Woodruff Key Fit

[anyoldname]

What type of fit would you expect to see (width wise) on a woodruff key?

I'm detailing up some drawings to hold a pulley onto a camshaft, our company standard is based on DIN 6888 but I'm not convinced we've interpreted it correctly, for a 5mm wide key I'm meant to use the tolerances shown below:

Key Way Width = 5 P9 (-0.012/-0.042)

Key Width = 5 h9 (0/-0.03) (Tols of standard part)

These values will give anything between 18 microns clearance and 42 microns interference and its the interference part that I'm questioning. If the key to pulley interface had this amount of interference then its going to take a fair old push to get the pulley onto the shaft.

There is nothing to suggest that the pulley slot should be any wider than the shaft slot.

Any comments or suggestions? Am I worrying about nothing?

I only have a copy of the standard in German so despite my efforts with Google Translate I may have missed something.

Cheers.

 

[TrackRat]

Sounds like a std. "hammer fit"... LOL

 

[thruthefence]

A quick glance at my "Machinery's Handbook",
"Bearings and other Machine elements"
Table 5
'British Standard woodruff keys and keyways',

shows the clearances shaft/sheave to be slightly greater for the sheave (or pully) for all sizes of woodruff keys.

eg, for a 1/16" T key, the shaft is cut to .062/.063, and the pulley to .063/.065".

for a 1/4" T key, the shaft is cut to .249/.251", and the pulley to .0252/.0253".

For the largest key in the table, the clearance is .002" for the pulley.

I guess these guys don't want to work too hard installing the pulleys.

Having said all that, the illustration for table 5 shows a tapered shaft.

 

[Tmoose]

I think the modern, and best way to do this is use easy to assemble but minimal clearance on key and key seat width for accurate angular positioning, and create nice flat perpendicular shoulders on the pulley/sheave and camshaft, and clamp them hard with a center bolt. Asking a tech to fit a key and sheave, or struggle with R&Ring interference fitted components is looking for trouble, as is asking a key to reliably transmit torque that is anything but strongly unidirectional.