Electronic Artisans ELS Article

[Mark CParticipant @markc]

Michael, I agree with that. Ian hits the mark regarding CNC, if you go to the trouble of getting it all working you can't be very far from a CNC system (I think JS more or less said the same thing?). If you want to cut a LOT of threads then there are easier options (or just get a real CNC). If you want to cut the odd thread on a lathe and single point tooling is a challenge then buy some dies etc

Mark (who is still managing to mix his number formats up….)

[Mark CParticipant @markc]

John, I would not contemplate trying to do any of this with anything other than absolute encoders. You need to track the relative position of the tool point on the surface of the work at all times. This is inline with Michael appraisal? The problem then moves to how fast you can move and still keep up – this is where JS came in with very large drives on machine centers. they can move and position at blindingly fast speeds (too fast to see without help) and still start and stop at the exact point on the work (probably within a few microns?)

Mark

[AjohnwParticipant @ajohnw51620]

The existing ELS uses a microswitch as a stop so shoulders and thread length aren't a problem. Less of a problem than on full cnc after a fashion as it can be referenced to the work and no need to know how far it is from the tool.

John

–

[AjohnwParticipant @ajohnw51620]

A KIS approach is needed Mark C and as a result it can only be so good. Lathes with gears aren't that good anyway for various reasons. There will be limitations. By most account Linuxcnc copes anyway as have other cheap machines without absolute positioning.

John

–

[Martin ConnellyParticipant @martinconnelly55370]

Ajohnw said "I'd be surprised if some one coded a system that just said the spindle is running at this speed so cut the thread and forget it."

That is exactly what Mach3 running through a Smooth Stepper does. If you stop the spindle when the thread is being cut the motion continues to the end of the thread before retracting for return to the start. There may be a difference when using other set-ups but since the index pulse is picked up and sent back to Mach3 by the Smooth Stepper I suspect it is the same with the parallel port.

Martin

[Mark CParticipant @markc]

John, KIS (KISS = keep it simple stupid ?) might be what is needed but in reality anything that is really usable is not going to be. You are going to get a glorified variable speed feed from the lead screw – might as well just stick a motor on the end with a control pot and cross your fingers…..If you need the help of electronics then the system is not going to be simple, there is going to be jiggery pokery going on – its just the nature of the thing.

Mark

[John HaineParticipant @johnhaine32865]

I think that's quite correct, Alan Savelio (cncyourmyford) demonstrated this for Mach3 at the Midlands show IIRC, or at least mentioned it, whereas with Linux CNC it behaves as you would hope. But hey, for many purposes it works! From what some people say Mach3 could never cut a thread in a million years, but I know it can 'cos I have seen it.

I'm not sure what the concern about referencing is? Yes, with CNC you have to reference the tool to the work, but it's no sweat. Once you've done that it knows exactly where to stop the feed and withdraw the tool – if the ELS uses a switch as a stop then you have to reference the work to the stop instead! (As well as the tool to the stop.)

I guess there's a possible issue with Mach3 and larger workpieces where the cutting torque will be larger, but you'd run it slower anyway so there's more drive torque available. And with CNC threading you can always take lots of light cuts, the machine is very patient!

[Martin 100Participant @martin100]

Posted by Ajohnw on 28/03/2016 11:45:39:

That chip with a label on it is almost bound to be a rom which suggests another micro. I also thought that these machines did have controlled spindle speeds. The usual DC speed controller works pretty well, even better with an encoder on it.

If it can work with one pulse per rev

 

On the Boxford TCL 160 (and maybe the 125?) it's more than one pulse per rev (32 minus one missing) and yes it is a processor on that board, a Z80 (or equivalent), can't recall if it's the lathe, the Boxford mill or even both that also have another processor on smaller board with an 8051 (or equivalent) for the turret changer on the lathe or the Z axis on the mill.

 

Boxford TCL 160 head assembly showing encoder disc.

 

 

 

Edited By Martin 100 on 28/03/2016 13:43:26

[Lester CaineParticipant @lestercaine30947]

I knew when I started this thread there would be some 'controversy' but on the whole the content has been most productive. I am essentially an electronics engineer and making the various components work is not a problem, but I think this thread does show that as of today there is still not a single all encompassing solution to a CNC lathe?

Where I struggle is with the mechanical side of the problem, and while on the whole these days I manage to get the mill working reasonably productively, It still takes a visit next door when something more complex comes up … or a run over to JS But this thread has started to build on my understanding of the mechanical side …

[John Stevenson 1Participant @johnstevenson1]

Posted by Michael Gilligan on 28/03/2016 12:00:30:

I may be in a very small minority here, but:

For my purposes [being to have full set of 'virtual' ratios available], I would like to see all reference to spindle speed taken out of the equation … i.e. the ELS should work properly when the lathe is being hand-cranked.

This is particularly relevant to short threads, such as found on various optical instruments; but I suggest that it might be the purer starting point in any case.

… I think this demands a high resolution encoder on the headstock spindle.

MichaelG.

.

 

EXACTLY

 

It's about POSITION, not VFD, slip or what ever kludge someone wants to throw into the mix.

 

All this red herring about how to control the speed is unnecessary. It's the same as threading with gears but one gear has 3 broken teeth and the thread jumps. Now everyone is coming up wit ideas how to get round the missing teeth instead of replacing the gear.

 

Speed is irrelevant.

 

Position is KING.

 

Watch this video by the formidable Andy Pugh.

 

**LINK**

Watch the lot if you want but the important bit is 50 seconds in. The bit where he rocks the spindle and the work follows. Note he's not doing a full rev so a one pulse per rev index flag isn't going to do sod all in a case like this.

 

Linux CNC can do this now and has been able to do this for quite a long while. What it isn't able to do is allow a simple setup to do this.

Unfortunately it's down to having a lot of hand holding or adopt the Tormach set of screens which at the moment also need help to set up on a non Tormach machine.

 

Something I have been saying for the whole of this post but it gets hijacked by people who think they know but don't.

 

{EDIT}

 

Thank you Martin for that information about the Z80 chip.

Hardly state of the art these days, so we could do it and now we can't ??

Something seriously wrong.

Edited By John Stevenson on 28/03/2016 13:55:58

[JAParticipant @ja]

I run a large bore imperial Myford lathe fitted with an inverter and variable speed control. After reading the article in MEW 240 I thought I could be cutting Metric threads without having to spend over £200 on a set of Metric conversion gears.

Now I am not so sure. Should I even be considering ELS as a cheaper viable option?

JA

[AjohnwParticipant @ajohnw51620]

The use of a Z80 based processor doesn't surprise me John. Wonderful beast especially working with assembler. I has been recognised as probably being the most optimised tiny chunk of silicon on the planet. At least one manufacturer was thinking of adding to other processor chips in the same package. The PIO chips were painful as the data sheets didn't contain sufficient info to get them working. That caused me a serious headache for several days once when looking at incorporating one in an electric vehicle controller.

The big slot in the encoder will be the index. The cheap way. Quadrature encoders can be used to determine the direction of rotation. Their main purpose in life. I suspect that if I was increasing the pulse count with them that I would look carefully at the waveforms first and maybe only double by using the rising edges and maybe even not do that. In use the phasing doesn't always matter that much other than detecting the direction of rotation.

On Linux cnc I did link to something that looked pretty. If people want to have a go I am sure people on the forum would help but do ask them to be gentle and explain that you have no experience at all. If things don't get to running it as some one would like the best place to ask about that would be the disto's forum. Ubuntu in this case. There are various files for organising that sort of thing and desktop icons can be created. This may involve using a text editor. A weird one may be suggested but stick to one that runs on the desktop – not the console ones. Those can do all sorts of weird and wonderful things but use isn't that obvious on some of them once they are running.

John

–

Edited By Ajohnw on 28/03/2016 14:53:15

Edited By Ajohnw on 28/03/2016 14:54:53

[John Stevenson 1Participant @johnstevenson1]

JA,

You don't need to spend £200 on a metric conversion kit for your Myford when you can do it all with just two extra gears at a fraction of the cost and to the same degree if not better accuracy.

PM me if interested as it's going further off this topic.

[BikepeteParticipant @bikepete]

Posted by Michael Gilligan on 28/03/2016 12:00:30:i.e. the ELS should work properly when the lathe is being hand-cranked.

Here's an example of hand cranking while thread cutting with ELS (looks like cross slide is also stepper controlled, so maybe closer to CNC… but it's still synching with a hand-cranked spindle) . Skip to approx 2:50 to see the hand crank.

Description says "Turning an ACME thread on an old ML7 lathe, by hand. The leadscrew is electronically tied to the spindle using an Arduino Due and stepper motor. There is a 1440 pulse per rev (with Z pulse) digital encoder on the spindle. The code is written so that any pitch can simply be typed in (here it was 3.175mm for an 1/8inch pitch). Imperial and metric threading is simply calculating the required pitch. The lathe backgear cluster has broken teeth, so the motor could not be used to provide a low enough RPM. Motion was done very slowly to ensure the thread cut properly (including repositioning the cutter, which was done automatically). The encoder was salvaged from old medical equipment. All details will be provided at no cost/obligation for anyone interested. CNC rocks."

EDIT – found an even better one:

Edited By Bikepete on 28/03/2016 19:21:05

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Bikepete on 28/03/2016 19:17:04:

Posted by Michael Gilligan on 28/03/2016 12:00:30:i.e. the ELS should work properly when the lathe is being hand-cranked.

Here's an example of hand cranking while thread cutting with ELS (looks like cross slide is also stepper controlled, so maybe closer to CNC… but it's still synching with a hand-cranked spindle) . Skip to approx 2:50 to see the hand crank.

.

BRILLIANT … Thank You

You have restored my faith !!

MichaelG.

[MuzzerParticipant @muzzer]

There's another excellent example of a simple ELS system based on an Arduino and an encoder.

There's also a forum linked to in the thread below the vid but it's in Russian. I'm sure with some perseverance it would be possible to reproduce it. Very nicely done.

It appears to be pretty a simple open loop arrangement where the encoder movement is converted to stepper pulses with a ratio that accounts for the leadscrew pitch and the required thread pitch. It's hardly rocket science – nice to see it done in a businesslike manner.

Murray

[Michael GilliganParticipant @michaelgilligan61133]

Thanks, Murray

We're definitely on the right track.

MichaelG.

.

Edit: via Google Translate

Edited By Michael Gilligan on 28/03/2016 20:44:28

[AjohnwParticipant @ajohnw51620]

The Russian one seems to be using encoders removed from printers with pretty high resolutions.

John

–

[Michael GilliganParticipant @michaelgilligan61133]

1800 lines, processed in x2 mode

3600 pulses per rev.

[Mark CParticipant @markc]

Mmmmmm, resolving down to 6 min's of arc, not exactly single pulse per rev then!

Mark

[AjohnwParticipant @ajohnw51620]

It leaves about 5uSec to do the step update sums not accounting for any jitter at 3000 RPM. Or twice that at 1500rpm. If some one happens to have that perfect Super 7 that can leave a mirror finish even feed rate gets important.

Might be worth remembering that the existing Artisian ELS uses an encoder and from hanging around on the forum not much in the line of complaints that I noticed anyway. It has a fault light too. Seems to be something Mach might not have.

John

–

[John Stevenson 1Participant @johnstevenson1]

Posted by Ajohnw on 28/03/2016 23:48:19:

Might be worth remembering that the existing Artisian ELS uses an encoder and from hanging around on the forum not much in the line of complaints that I noticed anyway. It has a fault light too. Seems to be something Mach might not have.

John

–

I was under the impression they all needed an encoder. ??

Come on John, it's about paying attention and reading posts – not post counts

[AjohnwParticipant @ajohnw51620]

I had noticed something about posts however there has been a lot of talk concerning one pulse encoders started by your um good self and the els uses more. People reading might think otherwise. To JS in case more posts arrive since reading the latest comment.

I mentioned not being keen on it earlier. The main thing in that area was buttons. I would have preferred BCD switches to set the pitch. At the time I looked I might have worked on the code if it was opensource but it didn't seem to be. The saddle stop it uses is clearly a good idea. I could even use it in conjunction with the micrometer stop on my Boxford however the lathe I had at the time didn't have one, no gearbox either.

John

–

Edited By Ajohnw on 29/03/2016 10:43:39

[Ian PParticipant @ianp]

Posted by Ajohnw on 29/03/2016 10:43:25:

I had noticed something about posts however there has been a lot of talk concerning one pulse encoders started by your um good self and the els uses more. People reading might think otherwise. To JS in case more posts arrive since reading the latest comment.

I mentioned not being keen on it earlier. The main thing in that area was buttons. I would have preferred BCD switches to set the pitch. At the time I looked I might have worked on the code if it was opensource but it didn't seem to be. The saddle stop it uses is clearly a good idea. I could even use it in conjunction with the micrometer stop on my Boxford however the lathe I had at the time didn't have one, no gearbox either.

John

–

Edited By Ajohnw on 29/03/2016 10:43:39

John

Are you saying that in general ELS systems always have (and always did have) multiline encoders?

Ian P

[Lester CaineParticipant @lestercaine30947]

Posted by Ajohnw on 29/03/2016 10:43:25:

I had noticed something about posts however there has been a lot of talk concerning one pulse encoders started by your um good self and the els uses more.

The ELS kit uses a single pulse sensor …

[Author]

Posts