Now that everyone is going with threaded bottom brackets...
#51
Senior Member
Join Date: Jun 2013
Location: location location
Posts: 3,035
Bikes: MBK Super Mirage 1991, CAAD10, Yuba Mundo Lux, and a Cannondale Criterium Single Speed
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 344 Post(s)
Liked 297 Times
in
207 Posts
My question was more at the factory level, rather than the brand level. If the plant manager at a bicycle factory in Taiwan, has their biggest customer, a major American brand, tell them “We want all our Al frames to have press-fit BB, and every single one to this tolerance which will prevent creak, and here’s your budget for that,” what can they do to meet the challenge, short of measure every single point of every frame that comes off the line, and ruthlessly toss every one that’s a micron out of spec?
#52
Senior Member
My question was more at the factory level, rather than the brand level. If the plant manager at a bicycle factory in Taiwan, has their biggest customer, a major American brand, tell them “We want all our Al frames to have press-fit BB, and every single one to this tolerance which will prevent creak, and here’s your budget for that,” what can they do to meet the challenge, short of measure every single point of every frame that comes off the line, and ruthlessly toss every one that’s a micron out of spec?
Problem in the bike industry is, the design team works for manufacturer A, but production is bid out to companies B, C, D on low cost. Manufacturer A doesn't perform QC and the fabricators has no incentive to spend more money on production. Testing and rejecting frames cost $. Maintaining your machines to good tolerances also cost $. A typical frame cost $60 to produce. So adding $5 for QC adds too much.
#53
Senior Member
My question was more at the factory level, rather than the brand level. If the plant manager at a bicycle factory in Taiwan, has their biggest customer, a major American brand, tell them “We want all our Al frames to have press-fit BB, and every single one to this tolerance which will prevent creak, and here’s your budget for that,” what can they do to meet the challenge, short of measure every single point of every frame that comes off the line, and ruthlessly toss every one that’s a micron out of spec?
The simple way to build carbon frames accurate is to build them around an aluminum BB insert. Then machine both sides of the insert at same time. Everyone knows this. It is not done because it adds weight. Eighty to one hundred grams of weight. Bragging rights about frame weight are important to marketing. Marketing wins over engineering every time. A slightly more difficult method would be a small aluminum insert on each side of the BB and then machine those at same time. Building a carbon frame accurate enough to directly press in bearings is possible. We know this because it is being done. If that would scale up is not known because no one has tried it.
To your exact query about making aluminum frames accurate enough to prevent creak (which means accurate enough to work) there is no good reason not to. It is simple. It is not an added cost process. Not machining them accurately only happens when an importer tells the sub to build a bike below the cost of building a bike. In small run shops it will be slightly easier to get the job done if threads are used. There may be some significant costs involved when large importers have been reliant on substandard shops building nothing but BSO.
#54
Senior Member
Ruthlessly tossing whatever fails to meet spec is exactly what Quality Control means.
The simple way to build carbon frames accurate is to build them around an aluminum BB insert. Then machine both sides of the insert at same time. Everyone knows this. It is not done because it adds weight. Eighty to one hundred grams of weight. Bragging rights about frame weight are important to marketing. Marketing wins over engineering every time. A slightly more difficult method would be a small aluminum insert on each side of the BB and then machine those at same time. Building a carbon frame accurate enough to directly press in bearings is possible. We know this because it is being done. If that would scale up is not known because no one has tried it.
To your exact query about making aluminum frames accurate enough to prevent creak (which means accurate enough to work) there is no good reason not to. It is simple. It is not an added cost process. Not machining them accurately only happens when an importer tells the sub to build a bike below the cost of building a bike. In small run shops it will be slightly easier to get the job done if threads are used. There may be some significant costs involved when large importers have been reliant on substandard shops building nothing but BSO.
The simple way to build carbon frames accurate is to build them around an aluminum BB insert. Then machine both sides of the insert at same time. Everyone knows this. It is not done because it adds weight. Eighty to one hundred grams of weight. Bragging rights about frame weight are important to marketing. Marketing wins over engineering every time. A slightly more difficult method would be a small aluminum insert on each side of the BB and then machine those at same time. Building a carbon frame accurate enough to directly press in bearings is possible. We know this because it is being done. If that would scale up is not known because no one has tried it.
To your exact query about making aluminum frames accurate enough to prevent creak (which means accurate enough to work) there is no good reason not to. It is simple. It is not an added cost process. Not machining them accurately only happens when an importer tells the sub to build a bike below the cost of building a bike. In small run shops it will be slightly easier to get the job done if threads are used. There may be some significant costs involved when large importers have been reliant on substandard shops building nothing but BSO.
There is no engineering challenge here. You don't need to posses black magic to build a quality product. it is an accounting challenge... Bearings last forever in all kind of application under much higher load. Except for water- and dirt entry, bicycle bearings should last 100,000 time longer than the bicycle.
#55
Senior Member
They would have to use more expensive tooling and maintain it better. and they would have to measure every frame and reject the ones that don't meet criteria. and if the design tends itself to less optimum BB, manufacturing would give feedback to the design team. that is how it works in all industries where PF is sued, which are most industries (cars etc.)
Problem in the bike industry is, the design team works for manufacturer A, but production is bid out to companies B, C, D on low cost. Manufacturer A doesn't perform QC and the fabricators has no incentive to spend more money on production. Testing and rejecting frames cost $. Maintaining your machines to good tolerances also cost $. A typical frame cost $60 to produce. So adding $5 for QC adds too much.
Problem in the bike industry is, the design team works for manufacturer A, but production is bid out to companies B, C, D on low cost. Manufacturer A doesn't perform QC and the fabricators has no incentive to spend more money on production. Testing and rejecting frames cost $. Maintaining your machines to good tolerances also cost $. A typical frame cost $60 to produce. So adding $5 for QC adds too much.
$60 sounds about right. Ready to believe your knowledge on that is more current than mine. $5 upfront to keep QC good also sounds right. Putting that against what these bikes sell for at retail makes QC a no brainer. Comparing $5 to what it will cost a retail customer to get it fixed later (and likely a less than ideal fix) would say it should be done right the first time.
Against that read these forums and all the endless rave reviews for Brands XYZ that we know pay no attention to QC. Nothing changes until consumers make more demands.
#56
Senior Member
Join Date: May 2018
Location: Fargo ND
Posts: 914
Bikes: Time Scylon, Lynskey R350, Ritchey Breakaway, Ritchey Double Switchback, Lynskey Ridgeline, ICAN Fatbike
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 474 Post(s)
Liked 554 Times
in
312 Posts
One small note: welding metal distorts the metal. Not my area of expertise, but I know it needs to be dealt with and dealing with it adds cost. Pole gets around this for some of their frames by milling them from a billet, but that is not a low cost item. As others have pointed out, the issue is the price point set for the frames.
As far as QC goes, an automated vision system can inspect each frame in a few seconds. I have been involved in a couple such systems for other products ranging in size from a gasket under a 12 mm diameter Belleville washer to an airliner wing. The costs are in developing the system and the shop floor time, The Non-repeating Engineering Cost needs to be amortized over the production of the frames and the shop floor cost is a direct manufacturing cost. As others have said, not difficult. It just adds cost.
As far as QC goes, an automated vision system can inspect each frame in a few seconds. I have been involved in a couple such systems for other products ranging in size from a gasket under a 12 mm diameter Belleville washer to an airliner wing. The costs are in developing the system and the shop floor time, The Non-repeating Engineering Cost needs to be amortized over the production of the frames and the shop floor cost is a direct manufacturing cost. As others have said, not difficult. It just adds cost.
#57
Senior Member
Join Date: Jul 2009
Posts: 5,997
Bikes: Colnago, Van Dessel, Factor, Cervelo, Ritchey
Mentioned: 8 Post(s)
Tagged: 0 Thread(s)
Quoted: 3978 Post(s)
Liked 7,425 Times
in
2,986 Posts
#58
Senior Member
One small note: welding metal distorts the metal. Not my area of expertise, but I know it needs to be dealt with and dealing with it adds cost. Pole gets around this for some of their frames by milling them from a billet, but that is not a low cost item. As others have pointed out, the issue is the price point set for the frames.
#59
Senior Member
So one overwhelming conclusion is that threadless would work best, if bike manufacturing tolerances were up to spec.
So that leads to the crucial question; what can bike manufacturers change in their processes to match the required tolerance levels? Is there a quick, straightforward fix that can be implemented to carbon mould assemblies, metal welding rigs, etc, that will guarantee that the BB shell is always perfectly round and aligned, and won’t flex under pedaling loads? And if so, why don’t manufacturers implement it?
So that leads to the crucial question; what can bike manufacturers change in their processes to match the required tolerance levels? Is there a quick, straightforward fix that can be implemented to carbon mould assemblies, metal welding rigs, etc, that will guarantee that the BB shell is always perfectly round and aligned, and won’t flex under pedaling loads? And if so, why don’t manufacturers implement it?
The next best answer is an aluminum insert. That mostly works, but aluminum is still softer than the steel bearing outer, so it can get deformed during installation and wear. That's essentially what a BSA or T47 BB is.
#60
Advanced Slacker
Join Date: Feb 2017
Posts: 6,214
Bikes: Soma Fog Cutter, Surly Wednesday, Canfielld Tilt
Mentioned: 26 Post(s)
Tagged: 0 Thread(s)
Quoted: 2763 Post(s)
Liked 2,537 Times
in
1,433 Posts
What is the problem than PF is trying to solve?
#61
Senior Member
Thread Starter
A few years ago, I dug into this because I didn't understand why bike manufacturers had tried to fix what wasn't broken.
As Raoul Luescher points out in one of his videos, an important thing to understand is that the threads in a threaded bottom bracket are a hold-over from the days of cup and cone BBs. Threads are extremely uncommon in bearing seats outside of bike BBs. Subsequent solutions (like cartridge and Hollowtech 2) were made to work with the prevailing threaded standards as that's what most bike frame were using, not because that was necessarily the best choice from an engineering perspective. Threadless came about in the mid-late 80s as an engineering-first solution from small manufacturers who had the freedom to break away from BSA as they weren't producing at scale.
Given that context, here's my understanding:
1. In the very early days (Klein, Merlin etc designs), they were trying to reduce threaded creaking (see my post above), allow the use of sealed bearings, simplify things and offer better bearing alignment.
2. Cannondale acquired Magic Motorcycle and used their BB30 design in an attempt to:
a. make BBs lighter by eliminating redundant material in the cups and using aluminum spindles
b. make the BB stiffer by increasing the spindle diameter
c. make the BB more robust by increasing the bearing size
d. introduce a significantly larger BB shell which offered larger attachment surfaces for Cannondale's fatter frame tubes.
e. Cannondale was trying (note: trying) to reduce manufacturing costs by simply line boring BB shells. Accurately threading a BB shell is not a trivial matter.
4. PF30 was introduced by Cannondale for two reasons:
a. PF30 uses squashable Delrin cups which theoretically allow the system to be more tolerant of bad bearing alignment.
b. use carbon cups and make things cheaper to manufacture. Bonding a aluminum shell into a carbon frame is tricky to do accurately and introduces the risk of galvanic corrosion
... IMO, introducing PF30 is where things really went off the rails as it introduced all kinds of other problems including allowing even worse tolerances, quick wear, narrow bearing spacing etc.
5. Shimano worked with Giant to develop PF86 as they wanted to continue using the wider bearing spacing and 24mm steel spindle from Hollowtech 2. It also is more tolerant of poor bearing alignment as there's more squashable Delrin in there.
6. Trek introduced BB90, FSA introduced BB386Evo, Cervelo introduced BBright, Colnago introduced ThreadFit (etc) all trying to solve the various problems of BB30 and PF30 while trying to avoid patent infringements.
As Raoul Luescher points out in one of his videos, an important thing to understand is that the threads in a threaded bottom bracket are a hold-over from the days of cup and cone BBs. Threads are extremely uncommon in bearing seats outside of bike BBs. Subsequent solutions (like cartridge and Hollowtech 2) were made to work with the prevailing threaded standards as that's what most bike frame were using, not because that was necessarily the best choice from an engineering perspective. Threadless came about in the mid-late 80s as an engineering-first solution from small manufacturers who had the freedom to break away from BSA as they weren't producing at scale.
Given that context, here's my understanding:
1. In the very early days (Klein, Merlin etc designs), they were trying to reduce threaded creaking (see my post above), allow the use of sealed bearings, simplify things and offer better bearing alignment.
2. Cannondale acquired Magic Motorcycle and used their BB30 design in an attempt to:
a. make BBs lighter by eliminating redundant material in the cups and using aluminum spindles
b. make the BB stiffer by increasing the spindle diameter
c. make the BB more robust by increasing the bearing size
d. introduce a significantly larger BB shell which offered larger attachment surfaces for Cannondale's fatter frame tubes.
e. Cannondale was trying (note: trying) to reduce manufacturing costs by simply line boring BB shells. Accurately threading a BB shell is not a trivial matter.
4. PF30 was introduced by Cannondale for two reasons:
a. PF30 uses squashable Delrin cups which theoretically allow the system to be more tolerant of bad bearing alignment.
b. use carbon cups and make things cheaper to manufacture. Bonding a aluminum shell into a carbon frame is tricky to do accurately and introduces the risk of galvanic corrosion
... IMO, introducing PF30 is where things really went off the rails as it introduced all kinds of other problems including allowing even worse tolerances, quick wear, narrow bearing spacing etc.
5. Shimano worked with Giant to develop PF86 as they wanted to continue using the wider bearing spacing and 24mm steel spindle from Hollowtech 2. It also is more tolerant of poor bearing alignment as there's more squashable Delrin in there.
6. Trek introduced BB90, FSA introduced BB386Evo, Cervelo introduced BBright, Colnago introduced ThreadFit (etc) all trying to solve the various problems of BB30 and PF30 while trying to avoid patent infringements.
Last edited by Hiro11; 02-12-21 at 09:54 AM.
#62
Senior Member
Join Date: Oct 2013
Location: Oahu, HI
Posts: 1,399
Bikes: 89 Paramount OS 84 Fuji Touring Series III New! 2013 Focus Izalco Ergoride
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 286 Post(s)
Likes: 0
Liked 74 Times
in
54 Posts
scott s.
.
Likes For scott967:
#63
Senior Member
Join Date: Sep 2014
Posts: 3,893
Mentioned: 20 Post(s)
Tagged: 0 Thread(s)
Quoted: 1062 Post(s)
Liked 665 Times
in
421 Posts
My question was more at the factory level, rather than the brand level. If the plant manager at a bicycle factory in Taiwan, has their biggest customer, a major American brand, tell them “We want all our Al frames to have press-fit BB, and every single one to this tolerance which will prevent creak, and here’s your budget for that,” what can they do to meet the challenge, short of measure every single point of every frame that comes off the line, and ruthlessly toss every one that’s a micron out of spec?
The idea behind statistical process control is that you watch for variations that are smaller than the tolerance limits, and feed the results back into the process through tiny corrections. You also improve processes to make them more consistent from piece to piece, so the quality measurements are meaningful. This kind of control is done all the time in the manufacture of things like motors, transmissions, and so forth.
Likes For Gresp15C:
#64
Senior Member
Thread Starter
Ah, the last two comments warm the cockles of my operations management heart.
#65
Senior Member
Join Date: Jul 2020
Location: Humboldt County, CA
Posts: 835
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 408 Post(s)
Likes: 0
Liked 431 Times
in
287 Posts
There's also the issue that, if you accept one (1) part from your supplier that doesn't meet some specification in the contract, you have just unilaterally renegotiated that specification for the rest of the contract. After you have done this, your supplier can sue you if you later reject a part that fails the contract spec but exceeds the spec of the worst part you have ever accepted.
What this means is that, for outsourced parts, quality costs twice. The first time, because your supplier will charge you more for the tight specs. The second time, because you have to do your own QA when you receive the parts, and the way that contract enforcement works means that your QA has to be at least as rigorous as the vendor's, and ideally more so, lest your vendor know more about your parts than you do.
--Shannon
What this means is that, for outsourced parts, quality costs twice. The first time, because your supplier will charge you more for the tight specs. The second time, because you have to do your own QA when you receive the parts, and the way that contract enforcement works means that your QA has to be at least as rigorous as the vendor's, and ideally more so, lest your vendor know more about your parts than you do.
--Shannon
#66
Senior Member
Join Date: Jun 2011
Location: Tampa, Florida
Posts: 5,585
Bikes: 2017 Colnago C-RS, 2012 Colnago Ace, 2010 Giant Cypress hybrid
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 408 Post(s)
Liked 122 Times
in
85 Posts
I had two 2017 CRS frames. The BB presses directly into carbon fiber. At the time, I was using Campy drivetrains. Campy cranks have bearings mounted to the spindles and empty cups that press into the frame. I greased the Campy cups when pressing them in, rather than using loctite. I heard a little creaking, but creaks can also come from other places, like the rear skewer not being tightened enough on a rim brake frame. If I still had Campy cranks, I'd probably try the BB infinite solution. https://www.bbinfinite.com/products/...ant=9864643843
Now using Shimano cranks with my AXS drivetrain, the thread together BBs seem to be the perfect solution and less expensive than the BB infinite.
Now using Shimano cranks with my AXS drivetrain, the thread together BBs seem to be the perfect solution and less expensive than the BB infinite.
__________________
HCFR Cycling Team
Ride Safe ... Ride Hard ... Ride Daily
2017 Colnago C-RS
2012 Colnago Ace
2010 Giant Cypress
HCFR Cycling Team
Ride Safe ... Ride Hard ... Ride Daily
2017 Colnago C-RS
2012 Colnago Ace
2010 Giant Cypress
#67
Senior Member
Disagree, at least in the context of Deming. What you need is a repeatable process that can meet six sigma. Testing in QC is kind of a losing proposition. But I suspect most of these places don't have process engineers nor capture the needed data on their processes.
scott s.
.
scott s.
.
I can tell you that when a shipment is not accepted for delivery, returned, no payment forthcoming, everyone sits up and takes notice. When that happens everyone knows what the score is. The supplier makes it right very quickly or the purchaser goes to plan B and supplier B. When there is no plan B the purchaser is out of luck. For most of the outsourced industry plan B is theoretical.
On the other side once it is established that there are no standards, that any large or small variance will be accepted and paid for, then no one in the production process has any reason to make anything but junk. Anyone on the line who says NO will just be laughed at. Any purchaser who has been accepting junk does not have a leg to stand on.
Can we talk about markups and margins? Asian bike frames sell at huge markups. The customer who pays $1000 or $3000 for a frame thinks he bought something exceptional. Very hard to tell that customer he bought a commodity from designers and marketers with little or no expertise and no business sense. The gap between expectations and reality is vast. All the incentives are perverse.
#68
Newbie
I've had nothing but problems with my Cannondale Synapse Ultegra. Two replacement bottom brackets along with a bunch of maintenance required to keep it from loud clicking
which of course drove me and anyone in the group ride crazy. Very happy that my new Trek Emonda bike has threaded bottom bracket.
which of course drove me and anyone in the group ride crazy. Very happy that my new Trek Emonda bike has threaded bottom bracket.
#69
Senior Member
Join Date: Apr 2012
Location: Nor Cal
Posts: 6,016
Mentioned: 17 Post(s)
Tagged: 0 Thread(s)
Quoted: 1814 Post(s)
Likes: 0
Liked 923 Times
in
569 Posts
I wonder what portion of this is the result of CF frames amplifying sound.
The same noise source that is overlooked in another frame material could be an anguish-inducing condition in a CF one.
The same noise source that is overlooked in another frame material could be an anguish-inducing condition in a CF one.