BT clutch release, analysis & functional history 36-84:

[nifty]

BT clutch release, analysis & functional history 36-84:

I have posted before on this subject, and in shooting from the hip, made a few errors. Rubone, please accept my humble apologies. In the interests of reliable, accurate information, I have prepared the below draft document.

Intent is to consolidate all available knowledge and arrive at an overall consensus. If I have anything wrong, please post. If others can add more information, please post.
Anything with a question mark ? I know requires confirmation. please help me eliminate all question marks?
Everything else could also require correction editing.

36-37/8 used 2446-36 and 2446-38 Clutch Push Rod which has a distinctive, reduced OD extension tip at bearing end, 2446-36 is approx 13-1/4” plus extension tip; tip supports ID of outer race of 2448-36 & 2448-38 Clutch release bearing (small bearing 1.154" OD x .523" thick, same as early 45”), tapered shoulder of pushrod corresponds to taper seat at inner race of bearing; 2149-36 Slinger, oil, is a sheetmetal butterfly/2 blade propeller installed between starter clutch & mainshaft nut (2149-36 solidly mounted to mainshaft is possibly H-D’s most effective oil slinger); 2436-36 release finger, is slotted (presumably finger slot also engages/self centralizes with protruding pushrod extension tip?). Also 37337-36 Clutch Hub Nut Seal, cork, 36-58? Provides rotational drive to pushrod, cork “grip” on pushrod constitutes torque-limiting drive device for pushrod. There are also current production -36 style AM pushrods with clip groove to convert early BT to 75up Shovel type Torrington release bearing.

note the pushrod extension

36-38 pushrod bearing & slinger.jpg (33.14 KiB) Viewed 1392 times

note slot for pushrod extension

2436-36 36-38 finger b.jpg (20.38 KiB) Viewed 1392 times

39-64 used 37285-38 (previously 2446-38) and 37285-41 Clutch Push Rod, similar to 2446-36, but sans earlier extension tip, approx 13-1/4” OAL; tapered shoulder of new pushrod corresponds to taper seat of new 37310-39 release bearing. The 37310-39 (previously 2448-39) conical release bearing contains many more & larger balls, and its increased OD does double duty as new oil slinger (propeller deleted). Also the pushrod was driven/rotated by 37337-36 Clutch Hub Nut Seal, cork 36-58? And later 12014 Clutch Hub Nut Seal, proper rubber lip seal 60?-84, seal “grip” on pushrod constitutes torque-limiting drive device for pushrod, also new for 39 was 33381-39 (later 33430-50T) starter clutch (kickstart ratchet on outer end of mainshaft) which incorporates a captive spring-loaded ball to provide rotational drive directly to corresponding groove in inner race of -39 release bearing, apparent intent was that if some other force acted on bearing, the ball would be pushed into starter clutch and thus starter clutch could spin harmlessly inside bearing (spring, ball and corresponding groove in inner race of -39 bearing form a torque-limiting drive device, however, I suspect the ball & spring were a work-around/afterthought to spin relatively heavy inner race with mainshaft and thus prevent pushrod & inner race wear from pushrod trying to suddenly accelerate inner race up to mainshaft speed whenever clutch released); there was also a new 37070-39 (previously 2436-39) release finger, slotted, approx 1/4" ???” wide slot, the upper & lower faces of this finger also do double duty by engaging cut-away outer race of -39 bearing to prevent its rotation (finger slot serves no purpose and presumably was provided to permit -39 finger to be backwards compatible to ‘36 with earlier 2446-36 pushrod). NARROW SLOT 37070-39 FINGER ONLY FOR -39 CONICAL BEARING, OR EARLY PUSHROD AND SMALL BEARING, 37070-39 FINGER MUST NOT BE USED WITH CLIP TYPE TORRINGTON NEEDLE ROLLER THRUST BEARING.

note sans extension

-38 or -41 pushrod.jpg (37.18 KiB) Viewed 1392 times

slot fits extension of -36 pushrod

OE slotted -39 finger Parkerised b.jpg (139.26 KiB) Viewed 1392 times

apart from copper, same as parkerised -39 finger

OE slotted -39 finger copper plated.jpg (37.53 KiB) Viewed 1392 times

-36 pushrod compatible with -39 finger & -39 bearing

-36 pushrod -39 finger -39 bearing.jpg (17.7 KiB) Viewed 1392 times

Even OE -39 fingers suffer from rubbing against stationary outer race of OE -39 bearing

wear from brg on OE slotted -39 finger copper plated.jpg (90.67 KiB) Viewed 1392 times

When engine running and clutch engaged, inner race of 37310-39 bearing should be
constantly spinning close to, or against internal balls & stationary outer race, which is restrained at cut-away by finger. (regardless of cable clearance, albeit when all adjustments correct, ball loading is low to none). Sounds wrong*, but that’s what they do. *and by modern thinking, it is wrong, but it worked so well it stayed in production for around 35 years and many original -39 bearings are still in use. AM versions of 37310-39 bearing are sometimes not so good.

When clutch is released/disengaged, end thrust of pushrod tends to lock the hardened
tapered shoulder of pushrod into corresponding hardened taper seat at inner race of
37310-39 bearing. If required, the integral small coil spring at starter clutch ball can compress and allow independent mainshaft & starter clutch rotation within 37310-39 bearing.

65-69 used 37285-65 pushrod, longer, (approx 13-5/8” to 13-3/4”, due to electric start and longer trans mainshaft), release bearing, pushrod seal, finger & starter clutch as for 39-64.

70-74 used 37285-70 pushrod which is longer again (approx 13-7/8” to 14”, due to alternator on engine and longer trans mainshaft), release bearing, pushrod seal, finger and starter clutch as for 39-64.

E75 new pushrod & corresponding Torrington needle roller thrust release changes. Almost immediately there were rapid & poorly documented changes to clutch release parts, true situation is now unclear.
Apparently, for a very short time there was a “Torrington” pushrod which does not have groove for snap ring.

There are also solid/slotless fingers.

witness mark from -39 bearing

OE solid finger copper plated b.jpg (141.32 KiB) Viewed 1392 times

OE solid finger copper plated c.jpg (145.45 KiB) Viewed 1392 times

Unless there are more parts of which I am unaware, solid/slotless fingers, can only be used with -39 release bearing, as for previous, slotless finger engages with corresponding cut-away of outer race of -39 bearing); Must ensure tip of pushrod does not protrude through -39 release bearing. (if run with insufficient end clearance the shoulders of pushrod & corresponding seat in bearing can wear until pushrod protrudes, then solid finger likely to graunch on protrusion and action is direct finger-to-pushrod, bearing bypassed.

In the slotless finger photos there appears to be a witness mark caused by pushrod hole in outer race of -39 bearing.
Anyone got a part number and any data for slotless finger?

There is also a 37070-39A release finger.

Donny Petersen at https://books.google.com.au/books?id=PH ... er&f=false
mentions a 37070-39A release finger, but his indicated usage is less than crystal clear to me and I can’t find a plausible explanation for this number in parts books, or by on-line searching.

41-54 BT parts book has only 37070-39 finger 41-54;
61-76 FL parts book has only 37070-39 finger 41-74; and 37070-75 finger 75up.

The forum OEM parts list at:
http://www.hydra-glide.net/kb/inf_ov_pa ... pageno=100
States 37070-39 finger is 39-57 and 37070-39A finger is 41up, no explanation, but I now think this is more accurate than Donny Petersen.

Extract from Donny Petersen:

-39A finger reference

extract clutch finger petersen.jpg (22.2 KiB) Viewed 1392 times

My interpretation of Donny’s writings:
37070-39 finger 39-69 (undeniably narrow slotted)
37070-39A finger 70-74 solid/slotless?? Donny’s “1979” to 1974 is surely a misprint.
37070-75 finger 75up wider slotted (“channel” “two-fingered”) and we know it was supposed to clear clip and pushrod tip. Actual slot width of 37070-75 finger would also be good to know?
Still only my interpretation, Donny might have got 39A wrong, something totally different may have occurred.
Not sure that Donny was 100% focused when he wrote above extract.

Using date clues from the forum OEM parts list, I’m starting to think perhaps 37070-39A is the copper plated* version of Parkerised 37070-39? i.e. -39A is a “revision” part introduced circa 1958 and physically identical and interchangeable, only difference is possibly specification of steel and heat-treatment/hardening process used, which generated the -39A part number?? Weird thing is the use of 37070-39A part number was dropped by around 1960, and subsequent parts books reverted to 37070-39. *Early OE Shovel rocker arms are similarly copper plated, the copper is a masking medium, only the exposed steel at wear points is fully hard, everything protected by copper is not hard.

At a stretch I can re-interpret Donny’s writings to mean both 37070-39 & 37070-39A are interchangeable and valid 39-74. But that would also mean that somehow Donny missed the slot in both fingers and the solid finger. Back in the day, in beautiful downtown Burbank, one might refer to it as the fickle finger of fate. But I digress.

37285-70A pushrod (apparently clipless, but used with Torrington needle roller thrust) quickly superseded by 37285-75A clip type pushrod.
37070-75 finger superseded by 37070-39B widest slot finger (this -39B revision finger introduced at some after 1976?).

April 08 1975 H-D TSB 687 effectively states:
As at mid February 1975, all BT had the new Torrington needle roller thrust roller thrust type bearing set up and associated parts. (as listed below)
37285-75 pushrod, which in 687 is drawn with finger end of pushrod short & reduced OD extension (approx .250” OD) and drawn with groove for clip. 37285-75 pushrod also drawn with pair of flats to drive/rotate “daisy wheel” slinger.
37311-75 small thrust washer (plain race, approx 0.254" x 0.670") 2 x 37311-75 used early 1975 models (which is late 1974 calendar year) 1 x 37311-75 used on late 1975 models (which is early 1975 calendar year)
37313-75 large thrust washer, outer (plain race, approx 0.254" x ??) used on late 1975 models (which is early 1975 calendar year)
11068 Lock Ring, which in 687 is drawn as Seeger circlip (ears for circlip pliers).

37312-75 Thrust bearing (Torrington needle roller thrust, approx 0.254" x 0.670" x 0.077” thick x 9 rollers x 3mm long; aka 17mm bearing, approx dynamic load 1150 pounds).

37228-75 Oil Slinger which in 687 is drawn as stamped sheetmetal “daisy wheel”, used 75-84.
37070-75 Release Finger, which in 687 is drawn slotted.

OLD PARTS deleted
37030-39 Release Finger, which in 687 is drawn solid, no slot.
I don’t trust any non blueprint drawings, but so far there is very little else to work with.


No 37030-39 finger known to me is solid, I suspect that the as-drawn 37030-39 finger previously bore a different part number and this detail was altered when TSB 687 was created.
I further suspect the 37070-75 Release Finger may have been given a wider slot to clear 11068 Lock Ring, without changing finger part number (slot width on paper sufficient to clear Seeger circlip, but real world assembly and “production” finger slot centrality with mainshaft, subsequently led to many clips being displaced in service)

L75? to 84 used 37285-75A pushrod, as for 37285-70A except, pushrod extension slightly longer, clip groove in pushrod; 37285-75A pushrod, when used with associated parts, is backwards compatible to 1970; initially used new “E” clip??, finally used with new 11096 “C” clip*, “C” clip is backwards compatible to 1970; 1977 saw both thrust washers/races changed to 37311-75A (bigger OD) but some of these not hardened, warranty issues, see TSB 710 May 6, 1977. Also new 37070-39B finger which has .416” wide slot to clear “C” clip, -39B finger is backwards compatible to 1939. 37285-75A pushrod subsequently used with upgraded 37313-80 Thrust Washers (Torrington plain race .256” x .880” x .095” thick). Possibly Baker in their own 4speed were first to subsequently introduce 24mm Torrington thrust and small centralizing bush (actual bearing OD 23,20mm), which are also backwards compatible to 1970 when 37285-75A pushrod used. 24mm Torrington has since been widely adopted by AM. OE 37070-39B finger has approx .416” wide slot and is machined from solid steel, Many AM replacement 37070-39B are sintered. *NOTE: As “replacements” for 11096 “C” clip, AM currently sell both Seeger circlips and E clips, these replicate the known to fail obsolete H-D parts. DO NOT USE!

IMO For Torrington type, use only 37070-39B wide slot finger, preferably solid steel, not sintered.
IMO For Torrington type, use only 24mm bearing upgrade. Approx dynamic load 2068 pounds.
IMO Function of all types is improved by finger being physically vertically centralized to mainshaft.
IMO If all good & correctly corresponding parts used, there is not much between a well set up OE -39 bearing, and a well set up 24mm Torrington needle roller thrust bearing.

Nifty

[nifty]

More pics, first post wouldn't accept any more
Nifty

Now the early pushrod, slinger and bearing have miraculously re-appeared in first post!
Enjoy it twice
Nifty

And more pics

[nifty]

rejected more pics
so another post
Nifty

[nifty]

Additional data:
Correction: replace “later 33430-50T” with later 33381-39T.

Not mentioned in TSB 687 is the 75up deletion of the captive ball & spring from 33381-39 starter clutch (no longer required because -75 daisy wheel slinger driven/rotated entirely by pushrod. Whereas 39-74 the inner race of -39 bearing was driven/rotated by ball & spring and large OD of -39 bearing served double duty as the oil slinger)
33560-75 starter clutch 75-84+ has captive ball & spring deleted (single keyway) If reverting to -39 bearing, should also revert to -39 starter clutch with ball.

Note: Also new for 1939 was -39 deflector
33429-39 DEFLECTOR, 0IL, mainshaft starter gear oil deflector/scraper 39-E79 all models with kickstart (piece of bent tin with dimple for countersunk screw, anti-rotation tab of 33429-39 is at front. Deflector catches & re-directs some of the oil flung off rotating slinger which 39-74 is the tin cover of -39 bearing, and 75-79 is the 37228-75 daisy wheel slinger on pushrod. Deflector directs drip stream of oil to bush of starter gear and back to -39 clutch release bearing, and to 37312-75 Torrington wafer bearing). And in so doing, also deflects/reduces oil which would otherwise be flung at release shaft upper bush, thus minimizing oil leak from top of release bush.

DEFLECTOR MUST HAVE SOME CLEARANCE TO OD OF -39 RELEASE BEARING or -75 slinger.

33429-80 DEFLECTOR, 0IL, 80-86, altered to work with 80up bearing retainer plate which is fastened by std SHCS, anti-rotation tab of 33429-80 is at top.

IMO The deflectors help keep release bearing lubricated, all types, but particularly Torrington type.

Nifty

[RooDog]

rejected more pics
so another post
Nifty

Could you please explain "sintered", as in powdered metal? Or something else?

[nifty]

Roodog
Yep, sintered is powdered metal, molded under pressure & heat, ala "Oilite" bronze bushes, fuel filters etc.
The smooth, squared-off fingers are that way because that's how they come out of mould/die.
Pic of new sintered -39B above is V-Twin
Nifty

[RooDog]

Thank You.
I know the tells for casting vs forged, and die cast vs billet. now I can add sintered/powered metal to my list of eyeball identification talents.....
....RooDog....

[nifty]

How not to assemble cover & contents.

What happened first? And where did the ball and spring go?

1. Mainshaft is missing pushrod bush.
2. -39 starter clutch is missing ball.
3. Witness marks from where ID of -39 release bearing has been touching OD of -39 starter clutch.
4. Wrongly positioned kick shaft washer (on inside of cover!) forced kicker gear into bumper (up-stop).
5. My bet is kickstart return spring was also over-wound, just to get kicker arm/lever upright.
6. End of countershaft has witness mark from end of kick shaft.
7. Kicker down-stop trashed (notch in retainer plate), probably from when starter clutch failed to engage with starter gear and rider kicked air all the way to stop, milliseconds before over-extending knee.

Nifty

[Mongrel505558]

How not to assemble cover & contents.

What happened first? And where did the ball and spring go?

1. Mainshaft is missing pushrod bush.
2. -39 starter clutch is missing ball.
3. Witness marks from where ID of -39 release bearing has been touching OD of -39 starter clutch.
4. Wrongly positioned kick shaft washer (on inside of cover!) forced kicker gear into bumper (up-stop).
5. My bet is kickstart return spring was also over-wound, just to get kicker arm/lever upright.
6. End of countershaft has witness mark from end of kick shaft.
7. Kicker down-stop trashed (notch in retainer plate), probably from when starter clutch failed to engage with starter gear and rider kicked air all the way to stop, milliseconds before over-extending knee.

Nifty

How was the rest of the bike?

[nifty]

re How was the rest of the bike?
I don't know, but I've got a fair idea.
Nifty

[nifty]

Another useful illustration I forgot earlier
Nifty

[RooDog]

Question.... In your opinion, or factory thinking:
In the late 1970s, and especially in 1980, why did the MoCo make so many design changes in the four speed, a design dating back to the late 1930s, knowing that the 5 speed tranny, a much more durable and fully rollerized, was in the immediate future, and in fact, issued in the 1980 FLTs and shortly into the FXR, and FXST models? Why did they spent time and resources on beating a dying horse?
....RooDog....

[nifty]

RooDog,
IMO $$$, pure & simple.
Moco tried to make 4speed faster to assemble by unskilled labor, eliminated loose rollers & their selective fits, tried to eliminate skilled labor from shift adjustment etc.
Reduced inventory by using a few more XL parts in plate-shift lid & MDG seal.
Switched to cheaper mainstream fasteners, cast clutch arm, pressure die cast plate-shift lid
Overall I think they would have probably saved more money by leaving alone
But beancounters were driving engineers to be seen to be working at increasing profitability, to appease first AMF and then post-buyback stockholders, even if it was smoke and mirrors to get the stockholders to think all this action would result in bigger dividends and so deter them from dumping their MoCo stock. (they got their pay-off during Evo era)
Only the MoCo could say how much the 4speed changes cost to initially implement, then cover the many warranty claims & re-designs of poorly thought out screw-ups.
Then there are the inconsistencies, like keeping the ratchet lid trans on FXWG, long after everything else had changed to plate shift lid.
I guess the original intent was along the lines of the 5speed is still a few years off, so let's up-date this thing as a stop gap.
Didn't quite work out, lots of turds, but can't all be bad, or no-one would use this forum.
Nifty

[RooDog]

Nifty....
Thank you for your thoughts on that matter. Makes perfect sense to me even if it seems unreasonable to make changes to a soon to be obsoleted item. I switched my $hovelhead over to a 5-Speed back in the 1990s, and only miss the kicker when doing pushrods. LOL...
Thanks again.....RooDog....

[nifty]

Correction
Delete: "Then there are the inconsistencies, like keeping the ratchet lid trans on FXWG, long after everything else had changed to plate shift lid."

Add: Ditching of the ratchet lid and change to plate shift was to get rid of ratchet housing, because it was in the way of soon to arrive rear belt and its bigger front pulley, everything went down-hill from there.

Any failed small parts, springs, pawls etc, with plate type shifter, can drop down into meshing gears.

Nifty