Please click here to link back to the last instalment of my CD Welder development.  This page is essentially development of applications for a working prototype and a number of design refinements.

10 March and I have started to think about pulse arc (and pulse cutting) applications.

Pulse arc welding is very short duration TIG welding.  In essence you form an arc in your shield gas but it only lasts for a few milliseconds.  It is the high temperature arc that does the melting.  This is fundamentally different from resistance welding which relies on I^2 R losses to generate the heat for the weld.

TIG welders use a high voltage low current RF discharge to initiate the welding arc.  With needle points it takes about 1 kV/mm of gap to initiate the discharge.   This technique will be difficult to implement with my welder due to the over-voltage and reverse bias protection across the MOSFETs, even at high frequencies.

Some commercial devices use a retracting tungsten electrode to initiate a single pulse arc under Argon (or carbon dioxide) lasting up to a few milliseconds.  While this works well it requires a fair bit of development and machining to make a solenoid retractor, even starting from a commercial TIG hand piece.  However it has the advantage of using commercial TIG consumables including electrodes and gas shields, and the gas path is established.

The first step is to develop a reliable detection circuit that will sense when the electrode is in contact with the work.  This turned out to be incredibly easy as shown diagrammatically in Figure 1.



Figure 1.  Electrode Contact Detection


Note that the welder cable has been shown as an RCL transmission line which is essential to understand the need for the Transil and free wheeling diode transient suppression.  It is these components that prevent the use of a high voltage RF generator to initiate an arc.  The only thing special about the trigger detection circuit is the constant current source.  This uses a high voltage Vce BJT to ensure that the transient suppression will kick in and prevent over-voltage failure.

The solenoid for electrode  retraction (top left of Figure 1) will need some work.  I figure that 100 turns at about an Amp will produce around 1 kg force if the iron core is close to the coil.  I can also mount a small rare earth magnet on the end of the electrode.  This needs some experimentation.

The other design challenge is ensuring a really good high current connection to the electrode.  I have a number of choices including springs, brushes or copper braid or a combination of these.  I have thought about coupling the discharge current through the solenoid but this is likely to make control of the electrode movement very dependant on the weld parameters.

The few commercial designs that I have looked at use relatively thin cables in an attempt to limit arc current to a few hundred amps.


Transils and Bus Plating

21 March and my new Transient Voltage Suppressors have arrived.  They were rather expensive but will individually handle 15,000 Amps over an exponential decaying 8/20 us pulse.  I have found the most suitable position for mounting one on the bus, but this requires an additional hole to be drilled - which means taking the bus to pieces again!



Figure 2.  New Transient Voltage Suppressors


One of the problems with using copper for the bus is that it rapidly tarnishes, particularly anywhere where it has been touched.  I have gone to some lengths to clean the bus elements and used rubber gloves during assembly but over time the tarnish eventually reappears.  Mark (builder of the welder from hell) has silver plated his electrode terminations.  As I’m going to have to disassemble the bus to drill that hole I may as well try silver plating while I’m about it.

Time for some bucket chemistry.  I don’t want to be electroplating if I can avoid it because the bus elements are quite long (and therefore require lots of solution) and they only need plating at component and physical terminations.  There are a few recipes for plating pastes that are relatively straight forward to mix up that do not involve cyanide.  All I need to do is find one that I can make that works.

My first attempt comprised a mixture of 2 g silver nitrate, 1 g ammonium chloride, 4 g of sodium thiosulphate (the original recipe actually calls for sodium bisulphate) 4 g of distilled water, and enough sodium carbonate (the original recipe calls for potassium carbonate) to make a paste.

With the mixture made I cleaned a section of the bus with a scourer, washed it with hot water and detergent, rinsed it in distilled water, dried it on a paper towel and masked off the end with electrical tape.  With gloves on (silver nitrate will stain your skin black as it converts to silver chloride) dabbed a piece of paper towel in the paste and scrubbed it over the area to be plated.  The silver took instantly.  The coating is very thin but very secure and on rinsing and polishing, holds a good shine (see Figure 3). 

Some ammonia was given off during the plating so good ventilation is required.  I suspect that this could also give rise to the production of silver nitride which is a rather unstable explosive so I won’t be keeping the paste very long.

I’m going to leave the test piece for a couple of days to see how it stands up to finger prints.  Then I can be getting on with what I hope will be my final disassembly of the bus.



Figure 3.  Silver Plated Copper Bus Test


24 March and after three days of exposure to finger prints and air my silver coating has developed some yellowing and surface oxidization as shown in Figure 4 (note that the copper is also heavily coated with oxide).  I suspect that the yellowing is due to oxidation of the thin silver coating, exposing the copper substrate however it could also be due to residues from my plating paste.  Although the surface discolouration is readily removed by light buffing with a paper towel this isn’t ideal - particularly as this results in exposure of the copper.  Unfortunately I don’t have the instrumentation to measure the associated degradation in surface Ohmic resistance.



Figure 4.  Silver Plating after Three Days


While re-plating restored the silver coating immediately to its former condition I don’t want to be doing this every few days.  Having renewed the silver coating I have kept my fingers off it and masked half of the plated area from air using electrical insulation tape.  The yellowing still appeared under the tape although it was significantly less than the exposed plated section. On doing some more research this yellowing appears to be a recognized issue with some forms of electroless silver plating systems.  I have a few more things to try:

  • Try a different electroless recipe (I have another three non-cyanide solutions than I can readily make).
  • Revert to a non-cyanide electrolytic process.
  • Use a cyanide complex (or send the bus elements out to an electroplating company).

I’m going to try Option 2, initially using the following plating solution.  13.7 g silver nitrate (original recipe calls for 11.5 g silver chloride), 36 g sodium thiosulphate, 2.2 g sodium metabisufite (original recipe calls for 4.25 g sodium bisulphite), 10.5 g sodium sulphate made up to 1 litre of solution.  I’ll also add about 1 g of carbon disulphide as a brightener (only slightly water soluble at 2.9 g/l at 20°C).  While I’d like to add a cyclic ketone I haven’t got anything suitable - but I might try some a millilitre or two of acetone as well.  The electroplating current density is 107 A/sq m (~10 A/sq ft) using a bright steel cathode.

There are some questions that you might be asking:; why I am bothering with this? and why are cyanide based electroplating processes better than wipe/dip processes?

My welder’s maximum discharge voltage is 20 Volts (currently set to 18 V) with measured discharge currents of about 10,000 Amps into essentially shorted welding electrodes.  Every 0.0001 Ohms in the discharge path is dissipating over 5% of the instantaneous discharge power.  Minimal discharge path resistance equates to more energy being dissipated in the weld junction.

The answer to the second question is literally more complex.  Electroless plating of silver using non-cyanide processes involves only the exposed copper surface and once this has reacted the plating ceases leaving a very thin silver coating.  Sodium (or potassium) silver cyanide complex in the presence of excess cyanide creates a silver buffer that maintains a very low silver ion concentration in solution throughout the electroplating process.  This is apparently necessary to obtain a coherent deposit.  The chemistry is explained nicely in  ‘Mellor’s Modern Inorganic Chemistry’, G.D. Parks (Ed.), Longmans, Green and Co., London, 1951, page 630.  Don’t you love old chemistry books.  They contain a wealth of information about chemical processes that are increasingly no longer available in the public domain.

28 March and I have completed two electroplating experiments based on the recipe above.  The first attempts didn’t perform well at all.  The plating was thick but dull and did not adhere.  The electroplating solution was initially yellow (indicating free sulphur) but gradually turned black due to the formation of colloidal silver sulphide during the electroplating.  I added sufficient sodium hydroxide until the solution went clear (about 5 g).  This increased the pH and conductivity of the bath. With a reduced current density of 30 A/ sq m I obtained a sound coating with a semi-bright finish after 60 minutes as shown in Figure 5.



Figure 5.  Electroplating after 24 Hours
(Note that the background is actually matt white.  If I get the background colour correct through adjusting the white balance then the plating finish is lost in the glare)


 I suspect that the surface brightness is somewhat reduced to the surface preparation of the copper.  This coating shows no signs of yellowing or degradation due to finger prints after 24 hours.  While this is certainly useable there are some other recipes that I want to try with a range of brighteners.  Unfortunately I don’t have all of the chemicals I need on hand and it will take a few days to synthesize these.  Please bare with me.

30 March and its been a busy week as the financial year draws to a close and I search for that 11 cent discrepancy in my profit and loss statement.  I’d rather be doing integrals, CFD or circuit modelling!

I have made some more progress on plating solutions and have achieved some excellent results using non-cyanide electroplating, but the outcome is somewhat fickle.  Sometimes the plating is bright and well-adhered (see Figure 6) and sometimes it has been simply lousy.  I had intended to continue my experiments this evening but over the past few days the silver in my latest bath has almost entirely precipitated (probably as silver sulphide) and the platings have become progressively dark and thin.  This is not good - the bath chemistry needs to be stable over time.  There is at least one more recipe that I want to try, as well as experimenting with plating currents, pre-plating preparation and strike, current densities and bath pH.  The process needs to be stable and produce a consistently good plating before I plate the welder bus.



Figure 6.  Good Plating.  The bad and the ugly have been documented and destroyed.
(Plating buffed with a paper towel on removal from the bath.)


As an aside my experiments are being conducted on a very small scale, each using no more than 200 ml of plating solution and just a few grammes of constituents.  I am recovering silver residues as silver chloride and any waste solutions contain only sodium and potassium salts (nitrates, sulphates, sulphides, ...) hydroxides and ammonium compounds with traces of other additives. Arguably I could be spraying these on the lawn as fertilizer but they are neutralized and find there way to a chemical recycling and disposal company.  While some of the base chemicals are corrosive I am deliberately limiting my experiments to chemicals that are readily available from hardware, gardening and photography shops, food additives, or compounds that can be readily synthesized from these.Plating-3

31 March and I am now getting consistent results based on a non-cyanide process described in Patent US 4,153,519.  It uses a thiocyanate (no, this is not cyanide) to maintain a low concentration of silver ions in solution.  I made a 200 ml solution containing 2 g silver nitrate and 15 g potassium thiocyanate.  The silver nitrate precipitated as a white colloid and rapidly dissolved.  I added 0.2 g of potassium iodide and a few drops of phenolphthalein indicator dissolved in acetone.  Then I added just a few drops of dilute ammonium hydroxide until the indicator just turned pink (indicating a pH of about 8).  I poured off 4 ml of this solution and made it up to 200 ml to make a strike bath.  The positive electrode was a lump of silver left over from photographic silver recovery.

You absolutely need to use a silver cathode with this plating solution.  I tried bright steel but it didn’t stay bright for very long.  The steel became plated with colloidal silver, the bath went dark brown and was ruined by contamination.

The procedure is shown in the side bar pictures.  Clean the copper as thoroughly as possible, degrease and rinse it.  Silver plating will hide small surface scratches and imperfections if the coating is thick enough.  However a smoother substrate will produce a better finish.  I used a fine grade of wet and dry carborundum paper followed by a detergent scrub, a rinse under tap water, followed by distilled water.

I connected my 12 V DC bench supply to the silver and copper, positioned these in the strike bath and passed about 20 mA for a minute.  This left a thin bright bluish silver coating on the copper (the photo doesn’t show the brightness of the coating due to the lighting).  Next it was straight in to the main bath with a current of about 150 mA for three minutes.  The surface area of this test piece is about 0.004 sq m.  Currents should be adjusted accordingly.

The copper becomes plated with a white lusterless layer of silver.  Rinse the plating under tap water and buff it to a bright finish using regular metal polish.  The fact that the coating can stand up to what is effectively an abrasive polish is testament to its adherence and thickness.  If the plating adhesion is poor then the polish will rip the silver off the copper.  However in my experiments the silver coating is  hard, has good adherence to the copper and is resistant to finger print corrosion.

If you need a thicker plating then either increase the  plating current and/or time in the second path or you can re-clean the initial plating and proceed directly to the second bath for a further coating.  When I tried this I ended up with a plating that I could use as a shaving mirror.  The plating was measured at about 15 um thick.

When you make a mistake (usually a cleanliness issue associated with finger prints) then you will need to clean the copper back to a bare substrate and start again.  Having gone through this exercise a few times I can assure you that when the silver coating has adhered it makes an excellent bond.

While you may think that I am done with plating I have decided to go back to my earlier recipe which showed great promise, but this time using the silver cathode.

3 April and I have not yet had time to complete my further silver plating experiments.  However the good news is that my current plating specimens shown in Figure 6 are as bright as the day they were made despite deliberate attempts to cover them with as many grubby finger prints as possible.  I have bolted a few elements together and after 4 days the connection is absolutely clean.  There should be no increase in the resistance of these connections over time due to surface oxidation.  While I still have my alternative recipe to retry, my next mission is to figure out a method for extending the plating down a full bus length (almost a foot) using basic apparatus.  I have some ideas.

Interlude While Moving House

8 April and I moving house in a couple of days to Wellington, the capital city of New Zealand about 600 km South of here.  The workshop has been disassembled,  the lab is pretty much boxed up and there won’t be much happen over the next few weeks until everything gets put back together at the at the other end.

23 April and we are back in Wellington at last.  My office, laboratory and workshop are still largely dispersed in various cardboard cartons but the computer networks are up and running and I have started to reassemble the workshop after a week of cleaning and minor repairs.  The foundations for my lathe and mill are finally set after levelling the lathe bed.  I’m using a precision 150 mm level accurate to 0.05 mm per metre (0.003°) but the machine will need a good shake-down before fitting the final shims.  The move has provided an overdue opportunity to lubricate everything, change the head stock gearbox oil and adjust the gibs to remove the slop that has developed over time.

25 April and the lathe and mill are reassembled and rewired with no bits left over.  Only one small piece, the carriage locking pin, had been lost in the move but the machine is  sufficiently functional to make a replacement.  I only have about 30 cartons left to unpack.

3 September 2012 and the design of the pulse arc stylus is progressing steadily.  While I have abandoned modifying a TIG hand piece because there is no space for a retraction solenoid, I have designed using standard ceramic gas shields and replaceable 0.5 to 1.0 mm diameter Collet chucks.  The most difficult part of the design is forming the electrical connection to the retracting tungsten electrode.  Brushes and springs have failed so I am trying ultra-fine copper braid.  I have a lot of machining to do yet.

Click here to navigate to the next welder page, How Much Current?