[Hobman]

Hi anybody made any CNC sinker orwire EDM from Camtronics plans?
I am looking for his help/suggestions.
I made two EDMs from Robert Langlois & Peter Rawlinson plans.
But these are non CNC machines.

Thanks.
Hobman

[roysol]

You may not get much response since a true full cnc edm machine is more complex than might appear at first blush. Certainly many more obstacles than a mill/lathe/plasma, etc. If you are determined to do it, and think it would be more fun than frustrating, I'll be glad to give you some of the problem areas for consideration. I should also point out you will not achieve the accuarcy or production levels of factory built machines.

Regards,

Roy Solomon

[Mcgyver]

Roy, I'm interested to hear what the challenges are and also what you think are the differences to expect on factory tolerances vs diy?

I'd heard Robert Langlois was at at one point working on a wire edm, don't know if it was cnc, but he's very knowlegable about electronics - be great to get some guidance or at least a heads up on where the problem areas are.

[roysol]

McGyver, The first problem would be driving the servos relative to the gap condition. Modern edm's use adaptive controls to monitor the stability of the burn and speed-up/slow-down/reverse-direction based on the feedback loop. They also use special routines to maintain the integrity of sharp outside corners, or remove sufficent material on inside corners. This includes changing power settings, flush pressure, wire tension, and wire path.
Which naturally leads to the next bits of fun, controlling wire tension and flush pressure(and having enough pressure properly fed in to the kerf). On the older machines, tension was created with a jockey pulley and a weight. This limits the range of adjustment, and precludes changing it on the fly via program control. Modern machines do this with a brake motor that creates tension electronically. Likewise with flush, the older machines use needle valves and volumetric flow gages to set flush pressure. You could need upwards of 12 bar depending on how high a part you want to be able to cut. Most of the new machines cut submerged in addition to the pressure flush. Which would lead to the next bits of fun, but I'll stop here for now to let you digest.
Still want to know more? Be glad to help.

Roy Solomon

[sbrpollock]

I Want To Know More! (There are probably a lot of us who would like to know more about this)

How about a primer on the whole EDM process?

[roysol]

Patrick, I am flattered that you think it sounds like I know what I'm talking about! Based on the number of views for this thread, you may be right. I don't have time for daily chapters, but perhaps a weekly post would be possible.

First some disclaimers. I'm just another guy in the shop, like most of us here. I happen to have learned a lot during my 25 years in the trade. My opinions are mine alone, not my employers, or any machine manufactureres. Everyone is welcome to take them for what they are worth, or ignore them.

The bible for edm is "The EDM Handbook" by E. Bud Guitrau, the true master!
There are also some very good articles and a forum at http://www.mmsonline.com/edm/index.html

Although this forum mostly favors plasma, there are at least a couple of other edm guys lurking, and I would welcome their additions to this thread. I would expect the discussion to be reasoned and respectfull.

That being said, if no one tells me to go to hell, and the thread keeps getting posts and views, I'll start adding more info. Any feedback on specific topics is welcomed.

Roy Solomon

[Mcgyver]

Roy thanks very much for stepping forward on this - I'm very interested to learn. you certainly made your point that there are many complexities here!

for others, here's a few more links that cover the basics.

http://www.jjjtrain.com/vms/other_edm_basics.html
http://www.agieus.com/Academy/

as a diy're, my priorities are accuracy and feasibility, not speed/production rate. Of course my priorities aren't everyone’s, but I'd guess most of those interested in a cnc diy edm are less concerned about production rates or they'd buy a commercial unit.

As some of these complexities are there to achieve combined objectives of speed & accuracy, if a home shop guy didn't care about speed, could some of the complexity be eliminated while maintaining accuracy? i.e. just run it really slowly at a lower voltage?

Roy, can you give some detail of the cutting movement/action....at what frequency does the unit spark, is the motion fairly analogue or is it spark/wait/step?

I suppose at some point it comes down to how accurately you can control the axis? If the best resolution you can get is .001, then that constraint will set the minimum voltage you can use?

With the low feed rates of an edm, I think it would be possible to make a home machine that you could control to a 10th of a thou. this is accuracy beyond normal diy stuff but with linear encoders reading to a few microns (my fav, http://www.newall.com/), high tpi screws and geared down servos, why not? If this was doable, and using a slower, low voltage approach, I'm wondering what class work you think is possible?

[roysol]

McGyver, you are correct that many areas can be simplified if productivity is not the concern. For example, the aforementioned tension method of jockey pulley and counter weight is much simpler than an electronic brake motor, and very effective. Also in that category would be submerged cutting and high pressure flush.

Moving the table accurately is about 1/3 of the final result. You will also need to guide the wire, (more on that later), and control the overburn.

The links you gave are good. I have run Agie’s and they make an excellent machine. The Virtual Machine Shop looks like a potential black hole time drain; I had to tear myself away from a lot of interesting reading. I did find two points in the edm chapters I would take issue with. One is they say copper wire is common; in fact brass is almost universally predominant, although there are some other materials being used in special applications. The other is their recommendation to use kerosene as a dielectric fluid in a sinker edm. This is never done anymore, kerosene is too volatile. Most machines use synthetic oil with a much higher flash point. I sent Ron Smith an e-mail regarding this; it might be nice if they could adopt a “wika” concept for open editing.

Now back to your questions. The state of the art is .0001mm steps, (4 millionths), or 10 millionths inch resolution. The general concepts for linear guides, rotary encoders, stepper motors, etc, is the same as the more conventional machines people are building, just to a higher order of magnitude. Your part will always be some amount less accurate than you can position your table, so this is the first order limitation. You don’t need much speed or thrust, so you can use whatever techniques (gear reduction, etc) that help towards that end.

As to the forward motion, the spark generation happens thousands of times per second, depending on the settings for off time and on time. At the same time, the gap condition (open gap voltage) is monitored and adjusted thousands of times per second. This is an interesting debate ongoing between machines with ballscrews that sense the gap more often, but have to deal with lost mechanical motion and limitations to acceleration, versus machines with linear motors that sample the gap less often, but can respond and accelerate the table quicker. Also toss glass scales into the mix. Some machines use closed loop feedback; others rely on the rotary encoders.

For a Diy system, you could use a forward feed low enough that you would not have to sense or respond to anything. Early machines were like this, and the next step in evolution was to stop the machine in response to a short circuit.

I think the next installment would focus on the wire drive/guide system, and the impact of dielectric fluid (deionized water) delivery/control.

Roy Solomon

[klpauba]

Are you able to mill a PCB with EDM? If not, why not?

[roysol]

Klpauba,
If you mean PCB as in "printed circuit board", the answer is no, because you can only EDM(the term would be "burn" as opposed to "mill") conductive materials. If you happened to mean PCD as in "poly crystalline diamond", the short answer is yes, this is being done by some people.

Roy Solomon

[Mcgyver]

thanks again Roy.

Resolution to 0.0001mm, one ten millionth of a meter is currently beyond my ability – but not by much! One could theoretically approach 1 micron resolution (1/1,000,000m), and cheaply to boot. I say theoretical because things like the rigidity of the machine, linear bearing clearance, etc seriously come into play when you’re talking 1/100,000 of an inch!

How to accomplish this bodacious claim? I think it would be possible with fine tpi lead screw, geared down servers, and a closed loop system with Newall linear encoders. These have 1 micron resolution and can be used submerged and supposedly are tolerant to magnetic forces (is there much emf or other interference generated by edm?). Heck, fine ready rod with a plastic nut and simple anti-backlash would do – the positional accuracy is from the encoders. This idea would have to be linear – rotary encoders or steppers wouldn’t work, at least not with ready rod

http://www.newall.com/LEDs/leds.htm

basing the system on the prepackaged accuracy of the Newalls and using really inexpensive components to drive the motion is a unique opportunity presented by the EDM characteristic that 1) motion slow and 2) involves no cutting force.

Now I don’t really think I’m going to be cutting to 1/100,000”, but that there is a strategy to get this kind of resolution and its only 10x of what is state of the art, gives me encouragement that a high class machine could be built (or at least as far as the first 1/3 goes!)

The VMS is a great project – bunch of guys putting a lot of work into something for the betterment of all. I’m sure Ron would welcome another contributing editor …or maybe we’ll just collect all your posts and send em to him

Mike

[roysol]

Hi McGyver, It sounds like you’ve got motion control covered. I’ll point out that the OEM’s use precision ground ball screws, and a final laser compensation for greater accuracy. The majority of factory built machines are 4/5 axis. This includes U/V for tapering, and a power Z axis. I’m guessing you’ll be satisfied with a 2 axis machine. The machine will create a fair amount of EMI. The Swiss machines come with full enclosures like you would see on a machining center, in this case, it’s just for dampening emissions to the CE(European Union) standard. I would not worry about this from a safety standpoint. You do need to use shielded cables for communications, such as RS-232, or CAT 5 cables for Ethernet. This is only a concern while you are cutting. I do not know what they do to shield data flow to/from the scales and encoders. I imagine you buy shielded cables, or do it yourself to standard cables. You won’t be able to listen to the radio while you’re sparking! :frown:

The next consideration for accuracy is driving and guiding the wire. Starting from the used end, you will be pulling the wire from the supply side spool. This is done with two steel rollers, one fixed, one spring loaded to pinch the wire between the two rollers. One roller is driven by a variable speed motor, the other idles free. The speed is increased to accommodate taller or more accurate parts. After these rollers, the wire can be collected on a take-up spool, dumped into a collection bin, or chopped into small pieces and collected. The new machines all use choppers.
The tension on the wire can be done several ways. The most basic and least consistent is to simply create drag on the supply spool. I’m sure you can imagine several ways this could be done, and the drawback to each. The most accurate method is to use a brake motor and a strain gage sensor to monitor and adjust the tension electronically. The intermediate method is to use a counter weight and jockey pulley arrangement. This requires another set of pinch rollers, a drive motor for these rollers, and a sensor on the jockey pulley to increase/decrease the speed of the take-up side motor to keep the jockey pulley in the center of the travel. This would be similar to a press feed supply loop sensor if you are familiar with that concept.

I have seen pictures of a diy machine with no wire guides, just strung between the pulleys. The pulleys are not very precise, and the distance between them too large for accurate work. Guides can be round, 3 point “V” contact, or open “V” design. The guide material has to be hard. OEM’s use diamond, ruby, sapphire, or carbide. Since your machine would not have U/V motion, the guides need to be mounted so the wire is square to the table, either by design, or with some adjustment. All the machines I've run have U/V travel, so I'm not sure how this is done by OEM's on a 2 axis machine. Anybody else want to chime in on this or anything else so far? Come on you lurkers, get off the back bench! :stickpoke If the guide is conductive, it has to be insulated. In conjunction with the guide, consideration has to be given to getting power to the wire, and ideally, flush to the workpiece co-axial to the wire.

I have attached a couple of jpeg's to help clarify.


Next on the list would be the dielectric fluid requirements, and how to achieve them.

Roy Solomon

[klpauba]

Yeah, I meant Printed Circuit Board.

The copper clad board is conductive, no? I would think that you could burn the traces through the copper coating clear down to the fiberglas board. If it could work, I bet it would be too slow.

Thanks, I appreciate your sharing the knowledge!

[Mcgyver]

Roy, another great post - keep ‘em coming!

you are being kind saying the movements all figured out, but with the lack of cutting forces i think some good stuff could be done. I agree a home shop probably won’t need the U/V motion but might want a rotary axis on the table.

The picture you gave shows the wire going through a round guide, then the electrical contact cylinder, then a guide that looks like its made up 4 balls in forced in contact with the wire. I'm guessing it has to be something this complex because unless it was an interference fit, just running the wire through a round guide would introduce movement or error = clearance between the wire and the guide?

With the example you gave of the wire being held between pulleys, once the wire is in tension between the pulleys, and the pulleys are on good preloaded angular contact bearings, are there additional forces acting on the wire? ie resonance from somewhere, or does the spark causes the wire to move, etc? I guess I'm wondering if, without the U/V requirement, it could work well between pulleys or does that approach doom the machine to second class work from the start.

Can you describe a typical case of wire size, feed rate, tension etc., say to cut a gear in 3/8 thick mild steel. Is the tension just to create shortest distance between two points, or are you actually stretching the wire – i.e., why the need to adjust tension, or is that done at set up for different wire sizes?

Mike

[mxtras]

Excellent back and forth, Mcgyver and roysol.

Mike - sub-micron placement with rotary encoders is not all that difficult as you likely know, but I agree that linears would be more suitable. Temperature compensation should not be an issue - if you expect TRUE sub-micron tolerances, then you are in a climate controlled environment - like it or not.

Roy - keep it coming. I for one appreciate you sharing your obvious expertice...either you are one hell of a sales guy or you do this for a living. I am thinking EDM is your life!

Ich liebe es!

Scott

[roysol]

Hello all, thanks for the feedback.

Scott, I have been in manufacturing most of my adult life. I have never done any sales. I currently am responsible for one wire and one sinker. I have 7 years experience on cnc mills, 5 on cnc lathes, about 10,000 hours on wire edm, and 5,000 hours on sinker edm. Philosophically, edm is not my life anymore than any job would be. But I do love my job!

Mike, I have some pictures I'll post next week of a diy machine that relys on the rollers for guiding. The guide I posted is more complicated than it has to be. I'll post some simpler ones also. The deflecting forces are the current, which creates "wire lag",the actual wire position is behind the gap, and flush pressure(which may not be that high on a diy machine). The tension adjustment is for lower pull during rough cutting, and maximum pull during trimming. Also to allow for different size wires to be used. In all cases you are NOT stretching the wire.

I mostly use .010 or .012 wire, sometimes .008, and I have cut with .004. They make machines currently capable of cutting with .0008" wire(no that's not a typo!).

The machine I run at 4 years old cuts 12 linear inches per hour. In 3/8 steel it would cut 32 inches per hour. The newest machines cut upwards of 35 square per hour, 885 inches linear per hour in 3/8 steel.

Cheers everyone, and have a great weekend!

Roy Solomon

[roysol]

That last line should read 85 inches per hour!

[InspirationTool]

Roy, with Sqin/hr fairly well established, what do EDM's usually ballpark in $/hr?

$50? $100? $200?

Thanks!

-Jeff

[FranH]

Your knowledge of edm is super.

A. Why can't a typical Bridgeport with its motor belts having additional gear reduction from IPM to inches per hour thus fitting in the edm world?

B. Arms to hold upper and lower wire guides to be fabricated to fit the table for xy motion.

C. The task is now to control the burn.

D. Not looking for a 4 finish, not looking for diminensions within thousanths. (.015 is fine)

E. Not looking for 4 axis.

F. Looking to rough a part closer to finish than plasma or waterjet as a prep for machining.

G. Run a large enough wire so breakage is not a problem.

H. Presently machining unfriendly material like nitraloy135M etc. So to rough remove material is the needed result.

I. Who is out there capable to do the electronics, wire feed, amperage, etc. ?

[CJL5585]

Roysol:

What happened to the EDM thread that seemed to be going so well? I would like to hear more about this.

Jerry