ELS for BOXFORD AUD

[Ex contributorParticipant @mgnbuk]

the Nema23 stalls under certain conditions.

Your Dip switch settings for SW5 -8 for the steps per rev suggest 6400 steps per rev according to the stepperonline description for a DM556T drive which yours looks like it may be – difficult to see from the picture & you don’t specify as far as I can see. A description of this given under item 3 on  here where 32 microsteps operation (6400 ppr) is shown as having only around 5% of the rated motor torque compared to full step operation.

Best torque is with full or half steps, so maybe you could try resetting both drive and ELS to run at, say, 400 steps per rev and see what happens ? Also up the current to the drive maximum of 4A, as the motor picture shows the motor is rated ar 4.2A and it will only give the rated stall torque at that current value. I would not expect that you could stall a 3Nm motor by hand – I have a  3Nm Nema 24 stepperonline motor driven with a stepperonline DM556N drive from a 36V no-name switch mode psu set for a low microstepping rate & can’t stop it rotating with a 30-ish mm diameter timing belt pulley fitted.

Nigel B.

[John Haine]

On John Haine Said:

On Robert Atkinson 2 Said:

……
A ELS on a conventional lathe has one disadvantage over a typical CNC system, you can’t ramp the speed up, it has to be syncronous with the spindle which you have no control over.
It seems to me that 840RPM is very fast for screwcutting.

Robert.

P.S. Hoping to fit a electronicleadscreweu single axis to my ML7 over the holidays. All the bits are in.

If that’s the case it is a poor design!  Even Mach3 does better than this.  The trick is to have a 1ppr index pulse and use that to trigger the acceleration from a fixed start position for each pass, returning to the same start after the cut.  You can also have an encoder to give finer control of cutting speed as in more advanced lathe CNC controls.

Maybe I wasn’t clear enough. My comment applies to thread cutting. If you accelerate the the cutter you get a variable pitch thread. You can accelerate before the cutter stsrts cutting but that is not always possible.

Robert.

 

[Robert Atkinson 2Participant @robertatkinson2]

Just seen Nigel B’s post and I agree entirely. Microstepping is not a good idea for this type of application. Set for half step.

Robert.

[John HaineParticipant @johnhaine32865]

Robert, I just don’t see what you are getting at.  An ELS isn’t going to be able to cut full depth at one pass, so for each pass it has to somehow move the tool back to the start.  Unless the spindle reverses (with the tool withdrawn of course) and the L/S turns backwards, you have to withdraw the tool, send it back to the start, feed in again and synchronise with the spindle.  Just like “hand”threading, the tool will start each pass clear of the end of the stock and has some clearance to accelerate before the next cut starts.  Provided there is at least a 1ppr spindle index pulse synchronisation is pretty trivial.  Obviously you can’t accelerate the tool once it’s cutting, though actually small adjustments to feed rate are made in more advanced controllers to allow for spindle speed variations caused by cutting force.

Just for information, the Z leadscrew on my S7 is driven through a 2:1 pulley from a reasonably beefy NEMA23 at 3A coil current, with microstepping set to 8 (i.e. 1600 steps per rev).  In all the time I’ve been running this (14 years) I cannot recall a single time I’ve missed steps other than when I’ve managed to stall the spindle with an over enthusiastic cut.

[Speedy Builder5Participant @speedybuilder5]

Thanks for the advice about pulses per rev, I changed this down to the lowest 400ppr which did not alter things very much, then I changed the Stepper ampage to the highest and started to get better results. So then I progressed to AutoTurning where the tool can be set to take successive cuts up to a shoulder with auto return.  That was Ok until I got up to about 840 rpm and then problems started to show up.

I have a new PSU on order from Stepper on-line (sometime in the New Year).  So I will wait for that and hope it will solve my problem.

Thanks once again for all the advice.  I am in the process of writing up my installation.  Anyone wanting a copy by PDF or Word, send me a mail by the messaging system and I will try and help.

Bob – Ps  Happy Christmas everyone.

[SillyOldDufferModerator @sillyoldduffer]

Just a thought, I haven’t tried it, but it might be worth putting a big capacitor across the power supply output.  One of these perhaps, 10000uF rated 63Vdc.

A PSU designed for a steady load like LEDs is rated to produce ‘n’ amps, and almost certainly includes a current limiting circuit that stops it damaging itself if overloaded.  But a pulse load calling for more than ‘n’ amps may not be an overload.  Having a duty cycle means the pulsed current is OFF for some percentage, thus averaging less than ‘n’ amps, but the PSU doesn’t know that and may be throttling back unnecessarily.

A big capacitor on the output stores electricity as if it were a flywheel. It recharges during the gaps, and can briefly deliver many more amps than the PSU.  This is useful when a pulse starts because the extra power would help prevent stalling.

The suggestion is a guess.   An oscilloscope would show if the existing PSU is sagging because it’s not designed for pulsing stepper motors, but I guess Bob doesn’t own one.  And a capacitor is a bit of a bodge too. A better answer is to do as Bob has done and buy the right kind of PSU!

Dave

[John Haine]

On John Haine Said:

Robert, I just don’t see what you are getting at.  <SNIP>  Provided there is at least a 1ppr spindle index pulse synchronisation is pretty trivial.  <SNIP>.

That’s the point. Most of these ELS systems do not have absolute postion sensing. They do not even have a index pulse. They rely on a fixed ratio between the spindle and leadscrew.  You can’t accerate a fixed ratio. All they have is quadrature encoders giving step and direction. The leadscrew follows the head.you can stop the movement of the lead screw but all the controller does is keep track of the relative angular position of the chuck by saving the count and resetting to the start point everytime that count, plus or minus the number of pulses per revolution occurs. You can stop or start the spindle but if you turn the following off you have to use thread index.
I admit I’ve not done the maths, maybe the cumulative error from acceleration is insignificant but that depends on many variables.

Robert.

[Speedy Builder5Participant @speedybuilder5]

I have been playing whilst awaiting a new PSU.

LH threading with normal chuck rotation is just a matter of selecting the opposite rotation of the Boxford tumbler gears (could also be done in the config of the ELS).  The main problem is allowing a start groove of 5mm for the backlash makeup which is non adjustable.

28tpi at 840 rpm (auto threading) single point tool. The finish could be better if I was using some cutting oil.

[Ex contributorParticipant @mgnbuk]

Have you seem James Clough’s (clough42) motor / drive solutions ELS related video here.

He undertook some dynamic testing of a range of motors and drives & plotted the results – if you don’t want to go the through the entire video, the conclusions & analysis starts around the 30 minute mark.

2 main takeways are that the drive can make a big difference to performance with the same motor, particularly at higher speeds, and that stepper motor performance drops off markedly with increasing speed. A motor that is giving 3.5Nm at 100 rpm is down to 2Nm at 500 rpm, less than 1.25Nm at 1000 rpm and around 0.75Nm at 1500 rpm with the better perfoming, more expensive, digital drive. The analogue drive used was much worse than this throughout the speed range, but very much so at at high speeds – a “dip” in the torque curve at 1000 rpm (resonance effects, maybe ?) showed only 0.4NM here.

Your testing at 8TPI has the leadscrew rotating at the same speed as the spindle, with the motor running twice as fast through the belt reduction. You may just be expecting more than the drive solution chosen an give at the speeds you were running at.

My only experince with steppers through work was taking them off & replacing them with (mainly) DC servomotors. Matchmaker, Acton and Beaver used Posidata stepper motor open loop control systems in the ’70s & Bridgeport had their own “Boss” system. All had the motors driving 0.2″ pitch ballscrews directly and were driven with very basic full step drives that used series resistors to control current (no chopper drives). They were rather pedestrian in operation and could not be pushed too hard or lost steps were the result. We retrofitted Heidenhain closed loop controls with (typically) 3.2NM dc servomotors driving the ballscrews through 2:1 reduction timing belt drives with the feedback encoders mounted on the back of the motors. The same machines operating under closed loop servomotor control were night and day different in terms of machining performance.

Stepperonline also sell closed loop brushless servo solutions the link is to an Ebay listing for a 3.2Nm (1000W) motor/drive kit. A lot spendier than the 3Nm stepper motor and drive, but the kit does include motor power and feedback cables and is 220V AC power input so PSU to buy. Compare the torque curve for that motor/drive combination in the listing to the clough42 results – full rated torque available on a continuous output basis up to 3000 rpm, tapering off up to 5000 rpm. And approx 3 x rated torque available on an intermittent basis. This is a full-on industrial solution to motion control, but you pay a proper industrial price for it !

Robert mentions the effects of acceleration on threading & that is a consideration even with closed loop CNC systems. Fanuc have a section in the appendices of their Operating manuals that details how to calculate the extra starting clearance required at the start of a thread  to allow the tool to reach the correct synchronised feedrate for an accurate pitch thread. It is dependant on the axis gain settings as well as the acceleration settings, spindle speed & thread pitch. I did go through this once & found that the extra acceleration distance required was quite small & reliably smaller than most operators would allow for clearance to the end of the part.

Nigel B.

[Speedy Builder5Participant @speedybuilder5]

I have an update on my installation.

Installed new PSU from Stepper-online which made no difference.

I installed a USB PSU for the controller so that any interference from the stepper PSU was eliminated – no difference.

I selected an option on the controller to show stepper RPM and here is the problem.  The Nema 23, 3Nm has a max RPM of 1000, but a recommended speed of 100 – 500rpm. When I repeated my tests, it was obvious that motor stalling occurred at stepper speeds in excess of 500RPM.

Although the supplier recommends stepper / leadscrew pulley ratios between 1:2 and 1:3 IF the torque is available from the stepper, then by  getting the pulley ratio nearer to 1:1 or less would solve my problem.  This is eluded to by Nigel B (Thanks Nigel).

 

[Speedy Builder5Participant @speedybuilder5]

So, I am thinking of changing the current Stepper/Leadscrew pulleys from 20:40,  to 32:40.  I believe that this would allow me to cut a 1/2″ 16 BSF thread at 563RPM giving a surface cutting speed of 73ft/min and keeping the stepper motor sub 500RPM.

Bob

[Speedy Builder5]

On Speedy Builder5 Said:

So, I am thinking of changing the current Stepper/Leadscrew pulleys from 20:40,  to 32:40.  I believe that this would allow me to cut a 1/2″ 16 BSF thread at 563RPM giving a surface cutting speed of 73ft/min and keeping the stepper motor sub 500RPM.

Bob

Give it a go Bob and report back, I’m always interested in the ELS projects.

Tony

[Speedy Builder5Participant @speedybuilder5]

Tony,

I have now fitted different pulleys for the stepper / leadscrew  32:40 and this has made quite a useful difference.  For example:- 12 tpi at 560 RPM which is quite scary  or 19tpi at 860RPM, even more so!

It almost makes turning threads in wood possible (8TPI at 360RPM).

When threading, there is no need to turn a groove where you want the end of the thread to stop as the system stops exactly in the same place each time and thus turns its own groove as the depth of thread increases.

Bob

[Author]

Posts