DOH! I was up til 4:00AM working on some paperwork so I was running on about 3 hours sleep this morning. I was taking business calls at the same time I was writing the message so I didn't take the rear triangle into consideration in my 30 second drawing, only the dropouts.
I hate it when I do that, it shoots my credibility! ;-)
I redid the drawing. With 17" (~43cm) chain stays the difference between 120mm and 126mm is about 0.4° per side not 3.6°... Glad I wasn't working on the new Bay Bridge this AM! =8-0
Everything else that I posted still stands, I took it off of previous well parsed posts of mine.
The next time I work on a badly misaligned frame, I'm going to measure the amount of offset in the rear dropouts. I've seen some on top brand frames that were out over 5° per side - in different directions.
Here's several websites that better explain the problems and processes:
<http://www.parktool.com/repair/readhowto.asp?id@>
<www.pdfonline.com/products/allpdf/samples/multi_column.doc>
You can make some homemade tools to check and align the rear dropouts too. Get some 2 extra extra long 3/8" bolts or all-thread, 4 large 3/8" flat washers and 4 3/8" hex nuts (10mm will work too). There's a website somewhere that shows how to do this.
While the dropouts are being aligned the derailleur hanger should be checked and realigned if necessary too. There's a special tool for this.
Scroll down near the bottom and click on GEAR HANGER ALIGNMENT TOOL <http://www.cyclus-tools.eu/index.php?id=rahmen&l=3>
Bike axle failure is something that usually happens over a long period of time - many cycles of flexing. Bending is usually related to one or more shock events.
Someone else mentioned that occasion of frequent R&R of the rear wheel repositions the stress points. Extreme wear and flaking on one part of a hub cone would be a good indicator of dropout misalignment 1n a wheel that's been infrequently removed.
PS: I have SolidWorks but most of the time I use a super fast 2D drawing program from Integraph called SmartSketch. I've tried AutoCAD before but the vertical learning curve is too steep for casual use. ;-)
Chas. Colerich Oakland, CA USA
Scott L. Minneman wrote:
> Your math/CAD/trig is definitely wrong. Seven-plus degrees would be a crazy
> amount -- the angle of the a chainstay to straight ahead is in that range,
> but not the small change seen at the dropout from this slight alteration.
>
> I'm seeing less than one-half of one degree of misalignment resulting from
> this amount of spread. I used AutoCAD and some pretty aggressive geometry,
> to make the angles worst-case steep. So, numerically, the angle is pretty
> darn small.
>
> The absolute numbers don't really matter, though...what matters is if people
> see things (axles, QRs, dropouts, stays) breaking as a direct result of this
> process. Some folks report problems, others don't seem to see any -- not
> unlike myriad other aspects of these machines that some folks seemingly have
> constant problems with and where others rarely experience issues. Can
> dropouts harmlessly be spread and smooshed together to accommodate slightly
> different OLN dimensions? It depends, right?
>
> At any rate, the misalignment Chas. Is talking about is fairly
> straightforward and very satisfying to correct, with the right tools. It's
> also easy enough to make some parts that'll reveal this alignment issue (if
> you don't have access to the alignment tools) and then you can improvise (be
> careful how you achieve the necessary leverage...this may be a job best left
> to professionals if you've noticed, over the years, that predictive physics
> and failure-free engineering isn't your strong suit).
>
> Scott Minneman
> San Francisco, CA USA
>
> Sent: Wednesday, December 16, 2009 10:27 AM
> Subject: Re: [CR] Question on Hub Spacing
>
> Here's another caveat... The majority of frames that I've checked with
> Campy, VAR or Park tools since the mid 70s have had misaligned dropouts!
>
> Dropouts even a couple of degrees out of alignment can put undue
> stresses on the hollow axles of QR hubs.
>
> Stretching 120mm wide dropouts 6mm to fit a 126mm wide 6 speed hub works
> out to 3mm (slightly under 1/8") per side. Doesn't sound like much but
> it comes out to 3.6° per side or a total of 7.2° of dropout misalignment
> when the rear triangle is not cold set and realigned.
>
> Someone check my math please - I drew it up on a CAD program rather than
> trigging it out so I could be wrong.
>
> Something has to give and it's probably not going to be the dropouts.
>
> The problem with this calculation is that since most dropouts are
> misaligned there's a slim possibility that stretching the rear triangle
> to 126mm might bring the dropouts into alignment! ;-)
>
> Another thing to consider is the OLN (over lock nut) width of the hub.
>
> I've seen specs for the following OLN widths on hubs used on single
> bikes with multiple cog freewheels made during the past 50 years:
>
> 117mm (ISO for 3 & 4 speed FW)
> 118mm
> 119mm
> 120mm (standard for 5 speed FW)
> 121mm (some Campagnolo 5 speed hubs & some Japanese hub)
> 122mm (ISO for 4 & 5 speed FW)
> 124mm (some Japanese and French hubs)
> 125mm (some Campagnolo hubs)
> 126mm (standard for 6 & 7 speed FW)
> 127mm
> 127.6mm (old standard for 6 speed FW)
> 128mm
> 129mm
> 130mm (standard for 7+ speed FW and freehubs)
> 135mm (standard for MTBs and some newer road hubs)
>
> I pulled these specs out of Sutherland's, manufacturer's literature and
> other published sources.
>
> I've frequently found the the OLN width has been modified with spacers
> or washers for whatever reason on many older hubs
>
> The much despised (at least in the CR circle) Shimano SIS indexing
> system pretty much forced the OLN width standardization to 120mm, 126mm,
> 130mm and 135mm.
>
> I like to adjust rear dropout widths to 120-121mm for 5 speed FWs and
> 126-127mm for 6+ speed freewheels. It makes for easier wheel insertion.
>
> Chas. Colerich
> Oakland, CA USA
>
> Ken Freeman wrote:
>> There should be a caveat thrown in here: regardless of Sheldon's research
>> and knowledge, cyclists and techs on several forums have thought they had
>> bent or broken axles after using a wider wheel in an older frame without
>> cold-setting the frame and aligning the dropouts. I don't know any of the
>> extenuating circumstances, but it seems one should be cautious about doing
>> this. With my NR hubs, I am reluctant to break or stress any original
>> axles.