The Sequel - Oh Blimey I bought a CNC Lathe (Beaver TC 20)

[awemawson]

Having packed up the RA17 and ordered it's collection I could get back to the coolant isse.

Firstly, I seems that the hose that blew on it's barb fitting is 5/8" bore Not 1/2" as marked  but was on a 1/2" barb, so not really surprising that that is the one that went first. I replaced the barb.

Secondly, as I was working out where to put a Pressure Relief Valve and what thread sizes I needed, I looked closely at the pump and guess what I found - a Pressure Relief Valve. It was screwed all the way in so at it's highest setting, but I don't know what pressure that corresponds to as it is un-calibrated.

I unscrewed the PRV as far as I dared so it is at a much lower setting and sure enough all works, and when the Washdown Gun is closed off you can hear the PRV working - so a result 

Sometime I'll put a pressure gauge on it and see what pressure it opens at. The wash is noticeably less vigorous.

[Pete.]

When I was working in the Dartford Tunnel cutting out the road deck we had trouble with those hoses bursting right next to the jubilee clips. We were using the fire mains as a water supply but right at the bottom of the tunnel the extra pressure was too much for the hoses, but only where the clips were. We changed the jubilees for double-ear oetiker clips and the problem went away. The theory was that when tightening the jubilees it was dragging the top layer of the reinforced hoses round and causing it to tear where the braid was moulded into the hose.

Ever since then I have always favoured using those o-clips instead of jubilees wherever I can.

[Pete W.]

          

Many years ago, when I lived in Romford, I organised a guided tour of the Dartford Tunnel.  (It was before the bridge was built - the Dartford & Purfleet Joint Tunnel Undertaking was the power in charge then.)

As part of the tour, we were taken beneath the road deck in one of the bores.  The whole tour was very interesting and the staff spared no effort to make the tour a success.

It's many years since I crossed the Thames there so I'm well out of date with developments.  Pete, why were you removing the road deck?  Temporary or permanent? 

[awemawson]

Now I'm starting to use this machine in earnest  I decided it was time to check and adjust the turret alignment.

Knowing that the turret had been driven into the tail stock at some time in its history, making that big dent in the cover that I had to re-fabricate, it was unlikely to be 'spot on' .

Two things are necessary:

Firstly that the tool disk front face is at right angles to the spindle axis, and parallel to the X axis movement.

Secondly that the bore that accepts the VDI tool spigots is accurately in line with the spindle centre line when X = 0 . This measurement can be 'out' either as X=0 has been incorrectly set for movement up and down of the X axis slide, or more likely that the turret rotational alignment has shifted in a crash.

Today's job was to check and correct the tool disk parallelism. I decided to use my Tesa electronic DTI, as it has the advantage of being able to be set to increasing sensitivity as the process proceeds, and also to have a long enough cable to be positioned at the operators console from where I'm moving the X slide.

The turret is held on by three massive cap head screws at the top and also the bottom, the bottom ones requiring that I remove the power tooling drive cover to gain access.

Loosening the screws then setting them to 'just nipping' I then installed the Tesa DTI head on a magnetic clamp on the face of the spindle (having locked the spindle to prevent rotation). Sure enough there was quite a large discrepancy as the tip of the DTI moved across the face of the tool disk - I didn't take accurate measurements but it was of the order of 5 thou in 2 inches!  Then much like you would align a vice on the table of a milling machine, it was a case of judicious tapping with a soft dead blow hammer aiming to rotate the turret on it's mounting face with the X axis slide. After several iterations I got the error down to about 1/10 thou in 2 inches. Tightening the clamp bolts increased this to 2/10 but that's where I left it.

To check the rotational location of the tool disk I need to rig a ground pin in one tool position and swing the DTI round it. The Tool Disk is bolted to the Hurth Coupling that actually fixes it, and it will be a case of slackening the bolts and a bit more iterative tapping !

If I can find a suitable pin, then that's possibly a job for this afternoon.

[awemawson]

One step forwards and three back - it seems that my various adjustments to the turret position have disturbed something within, and the turret no longer indexes properly 

The turret on indexing  brings the tool disk forwards, thus disengaging the front Hirth coupling and engaging the rear one that then allows the servo motor to rotate the turret to it's new position. Once there the system drives the tool disk back onto the front Hirth coupling locking it in place. BUT, it is not reliably getting to the right place, so when the Hirth moves to engage, it's two parts don't mesh correctly leaving the tool disk forwards and the system locks up awaiting the 'turret locked' signal, which it never gets.

Now it's a complicated control system. The PLC requests a tool change from the Baldor SMCC servo card, which controls the turret motor and takes in inputs from the Optical Incremental Encoder that is mounted on the servo motor shaft. Within the turret there is an Index proximity switch telling the SMCC card where zero is. As far as I can tell the program in the SMCC card knows how many encoder counts it needs to move to the requested tool position and homes in on that.

Now previously I had thought that the four way proximity switch block (that also monitors tool position) was what was being read by the servo, but now I realise that it cannot be, as when the tool disk is moved forwards for rotation, so are the protrusions on the 'hedgehog' that the proximity switches sense and are out of range. These four proximity switches are a confirmation to the PLC that the SMCC card has done what it was asked to do!

So I've been doing tests this morning, having first removed all the sheet steel covers off the turret to gain access yet again 

It turns out that the direction of rotation of the turret influences things, so going from tool 5 to tool 9 might work fine, but going from tool 9 to tool 5 will often give the error.

Things that could influence this are possibly a loose encoder coupling, or a bias offset on the servo system. Now currently getting at the encoder coupling to check its fixings is proving a problem. I'm guessing that there are one or two grub screws on each side of it, and there are screw capped holes to give access at both ends where I expect the screws to be, but the shaft of the motor needs turning to get at the screws, and the motor has a brake that is engaged with the power off. Head scratching time working out how to just power the brake, as it's too dangerous working with my head round the back of the turret if the main servo systems are alive 

It probably just needs a 24 volt supply to pull the brake off.

As for the servo offset, this harks back to something I thought I'd noticed months ago when I first got the turret semi-working, but I'm short of documentation on the Siemens Simodrive AC servo card.

[awemawson]

As I thought, the Turret Servo brake was simply powered off the 24 volt DC bus, so I jury rigged a lab supply to pull the brake off, and was able to see the coupling. Getting a screwdriver onto the screws was another matter 

The access hole is maybe 10 mm diameter, facing downwards at the tail stock end of the turret assembly. Trying to get a mirror and a torch and a screwdriver co-ordinated proved impossible. I got out my Wi-Fi endoscope for a good view of the coupling, but there wasn't room for the endoscope and the screwdriver at the same time. I then tried taping the endoscope to the screwdriver with the screwdriver leading the adventure, but the image got very confusing as I had to move the screwdriver. Eventually, with the endoscope porting the image to my iPhone, which was balanced inside the machine, and holding the camera bit close to the hole with one hand and poking with the screwdriver with the other EVENTUALLY I got it on each screw in turn and checked their tightness.

All very confusing, as the image has a significant delay, so you are seeing what you did, rather than what you are doing !

Anyway they were as tight as I could get them, so it was probably another red herring. Have a video of the coupling rotating

Having done all that I now cannot reproduce the fault 

Could be that things have warmed up a bit, but really I'd like to try and set the servo card so it balances nicely. I'll have to get some numbers off the card and go googling

[awemawson]

Possible reasons for the Hirth coupling not always being positioned where needed include the Siemens Simodrive Servo Card. However lacking any information on the card I was loath just to tweak away. Posting a plea for help on another forum has sourced the gen letting me experiment with a bit more confidence. The Beaver manual says very little about the turret, but what it does say is the the servo gain will need to be set quite high, but it's description of which pot was the gain is wrong hence not having played with it.

However at the moment I can't reproduce the fault 

I've increased the servo gain slightly, and subjectively the motion seems a bit crisper, but maybe that's wishful  thinking !

All this started when I was aligning the tool turret, and the tool disk face is now nicely set to match the X axis, but the tool turret rotation needs re-setting. My measuring dowel came in the post today, and it's out by 5 thou in about 5 inches, so not an enormous error. I now need to slacken the bolts securing the tool disk to the Hirth coupling, and gently tap it round however many degrees that error corresponds to, and hope that this time I don't expose more problems 

[awemawson]

Today is the day to set the Tool Disk angular orientation

The two part Hirth (or maybe curvic, documents refer to it by both names) coupling has one part fixed to the turret centre spindle, and the other bolted to the tool disk trapping a flange on the spindle. By loosening the ring of eight M12 socket cap screws it is possible to tap the tool disk round to adopt a changed angularity relative to the spindle. But what to use as reference 

With a 'normal' lathe tool holder clamped in the turret, what we are aiming for is the reference mounting surface for the lathe tool to be parallel to the X motion of the carriage. Fortunately I have a brand new, nicely ground holder that I mounted in position 6 (arbitrary - doesn't matter where!) with a balancing one in position 12, though this one wasn't brand new and it's surfaces weren't quite as  pristine!

Then setting the electronic DTI in the chuck, placing the finger on the relevant tool holder surface, rotating the lathe spindle to give me a zero, then locking the lathe spindle, I could track it up and down the surface seeing a conssiderable discrepancy. (about 30 thou in the 100 mm length of the holder)

Now I had prior to this   slackened off the ring of eight M12  cap screws and semi re-tightened them. It's worth noting that the internal hydraulic 'lock' cylinder is also holding the tool disk against its flange. Initially I thought that I was going to fall at the first hurdle, as the bolts were massively tight. Eventually bringing out a 4 foot tube as a 'torque amplifier' they all came loose, but I was grateful that my Allen Key was a decent quality Unbrako that stood up to the punishment.

Tapping the tool disk round with a brass drift, resetting and starting over, after several iterations I got the indicated error down to zero with an FSD of 1 thou on the gauge, which is quite pleasing. Rotating the turret round to the balancing tool holder and measuring that gave me an error of just under 2 thou over 100 mm which I put down to the fact it's a well used holder, but that discrepancy wouldn't be the end of the world.

I'd put all the removed tool holders in a 'Tote' to keep them safe, and when I went to move them out of the way realised how amazingly heavy they are - they ended up staying put! It made me have a greater respect for the servo system that rotates the turret.

Now I will have to go through the tool setting process once the tool holders are re-mounted, but that's fairly automatic now.

I did have one peculiarity during the process described above. I got to the stage where the adjustments had been completed, so the last action was to jog the tool holder past the DTI which was fine. I then left everything powered up while I boiled a kettle for a visitor. When I came back the machine looked ok, all lights correct, but it wouldn't jog. Even powering down and trying to go through the initialising process the issue persisted. Hunting for the error I removed the axis enable, axis brake and drive over travel relays relays one by one reinserting them as I went. After that everything was normal so what had happened I don't know. Obviously something to watch out for.

[awemawson]

All done and dusted. Tool holders re-fitted and tool off set measurements updated.

I weighed the tote of tool holders that I'd removed, 39 kGs plus the two I used for measurement that were still on the turret at 3 kGs each so 45 kgs being rotated by that poor old servo unit. 

No wonder it needs a reasonably high gain setting.

I was just about to re-run a previously run job to see how things measure up, then JUST stopped myself remembering that all the sheet metal covers are still off the turret and need re-fitting to keep the swarf, and probably more important the coolant, out of the works. Might get half an hour or so tonight to refit them as I believe I'm an orchestra widower this evening !

[awemawson]

I got the turret covers back on this evening - by heck some of the screws are in awkward places !

All seems to work. I ran a simple job (a shaft with two diameters and radius-ed transitions). Have a very splashy video. As is usual with these things the coolant stops you seeing much !

[Pete W.]

Hi there, Andrew,

Have you got room to fit one of those rotating glass disk gizmos (I think it's called a 'Kent Screen') on the inside of your door? 

[awemawson]

Pete, I've actually got one in a box somewhere, I must dig it out and see if there is room 

[PekkaNF]

Good work.

...
I decided to use my Tesa electronic DTI, as it has the advantage of being able to be set to increasing sensitivity as the process proceeds, and also to have a long enough cable to be positioned at the operators console from where I'm moving the X slide.
.....

That is very usefull looking indicator, small to install and display is a separate unit. I have been eyballing lazilily some LVDTs and diplays, but the (mitutoyo) price has been a deterrent.

Tried to google canditates, but I'm failing with proper search words.

[awemawson]

Yes Pekka it's a real boon having it. Picked it up on eBay many years ago. It uses a non standard battery, so I made up a paxolin block with holes to take AAA Duracells which solved the problem

I have a mains operated version that's a bit big and clunky but takes the same pin out sensors, and also takes two of them for high and low limits in a production set up, lighting red and green lights for 'go' and 'no go' !

. . .I've only got pictures of it's internals when I got it and mended  it ! It uses PROPER electronics not this modern stuff 

[awemawson]

So Pete-W I dug out the spinning VisiPort, but unfortunately it's not going to be an easy fit. Like all these CNC  machine sliding doors, there is a wiper seal stopping blood and guts getting driven down it's side, and were the VisiPort to be glued to the door, the seal would wipe it off again  - sad but true.

Still on a brighter note I've bottomed out why the machine just stopped and refused to do anything the other day. Absolutely NOTHING to do with the relays that I'd suspected. It was the key switch on the Operators Panel 

It turns out that the slightest touch on it, and it operates. Now this key is intended to stop people fiddling, so inhibits many functions. Measuring the contacts, even when 'closed' I'm getting 600 milliohms which for a gold plated contact is far too high. I think some of the Lithium corrosive gasses got into it from the original problem.

EBay has sourced me a very nice looking replacement, allegedly Swiss made so hopefully reasonable quality, brand new for £3.20 

[Pete W.]

Hear this everybody:

Never, ever, tease Andrew - he always has the resources to call your bluff!!                    

[awemawson]

The replacement key operated switch arrived this morning, so as it was a quick job it got fitted.

The extra depth proved not to be a problem, and the only complication was that the junction where the front and rear of the switch body clip together was within the depth of the (plastic) panel stopping the 'ears' of the clip expanding far enough to retain. After a stern word it gave in and got fitted and soldered in.

I decided not to fit the cover flap as they are a flipping (or flapping!) nuisance.

While the panel was off I replaced the D-25 socket retainers with the pillar style. The originals were for sliding clips whereas my cables have the retaining screws. Another job ticked off the list - I'd planned to do this wees ago.

[awemawson]

Well that was a short live pleasure!

Operating the switch to prove it inhibits the operations that it should, it fell apart in my hands  The main body of the switch has a threaded portion that has cracked.

Switch removed, wires soldered together and hole left in operating panel until replacement received 

(And no, it wasn't over tightened !)

[Alphawolf45]

 I have 3 of those Swisstac switches broken right now on my EDM.. And I know that the previous owner of the EDM replaced two of the 3 that has broke. I have decided that somebody chose the wrong type of plastic to make them from - too brittle.

[AdeV]

Operating the switch to prove it inhibits the operations that it should, it fell apart in my hands  The main body of the switch has a threaded portion that has cracked.

Nothing to do with...

After a stern word it gave in and got fitted and soldered in.

...was it?

 

[awemawson]

No, that was just the clip. (And no violence was involved!)

[charadam]

Serious question Andrew - is this something you should print, to ensure the material quality you need?

[awemawson]

The wall thickness of the switch body is the problem I think as its under 1 mm with threads in it as stress risers. I don't think I could print in the detail required.

I did however yesterday print a tubular spanner in ABS to make fitting the next one easier. Just a castellated tube. I'll pop in a picture when I'm on a proper PC not this iPad.


Later edit: photos added

[awemawson]

The replacement key switch body arrived in the post this morning and was fitted without drama 

[awemawson]

I'm still chasing the occasional 'wrong indexing' of the turret when changing tools, so have been chasing my tail looking at the turret servo system in some details.

There are three Siemens Simodrive servo systems in this lathe, all the same except that the drive cards have different current capabilities. X is a 20 amp, Z is a 30 amp, and the turret is a massive 40 amp. It came as a surprise that the turret was the biggest.

Having yet again pulled the covers off the turret to take a picture of the servo motor specification plate,needless to say in the most inaccessible place! I've been able to compare the settings on the card to the Siemens manual, and they correspond exactly. Then I thought that I was on to something. The same servo motor that indexes the turret also drives the powered tooling, and I thought that I had noticed some discrepancies in actual speed (measured by a reflective  rpm gauge) and the commanded speed.

So this mornings job was to tabulate the errors and try and work out what was going on.

I had a reflective sticky tab stuck on an ER32 collet nut. 100 rpm gave 350, 200 rpm gave about 400, 1000 rpm gave approx 1600 - what the heck?

Then, in a flash of inspiration I wrapped brown insulation round the collet nut, and stuck the reflective tab onto that. Guess what:
100 =100, 200 =200 etc etc up to 2000 rpm - absolutely spot on all the way through the range. But this itself is rather peculiar, as I've been fiddling with the Tacho scaling pot so I'd not expect it to have been so accurate

The main issue seems to be getting the twelve tools on the turret properly balanced - apparently unbalance is more of a problem than total weight (45kg) of the tools I'm told.

I was going to try and make up a spreadsheet of the various tool weights and try shifting them about to get a balance, then I realised, as the geometry of the tools differ, the point at which the weight acts also differ.

Then I was going to make an analog of the tool disk in aluminium plate with a central bearing, and manually shift tools about until it balanced, rather like balancing a grinding disk.

Then I thought . . . this is getting silly . . . .go and have a cup of tea . . . .

. . . so that's where I am now