Oh Blimey I bought a CNC Lathe !!!!

[lordedmond]

By ek that psi board has seen some heat on that resistor 


Con crats on the ready state

now for the big question dare you power up the spindle and make it do something 

well done 10 extra points for perseverance

Stuart

[awemawson]

A bit more progress. OK it seems that there is a sequence you have to go through to get the hydraulic pump going. Now it is ....

I can:

a/ Jog the X axis both +ve and -ve

b/ Jog the Z axis both +ve and -ve

c/ Jog the V axis only away from the main spindle (this is the auxiliary spindle)

d/ Rotate the C2 axis to its datum (this is the auxiliary spindle rotary positioning)

e/ If I try to start the main C1 spindle I drop my 30mA RCD trip that is in my 3 phase distribution box

f/ I can open and close the hydraulic collet closer on both spindles

I'm not particularly surprised at e/ as no doubt the leakage of an 18kW inverter drive is high and most industrial rcd's would be 100mA

The result of c/ is that I have now jogged V to the end of it's travel and it is reporting a V axis overload (that pesky proximity sensor like the one I fixed on the X axis) There is a software switch to allow you to jog out of this condition, but of course the issue is I cannot jog in the right direction so I will have to dig into the mechanics and try and wind the motor back manually to continue fault finding.

[vtsteam]

Ahhhhh, axis movement!!!! It's alive.......

Congrats!

[Swarfing]

Well you had 'READY', what you just said sounds like 'STEADY', on the way for 'GREEN' to 'GO' 

[awemawson]

So today's task: Reset that V axis overload clutch.

Obvious thing to do is expose the drive to the ball screw, and hand rotate it in the opposite direction until the overload clutch clicks back into place. Only one problem, the drive servo motor is low slung at what would be the tailstock end of a conventional lathe sitting just above the huge swarf conveyor. There does seem to be an access cover however this involves moving the swarf conveyor to get it off. As this took four of us a couple of hours to maneuver into position when I got the machine, I'm loath to do that.

[awemawson]

So maybe I can get at the ball screw from the other end? It's under this way cover:

[awemawson]

Yes sure enough, removing the suspiciously loose screws that retain it reveals the ball screw and LOADS of brass swarf.

[awemawson]

Cleaned it all up a bit, and grasping the ball screw easily managed to bring the V axis assembly back towards the headstock end. After a turn or two there was a satisfying 'clunk' as the overload clutch dropped back into position.

[awemawson]

So put it all back together, cleaned up the way cover a bit and tried starting up.

Sure enough V axis overload error is now cleared and I can bring the beast to a 'ready' state, get the hydraulics running and jog X & Z. Jogging V it won't go towards the headstock, and when trying to jog towards the tailstock it carries on going after I take my finger off the button. Emergency Stop button prevented another V axis overload error.

So my next task is to fault find the V axis - it could simply be a sticking 'go right' button, but I don't think so as same button is used for the Z axis jogging.

As those retaining screws were only finger tight I suspect someone has been in here fault finding just prior to me getting the lathe.

[awemawson]

It's all happening now! I decided to uprate the main workshop RCD to 100mA trip. Had one in stock so it was an easy choice. (*) Now the C1 main spindle axis is more than happy to turn without tripping the RCD. I suspect that the original 30mA one was perhaps over sensitive - my plunger EDM machine had also taken to tripping it and that is now ok.

V axis still only jogs in one direction. I have my suspicions though. The servo amplifier has an 'absolute positioning system' whereby it keeps it's absolute position in battery backed ram so you don't have to travel over reference points each time you power up.. The value currently held is a stupidly large negative number. I need to find out how (and where) to re-reference it to a sensible value.

But I think we are getting there . . . . . .  where ever there is 

(*) personally I'd be much happier at 30mA which will save a life, whereas 100mA will prevent a fire but still kill you. I'd opted originally for 30mA as on some three phase machines I have a standard UK 13Amp dual switched socket for convenience to power DRO's etc that is sourced from phase one and neutral.

[Pete.]

We have to turn up the RCD mA when we hire large portable generators. They don't like the big motor startups, even on the newest euro-start machines. It's caught a few of the guys out, had them opening up the commando plugs looking for water etc.

So you've ran the spindle then. What's left to do before it'll actually cut chips?

[awemawson]

Well I need to learn a lot for starters!

I think that the V axis issue is a referencing one. I'm not sufficiently familiar with the operation to have much confidence jogging things around without bits touching each other. I've not yet got the tool turret to index, nor the powered tooling to turn, but hey who wants to finish a good book.

[lordedmond]

don't put to much faith in those RCD's they can make you think you have all the bases covered but get across two phases one in each hand and its good night sweetheart 

yes they do a good job but they are not the be all and end all

most 240 vac equipment that has a delta suppression cap network ( most of the variable speed lathes ) fall into this will trip a domestic RCD due to the leakage on the caps ( the inrush can also trip them )

but just a thought when yo have run the spindle for a while and got some work out of it ( nice and warm ) it may well be OK on the lower RCD ( it may be damp ) but I will guess that its the odd order harmonic suppression circuit that may well be the cause

Stuart

[awemawson]

I don't think the rcd issue is leakage in the conventional sense of ohms to earth due to damp or poor insulation. The inverter / chopper spindle drive produces quasi AC by fast switching a DC rail. This inevitably produces a high frequency source that through the stray capacitance of the wiring gives sufficient 'leakage' current to trip some devices. The cheaper PC power supplies are notorious for this.

http://www.gtproductions.com.au/Mopping%20Up%20Earth%20Leakage.pdf

[lordedmond]

If its got one of those fitted then I am surprised that the 100ma one held up

OT

when I was at work I was with NWB as a supervisor in the facilities dept ( engineering ) we had a pretty hefty cable burn out about 150 amp  TP&N 4 core but only the neural was affected , this cable was only feeding a office section with many PC  mainly Dell's , the cable was duly replaced , but being me I was not satisfied that we had the problem buttoned down so out came the tong testers , red was 30 amp yellow was 32 amp blue was 28 amp so we were balanced ok , the shock came when we tested the neutral over 200 amp , out came the scope it was all odd order harmonics 7th and 9th

That was all down to the switch mode psu in the Dell's  they were overlaying the 50hz with a lot of HF and causing the cable to only carry the current in the skin ( high frequency skin effect )

keep up the good work you are very close now the winning post is only 1 fault away


Stuart

[Leblondmakino]

Yes defo sounds like a reference position issue on that axis..should be a way of re zeroing the encoder or it could be a soft limit setting that needs changing or re setting..

[awemawson]

Yes it was a referencing issue preventing the V axis moving towards the headstock end. Had a good look with a torch to find the proximity sensor and it's associated lump that it detects - can just see it looking down an access panel at the tailstock end. This showed it was set to reference at the extreme rightwards travel, and I was thus given the confidence to fast (not very!) travel it to close in jog mode then do a final approach in reference mode. Lo and behold the V axis value was then set to a sensible figure (it had been at something like -24000 mm) and I could then move the axis in jog mode in both directions.

All axis reference points are the same mechanically - a three terminal magnetic proximity sensor on a bracket looking for a simple cylindrical collar that is fixed with a set screw on a shaft parallel to the axis motion. Seems crude, but I suppose that the sensors are very repeatable.

I think that there is still a 'funny' to solve on the axis servo drivers - the absolute axis position is supposed to be retained on power off, and on two axis it is being lost despite the battery measuring ok, and also having been changed anyway. Not the end of the world if I cannot solve it - just means I'd have to re-reference after a power down.

Next thing to investigate is the tool turret - see if I can get a bit of time later today.

[awemawson]

Today was one of those days when circumstances stop you doing things We seemed to have an endless stream of visitors all morning expecting teas and coffees - don't these people realise some of us just need to slink off to the workshop! Seriously though it it nice to have visitors.

Eventually I managed to get to the coal face and look at this turret issue. Turns out no way is the tool turret going to turn unless the safety cover is closed and locked. There are three digital inputs for 'door open' 'door closed' and 'door locked' and although the 'door open' one is toggling as the door moves, the others don't. What is supposed to happen is that a relay latches on for 'open', is released as the door closes, and another relay latches on for 'closed'. When the controller sees the 'closed' state it fires a solenoid bolt, and a feedback contact for this bolt generates the 'locked' status. The 'closed' relay is briefly clicking over but not latching on.

I strongly suspect that the latch circuitry was originally routed through the bar feeder, which I don't have, so I'll probably end up having to link it out. Not proved yet so something for tomorrow when I get back from taking four pigs on their last journey.

Investigation was greatly hampered by having all the relay logic at the back, and the safety sliding cover at the front, so you cannot see whats going on as you slide the door Eventually solved this conundrum using the 'Indian Rope Trick' . I tied a length of rope to the door handle, and passed it right round the machine and back to the handle. Now I can slide the door from the back. A low tech solution to a high tech problem.

[awemawson]

My handy piece of rope has let me at last diagnose the door relay logic fault - we now have 'door open', 'door closed' and 'door locked' relays all doing what they should do, and about time too, I've been all round the houses on this one. In the end it proved to be a 'normally open' contact that was never closing - had to pull the relay (M1-K8 for the record) and dismantle and clean its contacts as I had no spare for that variety.

Now the door lock solenoid fires whenever an axis is moving or a spindle is turning - all clever stuff. Also now the jogging 'rapid' button works when the door is closed but is inhibited when it is open.

So on to the the Tool Turret. It is saying 'Turret is not locked' at the moment. This is the tool turret:

[lordedmond]

well done it will not be long now , it does look as if the major hurdles are behind you now


Stuart

[awemawson]

Apart from the obvious mechanical assembly that holds a twelve station tool carousel, buried in the works there are all sorts of goodies. There is a solenoid for locking, a feedback sensor to say 'locked', an encoder to say where the turret is rotationally, a motor to turn the turret either clockwise or anti-clockwise to select which tool to use, a motor to drive the 'driven tooling' (end mills, drills etc) and finally a valve to control the through tool coolant. Pulling off the two cover plates doesn't give very good access One reveals the wiring connections and the other the solenoid and encoder.

[awemawson]

The reported 'not locked' status is entirely correct - I can turn the carousel by hand. Oddly to my mind the locking solenoid seems to be biased to the 'unlocked' state, and powered to 'lock' this doesn't seem right to me. It all looks rather rusty and grotty, but the solenoid moves freely enough and if I push it into the 'locked' position, the digital input for the feedback sensor changes state. Also if I rotate the turret by hand, the encoder bits count up in binary correctly.

I now need to devise a way of powering the solenoid remotely so I can see what is happening. It's the old case of 'the actions round the front, and I need to be round the back' 

Have a picture of a rusty solenoid:

[awemawson]

well done it will not be long now , it does look as if the major hurdles are behind you now


Stuart

Thanks Stuart, yes it seems to be coming together at last. Still loads to learn about this machine. Amazing what facilities have been built into it.

[awemawson]

Amazingly  I found some web documentation on the tool turret: http://tinyurl.com/toolturret The numbers are the same but it is not quite identical. It has a useful cross sectional drawing that I'm trying to interpret.

A bit of investigation this evening has proved that the lock solenoid (which draws 2 amps at 24v DC) works, and if  jiggle the turret round to a correct position drops in its locking hole and provides feedback that I can read as a digital input. However the turret still has a bit of radial movement - looking at that drawing in the link above it implies there is also a toothed 'Hirth Serration' detent that possibly isn't seating firmly.

I've also proved that the motor that turns the turret works both forwards and backwards by manually operating its contactors. The relay that drives the solenoid is a plug in variety and suspiciously it has obviously been changed at some time as it is a different colour from the rest, however there is a tell tale LED on the relay drive digital output that is not lighting so perhaps some software condition has not been met to get it activated. More head scratching.

Here's a bad quality picture of the bank of relays:

[John Stevenson]

And I thought I had trouble when the drive belt snapped on the trusty ML7 ?