First time going deep – Gear Box.

Might be a lot of play between shaft and rollers, shaft worn undersized?
The layshaft is tight in the inner race of the roller bearing. The play is between the inner race and the rollers.
 
Sounds like something is not right. I'm going to pull the gears & shafts to investigate.
 
Using the setup shown above, I measured right around 0.005" max movement on the main shaft when it turns.
Wouldn't that number include any sleeve gear bearing and/or bush play, in addition to any shaft deflection?

Trying to do this measurement on the layshaft seems not very useful since I can easily deflect it at least 0.050" worth of roller bearing play with my finger.

It might just be me, but attempting to measure "deflection" at the unsupported end of a shaft seems an odd thing to do?
 
I had that same thought.
So how else is it done?

The usual method would be to set the shaft between centres (or on V blocks) and measure in the centre.


"Step 4: Measure for concentricity (run out) by placing the spindle between two centers (if not mounted on the mill) and using a dial indicator at the center area of the roll space, rotate the shaft around from one side of the keyway to the other and record the deviation."
First time going deep – Gear Box.


First time going deep – Gear Box.
 
Now that I have everything back out of the case, I put the mainshaft in my mill/drill press using 5/8" collet. While rotating the mill head, I measured 0.005" deflection at the mid-point of the shaft and 0.008" at the farthest point before the splines for the clutch center. As a control, I measured a 9/16" drill bit as well (had to change collet). It showed 0.0025" deflection. If I subtract out the control deflection, the result correlates with my in situ measurement of 0.005"

I don't have a collet that will fit the layshaft so I can't use this method on it.
I did try measuring deflection with the layshaft held vertically in the kickstart shaft bush. This method gave a deflection measurement of 0.010" on rotation, but I also measured 0.020" of play for the layshaft in the kickstart shaft bush. I'm not sure I can draw any conclusions from all this.
 
The usual method would be to set the shaft between centres (or on V blocks) and measure in the centre.


"Step 4: Measure for concentricity (run out) by placing the spindle between two centers (if not mounted on the mill) and using a dial indicator at the center area of the roll space, rotate the shaft around from one side of the keyway to the other and record the deviation."
First time going deep – Gear Box.


First time going deep – Gear Box.
Thanks for clarifying.
I think I will just take them to the machine shop on Monday and have them measured by a professional.
 
Might be a lot of play between shaft and rollers, shaft worn undersized?
Just to clarify, the 0.050" movement is angular, not perpendicular to the shaft, and it was measured at the end of the shaft farthest from the bearing. So the movement in the bearing is much smaller. I believe that when the layshaft is fixed at both ends, that angular movement be miniscule ( roughly the amount of play in the Kickstart shaft bush). When held in place, the inner race has no perceptible play between it and the rollers. I would hope so with a new bearing.

There is still the issue I brought up in post #16. I realize now that I definitely put the layshaft rollers in with the opposite orientation with respect to the inner race from how it came from the package. Does anyone know if this is a problem? The Old Britts article mentions the orientation, but it also says he doesn't know if it matters.
 
While the main & lay shafts are at the machinist's shop waiting to be measured for deflection, I dove into the clutch. Cleaned it up and measured the plates. The 4 friction plates are all 0.140" or 0.141" thick. Workshop manual has 0.142" as minimum thickness. Looks like I will be ordering new friction plates.
The only other anomaly I found seems to be only in appearance, but I would like to hear if anyone else has this experience.
When I spin the clutch center, it appears that it is off center. I will try to attach a video below. I measured the teeth of the clutch center to be concentric (within 0.002"), but the outer ring of the clutch center varies in thickness by about 0.050" from one side to the other. That is what makes it appear wobbly when it spins. I believe that this is not a problem because a) the functional part is concentric, and b) my clutch was working great with it like this before the tear down.
 
When I spin the clutch center, it appears that it is off center. I will try to attach a video below. I measured the teeth of the clutch center to be concentric (within 0.002"), but the outer ring of the clutch center varies in thickness by about 0.050" from one side to the other. That is what makes it appear wobbly when it spins. I believe that this is not a problem because a) the functional part is concentric, and b) my clutch was working great with it like this before the tear down.
This was a huge problem for Kenny Dreer when adding electric start to the VR880s by affixing a ring gear to the clutch basket in order to have the simplest direct-drive starter setup. Tom Reid did the conversions, he set me up with one on my VR880 replica; he provided 3 different tiny hand-tooled "keys" to carefully shim the critical starter mounting bolt hole in the transmission frame, so as to best align the starter gear to allow the ring gear to "wobble" slightly, without making hard contact with the tips of the starter gear teeth in the wells of the ring gear teeth.

He said that no two Commando clutches have the same wobble, but they ALL wobble to a degree.
 
This was a huge problem for Kenny Dreer when adding electric start to the VR880s by affixing a ring gear to the clutch basket in order to have the simplest direct-drive starter setup. Tom Reid did the conversions, he set me up with one on my VR880 replica; he provided 3 different tiny hand-tooled "keys" to carefully shim the critical starter mounting bolt hole in the transmission frame, so as to best align the starter gear to allow the ring gear to "wobble" slightly, without making hard contact with the tips of the starter gear teeth in the wells of the ring gear teeth.

He said that no two Commando clutches have the same wobble, but they ALL wobble to a degree.
Thanks for the information.
I was unable to upload the video, but I see that you know what I am referring to.
I am relieved to know that my clutch center is "normal".
 
I had that same thought.
So how else is it done?
[/QUOTE In checking for straightness, measurements by different means can contribute to what a conclusion may be.

The use of a true straight edge or perhaps a good surface plate may tell you something as well if there is a slight bent in the shaft. The staight edge along the length of the shaft to some degree at least should reveal if the shaft is good or not for staightness.
 
Last edited:
Thanks for clarifying.
I think I will just take them to the machine shop on Monday and have them measured by a professional.
The use of a straight edge or perhaps a surface plate may tell you something as well if there is a slight bent in the shaft. The staight edge along the length of the shaft to some degree may reveal if there is any bend in the shaft.
Got my sleeve bearing drift made.

View attachment 17061

I have the new bearing in the freezer, so I'll give the installation a try in a few hours.
View attachment 17062

Meanwhile....
Here are the old bearings:
View attachment 17063
Something feels funny about replacing British made bearings with Japanese, but this is what came with the AN Gearbox overhaul kit. The layshaft roller bearing was from India.

By the way, can anyone tell me if my old layshaft bearing pictured above was original equipment?



I can't tell you if its original equipment but it strikes me as a better designed bearing than the new one pictured. It has a shield whereas the other doesn't,
hence offering the bearing added protection from foreign matter entering the works.
 
Last edited:
Thanks for your suggestion.
I got tired of waiting for the machinist to get to my job so I ended up measuring shaft deflection myself using the 2nd method depicted by LAB in post #69 on this thread - V blocks & dial indicator. Using this method, I measured main shaft deflection to be < 0.002" and lay shaft < 0.001".
All the gears & shafts are back in the gearbox shell now. Ready to install inner cover & check end play as soon as I finish riveting the new kickstart pawl plate in place (how this whole thing started in the first place).
 
Geez... This is VERY thorough.
I just tore everything out of my gearbox, replaced the old stuff and threw it back together. Hope I don't blow up on the highway.
 
Back
Top