Resurrection of a CFEI 100 KVA Induction Furnace

[awemawson]

OK a bit more dismantling - the contactor is in pieces all over my bench 

Essentially there is a plate that slides that has three sets of contact pieces on both sides arranged so one way up it gives a 'normally closed' contactor, and in my use the other way up gives a normally 'open contactor'. These contacts on their plate slide impaling themselves on the 'bridge piece' thus making a circuit. There is an over centre spring loaded toggle arrangement keeping them slide to the operating side, but when the no volt coil or the manual lever hits the release they spring apart.

There is a spring loaded 'hook interlock' preventing the slider  returning to the operating position which when the contactor goes to 'reset' should be lifted out of the way. At the moment I cannot see what is supposed to lift this hook, and indeed it isn't lifting.

The toggle has two springs, and the one on the right sits in a die cast cup with guide lugs and travel limiting lugs. The travel limiting lugs have broken off 

More work to fully understand this device but I'm nearly there.

If I really have to I can either machine from solid, or 3D print and lost wax cast a replacement cup but I will probably work out how I can try the furnace without this contactor before c omitting the time and effort to what might be a lost cause.

Remember - all this started mid melt so the contactor was closed anyway so there may well be more serious problems once this one is over come.

[awemawson]

This morning I lashed the 'trip coil' back into the circuit as relay K1 needs it's continuity, and made temporary links round where the failed big contactor had been.

I powered it up via the 30mA RCD protected 16 amp three phase supply that normally feeds my guillotine as it's close enough to see & hear what's happening but far enough away for safety.

This was just for testing to see what faults are being displayed - and very sadly it's LOTS 

[youtube]Q4CFH_KETtA[/youtube]

Much more head scratching needed - sadly it's possible that this will be the end of this journey, as I have no circuit diagrams of the main microprocessor equipped logic board, nor have I ever seen another one after years of watching.

. . .early days yet, I need to retreat, lick my wounds, and cogitate!

[russ57]

In general, sudden onset of lots of faults is unlikely, although 1 fault could cause lots of damage.

I'd be looking for a power supply issue.
A missing rail, etc.

Or a loose / faulty connector. Is there any other components /modules near that corroded solenoid?

Otherwise a possible sad end to an epic journey... [emoji22]

-russ

[awemawson]

Yes I agree Russ.

Well out I went after lunch armed with circuit (just block diagram sadly) to test voltages and try and locate all the fuses. Poked and prodded before applying power and - guess what - it went ready once I'd given it water flow and pressure and pressed the reset button.

I'd been lead astray earlier by all the lights on for the power supplies and microprocessor on the front panel thinking we were awash with errors - but actually being 'green' they are showing 'OK' - the red ones show errors when on 

[youtube]9IflsxCpBQ8[/youtube]

Nothing had measured wrong - all supplies were there - no idea what has changed, and of course this may only be a temporary respite but at least it gives me hope. Obviously the program and basic microprocessor are still able to function.

I pinged the resonant frequency of the crucible, and did a very brief 'heat' (16 amp supply remember)

Having typed this out, I've just been back and repeated the process and amazingly it's still working. Again a very brief 'heat' as 70 amps from a 16 amp supply cable is not good 

Still need to sort the contactor out, but at least this makes it worth doing.

Now faults don't just go away - something is lurking - but at least the fundamental electronics seems OK (ish)

[awemawson]

So now the priority is to mend or replace the contactor.

Were it not for it's remote tripping facility it seems that there are quite a few just as 'manual isolators' - but the tripping is the only thing that the eStop button acts on and I'd like to retain that !

World wide searches haven't turned up any used ones, and although new compatible ones seem to be available their price rules them out at several hundreds of pounds.

The bit that is broken is a zinc or maybe aluminium based die casting - boxing it up is occupies about 50 x 50 x 65 mm. The hole for the big spring is 40 mm o/d by 36 mm deep.

Options:

A/ Machine one from the solid - time consuming but many of the features are there to minimise material use so wouldn't need to be incorporated.

B/ 3D model it, print in PLA and get it lost PLA cast in bronze.

I'm leaning towards option B - it worked well when I needed that drive dog for the Beaver lathe live tooling. I'm not even sure my copy of Fusion 360 is still operable - not used it for over a year and they have a habit of chucking you out !

As always comments and suggestions would be very much welcomed.

[vintageandclassicrepairs]

Hi Andrew,
I despaired when I saw the contactor in bits 
I wonder if there is moisture getting into the unit? due to the condition of the trip unit?
Does it have an internal heater to prevent condensation? if not I would add one (main fed)
I really hope you are sucessful in getting the furnace going again
If I had access to my old workplace I'm sure something could be found that would work? sadly that's not possible

John

[hermetic]

Could you not superglue it back together just to see how it works, my feeling is that the slope on the hook contacts the broken off pieces and is pushed to one side till it passes the "obstruction", and then latches, but it is very difficult to sus it out without having the mechanism in front of you! Could you cast one in aluminium using the repaired original as a pattern? At this scale the shrinkage may not be that important, you could add a few coats of high build paint to thicken it up a bit!
Phil

[mattinker]

So now the priority is to mend or replace the contactor.

B/ 3D model it, print in PLA and get it lost PLA cast in bronze.

Lost Pla's the way to go!

Cheers, Matthew
I'm leaning towards option B - it worked well when I needed that drive dog for the Beaver lathe live tooling. I'm not even sure my copy of Fusion 360 is still operable - not used it for over a year and they have a habit of chucking you out !

[awemawson]

Thanks chaps for your input.

Phil, those broken lugs take the full force of a very strong spring which is why they've broken off - sadly superglue won't be strong enough.

Yes lost PLA is what I'm shooting for.

This morning I steamed up my copy of Fusion360 and it forced me to do an update as I've not used it for a long time, and now it seems I'm on a 30 day limited trial.

Actually remembering how to use it is proving difficult - it's not at all intuative to me.

. . . . Where's Jules when you need a 3D model . . in fact thinking about it, he's not been arround in quite a while - I hope he's OK

[vtsteam]

Andrew, if only you had industrial quality cnc machinery and a foundry you could machine a pattern and cast and finish  it yourself. 

[awemawson]

Ho Ho Steve - you obviously have a career ahead of you on the stage 

So mangling my brain this afternoon I've produced an embryonic 3D model using Fusion360 - oh boy my head hurts ! It's printing at 20% fill (ie weak) using PETG as it was in the printer. This will let me see if it fits.

I've omitted the unnecessary features - extra  holes and other material reducing measures that were in the original - hence a model to see if it fits !

[ddmckee54]

Andrew:

If the part fits, you might want to try increasing the infill and print it again.  It might actually work for a while.

Don

[awemawson]

So the 3D print finished after 3 hours 21 minutes last night. It's come out well and it fits.

Not only that I was able to latch and unlatch the mechanism - as someone suggested, one of the lugs on the 'spring pot' pushes the hook latch up when reset. It's the upper single bar not one of the ones that broke off.

I was not at all surprised the the PetG print broke when I tripped the mechanism - in fact it broke just like the original !

I've subsequently noticed that the lugs that broke off were not 'merged' with the main body but rather are intimately close to it, so no reinforcement crosses their junction. Fine for lost PLA casting but no use as a component as is.

[awemawson]

Well you can't win 'em all 

Printing the PETG at a greater density gives it much more likelihood of warping and pulling away from the build table - I have a heated platen on this 3D printer but I forgot to turn it on.

I could see that it was beginning to curl away from the base but hoped that it would stay stuck to the end - no - it became detached and produced a birds nest.

Never mind - I'll do it again tomorrow with the heater on, but the print that I got here demonstrated dramatically more strength due to the 80% fill density (was 20%)  and the fact I merged the lugs into the main body so that they were one entity with the mesh layout running between them.

I suspect a good PETG print will have the strength to do the job, at least for testing, but it becomes quite weak at higher temperature so no good for the final item - that will be lost wax PLA cast in bronze. So PLA versions to be printed as well tomorrow.

[djc]

I wonder if it is possible to bypass or reduce the spring force on the ears of the die cast bit.

Drill a hole through the casing concentric to the spring and use a sliding, adjustable stop bolt (maybe with a rubber washer on the outside as a cushion.

Would lumiweld or similar repair the original metal part?

[awemawson]

I don't fancy introducing other bits into the breaker though I see your point.

My first thought had been Lumiweld, and indeed I have some, but decided not to do it at this stage incase it all droops into a puddle and I'd have lost my chance to check dimensions. When I have a working PETG model then I may well try it.

[howsitwork?]

Andrew

Having just returned to forum after a while away. I’m fascinated and terrified at the same time !
That’s some serious power you’re playing with but makes interesting reading . Stay safe .

And of course good luck !
Ian

[awemawson]

Yes it's all a bit scary Ian - definitely a 'one hand in you pocket' approach to probing needed !

Well there has been major progress on the 3D model of the bust 'spring pot' - I have a model that I'm sure will work when cast in something substantial like bronze.

However there has been another set back - I've found that the 'spring pot was not the only thing that has been damaged - and this time it's not going to be feasible to repair it 

The plastic moulding of the isolator has broken in two places:  firstly where the tube that slides the contact plate has over travelled, it slides in a groove and one end of the groove has broken away and the main body has cracked below it. Secondly there is a plastic formed collar around the operating spindle that has a stop lug on it that has broken off. The plastic is rather crumbly.

SO - I've come to the conclusion that even if I can have my 3D model 'lost PLA cast' and somehow re-build the plastic body of the contactor it will not be safe - so it has to go 

. . the search is on for a replacement . . 

[awemawson]

Now really I want a 'through door operated' 250 amp 3 pole 415 volt isolator with a 24 volt AC under volt coil'. Without re-mortgaging these fall into the unobtainium category.

Much more common are ones intended for toggle operation inside a cabinet with 'shunt coil tripping' - and I've ordered and am yet to receive an ABB SACE S3 225 amp isolator that is equipped with a 24 volt AC shunt. It's 'only' rated at 225 amps but then my generator is putting out about 150 amps on full load so I don't think that's an issue.

Now there are 'through door kits' for these breakers and I will certainly keep my eyes open for one but meanwhile just have to open the door to power up. It IS rather dodgy as the cabinet has loads of HV exposed bits - I may have to make up perspex covers.

The 'Shunt Trip' is needed for the Emergency Stop function, and applying 24 volts AC trips the breaker. In the original the 'under voltage coil' works in the opposite sense tripping the breaker if it is NOT powered so I will have to introduce a relay to invert the operation. Again 'under voltage coils' are available for this isolator but I've not found them for under several hundred pound which is a ridiculous price.

[awemawson]

As I will be re-working the circuit round the eStop I wanted to bottom out that peculiarity whereby the circuit diagram showed the ESTOP relay coil in series with the under voltage trip coil which obviously couldn't be correct as both are 24 volt ac devices powered from a 24 volt AC source.

Some work this morning with bright light and a meter has resulted in me finding the errors on the circuit and correcting it so at least I have a true starting point !

[vtsteam]

CNC Mill + touch probe + CAD + block of plastic = replacement plastic part

 

[mattinker]

As I will be re-working the circuit round the eStop I wanted to bottom out that peculiarity whereby the circuit diagram showed the ESTOP relay coil in series with the under voltage trip coil which obviously couldn't be correct as both are 24 volt ac devices powered from a 24 volt AC source.

Some work this morning with bright light and a meter has resulted in me finding the errors on the circuit and correcting it so at least I have a true starting point !

That makes more sense, real 24volt parallel !

cheers, Matthew

[awemawson]

Matthew it's not the first error I've found in the circuit diagrams for this beast. If the draughtman was given a hand written sketch of the eStop circuit I can fully understand where he went wrong confusing cable 304 and 301 and bringing 307 to the wrong transformer leg !

I've managed to source a pretty inexpensive ABB SACE S3H 250 amp contactor that has a 'through door actuator' but no door hardware or handle or shunt trip but seems the same body size as the ABB SACE S3 225 contactor that I have on order that DOES have the 'shunt trip'. So I am hoping to assemble a 'frankenmonster' from the two ABB S3 breakers and the door gear from the original Socomec breaker.

Mounting is going to be fun, as to avoid cutting another hole in the door the left hand side operated ABB breaker will need mounting a lot further to the right as the original has a right hand side operating handle.

The other issue is that I really want to move the supply to the eStop transformer to the switched side of the isolator which is going to potentially introduce timing issues ensuring that the new shunt trip isn't momentarily powered as the breaker closes.

[awemawson]

Well at long last I can make some progress with this repair.

EVRI / Hermes have lost BOTH contactors that I bought that were to be amalgamated to form one with a rotary operator and one with a 24 volt AC trip mechanism. The sellers of these used contactors fortunately both had additional items which I got sent by Post Office 24 and UPS - the UPS one arriving this morning.

This has allowed me to assemble a 'composite' having the trip and rotary actuator allowing the original Emergency Stop still to work, and the contactor to be operated 'through the door'. However this means I'm sticking with the original rather odd arrangement meaning that the emergency stop relay coil is powered even with the door isolator switch is off. As I'm generator powering the furnace this isn't too much of an issue as it only runs when the furnace is in use. Were it off the normal mains I'd somehow have to re-work it.

Now as you can see from the pictures the old and new contactors are vastly different in size, layout and termination. I have worked out a scheme where I think I can mount the new one so it's operating bar goes through the original door hole, and by re-wiring using 35 mm 'meter tails' with crimped tubular lugs can get it to fit - a bit of metal work to do for the actual physical mounting.

I've drawn it up in Autocad showing the outline and fixings for old and new putting their operating bars in the same location so I can lay out a 'hopefully' workable layout.

. . other duties call but I will return . . .

[awemawson]

Got a bit of time after lunch to make up a mounting panel from 6 mm polycarbonite sheet.

Dimensions seem about right but can't mount properly due to a local shortage of 3/16 UNC nuts ! - I'll probably make some but time presses being a Friday with cottage guests arriving - need to change into 'mine host' mode !

Ty-Wrapped on as the 2-3/4 x 3/16 bolts are only poked through the holes to locate it - still the square operating bar DOES poke through the hole in the door so it's in the ball park