Re-engineering the Shift Spring

[Jim in NZ]

Chuck, I thought that the coil binding you refer to was between the inside of the spring coil (25 in the drawing that swooshdave has kindly attached), and the boss on the ratchet arm (24).

The coil doesn't encircle the preselector shaft (27). So how can replacing the preselector shaft fix the problem?

Am I misunderstanding this?

You are doing a great service with your efforts to improve the spring. Please keep going, I'm sure there are lots of us who don't have the engineering expertise to do this, but are cheering you on!

[Scud]

 

The coil doesn't encircle the preselector shaft (27). So how can replacing the preselector shaft fix the problem?

 

 

Interesting question, which I hadn't though about until you asked. The short bend of the spring does ride on the stamped steel portion that is attached to the shaft (27). If the shape/size of that stamped portion was different, it could require less travel to install the spring. And if that's true, then this over-travel problem could be solved by grinding metal off that plate such that the spring is 10-degrees less loaded when installed. Perhaps the design flaw was not in spring, but in part that the spring touches...

 

@Chuck - feel free to grind away on the parts of my pre-selector if this suggestion seems worth exploring.

[Chuck]

Flickr is down right now, so can't show any pictures. I mentioned to Lucky Phil about the radius the end of the spring rides up on, and made the new spring a little longer so it wouldn't do that. Radiusing the edge and polishing it would help, too.

I made a for real .080 dia. spring today.. and an .071 dia. spring with 2.45 active coils. I would need to make a smaller mandrel to get the ID of the spring the same as the .080 spring.. it's .040" bigger. Figured I'd try it any way, and the preselector has never shifted better.    We're still very early into this thing, but I'm cautiously optimistic. It shifts up or down as fast as I can move it.

[Lucky Phil]

Chuck, I thought that the coil binding you refer to was between the inside of the spring coil (25 in the drawing that swooshdave has kindly attached), and the boss on the ratchet arm (24).

The coil doesn't encircle the preselector shaft (27). So how can replacing the preselector shaft fix the problem?

Am I misunderstanding this?

You are doing a great service with your efforts to improve the spring. Please keep going, I'm sure there are lots of us who don't have the engineering expertise to do this, but are cheering you on!

It was the boss on the ratchet arm that was too big on earlier models and caused the spring binding on the coils.

The sharp edge on the selector arm plate where it engages the spring tab can cause a stress raiser on the tab which may cause it to fail. I rounded off the selector shaft plate edge to help alleviate this. but Chuck may have found another solution.

I took my bike for a little blast yesterday for the first time in about 5 months ( been very busy) and reminded myself about how well this thing shifts from 2nd gear upwards without using the clutch. Its every bit as good as my GSXR1000 track bike, just snick, snick, snick really nice. I recommend it, just load the lever a little and flick the throttle under accelleration and it clicks right in there. 

 

Ciao

[Jim in NZ]

Yes, that's what I thought, thanks Lucky Phil.

I also rounded off that sharp edge many years ago, but none of my springs have broken there, they've all broken around the coil.

I'm about to replace another spring, and I don't want to spend around NZ$150 on a new preselector shaft unless there is at least some possibility that it might extend the spring life. Despite what MI say on their website, I haven't heard a plausible reason, so I'll just go ahead with the spring I think. Maybe next time, you guys working on this will have found a solution!

Incidentally, I have always put a kink in the long arm of each spring to lessen the tension it works under. Gearchanging is still OK (it doesn't need much spring pressure), but the bloody things still break!

 

Cheers,  Jim.

[Lucky Phil]

Yes, that's what I thought, thanks Lucky Phil.

I also rounded off that sharp edge many years ago, but none of my springs have broken there, they've all broken around the coil.

I'm about to replace another spring, and I don't want to spend around NZ$150 on a new preselector shaft unless there is at least some possibility that it might extend the spring life. Despite what MI say on their website, I haven't heard a plausible reason, so I'll just go ahead with the spring I think. Maybe next time, you guys working on this will have found a solution!

Incidentally, I have always put a kink in the long arm of each spring to lessen the tension it works under. Gearchanging is still OK (it doesn't need much spring pressure), but the bloody things still break!

 

Cheers,  Jim.

If you can Jim just wait for Chuck and Scuds spring. Knowing Chucks magic fingers work it will be a first class solution and unbreakable.

 

Ciao

[Jim in NZ]

"If you can Jim just wait for Chuck and Scuds spring. Knowing Chucks magic fingers work it will be a first class solution and unbreakable."

 

Yeah, thanks Phil, I'll follow them a while and see how they are getting on, as I'm not in a huge hurry.

My spring replacement is a preventative one. Last spring broke after 15,000km and I had to replace it on the side of the road 300km from home. Now the current spring has done the same mileage, and it is much easier to do it in the comfort of my garage at home before it breaks.

 

Cheers,  Jim.

[Chuck]

Jim said..

 

 

 Despite what MI say on their website, I haven't heard a plausible reason,

If you look at the picture on page 7 of the preselector shaft, you'll see (barely) how the end of the spring rides up in the radius at the end of travel. I made the short leg a little longer to avoid that. You could go in with a die grinder and move that radius back a little. While you are there,radius and polish the edge. The reason it is breaking in the coil is simply over travel of the spring. A less than optimum design. 

[Chuck]

As promised, how to make a spring.

At the end of Friday's installment, our hero was making the mandrel for the spring to sit on the bending fixture. As he was finishing it.. the cutting tool dug in at the end of the cut and made a crow digger. (technical term) 

Spent some time adjusting the gibs of the cross slide and saddle, and turned another. Center drill..

2017-11-18_03-48-17 by Charles Stottlemyer, on Flickr

and drill for a 1/4"bolt. If you drill 1/64" undersize first, the finish drill will be pretty close to on size. Plenty close enough for this work.

2017-11-18_03-48-27 by Charles Stottlemyer, on Flickr

All that is left is to use a parting tool to cut it off, and face off that side.

2017-11-18_03-48-42 by Charles Stottlemyer, on Flickr

Now we can bolt it to the fixture and make springs. Some die blue on the fixture, and a Guzzi spring gives the angle of the bend, 45 degrees, and I scribed a line at 90 degrees to eyeball the bend on the short arm.

2017-11-18_03-48-57 by Charles Stottlemyer, on Flickr

The boss is calling me to breakfast, back later. Don't touch that dial! 

[Chuck]

Put the lever over the long pin and make the first bend.

2017-11-18_03-49-10 by Charles Stottlemyer, on Flickr

That goes in the hole and is clamped by the screw with the wrench in it.

2017-11-18_03-53-19 by Charles Stottlemyer, on Flickr

Oh, BTW, the lathe is unplugged from the power source during this operation. Leaving a chuck key, etc. in the chuck is verboten. 

Lining up the groove in the brass piece

2017-11-18_03-53-36 by Charles Stottlemyer, on Flickr

and sliding the plastic on top. The groove is .005" shallow at the end, so clamp pressure is controlled by the front set screw.

2017-11-18_03-53-53 by Charles Stottlemyer, on Flickr

Using the chuck wrench as a lever, the chuck is rotated.

2017-11-18_04-05-32 by Charles Stottlemyer, on Flickr

The spring has to be over bent. You can see the two dashes on the back of the chuck that I decided was correct. It springs back about 90 degrees.

2017-11-18_04-06-00 by Charles Stottlemyer, on Flickr

Here's the new spring on the bending fixture. As predicted by the Machinerys handbook, the ID is perfect. All that is left is to put the lever over the tall pin and bend the 45 degree angle.

2017-11-18_04-07-52 by Charles Stottlemyer, on Flickr

Here's the first complete spring. The extra length will be cut off with the fiber wheel. We *know* the Guzzi spring at the bottom breaks, so the new one has a little more active coil. 

2017-11-18_04-08-03 by Charles Stottlemyer, on Flickr

Ok, let's try our other design with the .071 music wire..

2017-11-18_04-08-25 by Charles Stottlemyer, on Flickr

Using the spring calculator posted previously in this thread, this spring has considerably more "safe travel," and shouldn't break in service. To do it properly, I'll have to turn down the mandrel. The ID of the lighter wire is .030" or so bigger, but tried it anyway. 

Here it is installed. You can see the extra turn of wire here..

2017-11-18_04-08-38 by Charles Stottlemyer, on Flickr

It works so well if I were just making one for my bike, I'd bolt this preselector on and go with it. 

However, I'm fully aware that if you have spent some time and/or  money on a project, your opinion may very well be colored. Time for testing. The new scientific "fish scale" will be here early next week, and I'll be doodling on some kind of cam arrangement that mimics the spring travel several times per revolution of my low speed motor. Stay tuned. 

 

 

 

 

[Scud]

Hey, Chuck, thanks for taking the time to explain how you're doing this and post the pictures. Feel free to test the new spring I sent you to the failure point with your motorized cam arrangement - that would give an interesting baseline to test the new design against.

 

Gosh - failure in 10-15,000 miles... how many times do we shift per mile? Obviously, depends on riding style and traffic vs. freeway...  I've never counted, but I guess I could shift 20 times per mile on super-tight roads with elevation changes - and 0 times per mile when I'm chasing the horizon through the desert.  It seems like some of these springs are breaking after just a few thousand uses - while the other springs in the pre-selector never break.

[Chuck]

 

 

while the other springs in the pre-selector never break.

I had a shift return spring break in the Mighty Scura..    Maybe it was an anomaly?  The Mighty Scura *has* been shifted a lot. Pete had MPH go through the bike, and a new pawl spring was installed at that time. I've put 17000 miles (roughly) on it since. *Most* of that time has been spent in the Santa Monica and other mountains, and when not there, on the 5, 10, 405 etc. getting there, and shifting a lot. We have to simulate enough shifts per minute to test these things in a reasonable amount of time.. I ain't gettin any younger..  

Let's see, now.. 1 shift per second would be 60 per minute times 60 would be 3600 per hour. I wonder if that would over heat the spring and temper it? It will lose it's ju ju at 300 degrees F, according to my supplier. We may have to use spray mist coolant or something.. 

[Chuck]

Here's my antique electrocutioner's special low speed electric motor that a guy gave me (Guzzi content) many years ago. Since it's reversible, I've planned on using it to open my manual hanger door.. but have never got a round toit.

I gingerly plugged it in and tripped the switch. 46 rpm. That'll work.

This will give me an excuse to rewire this sucker.    

2017-11-19_11-42-20 by Charles Stottlemyer, on Flickr

[czakky]

Here this whole time I thought springs were made of magic! Thanks for de-mystifying this for us.

[Jim in NZ]

Chuck, the 0.071 3-coil job looks great. Being able to test it compared with the MG standard spring is beyond expectations! Looking forward to seeing how this works out. Thank-you for all your effort and time.

 

Cheers,  Jim.