Resurrection of a CFEI 100 KVA Induction Furnace

[awemawson]

I've rather come to the conclusion that I need to do away with the crimped lug entirely, form a cup shape from the 3/4" pipe to receive the 12.25 mm diameter of the 70 mm copper wire, and directly braze the wire to the cup / pipe.

This actually is as I thought that I had done it years ago, and maybe when I open up some more 'ends' that will be revealed, but the lug on the outside of the copper tube is too big and needs to go.

I've had one of the brand new short driver cables on the workshop floor for a while - these are the ones with the wrong conical ends fitted. So I decided to butcher one to see how it was done. Turns out that basically they also have formed a cup in pipe, but it's apparently thick enough to be crimped.

So I'll be doing much the same but brazing not crimping.

The internal 19 mm bore of the pipe gives 283.5 square mm  and  12.25 mm diameter of the copper wire gives 122.7 sq mm, so the copper takes up 43% of the bore, and the water 57%

[WeldingRod]

I've got to say, I like the short hose version a lot better!

Sent from my SM-G892A using Tapatalk

[awemawson]

Sadly not easy to replicate though I wish that I could.

You’ll notice that the cable is a flat plait, and the tube forming the crimp is very thick wall, neither of which I have

[awemawson]

What I do happen to have though is eight 70 mm 'Butt Splice' crimps that I maybe can modify.

Objective is to modify half the butt splice to be able to encircle the 3/4" copper pipe, but leave a gap for water to emerge and use the other half to crimp the 70 mm welding cable.

So, cut half way through in the middle, slit half the length and form the resulting mangled bit of copper round a steel bar held in the vice. Then trim off corners, a bit of filing, then a run through the sand blaster.

Looks like it might work 

[awemawson]

So now it's a case of cleaning up the welding wire by immersion of the end in phosphoric acid followed by a good rinse, crimping the modified butt splice, cleaning any flash from the crimping to save tearing / cutting the hose as it enters, and pushing it in.

. . . and in it went - easy peasy 

The other end isn't going to be quite so easy. The rubber Brewers hose needs pulling back about five inches, gripping while the cable is gripped and cut to length before crimping on another modified butt splice brazed to the copper pipe. Once cut if the Mole Wrench gripping the cable lets go you are up the Swany as the cable will retreat back up the Brewers Hose never to be recovered 

Once the first crimp is made things can relax a bit - I made a total of three crimps each end to reduce the entire length of the butt splice to 'crimped diameter, and had to tidy away quite a bit of crimping flash.

After that it was relatively easy to push the fitting into the hose once the Mole Grips were released, as the compression of the hose was helping this time !

So that is one hose assembled bar the Band-It clamps and serving with heat shrink.

. . . one down three to go and at least now I have a process that seems to work 

[awemawson]

Last night I mass produced modified butt splice crimps ready for an intensive day of cable / hose making.

First cable this morning went together really easily, the 3/4" copper end terminals slipping up the Brewers Hose very easily with a drop of Hellerman cable lube.

The next cable was an absolute nightmare. When I stripped the ends ready to pull the old cable out for re-use I was met by cable end in an appalling state - really badly frayed. And pulling the bulk of the cable from the hose it wasn't in the best condition. My first reaction was that this must have been one of the original cables that I re-used, but looking at the big reel of 70 mm welding cable that I still have from all those years ago they both have the same very fine wire format wheras the other two cables had somewhat thicker strands.

No choice though - this wire cannot be re-used so I set to and stripped the insulation from four meters of welding wire off the reel, and crimped a lug on the end to pull it through the hose with the lug hole Ty-Wrapped to the eye in the end of my  6 mm fibreglass pulling rod.

Now this was exactly how I'd pulled the other wires through, but this time it was an absolute pain and put up a huge fight - no real idea why. I got to within 21 inches of the end when the Ty-Wrap (a stout THICK Ty-Wrap) snapped 

Shining a torch down the hose I could see the lug, but how to get at it? In the end I went fishing with a hook bent into the end of a 1.6 mm welding rod. First attempt got it half way before the hook straightened - second attempt was sucessful.

Next job, crimp the end fitting on and pull the cable back, pushing the fitting into the hose. First inch was fine but after that I paid dearly for every millimetre of progress despite loads of Hellerman lube on the fitting. But we got there in the end.

Then it was a case of compressing the hose back down the cable six inches, clamping with a Mole Grip, and crimping the other end on, followed by releasing the mole grip and pushing the fitting up the hose - this end wasn't too bad.

So (apart from Band-It clamping the ends and serving with heat shrink) that's three down and one to go. but I've called it a day - our first set of 'Cottage Guests' arrive shortly, we having been closed down by Covid-19 so I need to be able to meet and greet at short notice.

I did however remove the ends from the last cable, and sadly it also is the finer strands, so that's something to look forwards to tomorrow 

[awemawson]

Well I'm pleased to be able to report that the last cable wasn't as bad as it could have been. although the fine strands were a bit frayed at one end there was sufficient 'good' cable to make up the remaining assembly.

And even better - the lengths do seem to be better than previously and dress together nicely on a trial fit.

To complete the last hose, having had yesterdays major issue pulling the copper through the hose, this time I first pulled a length of 6 mm polypropylene rope through, then used that to pull the copper, rather than trying to pull directly on the rigid 6 mm fiberglass duct rod (which is hard to grip)

Having done a trial fit to check the lengths are now correct (they are!) I then flushed out each hose using mains pressure water to get rid of the last of the phosphoric acid and any debris that may have entered. The end fittings are 3/4 BSP Female with a cone internally, but using a stack of rubber tap washers formed a passable seal for this temporary exercise. Pleasingly they are all free flowing (as they should be) but when I dismantled the original hoses I found bits of hose lining jammed in the copper pipes that had been scalped out as they were pressed in. To try and avoid this I spent quite some time and trouble fairing the copper fittings, even buffing them on the wire wheel.

I'm now off to re-fresh my memory how to use the Band-It tool and do one or two test pieces !

[awemawson]

Reading the very scruffy and out of focus instructions that I had secreted in the Band-It box I decided that there must be better copies 'on the web'. Well, downloading one from the makers web site it told me to cut the stainless band before wrapping and tightening resulting in a loss of about 6 inches of band for every clamp - blow that for a game of soldiers - I've 24 of these to do - that's a lot of wasted stainless band

So I 'winged it' and in practice it's not too bad to do - with the roll of banding uncut it gets in the way of cranking the tightening handle, but that's simply cured by deflecting it with your thumb as you crank. In relatively little time all twenty four Band-It clamps were applied, and I was just thinking 'time for tea' as I still hadn't received the delivery of large heatsink tube to dress the clamps to save cuts on the sharp bits when Adrian the ParcelForce man hove up delivering my order from RS.

Tea will have to wait - eight 100 mm long slices of heatshrink were cut and shrunk over the Band-It clamps so the hose assemblies are ready to re-fit.

I'll not refit them until I've replaced the Chiller flow and return hoses with the new ones that were delivered the other day - I'd thought that I could use the existing Jubilee hose clamps but of course the new hose is much thicker wall than what is on, so that doesn't work. I had considered using Band-It clamps, but they would be permanent so I ruled that out.

. . . so more Jubilee clips on order - as I said to Adrian - See you tomorrow !

[awemawson]

The larger Jubilee Clips turned up just before lunch, so I've been able to replace the flow and return hoses for the Chiller with ones with a more substantial construction, with a far thicker wall.

All went well, not much coolant split, and I was able to install the newly re-made cable / hose assemblies. Again this went well with no leaks - well not entirely true, I'd left one coupler nut only hand tight so it weaped, but easily cured.

However the Furnace Driver will not go 'Ready' as coolant is not flowing in the left hand flow gauge so it's magnetic float is not lifting with flow. Now this is indicated as an 'Ext 1' fault on the control panel, which is what happened just before that hose burst.

I know that the new hose / cable assemblies all flow satisfactorily as I tested them, so there must be a blockage elsewhere. The documentation gives no indication of the plumbing so I'm going to have to 'knife and fork' it. It could be the furnace body or the furnace driver - time only will tell, but I'm now allocated to other things so that's a job for tomorrow.

[awemawson]

Well I couldn't just leave it, I had to have another look to keep my sanity.

First thing to check - there is an input mesh / screen filter that I've never been able to get at properly due to the way the flow and return chiller pipe block access. So a bit of heavy dismantling involving my biggest Stilsons got the big connections  undone so that I could remove the whole assembly, check it and put it back in a different orientation so that it was accessible in future. Now this screen is on the input, and coolant was flowing, so unlikely to be 'the' fault but it needed looking at. Sure enough there was a fair amount of debris that was easily washed away under the tap, but it wasn't totally blocked.

Then I removed the four recently re-made hose / cables and again proved that there was free flow down them - there was.

I then turned my attention to the 'furnace body'. There are two coolant paths, "A to B" and "C  to D" - I believe that the continuous copper tube coil of the furnace has a deliberate block to coolant in it's centre, and coolant flows in at the top and bottom, (A&C) and out at the centre (B&D)

Blowing with an airline there is free passage from A to B, but C to D is blocked.

Why ? I have no idea. Perhaps a bit of debris stuck in there? Or perhaps one of the 'QD' connectors have failed? Either way it's going to have to come apart.

 . . As you can imagine this is a bit of a set back to the project - let's see what tomorrow brings

[hermetic]

could you try a vacuum cleaner sucking in the opposite direction to the flow?  I guess you will have tried the airline both ways, but just a thought!
Phil

[awemawson]

Yes Phil the wet and dry vac was much in use as it also saves dumping all the coolant on the floor. All bits that could be sucked out were (behave in the back Jones Minor!)

But now I'm even more puzzled than before - I went out after supper to strip off the QD connectors - they were all free and passing my 'blow' quite happily, yet earlier those lower ones weren't Have I dislodged something - no I don't think so as I'd have seen it come out.

I now think that the 'furnace body' has been exonerated, but it's VERY strange.

Now the output from the furnace driver has four 3/4" BSP fittings for electrical and cooling connection. A&B on the top line and B&C on the bottom. The electrical connections are a massive pair of copper straps 'B to C' and 'A to D' but coolant I think passes from A to B and C to D

It all gets VERY confusing as seeing the back of the panel where all this happens is neigh on impossible due to a mass of pipe and hoses, but one of them has a distinct kink - is the kink enough to completely block the pipe? I have (again!) no idea but that going to be my line of investigation tomorrow.

Oh joy I thought that this was nearly finished

Have a picture of my spare Furnace Body coil - you can see the two pipe connections in the middle - there is no passage between them, they connect to top and bottom threaded fittings. All irrelevent now if it's not the fault anyway

[awemawson]

A long and very confusing day :

 - I was determined to precisely map out the coolant water flow from inlet manifold (from chiller) through the Furnace Driver with it's monitoring flow gauges, out to the Furnace body via the Brewers Hose / Cable assemblies, back into the Furnace driver and eventually returning to the Chiller. Now within the Furnace Driver there is an absolute tangle of hoses as not only is the Furnace Body water cooled, but the power semi-conductors and tuning capacitors have cooling plates and a radiator / absorber for the low power electronics.

Now the pipes can be seen at a distance through the left hand door, and most terminate or originate on two manifolds under a shelf at the front again through a door. By grasping a hose by inserting a 'litter picker' between the 415 volt buss bars (power isolated!) I could wiggle it, but not quite see the other end. If I could remove that shelf . . . Seven bolts later the shelf was free EXCEPT whoever built this machine bored a hole in the shalf and ran a cableform through it. Every other thing passing through the shelf has a notch so the shelf would still come out - but not this one No room to make the hole into a notch. Desperate times, so I started to plan rigging up a web cam when one of the multitude of visitors that have come today proved to be a useful extra pair of eyes 

So - plumbing map made and various tests performed. One necessitated pulling the pair of flow sensors slightly off the right hand panel to disconnect it's hose. It has a proprietary low profile hose clip allowing it to go through a small panel hole not much bigger than the pipe itself. It had to be replaced with a Jubilee clip, and the panel hole filed to accommodate the clip going back through !

OK - Tests showed NO blockage so I turned my attention back to the Furnace Body that I'd partially dismantled yesterday. I re-tested and proved flow through both it's water circuits, so all VERY odd - what was the previous fault as these things don't just 'go away' 

While I had the Furnace Body apart I decided to have a close look at the Quick Disconnect fittings.  Shortly after I'd installed them years ago I became aware of a problem with their retaining ball bearings - some were corroding very badly - I assume being a fitting suitable for water that they are stainless steel. I came to the conclusion that despite the fitting having a brass body, at times the balls were carrying some of the huge currents hence the damage. My modification was to put a link of 70 mm CSA cable round the fittings, with suitable female hex pillars brazed to their adaptors.

Well of course those balls are still grotty so today I found out how to dismantle them, the intention being to order up some definitely stainless replacements. The balls that I took out were in a shocking state and accurate measurement wasn't possible. They were 'about' 1/4" or 6.35 mm but the corrosion and pitting gave readings from 6.1 to 6.6 mm! Just on the off chance I had a look in my 'might come in useful' bearing box, and for some obscure reason had a bag of 'about 1/4" ' balls - no idea what they are left over from, but they measured 6.33 mm and were magnetic so presumably not stainless. Well they fitted nicely, the QD worked like new and this confirmed me in my suspicion that the originals were 6.35 / 1/4". a bag of 100 now on order from Simply Bearing, delivery due tomorrow

So the plan now is to re-assemble everything, wet it up and do some flow testing (when the balls arrive)

 . . as I said, a confusing day 

[russ57]

I dont quite follow the flow diagram. The arrows don't 'flow', and one of the 2 blue 'one way' elements has contra flow..



-russ

[Sea.dog]

Blue is cold and red is hot, to state the obvious. I think that Andrew said that the pipe pairs are A&B and C&D. Therefore you have two cold lines in and two hot  lines out.

[russ57]

 :{
Should have enlarged it... I didn't see the 'manifold' words - there is obviously flow between the in and out manifold not shown...

-russ

[awemawson]

It is very confusing looking at the actual machine which is why I wanted to do the diagram to sort it out in my head.

The manifolds connect to flow and return on the Chiller, and as Graham says the pairs to / from the furnace body are A & B, and C & D

But then additionally there are the electrical connections. B & C are strapped together as are A & D and these two connections connect  the coil of the Furnace Body, and the internal capacitor bank and form a tuned resonant 'tank' circuit

Once drawn out it’s actually pretty simple.

[awemawson]

The new ball bearings arrived just before lunch, so I've just stripped and rebuilt all four quick disconnects.

I'd intended to clean up the inner brass-work with a bit of steel wool, but as I'm not changing the O ring seals decided against in case a strand of the steel got trapped under the rubber.

They came apart and went back together surprisingly easily - I used a tiny smear of Vaseline in each ball housing to hold the balls - I didn't want loads as although theoretically the coolant path should not come in contact I didn't want to risk Vaseline trace in the coolant.

Intriguingly the new balls measure 6.33 mm not 6.35 - just like the bag that I already had. Now my first thought was that the difference was due to using a digital vernier guessing stick rather than a proper micrometer, but as I happened to have my metric slip gauges handy I measured a 6 mm slip and it was dead nuts on. This digital vernier has always been surprisingly accurate and I had to apologise to it for doubting it's voracity

So is the 0.02 mm a looseness for normal bearing fits 

Sadly I can't do more this afternoon. Now lock down is lifted on holiday cottages our bookings are picking up and today is a 'double change over' with two sets of guests leaving, and two arriving, and the Covid-19 cleaning regime required is unbelievable and very time consuming.

Tomorrow should be better and I can then hopefully re-fit the Q/D's and hoses and resume testing 

[awemawson]

A bright and early start to put it all back together for testing - and also a nice bight day to go with it 

First I refitted the Flow and Return pipes for the Chiller completing the coolant circuit between it and the Furnace Driver.

Then on the Furnace Body I re-reinstalled the four Quick Connects and their associated linking straps that hopefully will preserve the newly installed stainless balls. Then the four new hose / cable assemblies from Furnace Body to Furnace Driver went back on, and that's about it . .

. . . time to turn on the coolant pump and see what happens. Two sharp 'click' sounds - I know what that is, it's the Flow Gauges - their magnetic torpedoes rise with flow knocking the top of the gauge, a good sign, actually a VERY good sign as previous symptoms were no flow in the left hand  one.

Now previously the pump bypass tap had to be fairly closed, (ie higher pressure) for the flow gauges to rise, but now I can open the tap fully (ie lowest pressure) and they still work. I do remember that as the pressure was adjusted from low to high the right hand one lifted first, followed as the tap was closed further, by the left hand one. So there must have been a partial blockage for some time - symptoms were there but I hadn't recognised them.

So, in for a penny, in for a pound - put 'local power' onto the furnace driver and see if it goes PRET (ready) - oh yes. I then intended to 'ping' the Furnace Body to confirm resonance but inadvertently had the key switch in 'CHAUFFE'  (heat), and I got the characteristic 2.4 khz humming and 15 kw indicated as power consumed. Quickly pressing stop, and feeling the 2.5 mm 5 core cable from local mains luckily no harm done, but I wouldn't normally try and actually HEAT something without having it powered by the big generator - phew got away with that one.

In conclusion, there must have been a  partial blockage previously, and like a blood clot moving about the body, got to somewhere vital and killed things. My various activities over the last few days have cleared the blockage but I'm no closer to knowing where it was apart from that is was in the left hand Flow Gauge / 'C' Port / Inlet Manifold loop UNLESS that input mesh screen was the issue all along. (but if so why was the right hand flow gauge not affected as well?)

Now I've disturbed a large number of coolant connections in this investigation (17 at a quick count!) so I've left the coolant pump running while I write this and have a coffee, let's hope no leaks and no floods when I return.

[russ57]

-russ

[awemawson]

Thanks Russ 

I'm glad to say, still no leaks despite having the pump running for several hours - I took to turning it off for a bit then back on reasoning that the start up is probably maximum strain - I also varied the pressure up and down. In retrospect it would have been handy to fit a pressure gauge to the coolant system to help notice any changes - maybe later.

I've spent the afternoon wrestling serpents again  When I fitted the long umbilical from generator to wall socket I didn't have the right sized cleats to fix it properly, so used some humongous Ty-Wraps. The proper cleats arrived several days ago but the coolant blockage took priority. I also decided to run it on the opposite side of the RSJ as it lay better there. Blooming heavy and although supposedly flexible it still has a mind of it's own.

Next job : sort out the earth bonding . .

[awemawson]

Getting ready for a CPC (Circuit Protective Conductor) - or Earth to people of my age  making session tomorrow I though it wise to check that the tubular crimp lugs were compatible with my crimper when using the 70 mm CSA cable that I intended to use, as these are unbranded generic ones.

As the cable sheath is Black I'm heat shrinking Green / Yellow onto the ends partially encapsulating the lug as support, and I wanted to check that the 'shrink ratio' brought it down to size.

All's well so let bonding begin . . tomorrow . . I've had too much excitement today 

In actual fact these 70 mm CSA tubular crimp lugs crimp much nicer than the tubular butt splices that I used in making the hose / cable assemblies, with far less 'crimp flash' where a bit exudes and leaves a sharp fin. Same crimper, same cable.

[awemawson]

OK Bonding Day !

Simple really, drill a few holes, tap a few holes, crimp a few tubular lugs and shrink some heat shrink sleeving on, what could be easier

Well very oddly my Cembre HT45  hydraulic lug crimper (of 1985 vintage) decided that today was the day to start playing up - it was somehow hydraulically locked and would neither open nor close. After a bit of an oily session on the bench dismantling its pump mechanism and its release valve, cleaning and putting it back together normal service was resumed - I'd only used it late yesterday afternoon to do that test crimp, and it was fine then - very peculiar.

So, slight delay starting, but I've decided to bond the Furnace Body, and Chiller Unit to a common point on the roof of the Furnace driver, and from there to the steel frame of the foundry building.

The very fine strands of the 70 mm CSA welding cable that I'm using make it nice and flexible, but it's a bit of a challenge to get the strands all into a crimp lug at the same time and also manipulate the crimper. I hit on the idea of compressing them with a small Ty-Wrap right at the end of the stripped length, then pushing the crimp lug on, which pushes the Ty-Wrap back towards the un-stripped cable sheath. When sufficient is in to be stable I then cut the sacrificial Ty-Wrap off pushing the strands fully into the lug. Unconventional, but it works for me !

So everything in the actual foundry building is now bonded together with ridiculously heavy cross section cable - better bigger than smaller 

Yet to do is bond the Generator, and the frames of the four buildings to each other.

I should perhaps explain:

A/ My Welding Shop is a steel framed 'lean to' added to the South side of my main workshop, and is made of bolted together RSJ's and angle iron, fitted to flying brackets welded to the steel frame of the main workshop.

B/ My Foundry is a steel framed building added onto the west end of my Welding Shop, and although again of RSJ and Angle Iron construction is fully welded but as far as I can tell is 'free standing' as far as the steelwork is concerned, and just joined by the roof so no real electrical connection between the two.

C/ The 'Stable' where my big generator sits is another 'lean to' added to the North side of the main workshop, and is made from RSJ uprights and  rafters as three frames, but they are held apart by timber purlins and free standing so again no electrical connection

So bonding this lot together gets a bit complicated. Certainly I will bond the generator frame to the adjacent RSJ frame of the stable, and try and get a cable through the wall to bond them both to the steel work of the main workshop. And I will try and bond the foundry, main workshop and welding shop frames together where they all come together at a corner. But I don't think I'm going to go as far as bonding all the stable frames together - not fully decided - we'll see!

[awemawson]

This afternoon, despite interruptions (*) I managed to bond the Foundry steel work to that of the Welding Shop and Main Workshop - this went quite smoothly when I was able to do it !

Then I turned my attention to the big generator. Its main earth terminal is inside the control box down at nearly floor level, but I wanted to bring the cable in from the top through a gland, so that it more directly could run to the nearest RSJ frame. This was a bit complicated as it involved removing the safety covers, the perspex one in particular is a pain to get in and out (I know - I made it, it's my fault  )

I still have to bond the two RSJ frames in the stable to both the generator and the steel work of the main workshop but that's a job for tomorrow - I've had enough!

(* a friend presented me with a bolt - "I'd like one almost like that but a different thread"  Well the bolt he had brought was 1/2 UNC and apparently it screwed in a bit but was loose. Now this is off a (possibly) Japanese grey import digger so the thread could be almost anything! It's a critical application as it's holding the main hydraulic pump so best get it right. Pump currently held by one bolt and is too heavy for him to get back in place single handed if removed for measurement. In the end I turned him a tapered piece of delrin and told him to force screw it in, back it out and bring it back for inspection - might work    )

[awemawson]

Another fabulous bright and sunny day, so I was up with the lark, and had finished the last of the bonding between the building frameworks by 8 am 

. . . perhaps now I can get on with something a bit more interesting and less neck cricking