Rotary Table Questions

[Susquatch]

A dividing head plate is calibrated for 40 turns per revolution, not 90, so to use one in a straightforward way... no. But on some divisions, it can still be useful with a little math - perhaps an engineer?

On a dividing head like mine I have to disnegage the clockwork to use dividing plates. On both my RT's the worm gear is always engaged, both plates and manual handle...

I think I 'spoke' poorly in my above post on the matter. Sigh. English! sorry. I think I said it backwards!

No worries..... Backwards usually works too...... Lol!

In my case, the worm gear seems to disengage and works as @Brent H suggests.

I think that means I just replace the handwheel with the dividing system.

A mathematician wouldn't bother with a dividing disk. They would just add the increment to the previous number and dial it in. Non-mathematicians can't do that because they can't handle addition in a base 60 numbering system. Assuming that is correct, I'd prefer to avoid mistakes which is what the dividing disks are supposed to do. (I think).

Why can't both trig, time and number bases all be in octal?

Which reminds me of another story. When one of my previous bosses retired, I was tasked with getting him a good gag gift. He was a member of Canada's metrication committee (the dummies who believed the US would follow Canada's lead and go metric too.). He was also an avid clock maker.

I decided to get him a metric clock. 10 hours in a day, 10 minutes in an hour, 10 seconds in a minute, etc.

There were none. Apparently the decimal base system doesn't apply to time or angles. Go figure!

So I designed a kit and bought all the gears and parts needed to make one and that's what we gave him! I don't know if he ever built it. Didn't matter. It was hilarious.

 

[Susquatch]

On my BP RT, the worm needs to be engaged at any point in the circle that is divisible by 4 and the handle needs to be in the UP position. Otherwise the markings on the hand wheel and table are out of sync no matter what you set the dial to and thus the set angles won’t be correct.

Yours might be different, but there are RTs where it matters where you engage the worm.

Yup, that's mine. I discovered that myself when I noticed that zero on the crank marks could not be aligned with zero on the table itself. But by disengaging the worm, I could turn it a bit such that they could be aligned.

 

[Susquatch]

I think I fixed the goofy markings on the crank that only short people could see. Loosening the little tiny Allen key allows the whole assembly to be turned so that all the markings are easily viewable from above. After that, another alignment of the worm gear and its good to go.

I'll play with my dividing plates tomorrow.

 

[Brent H]

https://www.sowatool.com/img/catalogue_downloads/USCAT17_Sec7_Accessories .pdf

might be of some use to you?

 

[Susquatch]

https://www.sowatool.com/img/catalogue_downloads/USCAT17_Sec7_Accessories .pdf

might be of some use to you?

I'll take a good look! Thanks @Brent H!

 

[Susquatch]

So here is my RT with the crank handle and front dial removed.



In the process I discovered that the outer screw on the handle is a lock for the front dial.

Obviously, it is already setup for divider plates. The three threaded holes bear witness to that.

Unfortunately, not mine...... My divider plates have a smaller inside ID and a smaller bolt circle and the index bar will never reach.

Nonetheless, I think I can make a bushing that fits the existing stub, attaches with long machine screws at the 3 existing plate mount holes, and then provides a new 2 screw mounting stub out closer to the handle pin for my plates. There isn't a lot of axial room for them to the handle stub, but if need be, I can extend the stub too.

The other alternative is to buy a new plate kit that fits as is.

 

[Brent H]

@Susquatch - that plate kit is crazy hard to find in Canada. Seems to be a UK or Australia thing - I located plates for my 12” RT in the US for $119 but they wouldn’t ship to Canada. Had to find a mule. It will probably be easier to mod the plates you have on hand. Plates for the larger RT’s are sometimes double sided and 1/2” thick.




The triangle of holes seems to be about 1.400” between centres.
Holes go from 26 to 99 on 2 plates double sided. If I had a 127 hole plate—— soon - soon

 

[RobinHood]

If I had a 127 hole plate—— soon - soon

Or an universal dividing head with differential dividing capabilities like a B&S #3; that would also work...

 

[Dabbler]

Re 127 plate...

You can print on paper a 127 section pie chart with fine lines and get within .004 on the 8" diameter.... or you could 3DPrint a plastic one....

or, do you have a 127 division gear on your lathe that can help?

 

[Brent H]

@Dabbler the 127 plate would be to make the gears for the old Utilathe

couple more pressing projects and then I will crank out a plate of 127 for all to behold .

 

[Susquatch]

@Dabbler the 127 plate would be to make the gears for the old Utilathe

couple more pressing projects and then I will crank out a plate of 127 for all to behold .

Wow, I'm impressed...... Not with making the gear. But rather with making the dividing plate to make the gear with!

 

[Dabbler]

Stefan Gotteswinter made one by doing a radial plot of 127 divisions on a laser printer and used his lathe and a loupe for indexing. He felt he could differentiate up to .002 on the diameter. I think the gear was for metric transposing on his Enco lathe.

You can find a 127 dividing plate for a rotary table in thingiverse for 3d printing. this one is configurable it might do what you want: https://www.thingiverse.com/thing:1908993

In any event, it isn't too advanced to design and build one in Fusion360 for printing.

 

[George]

Nigel at go create hobby machine shop on YouTube does a video about making gears and makes a dividing plate

 

[Susquatch]

Stefan Gotteswinter made one by doing a radial plot of 127 divisions on a laser printer and used his lathe and a loupe for indexing. He felt he could differentiate up to .002 on the diameter. I think the gear was for metric transposing on his Enco lathe.

You can find a 127 dividing plate for a rotary table in thingiverse for 3d printing. this one is configurable it might do what you want: https://www.thingiverse.com/thing:1908993

In any event, it isn't too advanced to design and build one in Fusion360 for printing.

Despite being a Pioneer in 3D printing during my career, I don't have a 3D printer myself ..... I've thought about it many times though.....

I'm just going to make an adapter to use my existing plate set.

For most jobs though, I'll prolly just "do the math". Besides, after fiddling around and measuring all my parts, I cannot imagine that a plate can be as accurate as manually setting the Rotary Table. Even with a plate done with 1 thou precision, there is the pin load, pin play, arm play, and stop interference to contend with. I watched a few you tube videos of people (Including a few pros) using divider plates. What a mess they made of the backlash and movement in the parts! One fellow even had to move the stops back and then held them with his fingers after that - did he get it right?

Assuming I am correct about all that, it seems to me that a mission critical job would be better done using math to create a table of angles and just dial them in. I am blown away by the low backlash on my RT. But even so, taking up backlash is no big deal to anyone who uses machines regularly and the plates do not look after backlash anyway. If anything, they add some!

I also have a rotary spindle with a 1 degree adjusting system and 5C Collets.



I used it and some math to make a small gear to raise and lower the head on my mill/drill. Till then I had to lock it, and reindex the old gear every 270 degrees to compensate for two missing teeth!

 

[RobinHood]

When you talk about making dividing plate(s) for the RT, are you referring to plates that would fit here?


That would make use of the “indirect indexing” method of dividing the circle.

If you need to have 127 divisions, indirect indexing will never work with a 127 hole plate unless your worm/worm gear ratio is 1:1. I believe the standards are either 40:1 or 90:1 for RTs and Dividing Heads.

I think you are planning on using “direct indexing”. The plate is mounted to the rotating member (spindle or table) and moves in unison (ie. 1:1), then it will work. How are you going to mount the dividing plate to the RT? It would have to be an awfully large ring to fit over the OD of the table.

In the picture below of your B&S #2 dividing head, the red arrow points to the direct indexing plate (they are removable) and the green arrows are all your indirect indexing plates.


If you are planning the use the dividing head, you need to make a direct indexing plate with 127 holes.

You are concerned about accuracy and backlash. If you make your template large (like 12” or so) and the gear you are cutting is only ~ 6”, your error will be reduced greatly as your are only cutting half way up the radius of the template. Your gear tooth spacing will be plenty close enough for slow speed threading.

 

[RobinHood]

Little bit off topic - but I’d does relate to accurately positioning holes on a circle (note the year this experiment was done!):

 

[Susquatch]

When you talk about making dividing plate(s) for the RT, are you referring to plates that would fit here?

That would make use of the “indirect indexing” method of dividing the circle.

Yes, that's exactly what I am planning to do and that's exactly where I want to mount them - albeit with an extended bushing to get the plates out closer to the crank and fit the smaller ID of my existing plates.

I guess I made a few assumptions about all of this. I assumed that one turn on my crank moves the table 4 degrees. So any indexing plate with "N" holes in it moves the table 4/N degrees for each indexing hole. All those tables that machinists use just simplify that math. How many holes a given plate has just breaks the 4 degrees into bigger or smaller increments

To be honest with you, I'm not all that sure I even want them. That's kind of what I was getting at in my previous posts. When I stand back and look at the big picture of what is going on MY TABLE, I realize that my 10" table provides pretty amazing precision all on its own without plates! Furthermore, the plates might actually reduce its precision!

You are absolutely correct, the bigger the table, the better the angular precision that CAN BE achieved.

You are concerned about accuracy and backlash. If you make your template large (like 12” or so) and the gear you are cutting is only ~ 6”, your error will be reduced greatly as your are only cutting half way up the radius of the template. Your gear tooth spacing will be plenty close enough for slow speed threading.

I'm not really concerned about backlash. My table seems to have zero backlash. At that huge radius, it's easy to see it physically move when the crank is turned. I still have to measure it though. When I do, I bet I'll find a little backlash.

I only mentioned backlash (and other errors) for the purpose of discussing the negative side of using a divider plate VS just using the tables own basic rotary system. My basic table will discern minutes of angle. At a 4" radius, that's only a thou or so. (didn't do the math just guessing).

But I suspect, adding all the movement (backlash, bending, play, error, etc) that is inherent in a dividing plate system will end up much worse than just using the tables built in scale.

As @Dabbler said a few posts ago, all this stuff is really just math. As long as I understand what the parts are all really doing, I can do the math as well as needed.

All that said, I like convenience as much as the next guy. If I can add a plate system to my table with very little expense, why not? It will certainly be faster than the math and less error prone too (I think).

 

[Susquatch]

Little bit off topic - but I’d does relate to accurately positioning holes on a circle (note the year this experiment was done!):

View attachment 17674

Very cool story!

Back in my day, we would have fired any engineer who required that kind of precision in a part though. Stacked tolerances and low cost are much more important than absolute precision.

An interesting story. One our suppliers made brake parts for lots of OEM's. They once told me that the Japanese auto companies had much more stringent specifications than we did. Not for the regular quality tests but rather for the endurance specs. We would typically specify minimum endurance requirements. So did the Japanese. But the Japanese ALSO specified maximum endurance! For most parts, we only cared if it lasted as long as the life of the vehicle. If it lasted longer that was a bonus. But apparently the Japanese were fanatics about controlling costs any anything that lasted too long was obviously too expensive!

 

[RobinHood]

Great story!

So it is no coincidence that components start to fail right after the warranty is up. Who would have guessed that?

Look at this fellow’s quest for a precision dividing head (it is a two part series). Alex is on a whole other level of precision…

 

[Susquatch]

When you talk about making dividing plate(s) for the RT, are you referring to plates that would fit here?
View attachment 17672

That would make use of the “indirect indexing” method of dividing the circle.

If you need to have 127 divisions, indirect indexing will never work with a 127 hole plate unless your worm/worm gear ratio is 1:1. I believe the standards are either 40:1 or 90:1 for RTs and Dividing Heads.

I think you are planning on using “direct indexing”. The plate is mounted to the rotating member (spindle or table) and moves in unison (ie. 1:1), then it will work. How are you going to mount the dividing plate to the RT? It would have to be an awfully large ring to fit over the OD of the table.

In the picture below of your B&S #2 dividing head, the red arrow points to the direct indexing plate (they are removable) and the green arrows are all your indirect indexing plates.
View attachment 17670

If you are planning the use the dividing head, you need to make a direct indexing plate with 127 holes.

You are concerned about accuracy and backlash. If you make your template large (like 12” or so) and the gear you are cutting is only ~ 6”, your error will be reduced greatly as your are only cutting half way up the radius of the template. Your gear tooth spacing will be plenty close enough for slow speed threading.

Just a few additional observations.

My universal dividing head is 9 degrees per revolution of the crank handle. 1 in 40 as per your note.

My RT is 4 degrees per revolution. 1 in 90 as per your note.

My collet indexer is direct only and is setup to do only whole degrees.

I'm still struggling to understand the value of all this numbered hole stuff or even why it's called 1 in 40 instead of 9 degrees per turn. It seems plainly obvious to me that 9 holes on a dividing plate for my universal head would yield 1 degree per hole. 18 holes would be a half degree or 30 minutes. 36 holes would be 15 minutes each, and 540 holes would be 1 minute each.

I have 3 plates (that fit it) and one that is a misfit. The three range from 15-20, 21-33, & 37-49 with a few skipped on each plate except the first. They all have six sets.

As far as I have seen, there are no plates with 540 holes, so there are no plates with 1 minute resolution. Yet, the regular crank handle on my RT will resolve 1 minute - with near zero backlash! There is wayyyy more backlash and error using those plates........ So why are they so popular? Surely it can't be just the convenience!

The more I think about it, the more I want a working set of dividing plates, and the more pathetic I look in the mirror......... Sheshhhh