Adding An Encoder To A Dividing Head?

[Tecnico]

There's a current discussion about using a dividing head to cut a gear with an oddball number of teeth (127) that can't be done with a dividing head unless you have a differential gear train.

It occurred to me that many members have DROs as part of their shop equipment and that if I'm not mistaken they can interface rotary encoders. TouchDRO's Yuriy talks about interfacing quadrature O/P rotary encoders on his site. UPDATE: I just found a posting where Yuriy says he has mounted an encoder to his DH with a 5:1 belt drive so he has 4000 pulses per revolution. Sounds interesting.

OK, I can handle the drive train adapter etc. but I run out of knowledge when it comes to interfacing the encoder to get the correct signal to the DRO, anybody have thoughts and/or advice?

D

 

[Susquatch]

OK, I can handle the drive train adapter etc. but I run out of knowledge when it comes to interfacing the encoder to get the correct signal to the DRO, anybody have thoughts and/or advice?

See if you can find the datasheet on the encoder chip. It will probably be one of the standard ones.

 

[Upnorth]

Do you really need to use an encoder? I think the easiest way to accomplish this would be to drive a stepper motor. As much as I don't like Mach 3 software that is what I would use to do this. Just drive it as a rotary axis. It will give you a readout of the angle on your computer. You will need a computer, preferably an old one. Also you will need a break out board to preserve your sanity over wiring up DB connectors. With Mach 3 the software engineering is already done for you. It just has to be set up to do what you want..

 

[Tecnico]

Do you really need to use an encoder? I think the easiest way to accomplish this would be to drive a stepper motor. As much as I don't like Mach 3 software that is what I would use to do this. Just drive it as a rotary axis. It will give you a readout of the angle on your computer. You will need a computer, preferably an old one. Also you will need a break out board to preserve your sanity over wiring up DB connectors. With Mach 3 the software engineering is already done for you. It just has to be set up to do what you want..

I know absolutely nothing about Mach 3 and steppers so it isn’t anywhere on my radar.

I have computers but what other hardware would be needed? Power supply, stepper interface board, stepper of some description, feedback encoder, software? Programming? Other?


: D

 

[Tecnico]

See if you can find the datasheet on the encoder chip. It will probably be one of the standard ones.

I don’t have any chip in mind, I’m starting from scratch on this one and really, without the vocabulary to know where to start.

Is this as simple as connecting the encoder output (assuming 5V square pulses) to the DRO input? What protocol is the DRO looking for & how does this get translated from the encoder?

D

 

[trials20]

Check out

RotaryController - Rotary Controller

Rotary Controller What is it? It's a device that allows easy, error free indexing using a modified rotary table, dividing head or similar Written by a software engineer with over 40 years experience, the controller has a sophisticated interrupt driven motion engine featuring smooth linear...

www.worldofward.com

He supplies stepper to dividing head retrofits, was going to try one but found a 4 axis mill instead.

 

[Susquatch]

I don’t have any chip in mind, I’m starting from scratch on this one and really, without the vocabulary to know where to start.

Is this as simple as connecting the encoder output (assuming 5V square pulses) to the DRO input? What protocol is the DRO looking for & how does this get translated from the encoder?

D

The thread below is a good read. I started out mechanical, switched to digital and then returned to mechanical. Project is on hold right now pending some good stock. No rush.

I might add digital later since I plan a 4axis dro. It's also a good read, if for no other reason than the learning along the way. Also sort of what you want to do.

Thread 'Compound needs a better degree dial' https://canadianhobbymetalworkers.com/threads/compound-needs-a-better-degree-dial.7852/

 

[Tecnico]

The thread below is a good read. I started out mechanical, switched to digital and then returned to mechanical. Project is on hold right now pending some good stock. No rush.

I might add digital later since I plan a 4axis dro. It's also a good read, if for no other reason than the learning along the way. Also sort of what you want to do.

Thread 'Compound needs a better degree dial' https://canadianhobbymetalworkers.com/threads/compound-needs-a-better-degree-dial.7852/

Lol!

At some point in it I mused that that thread could lead to interesting things for a DH. Forgot about that.

D

 

[jcdammeyer]

I've used my ELS to run my small rotary table. I set the jog distance to move a specific amount which correlates to the amount of rotation. I was going to add this as a feature. But Lester Caine in the UK distributed my ELS in the UK and also the DivisionMaster which did the same thing as a dedicated unit. Lester went to jail for child molestation so his web site is closed. Tony Jeffree who designed the Division Master is no longer producing it. But all is not lost. Apparently Steve Ward is making something similar.

RotaryController - Rotary Controller

Rotary Controller What is it? It's a device that allows easy, error free indexing using a modified rotary table, dividing head or similar Written by a software engineer with over 40 years experience, the controller has a sophisticated interrupt driven motion engine featuring smooth linear...

www.worldofward.com

 

[Bandit]

Done a bit of reading in machinery's handbook, seems nearly any number of teeth could be cut, but goes from "simple dividing to complex dividing" and fairly complex gear trains driving the head and advancing the crank and reversing direction of sector? arms while using 2 different rows of holes in the plates. Your birth date and horoscopes may also be involved in the complex complex stuff.

 

[Johnwa]

it would be a fairly simple standalone project. This video shows how to interface a linear scale. The rotary encoder should have the same signals.

 

[Susquatch]

Done a bit of reading in machinery's handbook, seems nearly any number of teeth could be cut, but goes from "simple dividing to complex dividing" and fairly complex gear trains driving the head and advancing the crank and reversing direction of sector? arms while using 2 different rows of holes in the plates. Your birth date and horoscopes may also be involved in the complex complex stuff.

Any division can also be easily done manually without using the sector gear system. Just do a spreadsheet and advance one at a time. It's a pain and subject to user error but totally doable.

Others seem to disagree, but I believe this is also the most accurate way to do it.

 

[Tecnico]

Any division can also be easily done manually without using the sector gear system. Just do a spreadsheet and advance one at a time. It's a pain and subject to user error but totally doable.

Others seem to disagree, but I believe this is also the most accurate way to do it.

When you say spreadsheet do you mean calculate the correct number of (full/partial) turns of the hand crank to arrive at your required angular position?

If you had a protractor on your hand crank that seems doable although tedious.

D

 

[Susquatch]

When you say spreadsheet do you mean calculate the correct number of (full/partial) turns of the hand crank to arrive at your required angular position?

No, I mean calculate the list of angles using a spread sheet. Then use the crank to turn the table to those angles. Just read the angles off the table index and forget the sector disks. They are only a convenience.

 

[Tecnico]

No, I mean calculate the list of angles using a spread sheet. Then use the crank to turn the table to those angles. Just read the angles of the table index and forget the sector disks. They are only a convenience.

Maybe I’m not understanding or maybe you’re thinking about a rotary table that has a scale of degrees & minutes on the hand crank?

I’m not sure if I see how you can accurately lay out something like a 127 tooth gear where the the angle per tooth is 2.83 degrees by just using the degree wheel on the DH spindle.

I could imagine that I could read to maybe half a degree by eying the pointer & scale but if I had a protractor scale on the crank and could read to .5 deg. then the error at the gear as cut would be divided by the DH gear ratio of 40:1

D

 

[Susquatch]

Maybe I’m not understanding or maybe you’re thinking about a rotary table that has a scale of degrees & minutes on the hand crank?

I’m not sure if I see how you can accurately lay out something like a 127 tooth gear where the the angle per tooth is 2.83 degrees by just using the degree wheel on the DH spindle.

I could imagine that I could read to maybe half a degree by eying the pointer & scale but if I had a protractor scale on the crank and could read to .5 deg. then the error at the gear as cut would be divided by the DH gear ratio of 40:1

D

I've discussed this here before but not everyone agrees. Nonetheless I remain stubbornly convinced that the scale on the table itself plus the scale on the crank is more accurate than a degree wheel and sectors.

I won't go into the nitty gritty here, but suffice to say that slop in the gear train more than eliminates any perceived advantage a user thinks they get with that handle and disc system.

Not all rotary tables or indexers are the same.

Mine has a scale next to the table and an additional scale plus vernier on the handle. One can easily set the degrees on the table scale and minutes seconds on the scale next to the handle.

The spreadsheet can turn 2.83 degrees into
2 degrees, 49.8 minutes,
or 2 deg, 49 min, 48 sec.

Smaller tables might not have individual minutes let alone seconds. But whatever minimum they do have will be a smaller and more precise refinement of position than whatever one thinks they are getting with a dividing wheel.

I'm not knocking dividing wheels. They have their place. Especially for eliminating errors as you cut 127 teeth one at a time.

Also, it's not hard for me to imagine a rotary table that has no degree scale on it.

But there is no way that I believe that an inferred angulur readout can be as precise as an absolute angular readout.

 

[Tecnico]

I've discussed this here before but not everyone agrees. Nonetheless I remain stubbornly convinced that the scale on the table itself plus the scale on the crank is more accurate than a degree wheel and sectors.

I won't go into the nitty gritty here, but suffice to say that slop in the gear train more than eliminates any perceived advantage a user thinks they get with that handle and disc system.

Not all rotary tables or indexers are the same.

Mine has a scale next to the table and an additional scale plus vernier on the handle. One can easily set the degrees on the table scale and minutes seconds on the scale next to the handle.

The spreadsheet can turn 2.83 degrees into
2 degrees, 49.8 minutes,
or 2 deg, 49 min, 48 sec.

Smaller tables might not have individual minutes let alone seconds. But whatever minimum they do have will be a smaller and more precise refinement of position than whatever one thinks they are getting with a dividing wheel.

I'm not knocking dividing wheels. They have their place. Especially for eliminating errors as you cut 127 teeth one at a time.

Also, it's not hard for me to imagine a rotary table that has no degree scale on it.

But there is no way that I believe that an inferred angular readout can be as precise as an absolute angular readout.

OK, we're on slightly different paths. I have a Vertex dividing head that doesn't have the degree scales on the hand crank so I can't do the degree - minute - second adjustment that you can on your rotary table.

I agree that you could do the indexing on your RT that has the hand crank scales although it would be tedious but that's how they did it before modern technology & CNC etc. came along. The "old guys" (starting with da Vinci) were pretty sharp! You will still have to carefully manage your gear backlash.

So, without the input crank scales and with something like the 127T gear angles not being available on the peg discs I have then I started thinking about different ways to get the 127 positions and things like encoders and input scales came up and away we went. Short of buying an RT (on the wish list) that's the path open to me. Well, that or spending $ and letting someone else have the fun of making gears......

The quest for knowledge & elightenment continues!

D

 

[Dabbler]

I think the easiest way to accomplish this would be to drive a stepper motor.

@Johnwa has built one, and it looks *very* nice!

But there are easier ways to crack this nut.

You can actually print out 127 divisions using a laser printer and cut out an 8" disk. Using a pointer and low magnification you can achieve 1/10 degree or better accuracy. All the errors cancel out, so the overall accuracy is still 1/10 of a degree for the whole circle. That is plenty good enough for a metric transposing gear.

https://www.blocklayer.com/circle-dividereng

is one of several web sites to set this up with.

 

[Susquatch]

All the errors cancel out,

What a great point. This is a gear. Not only that but just a metric/imperial transposition gear. If teeth wander around a wee bit it isn't the end of the world. As long as you get 127 teeth on it that all mesh well, it will work just fine.

I have a Vertex dividing head that doesn't have the degree scales on the hand crank so I can't do the degree - minute - second adjustment that you can on your rotary table.

Ah yes, I missed that. Big difference between a 10 inch table and a 4 inch dividing head. My dividing head will do degrees. I don't remember if it will do minutes.

Rounding to the nearest degree probably won't work really great for a 127 tooth gear. A third of a degree would be more than adequate.

One degree on a 4 inch head is 35 thou. That is a bit more than a 32nd of an inch. Getting hard to see well enough.

But I think my 127 gear has an OD around 7" though. At 7" one degree is around a 16th. At that size it's possible to mark off the teeth on the gear with a scribe on thirds of a degree. Prolly good enough for a transposing gear.

 

[jcdammeyer]

I just can't resist...

Degrees,minutes,seconds to decimal degrees converter

Degrees (°), minutes ('), seconds ('') to decimal degrees angle converter and how to convert.

www.rapidtables.com

235° 13' 22"
= 235° + 13'/60 + 22"/3600
= 235.222778°

If we have a rotary table with a 36:1 worm gear on it then 36 turns of the little handle is required for 360 degrees. Errr. That's then each turn is 360/36=10 degrees.

Now let's say we connect a 200 step/rev motor with a 10:1 micro-stepper drive and a 10;1 belt drive. That's 20,000 steps per rev or 10 degrees divided by 20,000 which is therefore a theoretical resolution of 0.0005 degrees. Therefore 446 steps would give us 0.223 degrees. Three decimal place accuracy for the 0.222778 degrees.

That works out to 235° 13' 22.8" or being out by 0.8 seconds. Not bad. Except... other than flex in the system with the toothed belt there's one other issue. Unless it's a closed loop stepper with a 1000 line encoder (quadrature 4000 divisions), contrary to popular belief, a stepper motor is only accurate to 1/2 steps not 10 micro-steps.

But the truth is, even with 400 steps and 10:1 we still have a resolution of 0.0025 degrees which works out to 235° 13' 21" Out by one whole second.

Do we care?