Resurrection of a CFEI 100 KVA Induction Furnace

[awemawson]

The 11:00 deadline came and went so I fired up the 110 kva generator and did a simple 'bar heating' exercise to put a bit of load on the connections and wiring - all seems well 

The same bit of 16 mm mild steel bar that I'd used before was raised quickly to glowing point - a matter of a couple of minutes or there abouts.

There must be some sort of power limiting algorithm running on the furnace microprocessor - it 'topped out' at  59 kW this time - I suspect that it is based on the charge in the crucible. I must experiment when I have the place to myself with greater and lesser charges and see when the yellow 'limit' light comes on as I turn up the power.

[edward]

Thanks Andrew - I love your threads and the forensic detail you go in to with your repairs.

There is no better reason for doing something than because you love it. I've been building bike wheels recently, mostly because I wanted to learn the skill rather than any actual need to do it, and when it all comes together its just so satisfying.

Your Hovercraft video gave me flashbacks to a trip I had across the channel on one of the big Hoverspeed craft with the venture Scouts when I was in my teens. It was really rough, half our unit were up the front enjoying crashing over the waves and the other half back aft somewhere, filling a lot of those paper bags with the contents of their guts. Luckily I've always had pretty good sea legs

[modeng200023]

Now I'm sitting here willing my cottage guests to go out for the day so I can start up the big genny - I've sort of decided that if they'e not gone out by 11:00 I'm doing it anyway !!!!!!!

Seems like a plan 
John

[awemawson]

Well they've actually all gone out now - one couple were still there when I did the previous run but they were indoors so may not even have heard anything - after all it was over in five minutes - that's the good thing with induction heating.

SO - let's do a second run with more mass in the crucible and see what happens. I added a 550 mm length of 33 mm hexagonal mild steel bar to the 16 mm round that was already in it, fired up Big Bertha and again ran to glowing heat - just a few minutes. Limit came on this time at 70 kW / 147 amps (had been 59 kW  / 122 amps previously)

I'm now thinking maybe the electronics is measuring the resonant circuit peak voltage and putting a limit on that - it's not entirely consistent but first run the voltage was  646 v at limit, this run it was 639 v. Need much more data I think.

Anyway I learnt two three lessons in this exercise.

a/ Driver goes into limit at different powers depending on charge in crucible

b/ Don't worry TOO much about the noise - it's much less at distance and I'm up close.

c/ And DON'T poke your tape measure in the crucible to measure a bar length - even when everything is OFF it's still hot enough to remove the paint !

[Spurry]

c/ And DON'T poke your tape measure in the crucible to measure a bar length - even when everything is OFF it's still hot enough to remove the paint !

So you are human after all! 
Would also echo Edward's earlier comments about the thread, 'twill be a sad day when it ends.
Pete

[awemawson]

So you are human after all! 

Pete

Only too human I'm afraid Pete - you've seen a few cock ups in this thread surely !

And there was I thinking I was boring the pants off you all with my ramblings

I've been looking in detail at what circuit diagrams I have for this beast - the power electronics is comprehensively covered, but the control electronics is treated as a 'black box'. It's a single large PCB called a Saphyr Monocarte with loads of  microprocessor support chips and both digital and analog circuitry. Fortunately the manual labels the various connections to it, and I've managed to deduce that it IS monitoring the voltage on the resonant circuit (Internal massive capacitors and external crucible coil) through a step down transformer, so presumably that is the 'Voltage' displayed on the control panel, and it looks as though it has a limit set at around 646 volts as that value features in several of my pictures when 'at limit' independent of which furnace body is connected. The 'Current' displayed is the 3 phase input current and fairly closely corresponds to the value that the diesel generator indicates.

. . . it's all a learning exercise !

[mattinker]

 

Great stuff!

[kayzed1]

Just imagine how many shillings you would need to feed the old meter to run that

[awemawson]

Well it wouldn't be TOO bad.  Including the standing charge we are paying about 18 pence per KWH. So running at 100 KW for an hour would be £18. BUT the heats I did today both were only minutes long. Say 10 minutes to do something useful, so £18 / 6 = £3

Not cheap but not earth shatteringly expensive. I've no idea yet a comparable figure is for running my generator 

[awemawson]

Before I start using the foundry in earnest there's quite a bit of sorting out needed - stuff has just been piled up all over the place to let me get on getting the furnace working - well now it does so time to start the sorting.

Just outside the foundry sliding door, in the welding shop, were two 1 metre wide shelving units that historically have held electrical installation materials, rolls of conduit wire, adaptable boxes,  isolator switches - you know - the stuff that accumulates and is too good to throw out as undoubtedly it WILL be needed. Over the past weeks I've been re-deploying it elsewhere - largely successfully but you always end up with stuff with no home 

In the clear out I found a brand new emergency 'maintained light' unit - sensible to install it in the foundry if it still works - it's NiCd 4 AH 2.4 volt battery was down to 1 volt so it's now on charge but 'the jury is out' whether it will actually work.

The idea was to remove the two shelving units, put a back on one of them to stop it racking and also stop bits falling off at the back, and put it in the foundry for crucible, pattern, moisture meter etc storage where the 18" square clothing locker currently is where the PPE kit lives. This necessitated moving a foundry bench (that still has a load of Mansfield sand in it) sideways six inches, a feat accomplished with help from a friend yesterday.

So today I looked at the pile of sheet metal for a suitable backing sheet, and found a full 8 x 4 sheet of galvanised expanded mesh left over from building the Plasma Table extraction system - a result as I'm unlikely to need the mesh in future and it saves a sheet of useful Zintec for other needs.

I'd intended to lay the shelving unit down with the mesh on top, and cut it with an angle grinder, but then I remembered an electric hand held shear that I'd bought some years ago but found to be pretty puny. Rated for 1.2 mm it really struggled, but it ate this mesh like nobodies business  It was the inability of this shear to really work 1.2 mm steel that prompted me to get a guillotine so I shouldn't really bear a grudge!

OK back fitted, shelving unit installed, quite a bit of foundry related 'stuff' put on it AND I now have a two metre length of wall space in the welding shop that has allowed me to push the Linisher, the Corner Notcher and the Band saw up to the wall out of the way - all of a sudden movement is far easier!

The second surplus shelving unit has found a temporary home in the Tractor Shed - it needs fixing to something to stop in racking though.

I just need now to find somewhere for the clothing locker - the corner by the roller shutter would be perfect apart from the roof leak above it that I don't seem to be able to cure 

[awemawson]

So, after a suitable time interval I checked the NiCd battery and amazingly it's come up to a fully charged state

So I plugged the main inverter PCB back in connecting the battery and mains. The fluorescent light tube lit and the LED was OFF - it should be the other way round !

I then made a close examination of the PCB, which only has four transistors and two diodes plus a bridge rectifier - getting it lit up with a torch to read type numbers (I'd intended to replace the lot!) I noticed what initially looked like a scratch in the PCB but actually was a crack. It's cheepo paxolin not fiberglass so not very strong. I identified no less than six breaks in copper tracks (remember this is new never used!). Scraping the conformal coating off the tracks I made solder 'bridges' over the damage and tried again plugging it in. Result - it now works as it should 

Now solder bridges are not going to be very reliable, and if I decide to use it I'll put wire links in to replace the tracks. I seem to remember I bought a 'job lot' years back when I was equipping the main workshop - they were genuinely unused but one was declared as not working so it must have been this one.

Barely worth repairing as I can buy a new one for £25, but quite satisfying 

[Sea.dog]

"Barely worth repairing."

BARELY WORTH REPAIRING?   

Wash your mouth out man 

[awemawson]

Well Graham it's two and a half hours into a three hour 'run test' and doing well - 2.4 volt battery down to 2.23 volts.

So you'll be pleased to hear that I WILL put proper repair links on the PCB along with a bit of epoxy glue, re-assemble it and stick it over the sliding door entrance to the foundry - pictures will follow later !

[Sea.dog]

  

[russ57]

Amazing it worked at all with that damage...

The score so far as best I can remember...

Buy and restore 100kva generator
Buy, restore and sell trailer for generator
Clear and renovate outbuilding for generator
Diagnose faulty  chiller
Replace chiller with total loss heat exchanger
Find, test, restore bore hole
Install bore pump and pump house
Replumb irrigation system to allow use of bore water
Replace coolant hoses
Restore air over hydraulics
Repair emergency light
Install earth bonding between multiple outbuildings
Clear room for furnace
Build external storage cage for O/A cylinder storage (part of clearing)
Reinforce shelving
Install 3phase 'domestic' supply for 'ping'
Repaired cold saw
Provide enormous entertainment and education for the unwashed

No doubt some I have forgotten.

I'm hoping this is not the end, but maybe you have something even better in store...


-russ

[awemawson]

Gosh Russ, you at least have been reading my posts 

. . . scary isn't it - just to play at melting a bit of brass !

[awemawson]

This 'non-maintained' (meaning it's not on all the time) emergency light is rated to last 3 hours. My test showed it to actually have a run time of 2 hours 50 minutes. If I can't get out of the foundry in less than 2 hours 50 minutes I have other problems not solved by lighting 

So I went ahead and did the repairs. Firstly I cut a strip of good quality fiberglass printed circuit board and super glued it across the crack to try and stop further movement. Then I put wire 'salvage' links between the nearest component pads either side of the breaks. There was one exception, where the track is deliberately very thin to act as a fuse - in this case I scraped away the solder mask and soldered directly to the track. All salvage links were then fixed down with a few drops of super glue.

I seem to remember when I worked at Ferranti on the process control computers boards were allowed to have up to five salvage links, but they had to be identified with a short length of red sleeving to distinguish them from 'modification links' - you were allowed loads of mod links! Mind you these were complex multi-layer boards with oodles of ic's on board.

Having proved it still worked I then mounted it over the sliding door entrance to the foundry running the lead to a spare 13 amp socket in a dual that also feeds the roller shutter.

Seems to work OK

[John Rudd]

Ferranti computers?....
Argus GD700 by chance?

[awemawson]

Argus 400, 500, 600, 700. first one I worked on (Wiggins Teape #5 paper mill in Dartford) was an Argus 400 with Argus 100 input output racks !

My era was mostly A500's & A700's in many variants and applications but I did have an Argus 600 in the office that I'd programmed as a digital clock - about all it was good for though Barclay's Bank used to run payroll on them at Juxon House in St Pauls Churchyard - a hideous architectural monstrosity next to the Cathedral.

We did everything from the 4 minute warning (Russian Hoards over the horizon, deep RAF radar bunkers) to mixing dough for Ginger Biscuits (McVitie's Waxlow Road Harlsden) 

[John Rudd]

I was working at a gas terminal in Barrow-in-Furness during the 80's where they had a 700 series supervisory system. I had the job of site acceptance testing and commissioning of the control system....quite strange how the operating system was loaded and run from discs...

[awemawson]

We had quite a bit of kit on gas and oil rigs. Your Barrow one would be part of the Morcombe Bay field. I had staff on rigs there but loads of other places as well from Bacton in Norfolk to Port Harcourt in Nigeria as well as Orkney & Shetland with monitoring at Dyce near the airport.

Glad those days are over - my phone never stopped ringing day & night 

[John Rudd]

I'm glad those days are gone too, the Easington terminal that I moved to after, had a Taylor system, again with Hawke disc drives.....Great when they had a head crash!! not!!    Especially at 2-3 am ...and got called to site to repair/rebuild....

[awemawson]

Disks were a major step forwards. Mobil's Coryton refinery had two Argus 500's with Burroughs 2 megabyte fast disk storage - 20" platter, 100 heads each side flying on compressed air but these were fast memory access not for loading - that was from a paper taper reader.

I remember one night having to change the disk enclosure (a three man lift due to weight) and the reload took (I calculated) 14.1 MILES of paper tape. because it was from Fortran source code that had to be compiled on the way in 

I HAND WOUND every inch of that tape - it was a long night !

(An other enclosure change on a hush hush site involved taking the platter the length of the M6 under military escort  so that they were able to witness it being cut to pieces with oxy-acetylene - a bit extreme, but it did contain the NATO fleet deployment details !)

[AdeV]

(An other enclosure change on a hush hush site involved taking the platter the length of the M6 under military escort  so that they were able to witness it being cut to pieces with oxy-acetylene - a bit extreme, but it did contain the NATO fleet deployment details !)

It would have been just as effective, surely, to attach the platter to the back of the landrover on a piece of rope and drag it behind? By the time you reached the M6 it would have been ground into dust.... much shorter trip!

I was lucky (unlucky?) enough to miss those early days of computing; by the time I entered the fray, desktop PCs were normal, albeit not entirely across the organisation. A little prior to that, whilst at North Staffs. Polyversity (it actually changed to a "University" in the middle of my course), I got some experience with a DEC VAX machine, which left me with an enduring soft spot for VMS. To the point I currently have 6 MicroVAXes..... As much as I'd like an 8800, I don't think I've got suitable storage space for it!

[awemawson]

Ade there were a myriad of ways that we could have ensured total destruction of the data, ranging from strong magnets, to acids to dissolve the nickle plate (magnetic layer) off the brass platter, but when you are dealing with these official bodies, and their 'expert' has made a plan they won't budge !

The easiest way would have been to force a 'head crash' on all 200 heads by over pressuring the flying mechanism . . .but no . . they probably wanted a nice day out . . but all they got was Wythenshawe in Manchester