Holes In Dividing Head Plate

[Reeltor]

I recently saw where someone needed to cut a 63 tooth gear but his dividing head (40:1) didn't have a plate with 63 holes. His rotary table (90:1) did have a 63 hole circle on a plate.

I'm just wondering why plates for a 40:1 dividing head don't seem to have a 127 hole pattern or in this person's case a 63 hole pattern. Is there a reason, other than when the plates were made there didn't seem to be a need for a 127 circle?

 

[randyc]

I've wondered the same thing and finally decided to do something about it, LOL. I developed a spread sheet that determines the closest standard hole plate for the desired number of increments within an error that the user specifies.

As an example, my dividing head has a 72 tooth worm gear and the plates supplied with it cannot provide exact division of 127. BUT my spreadsheet found that using a 30 hole plate would give me a division of 127.059. That represents a maximum error of .0013 degrees. I think that I could live with that, ha-ha-ha

 

[rrjohnso2000]

That spreadsheet seems quite handy. Is there a chance of sharing?

 

[randyc]

Sure, anyone that's interested can send me a PM with their e-mail address. Be aware, however, that unlike software companies, I do not provide technical support

User inputs are ONLY the blue text, any other colored text are either calculation boxes or instructions which you should read VERY carefully.

 

[Reeltor]

Boy, that was fast! Thanks for the spreadsheet, I'll have to have some quiet time to see how it works

Thanks again,

Mike

 

[randyc]

Boy, that was fast! Thanks for the spreadsheet, I'll have to have some quiet time to see how it works ....

You're welcome. As it happens, my e-mail was open and any forum PM also pops up on the e-mail hence the prompt response.

 

[Reeltor]

I've wondered the same thing and finally decided to do something about it, LOL. I developed a spread sheet that determines the closest standard hole plate for the desired number of increments within an error that the user specifies.

As an example, my dividing head has a 72 tooth worm gear and the plates supplied with it cannot provide exact division of 127. BUT my spreadsheet found that using a 30 hole plate would give me a division of 127.059. That represents a maximum error of .0013 degrees. I think that I could live with that, ha-ha-ha

Randy, never mind, I think I have it from the example in the sheet. I should learn to read

 

[randyc]

The instructions on the first sheet tell you how to determine the number of turns: 360 / (degrees per turn x number of desired spacings). Enter the number of turns in cell A4. Enter a tolerance in cell K4 - you can pick an arbitrary number like ".01" to start with. The spacing will be calculated and results will appear in the table. Take a look at the following:



If we want 127 spaces then the starting number of turns is 360 / (5 x 127) = .5669 since the integer of this fractional number is zero, then "0" should be entered in cell A4. After entering an allowable error of .01 in cell K4, the various combinations of hole plates and space number are shown. Four different combinations are shown: 127.385 for a hole plate of 23 and 13 spaces is one example. The best combination is hole plate 30 with hole spacing of 17.

So for this particular example, there are no WHOLE turns for each spacing, the dividing head is rotated just enough so that every 17th hole on the 30 hole plate provides one of the 127 spacings. To check this: 360 / (17/30 x 5) = 127.059

I'm sorry that this sounds cryptic but like most software, it is completely clear to the programmer. I'd not intended this spread sheet to be used by others so it is not well documented. I wrote this years ago and have never posted it on any forum for the reason that I do not have the time to explain how to use it to every person that is interested.

The best way to learn how to use the spread sheet is the same process as we often learn how to use machine tools: by using it. Experiment by entering some variables that correspond to known solutions. If necessary, after customizing the sheet to conform to your own dividing head, use the spread sheet to determine some solutions and then try them out on your dividing head.

 

[hman]

Randy -

Your spreadsheet inspired me to look at what I could do with my 90:1 rotab and its dividing plates. Best I could come up with was every 44th hole of the 62 hole plate ... 126.818 holes, or an overall error of about 1/2 degree (actually 0.516 deg.). But then I thought of an additional wrinkle ...

(1) I calculated what fraction of a degree one of the 62 holes on the chosen plate represented ( 360/(90*62) = 0.065 degree)
(2) I divided the overall angular error by the one-hole angle, coming up with 8 holes difference (0.516/0.065 = 8.000)
(3) I divided 127 by 8 and came up with about 16
-so-
By "cheating" on every 16th tooth of the gear, and advancing the indexing pin by 43 holes instead of 44 on every 16th tooth, there would be a maximum error for any tooth of that 0.065 degree, and no buildup of error on the "last" tooth.

Bottom line is that your 72:1 rotab (and your dividing plates) are much better suited to making a 127 tooth gear than is a 90:1 rotab (at least with the plates I have available). If I had (or decided to make) a 48 hole plate, I could duplicate your results.

Nevertheless, if you run into a different situation, where you're not satisfied with your calculated result, you can try playing the same game I demonstrated above.

Thanks again for the inspiration!

 

[randyc]

...By "cheating" on every 16th tooth of the gear, and advancing the indexing pin by 43 holes instead of 44 on every 16th tooth, there would be a maximum error for any tooth of that 0.065 degree, and no buildup of error on the "last" tooth...

John,

That's a GREAT idea - after your original inspiration you devised a simple, precise execution method ! I LOVE the fact that the spacing error is not completely cumulative, nice work

(Another trick is to allow fractional indexing - for example setting the sector arms to index 14-1/2 holes to get more precision than either 14 holes or 15 holes would provide.)

Thanks for posting your idea, it's excellent and I hope that anyone that reads this thread "gets it" !

 

[hman]

John,

That's a GREAT idea - after your original inspiration you devised a simple, precise execution method ! I LOVE the fact that the spacing error is not completely cumulative, nice work

(Another trick is to allow fractional indexing - for example setting the sector arms to index 14-1/2 holes to get more precision than either 14 holes or 15 holes would provide.)

Thanks for posting your idea, it's excellent and I hope that anyone that reads this thread "gets it" !

Thanks for your kind words.
There's a slight downside, of course ... having to keep track of which tooth you're on! It's no longer a matter of mindless repetition.

 

[Smithdoor]

Try compound indexing less than 10 sec better than CNC indexing
Dave

 

[randyc]

Thanks for your kind words.
There's a slight downside, of course ... having to keep track of which tooth you're on! It's no longer a matter of mindless repetition.

Of course, but a simple check list takes care of that.

As home machinists, we're unlikely to object to mindless repetition when the end result is easily within sight - production quantities are unlikely, right ? If I have need to use your excellent method - and that may well happen - I won't be concerned at all about keeping track of the indexing hole by memory.

 

[randyc]

Try compound indexing less than 10 sec better than CNC indexing
Dave

You're right. But some of us (me included) are limited by our equipment. My personal setup is a rotary table with dividing head plates so there is no provision for compound indexing. Good suggestion -

 

[Smithdoor]

This not hard to make

You're right. But some of us (me included) are limited by our equipment. My personal setup is a rotary table with dividing head plates so there is no provision for compound indexing. Good suggestion -

This is not hard to make a setup for compound indexing I am worked on a set of drawings using standard BS-0 plates.

Dave

 

[randyc]

We'll be interested in seeing them, Dave, are you going to post the drawings ?

 

[Smithdoor]

Sorry
The drawings have copy write
But will be on Kindle on Mar 23

Dave

 

[Smithdoor]

Now can be download at amazon.com
http://www.amazon.com/Machinists-Gu...=1427153054&sr=8-5&keywords=machinist's+guide

 

[chips&more]

A typical 40:1 dividing head with 3 hole plates does a good job of dividing up to I think 60 divisions. Above 60 divisions, it’s a hit or miss on getting the correct number of increments. I solved the problem by making my dividing head CNC. It didn’t take that long to convert, it’s accurate, fast at indexing and I don’t need a spread sheet (sorry guys). And no more counting this and that. And the infamous, don’t count the hole you are in (Boy, I can remember a BIG screw up because of that error!). And, and, and all that turning! No more, I just push a button now, life is good…Good Luck, Dave.

 

[randyc]

Well, that's an option that many (most ?) of us don't have Nevertheless glad that you found a more elegant solution !

 

[chips&more]

Well, that's an option that many (most ?) of us don't have Nevertheless glad that you found a more elegant solution !

No sure if it’s elegant, but thank you. I grew up seeing machines with handles and levers. Now, most machines have displays and buttons. Times have changed and I like some changes. The internet can provide a wealth of information and procurement of components. It’s all there and as HM can be put together very easily, with little cost, joining the button world and yield a huge reward. No more counting! I can’t tell you how many times I tripped up on counting, it was so monotonous!...Good Luck, Dave.

 

[Dinuka Shehan]

Please send your spreadsheet to me