The KH36 comes with spacers to keep the cranks from being pushed onto the hub.
It is my impression that this is not how ISIS is supposed to work. I think the hub interface is slightly angular, to create strong contact surface pressure when tightening the screw. With the spacers that cannot work, the pressure is just against the spacer.
I have clicking noises under strong load and I’m considering this as one of the sources (I think I have eliminated most other sources). I tried to mount cranks without the spacers. On the left that worked fine, on the right the screws holding the disc rub against the frame. This could be fixed by moving the disc mounting surface 2-3mm to the outside.
Before I machine the spirit crank to fix this, I’d like to have confirmation that my understanding of the ISIS interface is correct.
The ISIS interface is actually tapered. You will get both answeres here: To use with and without spacers.
Before machining your cranks: Which year is your hub? If I am not mistaken before 2015 the spirit hub was a 2 piece desing (axle + hub body). Both where pressed together, and a few hubs had this creaking noise. Since 2015 the hub is a one piece design.
The 1 deg included angle is to keep the whole system tight against the stop. If not for that, even if torqued up against the stop, play would develop between the axle and the crank. It would be very difficult for a unicycle hub to have the same stop as described in the spec, hence the use of spacers (as mentioned by the others). The angle means that the spacers may not be absolutely necessary in every instance (i.e., I don’t use them on my Schlumpf), but they help.
Then, your spacer is far too thick! As is specifiet in the ISIS specification, there shall be a press fit connection between the male and the female tapered spline of about 3 to 6 mm, normally around 5 mm. In practice this means, when you attach your crank arms by hand, there shall be around 3 to 6 mm space between crank shoulder and spacer (if not: take a thinner spacer!). Then you tighten it till there is no gap between spacer and crank left to reach the proper amount of press fit and then up to about 40 Nm which will tense up the screw enough so that it can not loosen itself.
The pedaling torque is transmitted by the splined interface. The taper ensures that this splined joint has enough pre-load and no play. The spacer ensures, that the crank will not move further on the spline with time (caused by widening of the crank) and allow the bolt to become loose.
If you need finer increments in spacer width, check out Shim Rings DIN 988 22x30 available in 0.1, 0.2, 0.3 and 0.5 mm. (see here) But they’re only needed to adjust the position of the disk brake rotor if you use and EDB. Cause as desribed, preload can be between 3 and 6 mm, a pretty wide range. If your gap (untightened!) without any spacer is less than 6mm, you don’t even need one.
Riders with Schlumpf until 2015 specification know perfectly why spacer is necessary to be safe, they hate to have a UPD because of loosing crank which push the button until hus is in neutral gear !
You’ll have to explain that to me… on the Schlumpf literature, Florian never mentions using spacers. He only mentions how important the torque is.
What kind of spacers do you put on a pre-2015 Schlumpf?
And this is the problem, only 2016 Schlumpf is from now compatible with spacer and safe regarding crank loosening risk !
Other point, 2016 Schlumpf axis is now at the same lenght than KH or Nimbus hubs, so no more risk to have crank interference with frame
2 big issues solved !
Bit of crank interface history to throw in here. ISIS is originally a bicycle standard http://www.isisdrive.com/isisdrive/ISIS_Drive_Standard_Document_revD.pdf which as pointed out earlier has a crank stop defined as part of the standard. Before ISIS there was Shimano Octalink which was a similar splined interface standard (though proprietary to Shimano), also with a crank stop which the crank tightens against. The difference being that Octalink had no taper on the splines - the result of which being that the interface tends to develop play.
As already mentioned, the taper does come into play even with the crank stop as it still results in a preload of the crank interface, preventing play when loaded. Because whilst the stop limits the movement of the crank up the axle, the movement of the crank up the axle doesn’t have to be stopped by the taper for the taper to be effective.
In all cases, for stop (which is machined onto the axle in the standard) read spacers for use with unicycle hubs.