sd5782 

I am working on a set of wheels that came to me on a very well used bike of which I know no history. They are very early low flange 36 hole, and perhaps model 7100/7110. Replacing bearings, out came 11 of 3/16”. I put 11 new ones back in but thought 10 was the normal, but all seemed well. Looking later at Velo base, it does say 10. I assume even though it spins smooth and went together fine, I should go with 10. Does anyone know if any took 11?



Early, low flange 36 hole

Dave Mayer

I thought 10 balls per side. I would take it apart again and see if the balls are packed too tight.

Wrenching at our local bike Co-op, I've run into a few cases where a hub exhibits strange behavior, in that it alternately seems loose, and every few revolutions it locks up. This is evidence of too many balls, or the balls not seated correctly in the cups. I've also seen cases where the hubs were packed with the wrong sized balls, or a seemingly random collection of mis-matched balls. Go figure.

HillRider

I can't speak to that particular hub set but several newer Shimano front hubs I've dealt with have indeed used 11 x 3/16" balls per side.

Iride01 

IMO, you are better off with one less than the correct number of bearings. With one too many the bearings will likely be rubbing against each and not spinning freely.

Hondo6

Shimano's exploded view for the HB-7110-F (small flange) and HB-7120-F (large flange) Dura Ace front hubs appear to show both using a total of 11 3/16" bearing balls per side.

https://si.shimano.com/en/manual/search?model=HB-7110-F

sd5782 

Thanks for the replies already. I did take out 1 per side although it didn’t look crowded with 11. I believe it spun better with 11, so as per Hondo6 I think I will go back to 11. It did spin SO nice with 11, and the quality of the cups and cones and whole hub is impressive.

oldbobcat

Sometimes, a ham-fisted mechanic will roll one of the balls out of alignment with the cone race as the cone is threaded on. The hub may seem to spin correctly for a while, but it might bind as the errant ball rolls farther out of place. As a precaution, especially when dealing with smaller balls, I'll hold the cone and the hub steady in one hand and tighten the cone by turning the axle. When I'm sure all the balls correctly seated, I install the locknut and fine-adjust with the cone wrenches.

smd4

IMO, you should always use the correct number of bearings.

sd5782 

Better searching shows these to be first generation and HS-731/831 hub set. 11 balls in front. I should have searched a bit more.

https://velobase.com/ViewComponent.a...f-35b1fba54611

Andrew R Stewart 

IME the symptom of one too many balls in a hub is the axle is no longer coaxial to the rim. If done on one side only the axle end on the side with the correct number will be pretty much centered. But the other side/axle end is off center. As this wheel rotates the off centeredness will rotate by 1/2 (usually) the sped that the rim does. So the rim will "look" to be out of true at one spot (say at the valve) and a revolution later look to be off at the seam (opposing the valve). The bearing with it's one too many balls rotates at half the pace the wheel does. When both sides have a ball to may it gets a bit more complex as the two off axis cones are now also off from each other by wherever the two "lifted" balls are positioned. Andy

FBinNY 

It's easy to ensure the correct ball count if you let the hub tell you.

Apply grease to the cup and load balls into a circle so that the ID is slightly greater that the OD of the con's small end. Generally, you'll have a sense of correct placement by eye, but can also pit the cone into place twist slowly then gingerly remove it and look at the balls where they'll be in use. They should appear to be filing the space and just about touching, yet not look crowded.

In most cases overloading the bearing will push the cone off center and cause the axle to wobble when turning.

As mentioned before, one too few doesn't matter much, while one too many does, so (only if you must) err low.

HillRider

I think all of these comments are moot. According to the Shimano data, it requires 11 x 3/16" balls per side and that is not unprecedented with other Shimano hubs.

jccaclimber

So you think a cost sensitive industry that makes parts by the million is paying for an extra part that makes their product worse?

Some exceptions abound, generally around trying to make a ruined headset mostly useable, but for others reading this, follow the manufacturer's guidance.

If you're swapping caged to full complement or vice versa that's a different thing than simply not putting all of the parts back in.

Iride01 

I'm not too sure that ball bearings are the thing you want to pin your cost saving ideas of manufacturing on. Ball bearings are some of the least expensive things there are since their production was perfected many years ago.

However to get back to the original thought. One too many bearings will usually not work at all. The bearings will not sit correctly in the race and will rub each other if they even turn at all. For most any application of loose ball bearing on a bicycle, hub, headset or BB, you'll get away with as little as only 66% of the maximum number of bearings required.

Will it wear faster? Maybe. But not as fast as one too many.

Hondo6

That may well be true. But for loose bearings (as in the hubs in question): if you know the number the manufacturer specifies to be correct, why on earth would you use a different number?

Iride01 

I've worked on bikes of others and used bikes I've bought for myself that didn't have the "correct" amount of bearings specified by the manufacturer.

My only purpose here was to say it's much better to have less than one too many. So any that are in too big a hurry to look up the correct number can just fill the race with balls and then just remove a few to ensure that the balls won't be pressed against each other. Or if balls spill on the floor and one or two can't be found then you can still ride you bike that same day and not have to wait till you get new ball bearings.

The bike isn't going to blow up or get damaged as long as about 66% of the race is filled with bearings.

Hondo6

Fair enough.

Still IMO best to use the correct number if known.

But if you're unsure, yes - one too few is far better IMO than one too many.

jccaclimber

I'm well aware that bearings are one of the best controlled COTS parts out there. Part of that is from not doing things that aren't needed. In the first post of yours that I quoted you said to use less than the correct number, which by definition is not correct. In this post you're comparing too many vs. too few. In that I agree, but that isn't the same. Knowing that 11 is the correct number, we both agree that 10 is better than 12, but 10 is not better than 11.
I also disagree that you could get away with 66% fill in a loose bearing situation. Using the OP's case as an example, 66% of 11 bearings is roughly 7 bearings. Using a half bearing excess as an example, the contact angle once all the bearings push to the top will be such that the forces will be more than 10x what they should be.

Once the bearings slip around to the top, which they will given any operating clearance, you'll end up with the attached image. This doesn't look like it should be supporting much on the bottom half. This example was drawn with 1/4" bearings and I didn't put in the actual contact angles as the races don't contact at the midplane, but neither of those changes this conclusion. I've seen bikes ride in to the local co-op like this and it ends up as the axle riding on the inside of the hub, race riding on the dust seal, not the cone on the balls.

Round up 3/4 of a ball instead of down by 1/4 ball so that it's 8/11 instead of 7/111 and I agree that it'll get you home, and even survive a while, but I don't know why there would be occasion to knowingly reassemble a hub that way for more than a parts run to the LBS.

Iride01 

Semantics. Depends on if you assume the correct number is only 11 bearings or if you assume the number is from 8 and 11. Since you are faulting me for that then I can assume you also believe that 8 to 11 bearings is the correct number.
You rounded down. In things like this I'd always round up any remainder. To not use the remainder means you don't have at least that percentage. So your 63.6% is glaringly not at least 66% of my requirement.

Though I may have made a mistake saying it as 66% the number of balls. I think it was 66% of the arc that the balls make contact with on the races. And even 66% might be an error of my memory. However there was a pretty detailed article about it in a industry magazine 15 to 20 or so years ago discussing what the minimum amount of balls should be for applications such as this. I doubt I'd ever find it again, however 66% is what I remember.