Spark erosion machines .

I built (bodged?) a simple device about 3 years ago with a lot of help from the members of another forum. I had no success using the eroder current to operate the solenoid so rigged it using a microswitch to provide the oscillation with a separate circuit for the electrode - works a treat - took the broken tap out of a 5" gauge Hunslet cylinder pdq.

 

The photobucket album I put up during the things evolution is here though it makes me cringe a little now.

I was having trouble getting on the site earlier this evening, so I wandered off to Google, and found this site www.modelengineering.com/3756%252 This has a number of interesting bits on it, including a EDM machine in its simplest form, the initial design is manually opperated, but it can be motorised, must go back and have another look at it, and some of the other bits. Ian S C

Ian, that link doesn't work for me, is there any chance you can spell it out without it being a link?

I did ask before but my post has vanished, maybe because it contained a link to another website?

 

Cheers, Andy

Posted by MICHAEL WILLIAMS on 12/03/2011 10:56:01:

 

Actually Eric Rumbo has - in MEW 117. And it would seem that the author ran into the same problem as Andy P, because he modified his with very simple electronics to do the same as Andy's - in other words separated the solenoid current from the spark, and made the solenoid bounce against a spring. That was quite ingenious. There is absolutely no need for a microprocessor for any of these designs - that would be total overkill, and since it would need interfacing it would probably end up more complicated than the timer-driven triggered-solenoid design anyway.

 

In terms of doing work though, the snag with the simple designs that rely on solenoids with significant mechanical mass to time them is that the spark rate is determined by this, and is woefully low. And that's the main reason that the more sophisticated designs exist, using stepper motors to determine the spark gap size. I can though envisage a modification to Eric's design that would in part get over this, whilst still retaining the essential simplicity. He uses a single spark per bounce, and the whole assembly has a repetition rate limited essentially by the system mass, so it only sparks at about 3 per second. In principle I can't see any reason for why you shouldn't be able to burst-fire the sparks so that you get a lot more per bounce. All you have to do is provide a much faster spark pulse rate, and instead of driving the solenoid directly from this, you divide it (simple to achieve) to achieve a similar bounce rate to the original design.

 

Admittedly this is a thought experiment at present, but it may well be worth investigating.

 

 

Try www.modelengineeringwebsite.com I think I'v got it right this time. It takes me a while to get into it as I'm on dail up.  Ian S C

Could you use the cone of a loudspeaker to oscillate the electrode, then use a simple 555 circuit to drive the speaker ?

Posted by Robert Humphrey on 28/03/2011 18:56:49:

 

Not in the existing design, no, but read on...

 

It's the moving mass of the arm and electrode that will realistically limit the speed the device can operate at. The sled needs to be large enough to enable reasonable accuracy and guidance, otherwise the device simply won't be able to negotiate things like the inner core of taps successfully. And I think that to get enough power to move that mass, you'd need rather more electromagnetic coupling than you would ever get into a loudspeaker coil without blowing it to bits - the cones the coils are intended to drive are of a very significantly lower mass!

This isn't an issue with the servo-driven ones, because the movement is far more limited, and not oscillatory as such.

 

But...

 

If you were going to build a device using something like a speaker drive system (with its suspension) that would work an EDM, then you might have some success if you were a bit sneaky about it. The idea would be that you remove the need to move the entire sled except for basic alignment, and just have a much lower-mass oscillating part at the end of it. Although if you did that, then the temptation would be not to have it bouncing about against a spring, but to use variable DC to position the electrode holder where you wanted it, in order to maintain the discharge. This gets rid of the servo motor control mechanism completely, and still leaves the electronics relatively uncomplicated. And it would increase the cutting rate considerably.

 

I will think about it some more...

I am currently building a spark erosion machine of the solenoid type based on this design:

 

 

The erosion rate of the unit, as published, is not high because the capacitor charges through the solenoid which has significant resistance and this limits the oscilation rate of the armature.

 

I have considered using a 555 timer to contol the oscilation rate and then supplying the tool with current from an independent supply. The problem with this approach is that the the pulse supply to the solenoid is critical. If it is set at too high a repition rate then the capacitor does not reach full charge. In addition the pulse duration is critical if it is too little then the armature is lifted and drops back before the capacitor is charged. If it is too long then the tool spends too long in contact with the workpiece and a lot of current is flowing but no sparking occurs.

 

I have adopted a different approach in that I sense the voltage on the tool. As the capacitor charges up the voltage on the tool rises. When it reaches a certain value then the solenoid turns off, dropping the tool and causing sparking. The trigger voltage can be varied by means of a potentiometer. Thus you can choose to operate at a slower speed with a high amount of capacitor charge or a faster speed with a lower capacitor charge. At some speed the current though the tool will be maximum giving the fastest erosion rate.

 

This is rather a simplistic explanation because there are other factors which influence the repitition rate. As erosion occurs the tool has to travel a greater distance for each oscilation so the frequency changes.

 

The circuit is very simple and I am currently refining it. When the unit is completed I will publish it on my website.

 

Mike

Posted by Michael Cox 1 on 28/03/2011 21:14:04:

 

That was why Eric Rumbo altered his prototype, and why Michael suggested that a different inductor/solenoid might be the way forward.

 

Eric did the same thing, only he used a 556 (dual 555), using the second section as a monostable. His only has two repetition rates, although it would be easy to make it variable. He introduced the spring because he found that his electrode was spending too much time in a shorted condition, and this was caning the power supply somewhat. I would be very interested to hear what happens with yours, though.

 

The simple SMEE design sparks rather faster, and they monitor the switching by the very simple expedient of running a LED in parallel with the electrode. What happens with this is that because the mark to space ratio varies according to how high or low the electrode is, the average illumination of the LED will too. The idea with that one is that you manually adjust the position so that it glows dimly, rather than brightly or not at all. Crude, but it seems to work! There are published details of this design here.

 

Incidentally, I can't make your link work at all... but I did find it with a search - here.

Just my $0.02 - your mileage may vary..

 

Any solenoid driven EDM machine I have seen has problems with burning accurate to size, and seems slower than other machines. The movement of the electrode when accelerated rapidly by a solenoid and spring is unpredictable and unless the equipment is heavily built the electrode will move all over the place. If you are just burning out a tap this inaccuracy is fine, but if you are trying to make an accurate hex or square hole for example, it is no good.

 

I have built and run successfully the Ben Fleming design EDM machine. It uses a simple resistor capacitor (RC) circuit for burning with a DC gearhead motor and leadscrew drive for the electrode holder. The electrode movement is controlled by a comparator which monitors the gap voltage and backs up the electrode if any shorting starts. The circuit does use 2 ICs but they are NOT microcontrollers and stepper motors are not used. These factors make Ben's EDM easy to build electrically and mechanically and like I said it works well and is fairly fast burning. I did make some mods due to some component tolerance problems with my set of scrounged / surplus electrical parts but many other people have reported successful builds of the machine around the world as is to Ben's book description. He supports questions and problem solving from builders via a Yahoo forum online. His EDM does not have a solenoid system or steppers - because you do not need them for simple reliable EDM.

 

JD

Hi Jeff,

Thanks for the comments. Try this link:

http://www.stockportsme.co.uk/Membersprojects/Spark_eroder.html

I have added a page to my website entitled " A spark erosion apparatus" that I hope is of interest to this discussion, see:

Mike

I've just joined this forum having spotted this thread whilst searching for more info on spark eroders. I've made one which fits on my BCA mill which consists of a small comercially available 6 volt solenoid and uses the standard resistor/capacitor discharge circuit. I've made this oscillate naturaly, without the solenoid, because of the different strike and quench voltages of the arc provided the electrode is in very close proximity to the workpiece. However the arc is unstable and you can get a contact weld. The solenoid helps to break this weld but can over respond and withdraw the electrode much more than the few microns needed to quench the arc. This mode of operation, which operates like a door bell, is inefficient but simple.

 

However I've used this design but change a number of things to get virtually continuous arcing. Firsty solenoids have a very non-linear position/force characteristic which means that once there is sufficient current to make the plunger move it will rapidly accelerate to the closed position as the magnetic path shortens. This non-linearity can be counteracted by using spring which applies the inverse function. Now when this arrangemnt is used the solenoid plunger moves a much small distance, but sufficient to break the arc, and works more like a voice coil positioning servo. The action can also be improved by introducing viscous damping using grease.

Another change was to shunt the solenoid with resistors so that the solenoid only carries a proportion of the total average arc current. In my case the solenoid of 12 ohm resistance is shunted by a 6 ohm 40W resistor. The use of a light bulb is a good idea as these have a non-linear resistive characteristic because of the large difference between the hot and cold resistance. This helps divert more current to the solenoid to help break a weld condition. I also use a bench supply where I can set the voltage and the current limit whichgives a more linear charging of the capacitor. One thing to consider is that the capacitor used in the circuit should be one designed to have a low ESR and able to handle a high ripple current.

Although this circuit has a more continuous arc the position of the electrode does need to be adjusted more often to keep the erosion rate n the sweet spot. Again by juggling parameters mine has a sweet spot of about 3mm which is OK since I have the down feed on the BCA. This downfeed could be motorised by using the average voltage between the electrode and workpiece to drive a DC motor to wind the electrode down. This means that when the electrode is not arcing the voltage is high and will drive the electrode down and the down rate will reduces as arcing begins.

My arrangement arcs continuously with just a slight vibration of the solenoid.

Clive

Hi Clive,

I have made some more modifications to my spark eroder and have fitted a motor drive to maintain a constant stroke on the oscillating spindle. My machine also operates at high frequency (circa 50 Hz) and the stroke of the spindle is very small, it quivers rather than reciprocates. Further details are available here:

http:mikesworkshop.weebly.com/a-spark-erosion-apparatus-mk3.html

Mike

The link should be

Mike

Hello Michael!

 

When I try to open the page in the shown URL I receive a message saying SITE NOT PUBLISHED

 

 

Best regards

Dias Costa

Hi Dias,

Sorry for the inconvenience but I missed out a hyphen in the link. Try this:

Mike

Hi Mike

 

I read your article on your Mk2 and like you I gained my inspiraton to try this from Derek Lynas's articlebut the link to the Mk3 didn't seem to work.

 

I'm looking to make an EDM that can sink a hole 0.2mm diameter into a broken balance staff from a carriage clock platform escapement. To this end my first attempt, which I've only made in the last couple of days, is quite small and light. My first trial yesterday was with the solenoid working in what I call the relaxation mode. However by restraining the movement of the electrode with my finger I found I could get it to arc in a more continuous mode with only a slight vibration. I believe this natural oscillation of the arcing could be due to the dynamics of the arc and fluid whereby when the arc is struck the pressure generated by the fluid rapidly vaporising causes the electrode to move and break the arc. This is only a theory and may only be true for a light moving system. I see that on your arrangement you've added a weight which essentialy reduces the sensitivity of the solenoid but I believe it adds inertia and the force is still linear so does nothing to correct the almost square law non-linear force/position characteristic of the solenoid. Since I can move my eroder head up and down I decided to explored the change of oscillation characteristics with soleniod plunger position and then juggled with the shunt resistor to get more continuous arcing. The solenoid I have used has a cone shaped plunger and matching female pole which produces a more linear force/position characteristic than a simple cylindrical plunger and flat pole piece. I can send you a low resolution video, with sound, of my eroder in operation if you are interested. The sound is more important as you can't actually see anything happening visually. This afternoons trial was to sink a 2mm diameter solid copper electrode into a tungsten carbide milling cutter. I took 10 minutes to sink a 1.2mm deep hole at an average current of 0.8 amps with a gap voltage of 23 volts. I've yet to connect the scope up to see what the gap voltage trace is like. To get a finer surface I reduced the voltage but at the lower current I have to change the shunt resistor. However I have noticed that at the lower voltage the average current can be as high as when operating at higher volts. At the moment my gap cap is only 100uF and I will be trying different values to see what effect it has. My main concern is the rather high rate of electrode erosion which was some 3 mm when sinking the 1.2mm deep hole. At the moment the electrode is negative with respect to the workpiece and I'll try a positive electrode. So many parameters to juggle with.

Clive

Hi Clive,

I hope the latest link to my Mk3 spark eroder now works for you. It works when I test it.

Your post is most interesting and it raises many new questions for me. The operation of the unit is complex and there are so many variables involved.

To get the unit operating at good efficiency seems to depend on the following:

1. Maximise the frequency. Low frequencies mean that the spindle spends very litttle time near the workpiece and therefore much timeis not being used effectively.

2. Minimise the "short" time when the electrode is actually in contact with the workpiece. During the "short" time no erosion is occuring.

3. Maximise the charge on the capacitor. Even at the highest frequencies that I have achieved the capacitor is always charge to near its maximum capacity as shown by my oscilloscope. This is true at almost any setting of my electrode voltage detector. This indicates that the electrode always spends sufficient time away from the workpiece to fully charge the capacitor.

Your ideas regarding spark shock induced oscillation are interesting. If this is true it should be possible to get continuous oscillation even without the use of the solenoid. I shall investigate this further.

Kind regards

Mike