How many bearings ?

[4F]

OK replaced my front wheel bearings (700 wheel) but have a query as to the correct no of bearings.

The bloke in my LBS said that 9 is the best / correct number each side however when I removed the older ones there were 10 each side !

I went with the advice of 9 but now wondering if I should have replaced the same number as I took out, or will it really not make a difference ?

 

[Mr Pig]

If ten fit comfortably I'd put ten in.

 

[llllllll]

yeah, don't put any less in than they came with. There should be no gaps between the bearings.

 

[4F]

Thanks, another one it is then

 

[robbarker]

The standard number is 10 3/16" balls in the front hub and 9 1/4" balls in the rear. So an extra one is a good idea!

 

[Yorkshireman]

yeah, don't put any less in than they came with. There should be no gaps between the bearings.

It's been my understanding (and practice) for the past 50+ years of bike fettling that there must be a gap (full race of balls less one or two) for the bearing to function correctly.

 

[briank]

Your front wheel only has two bearings - one on each side.

(Are cycling and pedantry inherently linked?

Ah, just me then.)

 

[threefingerjoe]

In most cases, you put in the same number of balls that you took out. Yes, there must be at least a slight gap. If I were doing it, and had room for another ball, while still maintaining a gap, I'd put in another ball. As someone stated, Shimano usually has 10 3/16" balls on each side in the front, 9 1/4" balls on each side in the rear. I just replaced the bearings on my own bike last night (Bontrager wheels) and was surprised to find 12 balls on each side in the front.

 

[rustychisel]

As said, it used to be accepted practise to put one less than would fit firmly, so as to allow a little room for the balls to 'shrug' rather than squash up.

 

[Mr Pig]

there must be a gap (full race of balls less one or two) for the bearing to function correctly.

I understand why there needs to be a gap, but do you know why it would be better to have as big a gap as over two bearing widths? I would have thought that provided there was 'a' gap, maybe half a bearing width or more, it would be better to have more balls in there to spead the load over. Any light you could shed on this would be interesting?

 

[robbarker]

There should be no big gaps in a race loaded with ball bearings, but the balls need to be able to rotate freely.

For further enlightenment on the wonderful life of ball bearings in bicycle hubs, read this fascinating collection of posts:

http://yarchive.net/bike/ball_bearings.html

 

[llllllll]

It's been my understanding (and practice) for the past 50+ years of bike fettling that there must be a gap (full race of balls less one or two) for the bearing to function correctly.

I'm surprised by that. My understanding is the more bearings you fit the better distributed the load is. Certainly all the hubs I've taken apart have had not gaps (mostly Shimano). By gaps I mean gaps big enough to fit a ball bearing in, they shouldn't be so tight that they can't rotate.

 

[Yorkshireman]

I'm surprised by that. My understanding is the more bearings you fit the better distributed the load is. Certainly all the hubs I've taken apart have had not gaps (mostly Shimano). By gaps I mean gaps big enough to fit a ball bearing in, they shouldn't be so tight that they can't rotate.

Crikey! It's over 50 years since this was explained to me I think the idea was that if the balls were too close (ie likely to be touching one another) they would grind/break bits off each other and cause the cups/cones to suffer. Also if the balls are too closely packed the cup/ball/cone interfaces don't 'settle' well (I think that the idea was to make allowance for very minor imperfections/differences between all three surfaces). I've just serviced both front and rear hubs (Shimano) as per the service instructions. The rear was fine so only got new balls (9 each race, but I could have got 10 in) and fresh grease (after cleaning the old stuff out). The front hub had knackered cones (due to me being a bit sloppy cone-tightening about a month ago - didn't take into account the extra bit of 'squeeze' due to the QR ), so new cones and balls, again I'm sure I could have got one more in there without causing the balls to climb up one another. I think the 2 balls less than 'full' applied to bearings using smaller balls (and a lot more of them - 16 or more) as in some of the older headset bearings - in fact I just replaced the headset on my Trek 700 with a Shimano one (caged balls) and if loose balls were used I would have needed six or more balls to get the same 'fill'. The link supplied by robbarker above certainly makes interesting reading .

 

[Mr Pig]

For further enlightenment read this fascinating collection of post

That was an interesting read, thank you.

Coincidently, today at work an operator had replaced a bearing on a printing machine three times, taking about an hour each time, before I was sent down to see what was going on. Each time a new bearing was fitted it was lasting about thirty minutes before it started screeching loudly and when the bearing was removed it was so rough you could hardly turn it.

When I looked at the bag of bearings he was using they turned out to be unbranded ones made in the USSR! I had a root around and found a Japanese one, he fitted that and it's fine! I'm sure the Russians make perfectly good bearings, but not these ones.

 

[bobg]

I'm with you Yorkshireman. My old Dad said the same - "you can get another one in but don't" Like you its not failed me in 50 years either