Stepper

[Mark CzarneckiParticipant @markczarnecki87382]

I have this stepper to go in an EDM head:
Dia. 2 3/4″ Long 1 3/4″ 48 steps 18v.
Its maximum unloaded speed is 176 rpm; above this it stops and shudders.
Is this normal or do I have an error someplace?

Thanks & Regards,Mark

[Mark CzarneckiParticipant @markczarnecki87382]

[John HaineParticipant @johnhaine32865]

That’s normal I think – they have a resonance and won’t go above a certain speed.

[Billy MillsParticipant @billymills]

Hi,

Sounds an unusual stepper, 18V suggests high inductance windings which would be slow with constant voltage drive. 48 steps is unusual these days too. Cannot say if 176 is sensible since more information is needed e.g. the motor type & drive.   Is this a new build or an existing machine?

 

Most  common steppers are 200 steps and around 5V voltage but it’s the current rating that is of more interest. At high stepping rates the rate of change of current is limited by the winding inductance and the applied voltage, you hit a steeply declining torque  when stepping quickly.This can then cause the motor to miss steps and “run rough”.

 

The way around this is to use a much higher voltage but to limit the current to the rated value by modulating the on/off time of the drive- that’s pulse width modulation PWM the speed improvement is roughly the overdrive voltage ratio- however it is a bit more complicated than that.

 

A second area of improvement is to increase the apparent number of steps from full stepping. Half stepping doubles the steps per rev by powering both coils between normal steps to produce a mid position. Microstepping allows a much larger number of steps by having many incremental steps of current to position the motor between full steps.  ( although microstepping is not a strictly accurate angular interpolation)  8 usteps gives 1600 steps/rev on a 200 step motor  ( 1600 steps is easy and a tad up on 48 steps/rev.

 

So a modern microstepping drive with PWM will drive the stepper much faster than simple whole step voltage drive and will not show the “resonance effect” where the motor fails to step at the driven rate. It also allows much finer control of rotation so that less gearing would be needed in an EDM application.

 

In an EDM application I can only assume that you are concerned with the  advance and backing out rate  since you would need a vast amount of power and cooling to  spark out the metal that fast!  But perhaps you have a  gear train slowing the backing out, If so switching to a low voltage stepper with a microstepping drive could produce a substantial improvement and  get rid of some of the gearbox.

 

  Arc Euro Trade supply a good range of steppers and some inexpensive microstepping drives, their website also will lead you to the “Jones on Steppers” link, one of the best intros into stepper motors.

 

regards, Alan.

 

 

[Mark CzarneckiParticipant @markczarnecki87382]

Hi, Alan:

 

 The motor plate says 14v,  27.5 ohms/coil. It was bought some ten years ago, and it

probably was surplus anyway, so who knows how obsolete the thing is!!.

  You’re quite right, the 176 rpm will be used solely for manual fast up/down; as is it

will give me some 45mm/min. Not supersonic, but I can live with it.

The stepper is driven by a 555, through an up/down counter to a TIP41; this is full 18v during the whole step (unipolar).  As planned, I’ll have a demultiplication giving me a 0.0065 mm/step. If this is too fast at the actual slowest motor speed of 40 rpm, I can always slow down the 555 further.

   Are the 0.1 uF caps across the motor coils necessary?

   Having no experience with EDM, my plan for operation would be to have the head

lower at the slowest rate, and then increase speed until it starts backing out often.

Is this the correct way?

   Re modern PWM drives and motors, I’m in Venezuela and getting material here

can be quite a hassle, so I tend to make do with whats available ( read scrap box )

Eg, I found a small (MXL) belt in my recycled collection, the locals had no pulleys

that size, so I had to make them. And so on…

 Thanks for your ideas and the link.

   Regards, Mark

[Billy MillsParticipant @billymills]

Hi Mark,

Best to run EDM  closed loop, monitor the gap voltage with two comparitors eg 339’s, when gap volts  too low -retreat, when gap volts too high – advance. If you set up the thresholds at about  2/3 and 1/3 then you should get  sensible stability.

 

It ain’t optimal but gets you into the ball park. There are a vast array of variables in EDM  so to get ANY cutting you have to stop reading and start sparking!

 

A simple set up is to pulse the counter when either comparitor says too much, the U/D line is driven by  the retreat comparitor to reverse back a step or two. The basic aim is to get that gap set at the distance that maintains a steady rate of removal i.e. an almost constant current into the workpiece.  This means that the advance rate equals removal rate.

The bit that some articles don’t mention  is the importance of good smooth movement without jerks or judder  to get nice clean sides and a good finish,  a sloppy  slide will work for a “proof of concept”, you need a nice slideway to control the gap between the electrodes- not just the position of the moving electrode!

 

As I am not looking at your circuit can only assume that the 0.1uF caps are across the stepper coils?  When your TIP41’s turn off the stored energy in the inductance of the coils has to go somewhere, this results in a massive damped sinusoid which could easily peak above the rated collector voltage of the TIP41 ie back EMF. The normal way of dumping this energy is to use a power zener rated about twice the supply voltage. This clips the peak  to a safe level  without  “shorting” the coil with the normal reverse wired diode . If you are a total cheapskate then you can use a bunch of 1N4002’s to connect the collectors to just one power zener around 30-36V wired to ground.That will lift your max step rate a little.

 

I am assuming that your set up is : a fullwave rectified supply of around 100-200V with a reservoir cap of around 100-330uF then a ballast resistor to limit the current to a safe value supplying the gap with a selected capacitor of  say 1 to 100uF across the gap.

 

 

ONLY USE a transformer isolated supply!!!!!  , keep the wiring from the gap capacitor and the gap  short and flat i.e. minimise the inductance. This will help control the severe radio interference that you will produce.

 

Your dielectric fluid will get hot and filled with particles, pro EDM speeds up the rate by flushing the gap and filtering the fluid. 

 

All for now, best regards, Alan

 

[Mark CzarneckiParticipant @markczarnecki87382]

Hi, Alan:

        Right on, my sensing circuit is just as you describe, it even uses an LM339.

I’m aware of the importance of an “unwobbly” tool slide; I’m making two 3/8″

columns some 70mm apart, with a central lead screw moving the tool holder.

 If I can get it not to jam, it will be a straight liner.

 Spark power is:  70v (supplied by two Sony Betamax Xformers in series;

luckily I was able to use all the secondary windings, so all the copper is in use!)

smoothed by 2000uF, through 15 Ohms 200W to the head, with a switchable C

up to 850 uF. Also there is a mosfet that shuts off charging when the tool shorts.

     Fluid will be circulated by a small pump, so my tools will have to be hollow. I’m

planning to use kerosene ( with an extinguisher handy!)

Will post pics when build is complete.

OT Gloat:  Just received a solid state Oscope: two channels,15 screens storage,

60Mhz, DMM function, all the digital bells and whistles,battery powered. Me

mucho happy.

Best Regards, Mark

 

[Billy MillsParticipant @billymills]

Hi Mark,

I would guess then that the two columns will hold the fixed electrode (job) at the btm. I have followed the same kind of tack but use two wide spaced leadscrews so that there is no risk of  non-parallel movement.  I used  studding, there is  no accuracy requirement-just smooth monotonic motion between the two electrodes. I insulate the top electrode.

 

The drilling of the tool and job is very sensible, why spend hours sparking metal that can be drilled away in seconds?  You also gain from the flushing. Kerosene is popular but so is water! . The drawback with water is that it is more reactive so it needs more care in some ways than kerosene.  Would smother a fire, not launch it!

 

The capacitor discharge concept was the first approach, it works well. The discharge melts a tiny bit off the job that then needs flushing out, the discharge time is of the order of 10’s of uS for small particles. So electrically it’s  a relaxation oscillator (like a cap and neon bulb,)   the C charges until sparkover, an ion channel forms , a tiny ball is detatched, the voltage falls too low to support the increased channel lenth and the V increases till the next channel forms. 

 

So you get a very rough sawtooth across the C. A big C has a lot more 1/2CV squared so the channel current is higher and more energy is available for detaching bigger lumps off the job- hence a rougher finish.  Now the peak current is High but the average low because once the channel is off the cap has to charge to the breakdown level.

 

 

This is one reason why cap discharge has been replaced by a pulse method, the supply is directly switched to the gap at well above breakdown voltage, the channel ionises quickly, the ball melts off and the voltage then removed. The switch is the simple bit, an oscillator sets the repetition rate, a monostable drives the gate of a bank of mosfets. This then gives  V, F & t as more variables to play with!  I am still learning about cap discharge.

 

Bear in mind that the peak current is very high, IF the inductance in the wiring is  not minimised then you end up making a spark transmitter and ringing will then time the sparking. An obvious way of minimising these issues is to parallel smaller caps and  wire them to the gap individually , feed through mini-ballast R’s.  Low Z caps are also good news, my fave is Rubycon ZL. You only need about 1uF for fine work at your voltage.

 

 Glad the you are set up to take shots of the V and I ( symbols choice show age!) against T, think that you will find a new way of randomly producing numbers but the average instant V and I is the best way of understanding what is actually happening in the gap.

Current sensing resistors can be a problem I either parallel surface mount 1 ohm r’s on a PC or use a length of insulated nichrome folded in half and twisted, inductance is a big issue here. Even Metal Film R’s are spiral cut into inductors

 

I have been ‘scoping for a long time, started using digital ten years back but still use Tek 7000 series a lot of the time. For one off events the digitals are great but if working with repeating signals I use a CRT job. Aliasing can get you into real capitol F trouble when working with high voltages, had a narrow escape once,   however for EDM work digital is just what you need, certain that you will enjoy your new tool.

 

By the way, I did spend some time worrying about backlash in the screw drive as this can mess up the controller- i.e. retreat gets delayed as teeth move over to eventually reverse the final gear then you have to wait for the controller to close the gap. Ended up with belt drive, minimising the stages to one and using microstepping, the microstepping is not accurate beyond 4 or so steps as who knows the transfer params of the stepper? certainly not the drive. However it does not matter- all we want is monotonic.

 

One last thought , with a M6 1mm thread, 5:1 gear and a 200step stepper that is 1 micron per step. Microstepping at  2000 steps gives 0.1micron/step . So you could use a PIC to count U/D pulses to indicate how far and how fast  things were going, that would be useful.  By the way, I spark mini punch tools, what is your aim?

 

Best regards,

Alan  

 

 

[Author]

Posts