spokewrench
Active Member
- Region
- USA
As of 1990, the Chinese owned more than one bicycle per household. These days, 4 times more e-bikes are sold in China than in the rest of the world combined. Chinese cyclists must know the importance of proper tightening of fasteners. Traditionally, bicycle manufacturers provided sheet-metal multitools with suitable leverage for each size of hex fastener. The use of Allen screws on Chinese bikes didn't make sense to me. With a folding set or bits, the same leverage is used for all sizes. Single keys often have handles that invite over torquing.
On my fourth e-bike, at least 10 Allen screws have torque specs painted in white on black. That's much better than trying to remember a bike manufacturer’s spec and hoping it's correct. Still, I wondered how painted specs were supposed to help most of China’s 200 million e-bike owners. Earning about $70 a week in urban areas and $20 in rural areas, few would have torque wrenches.
Then I remembered the Allen keys in the tool kits of two BMW motorcycles. They took the guesswork out of torquing. I’d tighten until the end of the handle started to hurt my hand. The results seemed consistent and appropriate for each size of screw.
I looked into it. Most L-shaped hex keys are not really Allen keys. To a mechanical engineer, each size Allen key has a handle of a certain length, with a tolerance of about 2%. In ANSI (inch) keys, the Long Series has been standard since the introduction of Class 5 bolts about a century ago. In ISO (metric) keys, the so-called Short Series is standard. The arm lengths of these two standard sets are similar.
In 1875, Allan Cummings patented the socket screw, but mass production wasn’t feasible. In 1906, Peter Robinson found that if a square socket were tapered and beveled, screws could be formed by machines. He got a Canadian patent in 1909.
The same year, William Allen of Hartford got a patent for a method to manufacture hex socket screws. In 1910 he got a trademark for Allen keys. In 1911, Standard Pressed Steel began producing hex socket screws and hex keys. They made good screws, but tool makers had such regard for Allen's keys that hex socket screws were known as Allen screws.
The biggest employment area for tool makers was designing industrial tools and machinery. Reliability was vital because unplanned down time was costly. There were no torque wrenches.
Nowadays, their second biggest area of employment is designing motor vehicle parts. I had a car whose lug wrench had a handle that was shorter than usual, and its end was a prying blade. I couldn’t turn the lug nuts as tight as I thought necessary. Later I saw the torque spec for those nuts. It was a third lower than usual. That wrench was designed to prevent over tightening.
Allen must have realized that a key with a short, square-cut handle could tell a mechanic when a screw was tight enough. A fine-pitch screw could have a socket for a smaller Allen key than its coarse-pitch counterpart. Allen could size a socket for whatever torque a tool maker specified.
Many Americans first saw Allen screws on American military equipment in WWII. If consumers had known they were to take the guesswork out of torquing, other manufacturers would have made hex keys with the same lengths as Allen keys. This could have brought price pressure in the market for the keys used to maintain machines like aircraft and bulldozers.
I needed real Allen keys to test my hypothesis that they would properly tighten the screws on Chinese e-bikes. I wanted H2, H3, H4, H5, H6, and H8. At the cheapest dealer I found advertising ISO, they would have cost $253 because I would have had to buy 310 pieces. The market for genuine Allen keys seems to be engineers running maintenance departments or shops for repair or assembly.
For $13 I bought a set of 30 from Amazon and found that they conformed to ISO specs. (They list outside lengths, while others list inside lengths.) I used the H4 to tighten a screw clamping the handlebars. I replaced it with a digital torque driver, noted the driver’s position, backed off, and retightened it to where the Allen key had left it. It read 4.5 Nm. The clamp read “6.0 Nm Max.” In general, you torque to 75% of max so that some elasticity remains. That would be 4.5 Nm. Perfect!
Next I tried the H8 in my M8 x 1 crank bolt. The bicycle manufacturer specified 40 to 45 Nm. I thought that was wild. I’d torqued by feel. I set up a hanging scale to measure the Newtons on the left pedal as I tried my Allen. I put 106 N on the .17 meter pedal, which meant 18 Nm. The bolt didn’t budge.
I’d expected the bolt to need about 25 Nm. I put the plastic handle on the key so I could turn harder. With 123 N on the pedal I felt the bolt budge slightly. By the way it moved, it felt tight enough. That translated to 21 Nm. It had moved so little that I thought I must have tightened it to about 18 Nm by feel when I installed it.
That was less than the 25 Nm I’d seen here. The bolt couldn’t be above Class 8.8 because stronger bolts withstand only one tightening. An 8.8 M8 is designed to clamp at 3580 pounds, which is 75% of its proof strength. With dry threads and a 1.25 pitch, it takes about 25 Nm. However, this pitch is only 1mm, so an estimated 20 Nm would get me there with dry threads. If I used thread locker or anti-seize, an estimated 17 Nm would provide the recommended clamping force.
The 18 Nm from the Allen key was right in the ball park. The maker of the bolt knew what he was doing. The bicycle manufacturer’s spec is about 2.5 times higher. Watch out!
I used weight transfer to check my H6 Allen key in the front through axle. I put a scale under the center stand. The weight of the bike was on the front tire and scale, which I set to zero. Now if I put clockwise torque on the through-axle, it would take weight from the scale.
Normally, one increases pressure slowly to torque a fastener. In this case I was hasty and could feel that I’d gone too far. I knew that because the pain was increasing when I wasn’t pushing harder. The scale read —2.40. I tried to hold it there for a few seconds, but it went up and down because the changing pain wouldn’t let me keep constant pressure by feel. I quickly let off the pressure. The end of the handle left a dent in my palm.
I tried it more slowly. When I felt the right point, the scale read —2.30. It was only 4% less than before, but I had no trouble keeping a steady scale reading because the pain was not increasing.
I multiplied pounds by 4.4 to get Newtons and multiplied that by .88 meter (the horizontal distance from the axle to the feet of the center stand) to get Newton meters. The axle was marked 8 to 10 Nm. My first try, which felt painfully high, was 9.3 Nm, only 3% above the center of the recommended range. The second was 8.9 Nm, or 1% below the center. Even when used badly, a proper Allen key can be plenty accurate. The bicycle manufacturer’s listed spec of 10 to 15, is 11 to 67 percent higher.
Torquing with an Allen key is like walking barefoot up a driveway that progresses from pea gravel to coarse crushed rock. With coarser, sharper gravel, pressure on soles increases because fewer points support your weight. You proceed slowly until you seem to have gone as far as possible and still be able to stop to chat.
The box of Allen keys is pretty big to keep on a bike. I bought a set of so-called Allen keys in a blue clip. They looked right but turned out to be about 10% short. There oughta be a law! Well, 10% under torquing is probably adequate in most cases.
On my fourth e-bike, at least 10 Allen screws have torque specs painted in white on black. That's much better than trying to remember a bike manufacturer’s spec and hoping it's correct. Still, I wondered how painted specs were supposed to help most of China’s 200 million e-bike owners. Earning about $70 a week in urban areas and $20 in rural areas, few would have torque wrenches.
Then I remembered the Allen keys in the tool kits of two BMW motorcycles. They took the guesswork out of torquing. I’d tighten until the end of the handle started to hurt my hand. The results seemed consistent and appropriate for each size of screw.
I looked into it. Most L-shaped hex keys are not really Allen keys. To a mechanical engineer, each size Allen key has a handle of a certain length, with a tolerance of about 2%. In ANSI (inch) keys, the Long Series has been standard since the introduction of Class 5 bolts about a century ago. In ISO (metric) keys, the so-called Short Series is standard. The arm lengths of these two standard sets are similar.
In 1875, Allan Cummings patented the socket screw, but mass production wasn’t feasible. In 1906, Peter Robinson found that if a square socket were tapered and beveled, screws could be formed by machines. He got a Canadian patent in 1909.
The same year, William Allen of Hartford got a patent for a method to manufacture hex socket screws. In 1910 he got a trademark for Allen keys. In 1911, Standard Pressed Steel began producing hex socket screws and hex keys. They made good screws, but tool makers had such regard for Allen's keys that hex socket screws were known as Allen screws.
The biggest employment area for tool makers was designing industrial tools and machinery. Reliability was vital because unplanned down time was costly. There were no torque wrenches.
Nowadays, their second biggest area of employment is designing motor vehicle parts. I had a car whose lug wrench had a handle that was shorter than usual, and its end was a prying blade. I couldn’t turn the lug nuts as tight as I thought necessary. Later I saw the torque spec for those nuts. It was a third lower than usual. That wrench was designed to prevent over tightening.
Allen must have realized that a key with a short, square-cut handle could tell a mechanic when a screw was tight enough. A fine-pitch screw could have a socket for a smaller Allen key than its coarse-pitch counterpart. Allen could size a socket for whatever torque a tool maker specified.
Many Americans first saw Allen screws on American military equipment in WWII. If consumers had known they were to take the guesswork out of torquing, other manufacturers would have made hex keys with the same lengths as Allen keys. This could have brought price pressure in the market for the keys used to maintain machines like aircraft and bulldozers.
I needed real Allen keys to test my hypothesis that they would properly tighten the screws on Chinese e-bikes. I wanted H2, H3, H4, H5, H6, and H8. At the cheapest dealer I found advertising ISO, they would have cost $253 because I would have had to buy 310 pieces. The market for genuine Allen keys seems to be engineers running maintenance departments or shops for repair or assembly.
For $13 I bought a set of 30 from Amazon and found that they conformed to ISO specs. (They list outside lengths, while others list inside lengths.) I used the H4 to tighten a screw clamping the handlebars. I replaced it with a digital torque driver, noted the driver’s position, backed off, and retightened it to where the Allen key had left it. It read 4.5 Nm. The clamp read “6.0 Nm Max.” In general, you torque to 75% of max so that some elasticity remains. That would be 4.5 Nm. Perfect!
Next I tried the H8 in my M8 x 1 crank bolt. The bicycle manufacturer specified 40 to 45 Nm. I thought that was wild. I’d torqued by feel. I set up a hanging scale to measure the Newtons on the left pedal as I tried my Allen. I put 106 N on the .17 meter pedal, which meant 18 Nm. The bolt didn’t budge.
I’d expected the bolt to need about 25 Nm. I put the plastic handle on the key so I could turn harder. With 123 N on the pedal I felt the bolt budge slightly. By the way it moved, it felt tight enough. That translated to 21 Nm. It had moved so little that I thought I must have tightened it to about 18 Nm by feel when I installed it.
That was less than the 25 Nm I’d seen here. The bolt couldn’t be above Class 8.8 because stronger bolts withstand only one tightening. An 8.8 M8 is designed to clamp at 3580 pounds, which is 75% of its proof strength. With dry threads and a 1.25 pitch, it takes about 25 Nm. However, this pitch is only 1mm, so an estimated 20 Nm would get me there with dry threads. If I used thread locker or anti-seize, an estimated 17 Nm would provide the recommended clamping force.
The 18 Nm from the Allen key was right in the ball park. The maker of the bolt knew what he was doing. The bicycle manufacturer’s spec is about 2.5 times higher. Watch out!
I used weight transfer to check my H6 Allen key in the front through axle. I put a scale under the center stand. The weight of the bike was on the front tire and scale, which I set to zero. Now if I put clockwise torque on the through-axle, it would take weight from the scale.
Normally, one increases pressure slowly to torque a fastener. In this case I was hasty and could feel that I’d gone too far. I knew that because the pain was increasing when I wasn’t pushing harder. The scale read —2.40. I tried to hold it there for a few seconds, but it went up and down because the changing pain wouldn’t let me keep constant pressure by feel. I quickly let off the pressure. The end of the handle left a dent in my palm.
I tried it more slowly. When I felt the right point, the scale read —2.30. It was only 4% less than before, but I had no trouble keeping a steady scale reading because the pain was not increasing.
I multiplied pounds by 4.4 to get Newtons and multiplied that by .88 meter (the horizontal distance from the axle to the feet of the center stand) to get Newton meters. The axle was marked 8 to 10 Nm. My first try, which felt painfully high, was 9.3 Nm, only 3% above the center of the recommended range. The second was 8.9 Nm, or 1% below the center. Even when used badly, a proper Allen key can be plenty accurate. The bicycle manufacturer’s listed spec of 10 to 15, is 11 to 67 percent higher.
Torquing with an Allen key is like walking barefoot up a driveway that progresses from pea gravel to coarse crushed rock. With coarser, sharper gravel, pressure on soles increases because fewer points support your weight. You proceed slowly until you seem to have gone as far as possible and still be able to stop to chat.
The box of Allen keys is pretty big to keep on a bike. I bought a set of so-called Allen keys in a blue clip. They looked right but turned out to be about 10% short. There oughta be a law! Well, 10% under torquing is probably adequate in most cases.
Last edited:
.
