Titivating a Wire EDM Machine

[awemawson]

Well I went ahead and tapped the holes. Horrible in soft aluminium.

On assembly amusingly it became obvious that one of the 'inner' holes (ie nearest the table) actually isn't accessable in any position of the table to put a screw in  (See picture 'Bottom-Left') Anyway I went ahead using some stock 6 mm pan head screws packed with washers so that they don't project below.

It basically works - I did have one nasty moment when moving the table about and it caught - one screw was sticking down ever so slightly - had that been on automatic programmed movement it would have torn the cover apart I suspect.

Here are some pictures showing the limits of travel to illustrate how the cover is pulled about over the base of the machine

[awemawson]

Now the cover rectangular hole is larger than the bit of the machine that pushes it about so there is deliberate 'lost motion' (or back lash if you like) - to illustrate this I've centred the table then show the cover able to slide a fair way back and forth - (same looseness is there side to side) :

[awemawson]

Now I don't think that this method of joining the two halves is a satisfactory design. Currently there is a single joggle allowing the original half cover to rise over my new one. It must have been the top element evidenced by the counter sunk holes in it.

However I believe that a 'double joggle' so that the top cover is offset 4 mm and the lower one by 2 mm allowing a 2 mm stainless steel strip to take the threads would be a far better design.

I'll see how she goes 'as is' as forming joggles accurately isn't something I've done before.

Crap-O-Cad drawings depicted below

[awemawson]

So what do you do if you've never joggled, haven't got a joggler, and want to joggle 

Well you do a bit of googling  then set to and make one - obvious really 

I need to joggle about 8" wide in 2 mm aluminium plate and have a 2x material (ie 4 mm offset) on the joggle. Finding a 22" length of hot rolled bar I halved it, then milled suitable steps in it to form the offset. Really the corners should be rounded, especially if the bent component is stressed, but this isn't and I didn't 

Then I drilled and tapped for a pair of 12 mm studs, so the joggler could be clamped in place while the (rather large) part is manoeuvred into place on the press.

[awemawson]

Then we need to try it out. First a small offcut in the vice, and check the offset is correct

[awemawson]

OK that was painless - now we need to do a full width test - actually this bit is a bit wider than the cover bits that I need to joggle.

Too big for the vice I think so lets use the workshop press

[awemawson]

That was about 25 tons - probably a bit enthusiastic as the mill scale marks came through on the aluminium 

Anyway it seemed to work ok

[awemawson]

So now I need to pluck up courage to do the actual bits on the cover, but first I need to convince myself that I can afford to lose the length from the cover that the joggle will gobble up - need to take some measurements

[awemawson]

Tests & measurements completed here are the results.

Total loss of length having joggled a piece of 2 mm aluminium = 1.78 mm
Extra free movement available = about 5 or 6 mm - phew !

The length loss at 1.78 mm is far less than I'd expected - odd things happen when you bend sheet metal !

The cover slides in the  (say ) Y axis and is limited by its down turned flange coming against the 'kerb' on the machine base, the cover needs to be able to move for the full axis travel plus a bit of leeway - that is where the 5 or 6 mm comes from and to increase it would mean less overlap of the joggles - the joining holes are already drilled but one cover could easily have elongated slots if things are tight. I can imagine that the cover arrangement is not perfectly square so leeway may need to be increased.

Doing this exercise reminds me that one kerb is still missing - bar stock not yet arrived !

[awemawson]

Unusual to have a metal delivery on a Saturday, but I'm glad to say the aluminium flat to make the missing 'kerb' arrived this morning in a torrential downpour.

Simple matter of cutting to length and accurately drilling and countersinking six 4 mm holes to match the exiting tapped holes in the machine - one of those 'easy' jobs that sometimes go horribly wrong, but today the gods were smiling down on me and it all fitted perfectly 

[awemawson]

Well that was a very short lived feeling of satisfaction 

Investigating the cover joggle I moved the table to the sound of graunchings and this was what I found

[awemawson]

The new kerb has been totally mangled in pretty much exactly the same way as the original was when I got it. But what is causing it 

Access is very difficult in the small gap between the upper stationary surface of the machine, and the lower moving surface of the table, but by using mirrors and a flexy light I found that one half of a shipping clamp was screwed to the table 

This is a short bit of angle iron, one web of which has been intent on removing that kerb when anyone fits it. Ah well at least now I know, and I can remake another kerb, but first I need to find a way to get at the shipping clamp to take it off. A very long allen key is called for. That angle web must have just cleared the machine upper surface.

.... off to order some more aluminium 

[Pete W.]

Hi there, Andrew,

I've been following this thread with interest.  I'm sorry to read that you've landed on 'a snake's head'!          

Ref.: long Allen key - if we were neighbours, I'd offer my set of Allen stubs that fit a 1/2" socket set handle.  They're that old that they're probably Imperial but then the same is maybe true of your machine?

I'm sure that, somewhere in my bookcase, I've got a Davall Gears catalogue in which they say they use a CNC wire electrode erosion machine to make the moulds for their plastic gears, including correcting for the shrinkage as the plastic solidifies!  That's some precise correction in a 64 DP gear!!!!!!!!!!!!!!! 

[awemawson]

Pete thanks.

In actuality it wasn't too bad getting the shipping clamp off as the cap screws were finger tight! Casting my mind back to when I got the machine in 2011 I'm having difficulty remembering who shifted and placed it. I think the chap who delivered it had been briefed on the clamps as the previous owner had fitted them and the shipper removed them, but it's entirely my fault for not checking. I remember putting the other ones in a plastic bag inside the machine for future use, so I'll add this one to that !

I have successfully cut gears with this machine, and also internal splines in gears. I generate them in a dedicated utility and save as a .dxf, then import it into FeatureCAM to generate the G Code as FeatureCAM amazingly 'knows' this machine and generates appropriate code. For instance it allows for keeping a 'tab' on the gear to keep it fixed to the stock, and then generates a stop, allowing me to place a magnet to hold it in place, then I can restart the machine for it to finish the cut. Remember that the cutting force is negligible, its the weight of the part that would make it sag without the magnet.

When I've re-calibrated it after its 'pulling apart' (at least the wire verticality will need re-setting) I'll see if I can cut a gear in something pretty hard such as tool steel, just as a test piece.

I've always intended to cut plain carbide indexable inserts into things like Acme thread profiles just for the fun of it but never got round to it.

Your comment about gear moulds and shrinkage: The machine has an inbuilt function allowing the part described in G code in memory to be scaled over a huge range, both increasing and decreasing, and to several decimal places. So shrinkage allowance would be dead easy.

(Incidentally this is a 1984 machine and everything is metric)

[awemawson]

Having made the joggler, that works splendidly, I've decided to jump another way  I wasn't happy with the extra height created, as it would leave barely 2 mm clearance above to to the machine table.

So on the suggestion of a member of the MIG Welding Forum, I've been experimenting pressing nuts into the thickness of the 2 mm aluminium sheet to make a re-inforced thread, and it seems to work really well

[awemawson]

On the actual cover I will use half nuts pressed from below to leave a ledge for strength - that was a full nut that started as 3 mm thick but got squashed to 2.3 mm > Needless to say I had to re-thread it !

I will need some M4 x 4 mm csk screws that I cannot find commercially so I shortened some 6 mm ones using a springy clamp made from two bits of 1 mm stainless and an angle grinder !

[Manxmodder]

Hi Andrew, this is a very interesting project. Something I have wondered about EDM machines is how much wire do they use during cutting and how expensive is replacement wire?......OZ.

[awemawson]

Well Oz, I got this machine in 2011 and I'm still on the original 8 kg roll

But in all honesty that's no real judge as  it's not been used much. I've bought various reels as they turn up on eBay but now I have quite a stock so am rather discerning on what I pay now - for instance one sold on ebay for £23 the other day, but with postage at £15 £38 was more than I was prepared to pay. I think it is not much more than that from a dealer.

***(later edit ... I'm wrong it's £115 plus VAT for 8 kgs  and it has 2400 metres on it )***

There is a heck of a lot of 0.25 mm brass wire on an 8 kg spool - I've also got several 16 kg ones 

[awemawson]

At long last I think I've sorted the sliding cover fixing.

The pressed in nuts seem to work a treat, and I ended up pressing in 'half nuts' which in M4 measured 2.09 mm thick - pretty well ideal for the 2 mm sheet that I'm working with. It was a bit of a fiddle balancing the cover on the press but I got there in the end.

With my truncated screws there is no projecting below the panel - so hopefully jamming will be a thing of the past.

[awemawson]

Replacement 10 x 8 mm aluminium bar arrived today so at last I've re-made that mangled kerb, and re-fitted the sliding covers. They seems to slide nicely with no bits catching this time (fingers crossed) - I've been up to the axis limits in all directions.

It's taken a while for two reasons - firstly deciding on the cover joining method and experimenting with it, and secondly waiting for the bar stock and stainless M4 half nuts to arrive. Covers will have to come back off in a few weeks for painting, but I'll treat that as a separate job as several panels need painting.

Meanwhile I've been refreshing my memory uploading and downloading code to the machine from a PC. As I have three CNC machines and the PC only has one serial port I use a switch box to switch the RS232 source - simple enough - but when I came to set up my DNC software (NClite V7) it insists on each CNC machine having a different COM port number  I could expand the number of COM ports using a USB adaptor but don't really want the timing issues that that may introduce. Awaiting response to my query from India where NClite is written 

[stvy]

Andrew,

Am enjoying this thread..... I have to admit I really like these re-build threads with these "above garden shed sized" machines.

One thing that this thread got me wondering is if this machine is the perfect tool to make a block of blank multifix tool holders..... Do you have a project in mind for it?

Steve

[awemawson]

Steve,

There is no definite object in mind. Today I've been doodling various gear profiles, just as a test piece. I'm loathe to start using it with water until I've finished one or two cosmetic changes. Several years ago I bent up replacements for several of the panels that were missing. Having now sorted that sliding cover it's time the tin work got a coat of paint ! I ordered up a RAL colour chart, and now it's arrived went to trying to match the grey it's mainly painted, but various bits are different greys  I brought in the heavy guns and got the wife to decide (RAL 7044 silk grey) and placed a paint order today. It all needs etch primer as most is galvanised sheet with some aluminium.

I've been spending my time trying to sort out my 'DNC' issues - DNC refers to the coupling of the machine to a PC via (usually) RS232.

Firstly physically: up until now I've have cables trailing over the floor - I've installed conduit uprights to take the cable up into the gods and thence wired back to my desk avoiding walk ways.

Secondly software wise. I have three CNC machines and was hoping to use a single serial port on the PC via a switch box. However the DNC software, which holds configurations for each machine (which have widely different configs) insists that each have a separate COM port number.

Rather unwillingly, as the software vendors are not answering queries, I've bought a 'USB to four Serial Port' adaptor 'Hydra Cable'. Installing it proved a nightmare with ports ending up with silly numbers. I solved that (hidden previously installed USB things needed purging) only to find that it's operation was inconsistent to say the least, with ports sometimes working and other times not.

http://www.ebay.co.uk/itm/330820909340?_trksid=p2059210.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

The 'Hydra Cable' is on it's way back to the supplier and a (hopefully) more professional version ordered:

http://www.ebay.co.uk/itm/361060354571?_trksid=p2059210.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Time only will tell 

Back to your question - Yes I'm sure that the profile of a Multifix holder could be cut relatively easily - It would be a very slow way of doing it - presumably they only need to be EN8 or some such decent engineering steel that could be cut on a CNC mill. Now if you needed it in tool steel in the hardened state it would be a good choice

[AdeV]

Multiple COM ports always were a nightmare on PCs. COM1-4 were typically available in the BIOS, but COM1 & 3, and 2& 4 used to share interrupts, so realistically you could only reliably use 2 at a time anyway...

It _used_ to be possible to buy add-on cards with additional COM ports attached (with their own interrupts & memory addresses), but try find one that goes in a modern PCIe slot. Assuming you've even got a computer with a PCIe slot these days - most laptops certainly don't have room for such things.

Q: Wouldn't you be better off running 3 salvaged laptops (=cheap on fleabay), and a KVM splitter? That's what I'd do. If you want a KVM splitter, I've got an 8-port PS/2/VGA one I no longer use going begging. Good quality Adder one too.

[awemawson]

Ade,

I'd initially intended to have a small form factor Compac (as I have a pile of them!) at each CNC machine, and whiz files around by Ethernet only using RS232 for the final link to the machine - but it's a space issue at the machines, so I decided to concentrate everything on one desk top (another HP / Compac SFF) actually at my desk.

We'll see how the replacement USB -> 4 x serial works 

[AdeV]

What expansion port(s) do you have in those SFF machines? If they'll take PCI cards, you can fit one (or two) of these: http://www.ebay.co.uk/itm/261083309079

And then you've got 4-6 serial ports (including the 2 on the machine)....