PDA

View Full Version : Question about dividing plates

ClintonH
03-06-2009, 06:21 PM
I'm going to be building a rotary table or dividing head soon and can't seem to find any info on building my own plates. There is a set with tailstock on sale at Busy Bee right now but they are designed for there 90:1 tables. I would like to build mine as a 40:1, will these still work or do I need to find 40:1 plates?
http://busybeetools.ca/cgi-bin/picture10?NTITEM=B2503
Thanks.

Clinton

SGW
03-06-2009, 06:48 PM
I imagine for nearly all divisions you'd want to do, the 90:1 plates would be perfectly satisfactory for a 40:1 ratio.

It's all about factors. With either ratio, you end up needing to turn the crank 1/2 a turn, 1/3 a turn, 1/4 a turn, or whatever fraction a particular number of divisions may require. For any given number of divisions, you'll need a different number of full turns and partial turns depending on the gear ratio, but in nearly all cases you should be able to get it.

For any given number of divisions, you can figure it out as:

(number of turns required) = (gear ratio)/(number of divisions)

For example, for 32 divisions, it would be 40/32 or 90/32.

I think Marv Klotz may have a handy-dandy program on his website to figure all that out.

JCHannum
03-06-2009, 07:36 PM
Index plates are dumb, and don't care whether they are on a 40:1 or 90:1 ratio index head. The BB plates are the standard 3 plate set used with 40:1 dividing heads with 15 to 49 holes.

ClintonH
03-06-2009, 08:42 PM
Thanks guys, I've never used either before so I have no clue. It just seemed like a good price so nows the time to buy. I'll go pick'em up for whenever i get around to building.:)

Clinton

Paul Alciatore
03-07-2009, 05:00 AM
This is something I had posted before in response to a similar question:

The holes in the plates are all about prime numbers. ...... In order to have an exact setup for a given number of divisions you need to have all of the prime factors of that number in either the circle or the worm. A 40:1 worm has 2, 2, 2, and 5. With the 21 (3 X 7) division circle, you can get any number of divisions that can be expressed by a combination of both sets of primes: 2, 2, 2, 3, 5, and 7. Other circles give other combinations and other numbers of divisions. But if the divisions you want contains a prime that neither the circle nor the worm contains, then you can not do it. At least not exactly.

But there is hope. You can make your own plates on the dividing head itself. This is because the dividing head acts like an accuracy amplifier when drilling a hole circle. The accuracy of the circle you are making is as many times more accurate as the ratio of the worm you are using. Thus, with a 40:1 worm in the head, your work will be 40 times more accurate than the circle you are using. All within the basic accuracy of the head of course.

You can take advantage of this by using a two step procedure to make your own plates. First make an approximate circle with the number of divisions you want. This can be layed out BY HAND or with a CAD print. This circle is only temporary so it can be just printed on paper and taped over one of the regular plates. Now make a blank plate and mount it for drilling. Use the temporary plate to drill the real one. If you have a 40:1 worm in the head, use 40 holes or lines on the temporary plate for each hole. Or whatever the ratio of the worm is, use that number of holes.

This new, shop made plate will be 40 times as accurate as your temporary, paper one so, if the temporary one is accurate to +/-1/4 degree, the permanent one will be within +/-1/160 degree or +/-22.5 seconds. Now use this shop made plate to cut a final plate which will be 40 times as accurate again or about +/-0.56 seconds. I really doubt that your indexer is that good or even anywhere close to it so any further refinement would be a totally wasted effort. This plate will be as good as any commercial plate you can buy.

If the worm ratio on your dividing head or table is even higher like 72:1 or 90:1 then the accuracy multiplier is even higher.

Circlip
03-07-2009, 06:35 AM
Don't matter whether you roll your own or go for shop bought Clinton,a plate with 40 equi-spaced holes on a PCD don't realise what the word "Ratio" means, likewise, if you have 40 equi-spaced holes on a 2" PCD the angle turned between EACH hole is the same for 40 equi-spaced holes on a 4"PCD.

Regards "Spaced out" Ian.

John Stevenson
03-07-2009, 09:03 AM
40 :1 chart here for the normal plates although this chart has 63, 25 and 127 on them which is not part of a standard set.

http://www.stevenson-engineers.co.uk/files/40%20to%201%20new.pdf

.

Norman Atkinson
03-07-2009, 09:58 AM
I have a division plate with one hole in it.

A bit like the 'I've got a dog with no nose' joke but if you know what you are doing, things get better and better- making division plates.

Oops

Norm

Peter.
03-07-2009, 02:11 PM
I wonder with all the cut-price DRO's/digital calipers that are able to be produced nowadays why someone hasn't applied the technology to produce a rotary table with a digital display for a modest price? Is there a limitation to the technology that prevents it?

doctor demo
03-07-2009, 02:59 PM
I wonder with all the cut-price DRO's/digital calipers that are able to be produced nowadays why someone hasn't applied the technology to produce a rotary table with a digital display for a modest price? Is there a limitation to the technology that prevents it?
Thanks Peter, I guess there is no point in going to the patent office....now that You put My idea on the net:D .

Steve

nheng
03-07-2009, 03:10 PM
Hi-res rotary encoders are still quite pricey. You could however, do what some have done in the past, convert rotary motion to linear motion.

You could wrap a steel cable or flat steel strap around the table like an old radio tuning dial, and attach it to a digital caliper. You might need to introduce a some tension in-line with this.

The problems start when you start cranking the table. An 8" table would need a 24" scale for almost full rotation. A 24" digital caliper can be had for around \$100 or less.

Doing it with a PC would be easier since you can introduce the angle factor:

angle (degrees) = (caliper readout) * 180 / (table radius) * PI

So, a caliper reading of 1" on an (exactly) 8" table platter would represent an angle of 14.324 degrees

Den

John Stevenson
03-07-2009, 03:17 PM
BW Electronics in the Uk does such a system

http://www.bwelectronics.co.uk/appnotes/rotary.html#top

It uses a fine stainless wire to pull an encoder round for rotary work ot pulls it out in a straigh line for linear work.

I bought a couple in the early days as they were quite cost effective but unfortunately prices have gone up whilst glass scale units have come down. I notice now he's quoting 0.003" accuarcy when in the early days he used to quote 0.001" ?

.

Peter.
03-07-2009, 03:22 PM
Hmm what about a system that uses an encoder wheel driven off a circular rack similar to how the dial calipers work? That ought to be quite accurate, compact and cheap to produce.

EDIT: whoops - just re-read nheng's post about cost of them :(

nheng
03-07-2009, 03:26 PM
Take a Shooting Start (Canadian) DRO and wrap the scale around the table. It is gear based, as far as I recall.

added after Peter's last - the hi-res encoders I referred to were those used in precise servo based systems for position feedback. High end ones are in the 20,000 count per revolution, maybe more.

nheng
03-07-2009, 03:38 PM
Clinton, Sorry for wandering off a bit here ... but it could be a way of generating precise angle spacing, perhaps on a par with a good x/y DRO.

Peter, There has been lengthy discussion of using PC mice in the past for encoding. I don't recall if it wandered into optical mice but there's an option possibly worth playing with. An interesting basic study of opto mouse movement is here: http://www.contrib.andrew.cmu.edu/~ttrutna/16-264/Vision_Project/

Peter.
03-07-2009, 05:19 PM
Yeah sorry for dragging your thread off-topic ClintonH. It WOULD be pretty cool to produce a rotary table with a digital readout though. Shame that electronics is one of my weakest points, I reckon that a simple controller and a mouse encoder could produce pretty good results, perhaps by gearing the encoder wheel for extra resolution.

ClintonH
03-07-2009, 05:39 PM
No problem on going off topic, it's all learning and that's what I need right now.
So would it be better to try and build my own or are these cheap enough that I should just pick them up and build additional plates latter as I need them? Anyone good with CAD and has some spare time..?;)

Clinton

JCHannum
03-07-2009, 07:11 PM
If you are considering purchasing plates, I would suggest these. They are made by John Stevenson and sold by his wife. They have the standard holes plus 25, 63 & 127 which adds the most commonly needed divisions not covered by the standard plates. He leaves a margin to permit additional divisions at a later date should that prove necessary.

http://www.metoolsonline.com/product_info.php?cPath=27&products_id=55

uncle pete
03-07-2009, 09:34 PM
ClintonH, You can if you want build your own plates, A book called "The Shop Wisdom Of Philip Duclos" sold by our host's here (Village Press) has a very good artical on building your own 40-1 dividing head and gives a few different ways to drill the plates.

A much better book IMO is written by George H. Thomas titled "Workshop Techniques" ISBN 1 85761 106 3 gives a very clever way to drill the plates on the dividing head itself, Altho this is for a 60-1 head. the principle would still be the same. If you buy this book I would also buy George's other book "The Model Engineers Workshop Manual" ISBN 1 85761 000 8, I probably have over 200 books on machining and if I had to choose the two best books, His two books are what I would pick.

But if you value your time the plates John Stevenson sells are the best deal going.

Pete

lazlo
03-08-2009, 12:20 AM
I wonder with all the cut-price DRO's/digital calipers that are able to be produced nowadays why someone hasn't applied the technology to produce a rotary table with a digital display for a modest price? Is there a limitation to the technology that prevents it?Thanks Peter, I guess there is no point in going to the patent office....now that You put My idea on the net:D .

I actually posted pictures of that exact setup here a couple of years back: rotary encoders output common quadrature signals, so I dropped an MT3 center/drawbar doodad I made into my Yuasa rotab, which is threaded 1/2-13. I shaft-clamped a Heidenhain optical rotary encoder on the drawbar, and hooked up the output to a cheap (\$40) Red Lion quadrature display. Instant rotab DRO.

You can see the top of the MT3 drawbar adapter here:

http://i164.photobucket.com/albums/u15/rtgeorge_album/Servo%20Plate/ServoPlate004.jpg

..and this is the Red Lion CUB5 quadrature display/totalizer I use:

lazlo
03-08-2009, 12:28 AM
If you are considering purchasing plates, I would suggest these. They are made by John Stevenson and sold by his wife. They have the standard holes plus 25, 63 & 127 which adds the most commonly needed divisions not covered by the standard plates.

Those are the standard Myford dividing head plates, which are 5.7" in diameter: they're too big for the Ellis and Carrol dividing heads, which use 5" plates.

The ubiquitous Chinese dividing head plates, like the BusyBee plates the OP linked, are 5", and will fit most home-shop sized dividing heads and rotary tables.

Norman Atkinson
03-08-2009, 05:23 AM