Measuring fork rake

[Mick F]

It's raining and miserable out there, I'm all snotty with a cold, and I'm at a loose end. Hence, I thought I'd resurrect this subject.

Within this thread viewtopic.php?f=5&t=58006 it was suggested by 531colin that you could remove your forks and measure on a table.

You need a flat surface (I use Granite kitchen worktop) and a packing piece thats the same thickness along its length (I use an offcut of worktop).
Put the fork steerer on the packing piece so its parallel to the worktop, the curve of the fork needs to take the dropouts up, they need to be parallel to the worktop, clamp it in place with G clamps. Measure from worktop to dropout, and from worktop to centre of steerer, and subtract.

However CJ came on and suggested an easier way.

Measure front centres (centre of q/r skewer to middle of crank dustcap) with both wheels against a kerb so you're sure the front wheel is pointed straight ahead, then turn fork around so it points straight backwards (you may need to remove front brake and mudguard and deflate the tyre down to clear the down-tube on bikes with tight clearances) and measure the front reversed centres.

Subtract the forks reversed dimension from the normal front centres and divide the result by two. That'll give you a value that's shorter than the actual fork offset by about 1mm (which is probably less than the uncertainty in your measurements) so add 1mm and that's it, to the nearest mm.

I asked about the "extra 1mm" and it was to do with Pythagoras.

I measured mine using CJ's method and measured 42.5mm plus the 1mm gave me a fork rake of 43.5mm.

Today I removed my forks and measured in accordance with 531colin's method. I clamped the forks to the table and pulled out my front axle - easy job - and used it to determine the centre line of the front dropout. I even fitted a new battery in my digital callipers!

I've measured and measured and measured, and consistently found I had a rake of 41.7mm.

Of the two methods, 531colin's gives the most confident answer and therefore more accurate.

[niggle]

I suppose the 'Pythagoras' bit is to do with the fact that reversing the fork will raise the front end slightly. As you say 531colin's method is the most accurate due to being a direct measurement rather than a deduced one, as long as you are using accurately measured data, eg. your worktop is perfectly flat and you have been able to get an accurate fix on the axle centre.

One thing that 'springs' to mind with lovely compliant steel forks like yours is if the rake might be affected by the rider's weight, i.e. do they sag slightly, thus pushing the fork legs forward and increasing rake? I do not suppose this would happen under the bike's weight alone, but could it affect issues of rake choice when replacing such a fork with something stiffer like carbon or a disc brake specific fork?

[Brucey]

an accurate (and straightforward) method is a variant of CJ's. Strip the front end of the bike enough to be able to turn the forks 180 and take measurements not of the wheelbase, but along a line at right angles to the steerer axis, and divide by two. No Pythagoras required.

There is no fixed datum for this measurement, but you an usually contrive something by working off one crank having strapped the other one to the seat tube or something. Even if this measurement is taken at a slight angle the error will be tiny provided the forks are turned so that the axle would be at right angles to the line.

You can gauge the correct point along the crank to take the measurement by turning the fork 90 deg and putting a straight edge across the fork ends. Again if the datum point is wrong by a few mm it won't cause a significant error; the crucial thing is to use the same datum point each time.

If you are happy to strip the bike completely and have suitable equipment the other method works quite well. However because it doesn't use the headset bearings as a reference, it is subject to small errors through this and any steerer deformation.

A further variant is to lay the bike on its side and take measurements to the ground with the forks turned each way 90 degrees from straight ahead.

Slight steerer deformation is more common than you might think on used forks, and non-concentricity of headset bearings isn't unusual either. Any method that leaves the forks in the frame may seem to make it more difficult to take 'accurate' readings but at least those readings are 'real'.

cheers

[531colin]

Now you have the forks out, Mick, the best way is CJ's second variant;
Measure dropout to worktop, reverse forks 180deg, measure again, halve the difference.
In any case fork offset "between 40 and 45mm" is accurate enough.

[Mick F]

Yes Colin, I tried that.
The clamping affair I have made it easy to get the steerer parallel - but too low to turn the forks through 180.

I sighted the forks against a known level in the kitchen - the sink! - to get them correctly levelled off at the ends. The clamp had a couple of bars that the steerer bedded onto so I knew it was level.

Using the front axle I could measure the diameter and halve it to find the centre and use that figure when I measured from the top of it.

I am confident in the figure of 41.7mm but am well aware that it matters not a jot for me! I'm not about to change anything!
(but at least I now know what the figure is)

If the forks "give" when I sit on the bike, I don't know - well I do know because they must - but how much and to what affect, I don't know. I don't know how I could measure it anyway.

[CJ]

I suppose the 'Pythagoras' bit is to do with the fact that reversing the fork will raise the front end slightly. As you say 531colin's method is the most accurate due to being a direct measurement rather than a deduced one, as long as you are using accurately measured data, eg. your worktop is perfectly flat and you have been able to get an accurate fix on the axle centre.

No it isn't and 531Colin says so! The devil is in the assumptions: that the worktop is flat, that the top of the vice is parallel to the bottom, that the bottom headrace seat is perfectly concentric with the steerer tube. Most of these things are near enough true for everyday purposes, but not for metrology purposes, especially in this case where any small deviations will be amplified by the length of the fork. The trick or challenge of metrology is to devise a method of measurement that automatically cancels out the unknowns. Which is what you do by rotating the fork 180 degrees, measuring again and halving the difference.

One thing that 'springs' to mind with lovely compliant steel forks like yours is if the rake might be affected by the rider's weight, i.e. do they sag slightly, thus pushing the fork legs forward and increasing rake? I do not suppose this would happen under the bike's weight alone, but could it affect issues of rake choice when replacing such a fork with something stiffer like carbon or a disc brake specific fork?

Indeed that will be so. If you measure front centres with and without a rider on the bike you may observe a significant inrease, most of which is probably due to an increase in fork offset and part of which will be bending of frame tubes producing a slight relaxation of head angle. Unfortunately I haven't yet thought of a measuring trick to separate these two effects!

[CJ]

Yes Colin, I tried that.
The clamping affair I have made it easy to get the steerer parallel - but too low to turn the forks through 180.

So put something thick underneath it!

[Mick F]

Yes CJ, I thought of that too!
Too late to do 'owt now coz the bike is back together.
Had I had a large V block and a surface plate, it would have been simple.

Actually, in the kitchen, it was easier to leave the forks clamped and accurately positioned (took me ages to get it right) and then measure and re-measure to acheive consistent results.

[Brucey]

consistent, yes; but CJ's (and my/others) point is that they might be consistently wrong....

BTW I would say that if you alter fork offset;

~0-3mm = no significant difference

~3-6mm = a small difference, may not be noticed by the rider

~6mm or more = a significant difference, will be noticed by most riders.

I would be interested to hear if this matches with the experience of others.

cheers

[Mick F]

consistent, yes; but CJ's (and my/others) point is that they might be consistently wrong....

They?
No. Not me.
I was as accurate as I could be, and any inaccuracies were minor.

[PH]

The big question is how can you put your forks on the kitchen table and stay married?

[Mick F]

She was at work.
It's me who cleans up ..............


...... but don't tell her!

[CJ]

I was as accurate as I could be, and any inaccuracies were minor.

But your assumptions - that everything was flat and parallel-sided - were major.

Deviations from flatness of the worktop, parallelism of the vice and concentricity of the fork steerer, that would all be too small to see or to measure with the equipment at your disposal, will be amplified by the length of the fork to result in errors that may be as much as several mm at the fork tips.

Even a surface table and a V-block would be vulnerable to lack of concentricity of the crown race seat. I would cancel that out by resting the race seat on one end the support and turning the fork 180 degrees to take the two measurements. The fork will then, of course, definitely NOT be parallel to the worktop for either measurement, but the axial alignment error is cancelled by the turning and the error resulting from the slight deviation of the offset from the perpendicular will be no more than a tiny fraction of a mm.

[Brucey]

+1 on that

[Mick F]

No.
I spent ages getting the steerer level and the fork end level too.
The table is level.
The table has a machined surface, and I do admit that it isn't perfect, it must be within tiny fractions of a mm.
The headset is circular and the fixed races are fitted solidly to a good quality frame and fork set. I very much doubt they would be eccentric or out of line to alter the static measurement to any degree at all.

Whereas I appreciate that my measurements aren't within high scientific tollerances, they are within +/- 0.5mm.

That'll do for me.