Resurrection of a CFEI 100 KVA Induction Furnace

[Sea.dog]

Steady on, it's not like you to get tetchy 

[awemawson]

Sorry Graham if I seemed to bite your head off - it wasn't intended.   

. . . but it has been a frustrating day not achieving anything like as much as I'd hoped

[awemawson]

The plot thickens !

It seems that the spare furnace coil has one set of threaded connections and one set of clipped connections, so not easy for the pinging test 

Digging out the ends of the long pipes (ends were wrapped in plastic woven ex army sand bags) they do indeed still have the Quick Disconnect fittings, so it looks like a bit of pipe extraction is called for tomorrow and I'll hope to get them connected.

[awemawson]

So I woke up this morning with a bright idea ! Take the QD fittings off the long hoses, fit them to the short ones, and you don't have to struggle manipulating long heavy hoses 

Well the best laid plans of men and mice . . . . turns out that the short hose have a cone fitting and the long ones a face fitting . . .ah well it was a good idea while it lasted !

So I dragged the long hoses out from the tractor shed, tried to empty as much water from them as possible then, like fighting an anaconda, got them into a wheel barrow for their journey to the foundry.

While the QD couplings were off the hoses I gave them a bit of a clean up while passing my buffing spindle with the new bearings - well you have to don't you 

Then it was a case of identifying what went where. The Driver end was easy enough  as the terminal plate was marked, and I'd obviously labelled the pipes with a Sharpie before removal all those years ago, but only one mark had survived. Far end of pipe labelled to correspond and I started dismantling the terminal box on the  Inverting Furnace Body.

I'd forgotten that my solution to a non-conductive, non-magnetic strong frame was to make a sort of 'Fish Tank Frame' from epoxy fibreglass - I'm quite impressed with my effort - I even have thread inserts  to retain the panels. These were asbestos cement when I got the body, but I replaced them with cement 'tile backing board' - it's a bit friable and if this project continues I'll give them a coat of Sodium Silicate solution to firm them up a bit.

Getting the QDs onto their respective fittings was really quite a struggle - those pipe fight back at every opportunity. One of the QDs shows evidence of the retaining (stainless) ball bearings having at one time carried some of the current. That's what that huge copper strap and 70 mm cross sectional  jumper is for as it shorts across the QD - IIRC it was a retrospective mod when I had the problem before. But those fretted balls made locating the fitting quite a problem. It's OK for the ping test but if I get as far as pumping water through this beast they will have to be replaced.

So - drives cables / pipes (whatever you want to call them) are now connected and I need to work out how to safely feed the big blue box with a bit of 415 volt three phase for the ping test.

[modeng200023]

Andrew, that really seems to be a beast waiting to be taimed. One thing is for sure, it makes very interesting reading about your struggles trying to tame it.
John

[awemawson]

John, at the moment I think the beast is winning !

This afternoon I wired a temporary 16 amp 415 volt three phase feed into the machine to try and run up the electronics and 'ping' the coil in the crucible to establish resonant frequency.

It starts the initialisation process then comes up with an error saying DEF - obviously more investigation required !

[John Rudd]

Well it seems no amount of button pressing is going to fix it anytime soon....

So that Green button should latch something? Or it is and the latch is dropping out?

Is the DEF error a shortened version of some French word as most of the wording on the panel looks French?

[Pete.]

I reckon that seeing as the chiller is broke it means defrost 

[awemawson]

One issue I have sorted - I found a picture of when I pinged the coil back in 2006 and the L1 L2 L3  phase sequence indicator on the front panel is illuminated and the system says Pret (ready). So I have reversed two phases and now the indicator is illuminated 

Unfortunately the system still says DEF which I suspect refers to Default values, so maybe it's lost a few parameters (which will be a nightmare to re-create!) however it's back up battery is still showing a steady 3.7 volts

The "electronique" indicator is not showing, but is on my picture from 2006 but I can find no reference in any of the documentation to say what it means. Note that the "micro" indicator IS illuminated which I take to mean that the microprocessor is running.

Note that the "electronique" indicator IS NOT illuminated during initialisation in the 2006 picture.

Note: Pictures below are from 2006

[mattinker]

Défaut = fault. Par défault mean default!

Have you got any of the French docs?


Cheers, Matthew

[awemawson]

Yes Matthew I have a manual which includes diagrams of all the power electronics but NOT the main microprocessor board sadly. The text descriptions are fairly sketchy as these things often are, but I had thought I might photograph a few pages and pop the over to you to peruse if I may.

Or I may run it through scanner and turn it into a .pdf - might make it easier - I'll kick a few things about tomorrow.

 

[mattinker]

Yes Matthew I have a manual which includes diagrams of all the power electronics but NOT the main microprocessor board sadly. The text descriptions are fairly sketchy as these things often are, but I had thought I might photograph a few pages and pop the over to you to peruse if I may.

Or I may run it through scanner and turn it into a .pdf - might make it easier - I'll kick a few things about tomorrow.

Great, could you include a photo of the manufacturer's plate in the pdf?

[awemawson]

Well a day of two halves with greatly differing results !

I spent the morning scanning the manual into 'searchable pdf' format to send to Matthew - I have two versions - one came with the machine but is mainly schematics of the power electronics, and a second that CFEI sent me back in 2005 that has the same drawings but also quite a bit of text that looks to be mainly 'over view' type details in French. I started on the original manual as it is only 20 pages - when I opened the scan as a PDF it is NOT searchable - and I've tried all sorts of options in the Plustek application - total of 4 scans so far and still not searchable 

So I set too thinking and closely examining those pictures from 2005 and what was happening today. Does a red indicator ON mean a fault - or does it go OFF for a fault. Should that 'phase sequence error light' be on or off 

Well studying the circuits that I was scanning showed me that the two red lights indicating faulty 'Temperature' and faulty 'Ext 1' corresponded to two lines on page 10 feeding 24 volts to inputs on the main logic board. The temperature line was a series of  'normally closed' temperature sensors and interestingly that bit of the circuit has pencilled alterations as well. The 'Ext 1' one was apparently a simple link.

So out with the meter, and both those inputs are NOT at 24 volts - presumably an error 'up the chain' somewhere.  Make up short temporary links, put my fingers in my ears and :

WE CAN PING

Obviously I need to chase the faults 'up the chain' and see what's causing it, but certainly the microprocessor board seems to be OK and the power electronics must be reasonably healthy to kick the furnace coil into oscillation 

(It didn't help that the circuit diagram has '0 volts' and '24 volts' incorrectly drawn in each others connector block ! )

[John Rudd]

Great to see you making progress ..

Looking forward to the next instalment.....

[Pete.]

My phone has an 'image to OCR' function Andrew if you want to send me the PDF I'll attempt to make it searchable for text.

[awemawson]

Thanks Pete but I've not given up with it yet - I have other bits of software to unleash on it still !

So - what happens if I put a bit of steel in the crucible and 'ping' it again? Theoretically the frequency should be lowered - but is it ?


. . . . YES 

(This just confirms that it IS actually measuring the frequency and not just throwing up random numbers !)

[Sea.dog]

Fantastic news Andrew. In my book when a fault lamp is illuminated it indicates a fault. Have you determined what the three sensors do? (I see that originally it was two, but they've been scratched out and a third drawn in)

[awemawson]

Graham, yes I agree on for fault, but the picture of it 'ready' back in 2006 had the phase rotation light 'on' - in fact I have proved today that it has to be 'off' in fact so not sure what was going on then ! (*)

The various sensors all seem to be thermostatic one on things like the water source, internal big coil, the tuning capacitors etc.

The simple link in the other chain is something to do with the key switch that switches between the ping mode and normal heating.

More investigation needed tomorrow - at the moment I'm mentally exhausted 

(*  Later Edit :No it doesn't - only the first picture from 2005 while it is initialising shows the phase rotation fault indicator 'on' . I must have swapped phases for the other pictures !)

[hermetic]

Excellent news Andrew, I have complete faith in you! On the chiller, lots of food industry factories use brine tanks for fast freezing, also ice cream and chilled desert makers, have you any contacts within aforesaid industry that might be able to link you to a chiller unit! You have probably already considered this, but just in case......Good luck! I can smell hot metal even this far north!
Phil
East Yorkshire

[awemawson]

Thanks for the thoughts and suggestions Phil,

Matthew - I have emailed you the searchable .PDF scan of the manual and also a picture of the machine plate as you suggested. The file is pretty big at 8 Mb so I hope that your eMail server doesn't bounce it.

[awemawson]

So this first of the 'fault indications' that I linked out yesterday is solved 

Taking my handy Telecoms cable tracer to the job I followed the wiring to where, according to the circuit diagram, there is a simple link between the 'Saphyr Monocrate' digital input, and 24 volts. No! The cable wanders all over the place ending up on an undocumented 110 volt contactor where this 'link' is the one and only used contact. But what drives the contactor?

Again using the cable tracer lead me a merry dance ending up on a  3 core cable Ty-Wrapped to hoses deep in the back of the one panel that doesn't open Poking a camera in showed some sort of junction again Ty-Wrapped to the underside of the hose with a black cable emerging next to the hose to / from the water flow gauges. Ah - the penny dropped - the flow gauges are not just visual, they have a magnetic float that operates a reed switch THAT OPERATES THE CONTACTOR 

Some rare earth magnets and masking tape simulated water flow, and sure enough that's one "fault" knocked on the head  I must have been through this process all those years ago I suppose, but it's lost in the mists of time.

A quick cup of coffee and on the the next fault chain - the temperature sensors.

[John Rudd]

So this first of the 'fault indications' that I linked out yesterday is solved 


Some rare earth magnets and masking tape simulated water flow, and sure enough that's one "fault" knocked on the head  I must have been through this process all those years ago I suppose, but it's lost in the mists of time.

I love a bit of fault finding, keeps the old grey matter working....

So I hope you are going to document all your findings this time...

[awemawson]

Well John this is all VERY odd !

I've spent quite some time cable tracing - some of the temperature switches are literally impossible to get at, they must have been fitted before the machine sub-assemblies were joined, and others required quite a bit of contortion to get at.

BUT I found a switch on the water inlet manifold that was open circuit and accounting for that  series line of about five switches not being pulled to +24 volts hence the error. I pulled the wires off the switch and linked them proving the 'fault'

However, when I finally got a spanner onto the flats of the switch and got it out, it turns out to be a pressure switch monitoring the inlet water. So hiow the heck all those years ago did I manage to 'Ping' this beast with no water source and certainly no pressure on that switch - VERY ODD.

Anyway at least I've found what's doing it - now I need to decide how to proceed.

[awemawson]

So having tested that it works, I returned the water pressure switch to it's place on the water inlet manifold, but left a very obvious jumper link across it to prevent me forgetting I'd done it.

This afternoon I dug out the hoses that used to run between the (now defunct) Water Chiller and the Furnace Driver - you may remember, when I first got them from their rotten storage pallet they had been eaten by rats. Well it turns out that the damage was confined to one pipe only, fortunately the longer one (previous location had dictated unequal pipes) and less than 24 inches from the end. I was able to re-terminate that hose with the existing fittings and make two equal length hoses. Tough to press in those barb fittings, but much aided by Lambing Birthing Gel - basically agricultural KY Jelly 

So if I ever get another water chiller at least I have a starting point for the hoses.

Hence now the hunt is on - I need a water chiller that will provide water thus:

 - 25 litres /minute

-  3 bar minimum pressure

-  5 bar maximum pressure

-  40 deg C Maximum temperature

-  5  deg C Minimum temperature

The previous one consumed about 11 kW and provided 39 kW of cooling - I always felt that that amount of cooling was barely sufficient however I had interposed a second heat exchanger in the form of a domestic central heating system copper hot tank as a sort of negative thermal store, but I now suspect that the actual problem may have been thermal transfer (or lack of it!) between the main tank and it's internal coil.

So those of you who scan commercial on line auctions can you please give me a 'heads up' if you see any big chillers listed.

[awemawson]

Happy Easter everyone.

At a bit of an impasse until I can sort out the cooling so I've turned my attention to sorting out the foundry itself.

This morning I have removed EVERYTHING not directly foundry related, (and some that are!) to generate floor space, and have moved all the equipment up to one end of the building. I was even able to open the small pedestrian Roller Shutter Door for the first time in years - if you look through it, the round plastic water cistern is what I used to use as a duct sitting upside down on the Water Chiller with a large rectangular hole in it's side holding galvanised duct to the outside world. And inside it I'm glad to say I've found the safety cover that keeps fingers off the Driver output terminals.

This has allowed me to sweep half the concrete floor, vacuum it with the workshop vac and give it a coat of Sodium Silicate to hopefully seal it a bit and stop the dust rising. Google gives me contradictory advice regarding dilution ratio, but I used 1 part very syrupy heavy Sodium Silicate  to three parts water. That half of floor so far has used 2 litres of concentrated Sodium Silicate.

Will it work - well again there's contradictory advice 'out there' - so long as it dries it's can't hurt at least 

. . . so now off to plot and scheme cooling methods - joules / kilowatts / degrees Centigrade and all that stuff !