Titivating a Wire EDM Machine

[awemawson]

So the first thing to do was cut a male spigot on the polycarbonate rod -

[awemawson]

So having made that, parted it off, and made a second one (I had decided to make two new spools) I had to cut the female thread in the end. This was not only needed as half of the hubs, but also to hold the male threads for their 'second operation' of facing off and boring to size.

I realised that screwing the male into the female, then boring and  turning the female would tighten the thread, so I made sure that the two mating faces were as clean and smooth as possible. Even so they locked up good and proper  To get them apart I ended up knurling the unthreaded bit so that a strap wrench would have half a chance of gripping. I also used a thin parting tool to cut a slot between them in an attempt to relieve the tension. Eventually they came apart. The knurled bit is where the hub will be bonded into the cheeks so won't detract from the finished spool.

To avoid the same issue on the second one I hit on the idea of wrapping a soft copper wire round the junction between the two components, so that when turned on the lathe the wire could be pulled out with pliers to relieve the tension, and it worked quite well.

[awemawson]

So tomorrows job is to start on the cheeks. I'll cut the profile and windows on the CNC mill, then taper the cheeks on the lathe

[awemawson]

Bit of a late start today as I had to fell a tree that was threatening to push a wall over 

Got the spool cheeks drawn up in CAD and into FeatureCAM to produce G code. Made a test cut / protototype from an off cut of Perspex that had been originally the splash guards on this machine. Perspex doesn't machine quite as nicely as polycarbonate but is was just to prove the process.

I wanted to engrave the word 'FANUC' on the reel cheek as per the dxf Phil kindly generated, but it produced more code than the TNC355 can gobble up - I'll run it as a drip feed second operaton - but at least this made a good prototype.

The perspex one may even get pressed into service to repair the original!

[awemawson]

A bit more experimenting this evening, and I've reduced the G code of the logo from about 1500 lines down to 235 

This has let me download it to the TNC355 Heidenhain controller and cut a test engraving, which has come out acceptably. Thanks Phil 

[philf]

A bit more experimenting this evening, and I've reduced the G code of the logo from about 1500 lines down to 235

This has let me download it to the TNC355 Heidenhain controller and cut a test engraving, which has come out acceptably. Thanks Phil

Looking good Andrew.

I'm surprised that it translated into 1500 lines of code. Did it turn all the rads into a series of flats?

About 20 years ago I created some code so that our toolroom could engrave serial numbers using their CNC mills. They would CNC incredibly precise microwave components and then have to engrave them on a benchtop Taylor Hobson pantograph engraver.

This made the engraver redundant and it ended up in my workshop.

Cheers.

Phil.

[awemawson]

Phil, the first attempt was removing all the material bounded by the letter outline, whereas the second was a simple line outline.

[awemawson]

So today I churned out four cheeks on the cnc milling machine

[awemawson]

Each had to have one side (it's inside) tapered to help guide the wire. This was done on the manual lathe

[awemawson]

So then it was a case of a trial assembly, then gluing them together

[awemawson]

So this little mini projectette is almost done.

There is a tapered sleeve that the wire actually winds on that I need to make. It has a parallel bore but a steeply tapered outer diameter so that it can be pushed out of the wound wire easily to recover it for re-use. It is what sets the width of the spool, the screwy bits clamping it against the cheeks. The original has been broken into four pieces and bound together with tape! Two of them to make.

And also the spool is driven by a sliding peg that sits in a slot in the hub - there are four slots on each side of the spools, but only one is actually used. So that's another little task for when the glue has hardened - the makers claim 24 hours.

[awemawson]

So this morning I made the two tapered rings. These go on the hubs of the spools, the idea being, when you want to remove the scrap brass wire, the ring comes off with the wire and as it is tapered will then push out of the wire easily. Remains to be seen if this theory works - the original was always a pain to remove, but it had been covered in masking tape to hold it together. I increased the taper to 8 degrees from the original 5 degrees to help removal, and also made the rings a couple of mm thicker so hopefully they wont suffer the same cracking fate as the original.

I only just had enough of the 180 mm x 70 mm polycarbonate rod left to make them - the chucking bit was really rather short  But it worked 

[awemawson]

This afternoon I've milled the drive slots and assembled the spools and installed one on the machine - seems to work which is just as well after all this faffing about 

So I think, barring accidents that just about wraps up the mechanical side of things.

[awemawson]

Next job is to upgrade the memory. As previously stated this (and other Fanuc 6 controlled machines) uses 'Bubble Memory'. When I got the machine is came equpped with a 16K byte memory which (rather charmingly) they refer to in the manual as 40 metres of paper tape equivalent.

I got my sticky fingers on a 128K memory board shortly after I got the machine and installed it. The process is a bit complicated for two reasons. Firstly the bubble memory contains all the machine parameters including those needed to drive the RS232 interface, and it's that interface through which you need to re-load all the parameters. A bit of a chicken and egg situation! This is circumvented by manually inputting just the bare comms port settings and then re-loading. But before you do this you have to manually load all the 'Defect Loops'. When these memories were made not all the internals were perfect, and any defects are noted on a table on the card.

But now another bubble memory drifted into sight recently (ebay in the USA) - this time a whopping 512K byte or 1280 meters of paper tape  This is the maximum memory size for this machine. It arrived yesterday so I must revise the procedure, take numerous back up copies of parameters, cutting data, and ball screw error corrections, and go through the process again. At least this time all the data is retained in the memory that I remove, as bubble memory is 'non volatile' rather like core store used to be. So in a worst case senario I can re-install the original and start again.

[awemawson]

Well to cut a long story short - no, the replacement Bubble Memory didn't work. I'm sure that the seller sold it in good faith but it fails to clear when told to. Currently in transit back to Knoxville, Tennessee curtisy of the ebay returns system, foc to me, and paypal are arranging a refund.

However I sourced a second one here in the UK, and that worked first time and is in and running 

[DMIOM]

Maybe because there's a ghost trapped in the bubble? - (or is it a version of 'the genie in the lamp'?)

[awemawson]

One of the other minor irritations that's been outstanding is a non working bulb holder. The "+V" axis motion button is a simple momentary 'press to make', that incorporates a bulb holder, the 24v flange bulb illuminating when the V axis has travelled to the 'Reference Position'. The contacts of the holder had disintegrated and none of my rescued spare buttons were suitable.

Compatible button available from RS for £26 but you pay extra for the bezel / label assembly, OR for £8 I can buy ten of them from Hong Kong, however these have a 24v LED built in.

Well as you may guess, I got them from Hong Kong and they arrived by today's post. Pretty tortuous fitting as it had to be done in situ as the control panel doesn't easily remove. However only four soldered connections.

Got it in, made a reasonable job of it, tried it, and the button is illuminated ALL the time 

Fortunately this bit of the machine I have circuit diagrams for, so off I went searching. Turns out that there is a bit of very clever design here. North end of bulb (or LED) goes to +24, south end goes to driver circuit which brings it down to 0v, but there is a 110 ohm resistor also going to 0v. Now I'm assuming here, but I think that it's true. That resistor is intended to keep the bulb filament warm but not glowing. I've seen it before in high reliability equipment  if you don't let the filament go totally cold both the bulb and the driver will live for far longer.

Well of course the 110 ohm resister was easily low enough value to light the LED which is happy at a few milliamps. Cure was simple, unsolder one end of the resistor and sleeve it. Now full functionality restored albeit with a slightly brasher colour to the light when lit !

.... however I'm chuffed to have found that   

[vtsteam]

I somehow missed this thread, Andrew -- another classic!

I remember reading about the fantastic bubble memory in Byte magazine-- early non-volatile memory, and wishing I could afford it. Oh think of the possibilities!!!

Great sleuthing the cause of the lamp issue, and those wire reels are first class.

 

[awemawson]

Thanks Steve for the kind words 

I'm not surprised you missed the thread seeing that you seem to hibernate for months on end and disappear off the face of the earth 

[awemawson]

Well at long last I've finished another time consuming 'sub project' that sprang from this re-build. Documentation

As I had obtained a brand new copy of the Operations Manual re-printed for me by Fanuc in Japan, it made my VERY tatty copy of the Maintenance Manual shout out for attention. The problem is that using the machine one's hands tend to be wet, which makes handling manuals in a way that will allow them to survive isn't easy.

I decided to scan both manuals into searchable PDF files with a view to having them available electronically as well as printing out and laminating key pages. Now I have a very nice HP scanjet scanner - (in fact I have two of them !) but it didn't survive the upgrade from Windows 2000 professional to Windows 7 Professional - drivers not available  So needing to buy another scanner lead me to discovering the PlusTek Opticbook 3800:

http://plustek.com/usa/products/opticbook-series/opticbook-3800/

Not the cheapest, nor the most stylish scanner in the world, but designed specifically to scan books and manuals, so is laid out with the glass very close to the front edge, allowing scanning right into (within 2mm) the fold of a book. Also comes bundled with some very useful scanning software. Even better, a brand new one popped up on ebay at less than half price 

So I set to scanning - a laborious job - not too bad on the new manual but appallingly tedious on the tatty one that is literally falling apart. A bit of an issue cleaning up the scans (covered in another thread) :

http://madmodder.net/index.php/topic,10340.0.html

But having over come that I ended up with two humungous files. One of 219 pages and the other 313 represented as about 35 mByte each as pdfs.

Then I set myself the task of making the index pages dynamic, by inserting 'links' in the pdfs in the appropriate places to jump to the right page in the document. Three pages of index lines in the Maintenance Manual and another seven pages of links in the Operations Manual. Well this morning saw the end of that task thank goodness and I've actually been able to load them up to my iPad to view under iBooks AND the dynamic indexing works - quelle relief!

All I need now is a cheap waterproof iPad or clone for the workshop 

[vtsteam]

The world of Fanuc wire EDM titivators owes you a great debt of gratitude for that, Andrew! 

[awemawson]

He he ..... that would be me and ................... probably nobody else 



.... but if any one in the future needs a copy get in touch

[dsquire]

He he ..... that would be me and ................... probably nobody else 

Andrew

Don't sell yourself short. I appreciate all the time that you have spent researching and repairing your EDM. I admire a man that can take something apart to see what makes it tick, then put it back together so it does tick. I look forward to more and when the widgets are finally rolling off the assembly line. 

Keep on having fun with it Andrew. That's what retirement is all about.   

Cheers   

Don