinion Gears

I�ve decided to assemble a photo serial of my reversed band & pinion setup from Subarugears as I anticipate at that place are many others out there whose projects will halt once they reach the step regarding setting upwardly the band & pinion gear mesh. I have not constitute any sufficient resource online that speaks directly to the real-life setup concerning this step. Though Subarugears and others here on the samba have washed this step, I take felt that the level of detail specific to this phase has not been at the granularity I had hoped for. Seeking a sort of �mesh-map� that is more than but an artistic rendition of general instance examples skillful and bad, I�ve decided to assemble one with photos sourced from my specific trials and tribulations.

For starters, here is a bit of background on my situation and how I arrived at my decision to go with the Subaru 5MT as my gearbox of selection:

Beginning and foremost, mayhap I am a chip more neurotic regarding the gear mesh as my Vanagon has a history of eating gearboxes. Be it the appetite of the EG33 Subaru SVX ability establish, the atomic number 82-weightedness of my correct foot, or the alluring roar of the Boxer H6 when this combination comes together� big smiles happen, that is until something else in the transaxle gives upward and brings the abominable camper to its knees. Offset it was a thing of sheering all of the teeth off of quaternary gear (in the original aircooled gearbox)� then it was grinding the teeth away from the ring and pinion of the 094 late model, Rancho-built gearbox (obviously due to the pinion nut backing off and allowing excessive slack/backlash, which they so kindly replaced)� at present, I cannot tell for certain as I oasis�t opened information technology up all the same for fear of what I may find (on the newest Rancho-built replacement gearbox).

Disclaimer: Don�t let me give you the incorrect idea of Rancho Performance Transaxles� their staff has gone above and beyond for me in taking care to replace my damaged gearboxes, despite my non-stock drivetrain configuration (which they are/were aware of). Their R&D department has used cases like mine to refine their builds so that others in the future will benefit from beefier parts of identified weak points, inhibiting potential failure.

Regardless, the time and money investment in gearboxes has been much greater than anticipated. So, for any of you who are toying with the thought of a Boxer H6 in your Vanagon, consider yourselves warned. Yeah, it will be most fantabulous being able to pass virtually everything on the road in your brick-mobile, though there will exist many other necessary (and costly) upgrades required to continue you rubber from the vehicle�s Jeckle & Hyde esque mood swings (ie: brakes/suspension/bushings) every bit well as keeping the vehicle condom from itself (ie: transaxle upgrade).

With that said, my Vanagon had noticeably large gear gaps in the latest 094 gearbox, then dandy time and try was spent selecting the platonic gearbox and gear ratios when I had decided to become with the Subarugears Reversed Band & Pinion kit.

My previous setup consisted of a factory Vanagon 1st -2nd, with Weddle 3rd-4th and the 4.868 Band & Pinion, turning 225/75/16 serial tires (~29� bore).
Hither are the particulars regarding the ratios (for the gear heads out there):
1st � iii.78
2nd � 2.06
3rd � ane.14
quaternary � .70
R&P � 4.868

Overlaying the previous air-cooled configuration (with 27" diameter tires and 4.57 Ring and Pinion) with the 094 Rancho upgrade (with 29" tires and 4.868 Ring and Pinion) looks like this:

Though these gears might take been ok for a Vanagon with a tire bore up to about 27�, they were not ideal once I hit the hills as 4th was way too tall and the gap was excessive between 3rd and fourth despite the hefty ability proceeds from the SVX H6.

Afterwards much scrutiny, I�d arrived at the JDM Turbo Forester Gearbox of the model year 1999-2000 range (transaxle code TY55VB1AA). This gearbox was built to handle the ability of the WRX/Turbo Forester so it should be able to take the ability of the SVX without compromising its integrity. This gearbox was also ideal equally information technology has the male differential output shafts which are much easier to accommodate the Subarugears axle flanges without having to purchase stub shaft adapters separately. Unless you lot're looking to upgrade the differential right out the gates, this is a toll saving point to consider. The TY55VB1AA ratios wait a petty something similar this:
1st � 3.454
2nd � 1.947
third � 1.366
4th � .972
5th � 0.738
R&P � iv.868

Overlaying the previous 094 Rancho configuration with the Subaru 5MT upgrade looks like this:

For my purposes, these ratios were ideal. I kept 1st and 2nd inside 2mph of the stock Vanagon gears @ 4k RPM meaning I volition notice very picayune (if any) difference in gearing while out exploring woods roads. tertiary and fourth will be the largest improvement over the previous setup as the Vanagon would previously be turning nearly 63MPH @ 4k in 3rd, with the Subaru Gearbox information technology�ll be about 52MPH in 3rd and just shy of 74MPH in 4th, and so my relative gearing proceeds is substantially +/- 11MPH @ 4k RPM on either side of the previous third ratio. I believe this volition give me optimal range when I am traveling the mount highways as tertiary was previously too brusk for highway travel (revs were too high) and 4th was likewise tall. Merely as Goldilocks found with the third basin of porridge, this combo should be juuuust right. As I found 5th to exist a scrap on the tall side, compromising with a .738 5th should bode nicely with the overall gearing in this configuration.

Now that you�ve been primed with the background of my state of affairs, let�s go into the task at hand.

I don�t plan to become into detail regarding the required example modifications as they are pretty well documented already and are a sort of trial-and-fault type deal. In one case you�ve flipped the unequal, you lot�ll exist able to run into where contact is existence made. At that bespeak, grind the contact points, recheck clearance, grind the contact points, re-check clearance, and so on. Subarugears documents that flake of work here (http://www.subarugears.com/Casing/Casing.html). I�ve found that applying a thin layer of gear paint to the surfaces that are not immediately identifiable one time the majority of clearance has been fabricated will assist to highlight the contact points as they volition be shining bright in contrast with the paint. Just my $.02 on that note.

Specifically, once the case is clear, the ring gear is installed on the unequal and the pinion shaft is installed within the output gear stack, we will arrive at my major point of disharmonize. At this stage of the install, ane would hope everything should magically fall into place� A bit of wishfull thinking? Peradventure. Reality however oftentimes times proves to exist a harsh mistress, rivaling figures of demon lore.

Following the protocol described in the assembly video on the Subarugears site (http://world wide web.subarugears.com/Assembly/Assembly.html), I first measured my pinion depth with ZERO shims to see approximately how many tenths of a millimeter worth of shims would be required to bring my pinion depth to the ideal of Zero. Well, for starters, my pinion depth guess (as provided by Subarugears in their install kit) didn�t quite fit in the alignment holes as perfectly equally I retrieve they should have. The dowels of the approximate were just a bit out of stage and required a chip of persuasion (a la ball-pin hammer) to get the dowels to seat into the instance half. Once �installed�, at that place was a slight upwards bow in the gauge that made the middle read point out of airplane with the example one-half. To accept my reading, I lightly pressed the middle of the plastic behind the read bespeak to bring it back to the plane of the instance halves. Once it was as close equally feasibly possible to flush, I slid the �punch indicator� between the �read betoken� and the pinion head to determine the required pinion shim thickness. Due to the bowing of the gauge, my readings were anywhere between .7mm and .9mm at whatsoever given moment when measuring with Naught shims, so there was a bit of range to the determining of the initial starting betoken for the pinion depth on this project.

My gearbox was delivered with a manufactory shim array of approx. .63mm, I needed to place an order for more than shims that could get me to go me to the upper end of my ballpark measurement of .9mm. I ordered up a few shims, 2x of the 32295AA091 (.300mm) variety and 1x of the 32295AA071 (.250mm) variety. Since Subarugears advises that most cases would need about .45mm-.50mm of shims to zero out the pinion gear, I erred on the lesser side of .9mm equally I figured my wonky pinion depth measurement likely was non telling me the whole truth. With the pinion depth approximate�s range of 0.00mm-one.00mm, .9mm seemed excessive to me, but at that point, I was non at liberty to challenge the approximate.

Once the new shims arrived, I jumped to beginning the assembly process. In went the fresh shims at .85mm (actually reading closer to .83mm with my micrometer), then the pinion gear gets dropped in and aligned with the pinion bearing dowel and bolted in place via the pinion keeper bearing/plate on the pinion shaft reverse the pinion geared. Notation: After the many installations yous�ll soon read about, I had stopped installing the input shaft despite education to do so as I had felt that information technology offered no benefit at this signal of the install and only added boosted time and piece of work to the process.
Lots of momentum at this point; All is proficient in the neighborhood.

Next, I flipped the bearing races from one side to the other (this keeps them associated with their original bearing races [mail diff reversal], I also have learned to turn-in the diff cup of this case half and then that there is excessive backfire to start with [reasons will soon unfold]),
Then we driblet the differential with the reversed ring gear downwards into the bearing race and into the screw of the pinion gear� Ah-aye, gear cog lovin�.
At present, we�ll back the diff cup off of the other instance one-half (the half clearanced for the ring gear reversal) so that there is no preload on the differential bearings. Install the case half and bolt in place, snugly-tight. Boom. Movin�. At present fourth dimension to conform the backfire�

Instead of flipping the transaxle over to the contrary side every bit the factory documentation suggests, it is much easier to get out information technology in the orientation it was upon associates for the side by side adjustment. Doing then will allow for us to drop the differential unit of measurement downwardly into the spiral of the pinion gear via gravity as opposed to pushing information technology up into the spiral equally nosotros would if we were to follow the instructions verbatim. (ix.81 points for working with the Yard-force!)

It didn�t occur to me at starting time that the nature of this ring and pinion reversal would imply most of the instructions would, in fact, need to be translated into the exact reverse of what the text/graphics are saying. Keyword: Well-nigh (but not all), then don�t get too ambitious with reading into the opposites. This adds some other layer of �FUN� to the procedure� a sort of ENIGMA machine (https://en.wikipedia.org/wiki/Enigma_machine) requiring a drove of the sharpest minds of England to decypher. Well, maybe information technology was non that involved, though the varying caste of opposites in this exercise are far from directly forward.

So, nosotros�re dropping the differential/ring gear down into the hypoid spiral of the pinion gear by turning the loving cup on the bottom side of the case (side opposite the ring gear) counter-clockwise while nosotros are turning the pinion gear by hand, feeling for a slight resistance in the input shaft, at which point we�ve reached the Goose egg depth or Nada backlash point. While enroute to the Zip depth point, I tend to hand tighten the cup on the superlative side (ring gear side) of the case to relieve the slack in the differential bearings created from our diff-drop do. This helps to keep the differential unit perpendicular to the pinion shaft equally we approach the ZERO depth betoken.

The Cipher betoken approaches very quickly once evident. I�ve found that there is little effort required in turning the pinion shaft for well-nigh of the diff-driblet stage, then a slight plough of the begetting cup in the drop direction (say .5-i tooth worth) and a slight resistance is felt, then some other slight plough of the begetting cup in the drop direction (another .v-i tooth worth) and now the pinion shaft is leap by the band gear and nothing wants to plow with ease. At this point, back the ring gear out to the betoken where slight resistance is felt, have any slack out of the ring gear side bearing cup and nosotros�re prepare to motion on to the next step.

On the ring gear side, back the diff bearing cup off one.5 teeth. (NOTE: the bearing cup lock plates are adjustable to the .5 molar level by flipping the lock plate to its other side. Quite nifty little buggers.)
Opposite the ring gear side, tighten the bearing cup 1.v teeth. At this bespeak, the backlash SHOULD be within the acceptable range. Finally, tighten the bearing cup reverse the ring gear side of the differential an additional 1 tooth to set the differential bearing preload.

I�ve tried to apply the damn dial gauge every bit chosen for at this indicate, though the service certificate suggests to have the reading off of the band gear teeth through the bleed plug pigsty (a exercise no longer valid with the ring gear reversal as it is no longer in-line with the drain plug pigsty). The suggested read range on the gauge is in the realm of .13mm to .18mm or .0051� to .0071�. These are the values specified past taking the measurement from the RING GEAR� which we�ve determined we can no longer do as a result of ring gear flipage.

Let�s get into some maths� assuming that the ring gear has a diameter of approx. 7� (for the sake of illustration as I practise not have the band gear in front of me at this fourth dimension for hard mensurate), that generates a radius of 3.v�. No surprises there. Now, since the ring gear�s outer diameter is where we are supposed to be taking our backlash measurement, whatever the resulting read returns is a mensurate of the ARC LENGTH (or as in our case, the Backfire) on that outer edge of the ring gear circle.

For the curious and/or an ARC LENGTH concept refresher: http://regentsprep.org/Regents/math/algtrig/ATM1/arclengthlesson.htm

Knowing our radius to be iii.5� and our target ARC LENGTH (or Backfire) reading to fall inside the range of .0051� to .0071�, nosotros can make up one's mind the inner angle range responsible for rendering the desired ARC LENGH range using a unproblematic formula:
RADIAN Measure (inner angle) = ARC LENGTH (or BACKLASH) / RADIUS (of ring gear)

Then, then RADIAN Measure (lower extreme) = .0051�/ 3.5� = .00146 radians
And, RADIAN Measure out (upper farthermost) = .0071� / iii.5� = .00203 radians

These are the EXTREMES of the relative inner angle with respect to the � the DIAMETER measure (the RADIUS) of the band gear. The Subarugears set upwards video suggests that the same BACKLASH range be taken from the splines (or the pinion nut) of the pinion shaft, however the relative diameter of the pinion shaft to the ring gear is DRAMATIC! We�re talking ane� versus about vii� here. There�s no style that the backlash reading taking at a i� diameter is going to exist the same as that taken at a 7� bore, considering� SCIENCE! (well, Mathematics rather� merely I digress.)

Rearranging our formula a bit, we can solve for what the ARC LENGTH (Backfire) should exist when taken on a diameter of 1�.

RADIAN Measure (inner angle) = ARC LENGTH (or Backfire) / RADIUS (of pinion shaft)

Multiply both sides by the radius and nosotros accept:
RADIUS (of pinion shaft) 10 RADIAN Mensurate (inner angle) = ARC LENGTH (or Backlash)

And pivot the formula around the �=� sign:
ARC LENGTH (or BACKLASH) = RADIUS (of pinion shaft) x RADIAN Measure out (inner angle)

Inputting our known angle values from above:
ARC LENGTH (or BACKLASH) = .5� x RADIAN Mensurate (inner bending)
ARC LENGTH (or BACKLASH lower extreme) = .5� 10 .00146 radians = .00073�
ARC LENGTH (or BACKLASH upper extreme) = .5� 10 .00203 radians = .00102�

Equally you can see, a Backfire range of .00073� to .00102� is HUGELY dissimilar than a Backfire range of .0051� to .0071�, like a 7x level of hugely different!! At this point, I decided that dealing with measuring the backlash was going to conflict with the Subarugears� suggested practice and past following the suggested protocol, I�d likely be running into problem downward the line, as following the suggested route volition put me dramatically out-of-spec, far more than if I were to do it by feel post-obit the manufactory documentation minus the now impossible to practice band gear backfire measure. Having such a sour past with Ring and Pinion failures, I exercise non want to chance destroying a part that costs $1.5K due to procedural oversight. Unfortunately, it seems that anybody has been following this practice in their DIY setups� Hopefully this commodity will inspire a refresh and/or detailed documentation explaining this pace a scrap improve for others out in that location who are eager to utilise their production.

At whatsoever charge per unit, I�ve elected to run this course in backlash aligning past feel as it seems that the reads I�ve been getting from the contact blueprint (as viewed in later steps) suggest that my backfire hasn�t been obscure. Moving correct along now� Everything has been assembled, and the pinion depth and backlash accept been reset to one another, guess what we get to do?? Nosotros get to take everything apart once more!!

This time, we will not change whatever of our currently defined settings every bit we are going to bank check those settings past reading the contact mesh pattern. To practise so, we volition paint virtually iii-4 gear teeth every quarter turn of the ring gear with special gear pigment. At first, I used the Dykem �Steel Blue� machinist paint, though I establish that the moving-picture show was too thin and information technology was quite difficult to decipher the contact blueprint with this medium every bit information technology was relatively transparent and had inconsistencies with thickness across the gear teeth. Now, I�g using the Richmond Gear Pigment bachelor through Summit Racing for less than $x for what can easily be a lifetime supply for those who don�t intend on rebuilding gearboxes anytime in the foreseeable time to come. I call up information technology would be a nice improver for future releases of the Subarugears Ring and Pinion kits to include a bit of gear paint as I�d easily spent an additional $xx between the Dykem and Richmond Gear paint trying to track down the all-time product (and at present I�m left with style more than I demand). A sample well-nigh the size of a hot-sauce/ketchup/mustard/mayo/etc. parcel would be plenty for the DIY job and one less thing to track down for the DIYers once their kit is in manus... Just a thought.

With the quarters of the ring gear painted up, we are set up to re-assemble the transaxle.

At present that everything is dorsum together, we�ll spin the input shaft both clockwise, then counter-clockwise enough times for the differential/ring gear assembly to make a few good full revolutions in each management. This method, withal applies a relatively minimal load on the contact point between the gears. To combat this, I also recommend spinning the output splines from the differential a few full cycles in both directions as this effectively multiplies the load nosotros�d administered through turning the input shaft past a factor of the given ring and pinion ratio (in my case 4.868 times the load). This helps to simulate the effects of the transaxle being loaded by the forces exerted on the gear tooth contact patch under normal, in-vehicle functioning, which in turn gives a more accurate/dynamic gear mesh smear blueprint on the gear teeth.

After your forearms have had themselves a solid workout and the gears accept been essentially cycled, approximate what we get to do (you might be seeing a tendency by this point)� that�southward correct, have it all autonomously over again!! Horray for back-up!! Pardon my sarcasm hither� you lot have to understand that I�ve done this now more times than I tin can count and it is very well to the bespeak that I can (and am currently) doing these steps over and over again in my sleep. In all fairness, this footstep is necessary and will surely be repeated at least a handful more times in the consequence that you are lucky and arrive at the platonic mesh blueprint sooner than I did, which is what this whole commodity aims to foster in the first place. If information technology tin help anyone salvage time on their build/associates, and so I�ll chalk this discussion up equally a success. I�ve spent countless hours assembling, disassembling and reassembling this gearbox when I had initially expected it to be quick and rather painless process (equally the assembly video and other Samba threads seemed to suggest). Having already converted my Aircooled �81 Vanagon to Watercooled Subaru SVX power, modified the ECU wiring harness, rewired the entire vehicle with a late model harness, converted to a late model transaxle and shift linkage, and upgraded nearly everything underneath the Vanagon involving brakes/steering/intermission myself (and occasionally with the aid of friends)� I was entirely surprised when the reality that this upgrade ended upwardly causing me the most grief of all. At any charge per unit, where were we� oh yes, rinse and repeat.

Upon disassembly, you might see a contact wearable pattern that looks something like 1 of the combinations on the chart below:

As yous can see, I�ve run through this gauntlet of gears quite a few times at present� In this latest cycle, I�ve been taking photos of the patterns at approximately .2mm pinion shim adjustment intervals. As I approached the �Ideal�, I realized that a chart like the one above would have been hugely helpful by expediting my assembly past giving both a target and a relative map to get there. I at present figure that after the initial gear mesh read, I�d likely demand only 3-5 more assembly cycles to arrive at what we could categorize equally the �all-time� mesh pattern available while working in the finite 0.0mm-1.0mm (or so) realm of adjustment. My promise is that someone out there could run through their initial adjustment, lucifer up their pattern to the chart, and then adjust their pinion depth accordingly by calculation or subtracting the approximate shim departure from their initial gear mesh reading to one on the chart that is more ideal without having to go through the insane corporeality of associates cycles that I have thus far. So once more, if nobody sees the value in this gear mesh lookup tabular array, that�s fine� I know now that information technology will be of tremendous help for me in the consequence that I end up doing another Subaru 5mt assembly in the hereafter.

Unfortunately, despite my arduous efforts, I yet practise not feel that I have arrived at an acceptable gear mesh pattern for this Ring & Pinion gearset.

Comparing to the factory-spec illustrations:

And to a very helpful YouTube explanation of the mesh blueprint crusade and effect relationship:

Link

�I am not convinced that I have arrived at the correct blueprint, though at this point, I am unsure of what to practice short of machining the shimmed side of the transaxle case down to permit boosted pinion depth since my readings have never indicated that I've arrived at a point anywhere virtually "besides deep". It appears that the measure of .00mm has the best read on the Coast side, though the Bulldoze side is still suggesting that the pinion shim is too thick... Well, with naught shims in the picture, how this is possible is beyond me. I cannot tell if my transmission instance is defective, the Subarugears Reversed Ring & Pinion is lacking, or if my power to sympathise/implement/follow directions is defective. All I know is that my project has now been stalled out and I've been not been able to brand sufficient progress since February of this year. My hope is that this thread will generate a word about this comparatively documented pace and hopefully bring in thoughts and opinions from those of you who have either installed the Subarugears 5mt Reversed Ring and Pinion in your Subaru transaxles, those with feel setting upwards ring and pinions and/or those of you looking to accept this road for a transaxle replacement. Whatever input volition exist greatly appreciated. Cheers!
_________________
1981 VW Westfalia SVX 5MT
1990 VW Westfalia Subaru Franken two.5 (Automatic)
1989 VW Syncro Panel Tiico ii.0 (presently to be EJ22)

DOWNLOAD HERE

How to Determine Replacing Ring and Pinion Gears UPDATED

Posted by: barbaraeiver1991.blogspot.com