Societe Genevoise make/made a range of jig borers for which this is possibly an attachment. Try lathes.co.uk for more detail.

This is worth a look:

https://www.nielsmachines.com/en/sip-rotary-table-index-kit.html

MichaelG.

.

Edit: __ Just to put the build-quality into perspective:

https://drive.google.com/file/d/0Bxr59DtuJda_NkREYkpSTXdkUGM/view?resourcekey=0-Pa4O7WxM81Rie7_DyeNchA

The answer might be in the link provided by Michael.   The pictures of the unit being sold by neilsmachines look identical to Brian’s photo, and  neilsmachines say it’s for these SIP dividing tables:

Sip PD-3 Circular Dividing Table ø350mm

Sip PI-4 Tilting Rotary Table ø300

Sip PI-5 Tilting Rotary Table ø450

Sip PD-4 Circular Dividing Table ø450mm

Sip PD-5 Circular Dividing Table ø600mm

Sip PD-6A Circular Dividing Table ø800mm

Brian’s set is missing the settings list and one of the 6 bolts next to the T driver.

The missing list is a good example of the extra goodness one gets with an expensive indexer.   Not a bit of cheap paper stuck inside the box, instead a bombproof enamelled plate is provided.   Unfortunately, this wonderful feature was no protection against the Slack Alice who forgot to put the plate back in the box.   (Good news: the photos of this double-sided plate on neilsmachines.com should allow almost all the values to be recovered.)

This is the neilsmachines box:

Brian’s example:

Dave

I occasionally worked on a SIP jig borer which was supplied with a large rotary table and this attachment, I do recall the table being used but not this attachment, nice to have but seldom used, lovely properly engineered machines serviced and calibrated every year. When the place shut down the two machines either went east or were scrapped, makes me weep thinking about it.

Tony

Even though SIP kit is as good as it gets,  let’s not lose sight of the dubious utility of this particular accessory.   It’s a mechanical indexer, a device that helps the operator keep track of the number of turns, and part turns, he needs to spin the handle in order to turn a rotary table by an angle.    Gears, drilling ‘n’ holes around a PCD, cutting arcs, and anything else that requires a job to be moved through an angle accurately.  Before electronics, these devices were vital!

These days, there’s huge advantage in replacing the mechanical indexer, which isn’t accurate in itself, with a stepper motor and microcontroller.  No need to understand complicated mechanical instructions, fit the right index wheel, and then count turns whilst cranking a handle.  Instead, the microcontroller is simply told the angle or number of divisions needed, it does the maths, and then turns the rotary table correctly, never making a mistake!

Dave

The link gives figures that mean the angular error of the stepper motor output shaft is going to be ±0.009° (non-accumulative) from demanded position based on 200 steps and 10 micro-steps per step. If you have a table with a 90:1 drive ratio this will give an error of the table position of 0.0001°. A professional setup might include a 10:1 harmonic drive between the motor and the table to achieve this without micro-stepping or gearbox backlash. Add in closed loop on the motor as well. Does not seem too bad to me.

Martin C

On 6 July 2024 at 17:49 Martin Connelly Said:

The link gives figures that mean the angular error of the stepper motor output shaft […] Does not seem too bad to me.

Martin C

I didn’t say it was “too bad” Martin … but I do note that you have fallen into the trap of ‘claiming’ the 90:1 ratio in your estimate … which is of course equally valid for the indexing plates, and therefore cancels-out.

No, I don’t know how good the SIP plates are, but I am very sure that they are better than +/- 5%

MichaelG.

Well Michael, the positional error of ±0.009° of the motor output shaft at a radius of 50mm is equivalent to a positional error in the holes on a dividing plate of ±0.00785mm. Are they that well positioned? If you are at a radius of 100mm, which is a big dividing plate then the error would be doubled to ±0.0157mm. The stepper motor error is ±5% of the 0.18° step due to 10x micro-stepping a 200 step motor. So what do you mean by saying you are very sure the holes in the SIP plates are better than ±5%, ±5% of what? The final output error after going through a 90:1 gear ratio is very important to anyone who wants to know what the result will be at the point where the machining takes place.

Since modern machine shops use CNC for this sort of positioning are you suggesting they are not working with a suitably accurate system?

Martin C

[ early morning update ]

Although it does seem intuitively obvious that a maker of Jig Borers would produce very accurate index plates … I completely failed, last night, to find any substantiating evidence for my presumptuous  assertion regarding the SIP items.

My guess is that each of their holes will have been positioned to something better than than +/- 20 seconds of arc … but without evidence any further discussion would be futile.

MichaelG.

.

.

Edit: __ This is the best explanation of Stepper Motor accuracy that I have found :

https://www.linengineering.com/news/methods-for-increasing-accuracy-in-stepper-motors

On 6 July 2024 at 15:43 Michael Gilligan Said:

[ just wondering ]

Have you seen angular accuracy of a typical stepper motor, Dave ?

MichaelG.

.

Edit: __ This is the first statement that came readily to hand:

https://www.geckodrive.com/support/accuracy-and-resolution/#:~:text=A%20step%20motor%20is%20a,a%200.18%2Ddegree%20error%20range.

I plead not guilty.  I didn’t claim that a stepper was more accurate than a mechanical indexer, only that they’re a more convenient way od  getting the same result.

These days, there’s huge advantage in replacing the mechanical indexer, which isn’t accurate in itself, with a stepper motor and microcontroller. No need to understand complicated mechanical instructions, fit the right index wheel, and then count turns whilst cranking a handle. Instead, the microcontroller is simply told the angle or number of divisions needed, it does the maths, and then turns the rotary table correctly, never making a mistake!

The bit about the microcontroller never making a mistake is important in this context, because no matter how well-made, mechanical indexers are very prone to operator error!

In a rotary table application, stepper inaccuracies are emphatically reduced by the table’s worm drive, at least 40:1 with 90:1 being common.    It is the build quality of the worm drive and rotary table that mainly determine the table’s positional accuracy, not the indexer.  Be interesting to compare an inexpensive HV-6 clone, with a SIP;   I’d expect a SIP table in good condition to be 5 to 10x more accurate, with the very best perhaps managing 15x.     Also interesting to compare how robust they are internally!   I remember reading a sad tale of woe from a chap who had invested in an expensive rotary table after his HV failed during a hard slog hacking out the spokes of his model traction engine.  Guess what, the expensive table failed in exactly the same way!   Problem was that the expensive table was more accurate and precise, and the assumption it would also be more robust was wrong!

Back to steppers, whilst true that ±1.8° is the potential error at 200 steps,  most of the time the actual error is 0°.  And, unless the motor is overloaded, the occasional plus errors are cancelled by the occasional minus errors.  ±1.8° is the worst case, not typical.   In practice, steppers behave much better than that.    More subtly, the angular error is also effectively reduced by micro-stepping, and I’d expect a rotary table to be driven at 3200 steps per rotation, not 200.   Explained in Michael’s link https://www.linengineering.com/news/methods-for-increasing-accuracy-in-stepper-motors though I had to read it three times!

Nealb makes a good point about servos though.   Though steppers are an inexpensive way of driving a rotary table, where the worm almost guarantees the motor won’t be overloaded,  steppers are less accurate in heavier applications.   In that case, much better to fit a servo motor:  these report their actual position, allowing the controller to make adjustments when necessary.

Dave

Is there a simple formula or algorithm for handling the number of steps when it isn’t a round number to avoid cumulative error. Emphasis on simple for dummies not a pointer to the PhD theses of professor Bigbrain. I think it must be similar to phase lock loop swallow counters that I used to programme 30 year ago but have rather lost the ability.

Normally I just run a spreadsheet if I haven’t got the right hole plate and can see where to adjust but following a different thread a month back I am looking at a stepper design.