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

[awemawson]

The hunt for the earth leakage goes on!

Consistent approx 300 ohms 'phase to chassis' with the huge breaker giving the KTK Mentor it's 415v 3 phase closed , but clears with breaker open. First job, open up the KTK Mentor DC Drive.

Well I really thought that I'd found the fault - several 0.1 uF capacitors in a very sorry state. So I disconnected the terminals putting power on the device, but noticed that there were much thinner cables going elsewhere.

Measuring to the cable ends the 300 ohms problem was coming NOT from this bit of the Mentor, but the small cables - but where do they go? A major bit of trunking dismantling and cable tracing found that they went down to the bottom of the KTK Mentor via some smaller fuses.

Isolating the fuses the fault persists (so not this bit of the Mentor drive), but hang on, there again is another cable coming off these terminals back into the trunking - but to where?

Answer - it goes to a  small motor starter unit, who's output cables plummet down to the main termination rails at the cabinet base FOR THE FAN FOR THE MAIN MOTOR. Faulty fan perhaps, yet I know it turns as that was one thing I got going on site before I bought the machine.

But hang on what is that white domestic looking three core flex doing connected there and why is it's blue neutral wire connected to earth.

Answer: some plonker has decided to add a Ventaxia Window Fan to the roof of the cabinet that houses the turret electronics and a pair of 24 v power supplies. This machine has no neutral feed, so Mr Plonker has used Earth instead  I've temporarily isolated the fan.

So the elusive 'earth leakage fault' has proved to be someone in the past doing silly things. Now the machine happily starts up on my 100 mA RCD protected supply without tripping, and I can consider returning the Sinumerik 820T controller to it's proper place, but before then I must decide what I am doing about relocating the batteries that caused the problem in the first place - replacements arrived today.

. . . mind you those 0.1 uF capacitors will need changing before long

[charadam]

Andrew, I am lost in admiration at your fault-finding ability and sheer breadth and depth of know-how.

I must however sound a note of censure - your boots man, they are a disgrace!

I offer in evidence post #106.

[awemawson]

You are  NOT the first MadModder to comment on my footwear 

They are 'Crocs Bistro' of the unvented side variety in white (yes honestly!) intended for the catering industry and I get through three pairs a year. This particular pair were actually thrown in the bin a few weeks ago, but I rescued them when I was Creosoting those chicken sheds, as the stuff goes everywhere. I do like to properly wear them out and get my monies worth.

I wear Crocs all day every day about the farm, or if I'm walking the dogs in the woods or countryside, and as I take UK size 13's which are difficult to find I buy them well in advance if any turn up at a sensible price - I do have four brand new pairs in stock for future use 

[awemawson]

Now of course the mistake Mr Plonker made fitting this fan was to draw 240 volts from one phase referencing it to earth. All well and good if the earth current isn't being monitored by a safety device (The RCD)

What he SHOULD have done in this situation where no Neutral is available, is to use a small transformer primary between two phases (415 volts) and secondary giving a floating 240 volts. The fan can still be earthed (though it wasn't !)

But what size transformer? Well I can't find a model number on the Vent Axia so I needed to measure the current draw rather then look it up. So I temporarily wired it to 240 volts leaving a long enough neutral wire for my Fluke clip on meter to measure the current.

Answer : 200 mA . Now I'd expected a start up surge, but actually it starts at 100 mA then increases to 200 mA. so a 50 VA enclosed 'panel transformer should do the job nicely - off to eBay to find one 

[Pete.]

I think I've seen one of those in a disused panel at work Andrew. I'll have a look tomorrow.

[awemawson]

Thanks Pete, I'll wait with bated breath 

[awemawson]

I spent a bit of time this morning studying the KTK-Mentor DC Spindle Drive power board to see how feasible it is to replace those dodgy capacitors, and indeed try and determine what they are through the 'stuff' they have exuded.

The capacitors turn out to be PME261JC6100KR30 by Rifa and are 0.1 uf at 1000v DC / 500v AC - available in stock from RS but at quite a price. By Googling I discovered a place called Mercateo that I've not come across before that were noticeably cheaper - order placed.

So looking at the power board for the drive it is pretty daunting with loads of connections all waiting to trip me up, however sitting down and studying pictures of it, and working out which bit did what in fact it doesn't look to be too bad (famous last words!)

I took one of my pictures and annotated it to get things a bit clearer in my mind.

[awemawson]

It turns out that this lathe was at one time equipped with two probing systems. As well as the tool probe which is hard wired, there has been a Renishaw LT02 optically coupled work probe that sits in a pocket in the tool turret and can measure a work piece.

The Optical Receiver and interface unit are still in place, and amongst the scanty documentation is the installation manual and test certificate BUT NOT THE PROBE.

. . . no, not the probe 'cos the chap who I got the other probe from flogged it on eBay back in April 

(amazing what a bit of data mining reveals!)

Come on empty your pockets - who's got a Renishaw LT02 probe for me

[AdeV]

You've already found what the cap is... but I was going to suggest it looks a LOT like an AC smoothing capacitor. Sharp fitted them (or one very like them) to all of their 1980s computers... and these days, most of those computers (often having spent years in cupboards and attics) usually blow them fairly soon after being plugged in. As a side note, somewhat oddly, they're connected directly to the mains before the switch.... so an old Sharp can give a scary-sounding  BANG and a cloud of magic electrical smoke even when it's turned off!

Chances are, that cap is entirely surplus to requirements, and it's rather sorry state is due to the fact it's already gone bang & let out the magic smoke.

Still... better in than out (as no-one ever said). 

[awemawson]

Reading up on RFI suppressor capacitors it's actually an interesting and more complex situation than you might imagine.

OK we are trying to suppress RF noise and the capacitor does that, but it has to be able to stand up to the environment in which it finds itself. All mains systems suffer from voltage spikes. Either back EMF from other devices or noise on the transmission lines caused by network switching or lightening for instance.

The chosen capacitor needs to withstand these spikes, but the fact is that actually they don't. There is usually a breakdown of the insulation that vaporises a tiny bit of the metalisation that forms the plates. Insurance rated capacitors like these ones have to have self extinguishing insulation and a 'self healing' characteristic whereby the capacitor continues to function but at a slightly reduced capacitance. Apparently on average equipment sees 18 of these spikes every day 

A bit about it here:

[awemawson]

THE RFI suppressor capacitors arrived this morning, so no excuses - replace them !

Firstly - definite mains isolation - this thing takes no prisoners with high voltage DC - so not only turned off but unplugged as well. I was surprised to see that the front control panel hinge pins are self tappers - seems crude.

Then it was just a case of marking anything that might confuse on re-assembly, and un-bolting and un-screwing everything until finally the board could be removed from the power semiconductors that it mounts on

[awemawson]

Then it was a case of un-soldering the old capacitors, cleaning the board up, and soldering in the new. The old capacitors were very degraded, with blobs of I presume solder escaping from the cracks in them. These could not have been left unchanged safely I think.

The gunge that escaped took a devil of a lot of cleaning. Carburettor cleaner, and IPA didn't touch it. It was slightly soluble in acetone with vigerous rubbing so that's what I used.

So, cleaned up, repopulated, bolted back in and tested with the power on. It goes to the ready state as it did before, but this is no real proof that it's functional - just that it's not dead !

Anyway then I was dragged away to unload a pallet from a lorry. All those problems the other week with my hydrovane showed me how vulnerable I am to loss of compressed air - so this one, a three phase '502' on a 90 litre tank , was liberated from eBay and travelled down from Norwich - best go and test it's OK

[PekkaNF]

Good job with RFI capacitors.

How do those MOV:s feel?

[awemawson]

Thanks Pekka 

The MOV's don't seem to have suffered.

I received a text this morning from the chap who used to operate this lathe telling me that there was an issue changing to low gear which they never got round to fixing, and if run in high gear for too long 'something in the back got very hot and had to be left to cool'

Now the 'something' can only really have been the KTK Mentor spindle drive so maybe those poor RFI suppression capacitors died from heat stroke 

[Pete.]

Knew I'd seen one somewhere, just took me a while to remember where and get round to finding it...

[awemawson]

What a gentleman 

[vtsteam]

Good man, Pete! 

[awemawson]

So with the mention of possible unresolved gear changing problems when still at the University (it HAS gears - news to me  ) time to see what this beast has by way of transmission.

Main DC DRIVE motor by Mawdsleys 26.5 kW / 5000 RPM with attached blower motor. Drive goes (I assume, not dismantled yet) by drive belt in a big cover to the co-axial gear box by Andantex. This gear box has a lump bolted on the side with a Parvalux style 90 degree motor that I assume moves the gear box from one ratio to the other.

The gearbox label is not very clear where the ratio has been stamped, but I am reading that as "1 & 3.2" which seems sensible - straight through = 1, and via presumably a planetary system giving the 3.2

It seems that Andantex are still in business however nothing is showing up for this particular gear box. Looking at the 'type'  box where it is stamped "BVR  353" implies it's a special for Beaver, but probably a variant of a standard one - possibly the gear change arrangement was customised.

Looking at note that I have for commissioning a later TC-20 lathe there would seem to be two 'M Codes' involved 'M41 - Low Gear Request' and 'M42 - High Gear Request'

I would suspect that if there is a gear changing problem it most likely is with the Parvalux motor and it's electrical drive system so long as the main box hasn't been mangled by changing gear 'on the fly', and the ladder logic in the controller should prevent this (hence 'Low Gear REQUEST' rather than 'Low Gear')

. . . again time only will tell.

What I can't find at the moment is the spindle encoder - this lathe has a 'C' axis whereby the spindle can be orientated anywhere in the 360 degrees of rotation for operations using 'live / driven tooling' in the tool turret

I am negotiating with someone who claims to have all the wiring diagrams for this vintage of TC-20, and I'm loath to replace the controller and power up things like the main drive and gear change system until I've given the drawings a meticulous inspection. He's been on holiday returning Monday so hopefully . . . .

[awemawson]

Well I found the spindle encoder ! It was hiding behind the tin work that protects the hydraulic chuck open / closer.

[awemawson]

A bit more data mining, and I think I've found the Gearbox that the Beaver one is a modification of. It's the MSD-size 353 in the attached  screen grabs of their catalogue.

Visually it's the same, the dimensions tally, but the ratio is different as mine is 1:3.2 whereas the standard is either 1:1.38 or 1:4.94 and the gear changing arrangement is different.

Now the model size is '353' and the data plate quotes 'BVR 353'  - so a few customised tweaks for Beaver I think 

[vtsteam]

Chased that one down. 

[awemawson]

I think the combination of digital cameras (iPhone 7 in my case) and Google have made this sort of thing so much easier.

I was puzzled by the gear changing motor, as the bit of the label that I could see (upside down) implied that it was a single phase motor - but I've traced it's cable back to a standard interlocked pair of contactors wired for three phase reversing operation as I'd expect - but not with a single phase motor 

I managed to stick my iPhone into a place where I could get a decent picture - and all is revealed

It can work single phase with a 4 uF capacitor or 3 phase !

[awemawson]

Today's foray into the electronics of this beast was to resolve some oddities regarding the pair of 24 volt 7.2 amp power supplies that are housed the the rear left hand cabinet with the turret drive electronics.

First Quandary: They seem to be wired in parallel - this isn't good practise as being regulated they will be bound to be trying to achieve slightly different voltages - although equipped with remote sensing inputs, these are directly wired to the PSU's output terminals hence not used. The effect will be one power supply taking a larger share of the load than the other.

Second Quandary: The -ve output cable is numbered '102' and the +ve output is labelled '100'  - cable '100' doesn't feature anywhere else on the machine  (*) and +24 volts on this machine is cable '103' - OK a simple wire labelling error - careful tracking proved it should be labelled '103'

Third Quandary : and this is the BIG one. With no load the pair of PSU's deliver 24.2 volts, but on load this rises to 33 volts  It was this strangeness that got me hunting in this  area in the first place. Now at the moment the 24 volt rail is only being used for relays etc which won't have too much of a problem with the over voltage, BUT also the Baldor SMCC microprocessor card that manages the turret movements for tool changing runs off this 24 volts and had lights flashing all over the place when the fault was on - I hope it hasn't damaged it.

I could reproduce the fault with a  10 ohm load resistor as I disconnected the PSU's one by one in situ - the left hand one was U/S, the right hand one seems OK

Removing one PSU for testing the Baldor card seems to have calmed down now it only has 24.2 volts attacking it from the remaining PSU, and it's green ready light is coming on - so fingers crossed.

Bench testing I loaded the faulty one with a 48 watt 24 volt lorry bulb (actually a spare from my Startrite Bandsaw!) so drawing about 2 amps, and sure enough the output rose to 33 volts!. Simple transformer, twin diode rectifier and big electrolytic capacitor produces about 35 volts. There is a series pass regulator comprising three 2N3055 power Transistors with load sharing resistors in the Emitter circuit, and this group of three is driven by a fourth 2N3055 making a 'super alpha pair' (well quad in this case!). The base of the fourth 2N3055 is driven by a UA723CN voltage regulator IC

I suspect the 723 chip, but equally it could be one of the 2N3055's breaking down - I was amazed to find I have neither in stock - used to have masses of these things !

Although I've ordered some components, and I will repair this PSU, I am going to replace the pair of Kayser 24 volt 7.2 amp supplies with a single switch mode supply rated at 20.8 amps - over rated so it will be under run - which eBay has provided.


(* not quite true as alarmingly the three phase 415 volt input cable to the KTK Mentor Spindle Drive are labelled 100, 101, & 102 - some draughtman made what could have proved a expensive error )

[vtsteam]

This machine seems to have a lot of surprises under the hood. The manmade type! Good detective work, Andrew. 

[awemawson]

I suspect most machines of this age have a few issues lurking un-noticed, after all it IS 29 years old - almost older than ME 

Being a bit OCD I'm going through everything with a fine toothed comb, firstly to try and understand what I've got and secondly to try and iron out as many of the issues as possible before the controller goes back in. When that starts taking control of hugely powerful X and Z axis servos, to say nothing of the 26.5 kW spindle motor I want peace and quiet to reign and minimise that nasty expensive bangs !

There are still areas on the machine I've not been able to access. There is a curious set of 'lazy susan' type linked bars behind the tail stock under covers that just let me a slight glimpse of them - I suspect that they support the swarf covers but I can't get in to see properly.

I'd like to be able to remove all the swarf covers but I don't think it's feasible. There is evidence of a bit of rust under the moving carriage but I can't get at it to clean it up, and it's not nice to move the carriage to do so. I have been manually triggering the automatic oiling system so areas like this are at least lubricated before thing moved.

Progress on the wiring diagrams - they are being scanned to .PDFs and expected in a day or two