Fitting a electronicleadscrew.eu ELS to my ML7

[Robert Atkinson 2Participant @robertatkinson2]

As mentioned elsewhere on the forum I’m fitting an ELS from http://www.electronicleadscrew.eu to my Myford ML7. The kit, single axis in my case, is only the controller and optionally an encoder. This leaves the power supply, motor, motor driver and all hardware up to the user.
As I’m very comfortable with the electronics I’m starting with the mechanical bits.
Bear with me, this may take some time…

I’ve decided to use minature toothed belts for both the encoder and motor connectons to the lathe. I settled on 6mm wide GT2 (2mm pitch) beltolimex  and pulleys by Olimex and supplied by Digikey. I use Digikey for a lot of my electronic components. The Olimex parts are low cost.
I chose a 60 tooth pulley to go on the end of the spindle coupled to a 30 tooth on the encoder. So the first task is to bore out the pulley to fit in place of the 25T gear on the end of the spindle. The end of of the spindle is 7/8″.

So this is my first ever attempt at boring. Just to add a bit of pressure it is to a size.
First issue is that the pulley is flanged so I can’t just stck it in the chuck. So first tak is to make a collet. I decided on delrin as I had a bit of scrap plate handy. Hacked it into rough hexagon with a hacksaw and stuck a M12 bolt through the middle. also put a hole approximately on what would be the inner diameter. Held it in the 3 jaw by the bolt head and turned the outside to a few millimeters larger than the pulley flange diameter. Took it off, removed the bolt and chucked it in the 3 jaw marking the jaw 1 position. Bored out the center to the OD of the pulley teeth (34mm). Took out of chuck and silt it opposite the remains of the 10mm hole.
The Delrin is flexible enough at the thinned out bit made by the 10mm semicircle to snap over the pully flange. Aligned the mark with jaw 1, chucked it and run-out was acceptable. Drilled out the existing 5mm bore to 1/2″. Then started boring using a HSS tool and brush applied WD40 as lubrication.

Even if I say so myself the finish isn’t too bad for a first attempt:

Next job is to remove the tumbler gear ans other parts of the change gear equipment and make a mount for the encoder.

Robert.

[Speedy Builder5Participant @speedybuilder5]

Hold it on the flanges and take light cuts. If you really want to, cut up 3 bits of soft aluminium and put them between chuck teeth and the pulley flanges.

 

[Speedy Builder5Participant @speedybuilder5]

Have a look at this thread, similar to à myford setup

ELS for BOXFORD AUD

[David SeniorParticipant @davidsenior29320]

I used the 4-jaw to hold by the flanges on mine, but they were steel. Your flexible ‘collet’ is a good idea and should give good concentricity with the teeth.

Dave

[Speedy Builder5Participant @speedybuilder5]

Buy your power supply from sane vendor as the stepper control/motor.
I learnt by buying a PSU intended for lighting, but although the rated power was OK , the duty cycle wasn’t up to it. Also are you going for open loop motor or more expensive (accurate) closed loop??

bob

[NealebParticipant @nealeb]

It sounds as if you are gearing up the encoder from the spindle, which raised a question mark in my mind. As you are going to use this for screwcutting, the control electronics needs to keep track of exact spindle position, not just speed. This is usually done by picking up the index (Z) pulse once per rev from the encoder. However, with the gearing up, there will be two pulses per rev. I’m sure that this is OK if the controller can keep track. i have recently rebuilt the electronics on a cnc lathe which I bought for the mechanical bits only and my controller seems to assume one index pulse per spindle rev.

I also use a linear (transformer, rectifier, capacitor) power supply for the steppers and just use SMPS cheap units to feed the controller, etc. The linear supply can handle the peaky pulse demands of a stepper driver without over-current trips getting in the way, while the stepper and driver don’t care about ripple on the supply that would bother other more sensitive electronics. A driver rated for, say, 24-48V isn’t going to be bothered by a few volts of ripple as long as you stay in that range.

[mgnbukParticipant @mgnbuk]

This is usually done by picking up the index (Z) pulse once per rev from the encoder.

All the industrial turning controls I worked with (mainly Fanuc, but also Siemens, Bosch, Heidenhain, Osai & Fagor) used the Index pulse as well as both counting channels for spindle encoders, Nealeb. All specified 1024 line encoders driven 1:1 with the spindle. But the electronicleadscrew.eu controller appears to just use the A and B counting channels here. The documentation does say ” any ratio can be used, but better closer to 1″.

I think SpeedyBuilder5 is labouring under a bit of a misundestanding WRT the “accuracy” of a closed loop stepper. In this application I don’t believe it will improve the situation for a motor that cannot keep up with the commanded rate  due to the loads imposed upon it.

A closed loop stepper drive will notice that the motor is not keeping up, and by how many pulses, due to the encoder – if the motor loading reduces, these missed steps will be returned so the motor end position ends up correct (no missed steps overall). If the motor cannot keep up and the error continues to increase, when it reaches a value set in the drive parameters the drive will enter an error state and signal this to the controller to halt proceedings. But dynamically during operation in this case, the leadscrew would be getting increasingly out of sync with the spindle, cutting a thread with a pitch error. Stopping the drive during a thread because the error has become greater than the fault threshold will probably not end well for either workpiece or tool !

Nigel B.

[Robert Atkinson 2Participant @robertatkinson2]

Hi,
Happy new year and thanks for the comments. In rough order:

The flanges are aluminium and not really strong enough even for aluminium. As I have at least one other 60T pulley to bore it was worth spending the time to make the delrin collet.
I’ve seen and had commented on the AUD thread.

I have already chosen a stepper motor and driver from my “stock”. It’s slightly unusual, a Vexta 5 phase unit. Power supplies are no problem. Initial testing will be carried out with a metered variable bench supply with current limiting. Final supply will probably be a “soap on a rope” style unit like those used with laptops. This makes safety compliance easy. Again I’ve got plenty of “stock” units.
As I said the electronics are not a problem for me, it’s part of my day job. I’ve designed motor drives from scratch. Most of this was years ago. Due to a unusual project at work I designed a small PIC based board that controlled two steppers and a 50W brushed DC motor last year. This was from scratch with discrete components no dedicated motor drive ICs. I also wrote the code.

These ELS modules do not use an index pulse like a CNC. There is just a quadrature output from the encoder. The leadscrew just follows the spindle rotation geared by a programmable ratio. This works even if the spindle is turned by hand. If the ELS following is turned off during threading (manual mode) then you have to use a thread indicator just like change gears.

While a linear power supply is a very good choice for motor drives, as mentioned above I’ll probably use a fully packaged and approved switchmode supply. This will need additional filtering / energy storage at the power input to the stepper driver.

Robert.

[Robert Atkinson 2Participant @robertatkinson2]

Nigel B.
Your post came in while I was typing. I agree entirely. In a previous job I designed the electronics of linear drive systems (mixture of Rennshaw encoders, Baldor coils and magnets and Rockwell drives). The encoders had a high precsion (less than the 1um encoder resolution) index “mark”. This was used for initial homing. On some applications with long process run times or / and temperature compensation for thermal expansion the axis were re-indexed periodically. If the running position count was significantly out at re-indexing a fault was flagged.
Also closed loop steppers help with minor errors, but as you say, if the size of step * allowable error count is significant compared to the thread pitch you get pitch errors. It just not as noticable as a fully stalled stepper. It all depends what your requirements and expectations are.

Robert.

[Author]

Posts