Rewind Dynaco mains Tx for 240V

[retailer]

Having just returned from a few weeks away doing nothing I'm eager to get to the Dynaco SCA 35 kit amp I recently purchased - the mains transformer is the US 110V 60Hz version and I plan to rewind for 240V operation, I found a small stack of matching laminations so I have enough to add to increase the core area to cope with the decrease in mains frequency, the chassis and transformer mounting holes are elongated and I worked out I can go from a 50mm to a 65mm lamination stack and still mount the transformer in the original holes. I'd like to make some small changes to the winding layout the original power supply has 2 non centre tapped filament windings, one for each channel, and a split secondary, approx 315V - 0 - 315V feeding a pair of silicon diodes, giving a HT of 370V I'd like to change this to a single winding with a bridge rectifier, this will make the winding of the transformer a bit easier for me and will allow me squeeze in a shield winding between the mains and secondary windings, in addition I'll try to wind the filament windings as bifilar windings which will give a true centre tap which may allow me to eliminate the 'hum bucking' 1k pots - does any one see any issues with my proposed idea.

[dougietamson]

Originally Posted by retailer

Having just returned from a few weeks away doing nothing I'm eager to get to the Dynaco SCA 35 kit amp I recently purchased - the mains transformer is the US 110V 60Hz version and I plan to rewind for 240V operation, I found a small stack of matching laminations so I have enough to add to increase the core area to cope with the decrease in mains frequency, the chassis and transformer mounting holes are elongated and I worked out I can go from a 50mm to a 65mm lamination stack and still mount the transformer in the original holes. I'd like to make some small changes to the winding layout the original power supply has 2 non centre tapped filament windings, one for each channel, and a split secondary, approx 315V - 0 - 315V feeding a pair of silicon diodes, giving a HT of 370V I'd like to change this to a single winding with a bridge rectifier, this will make the winding of the transformer a bit easier for me and will allow me squeeze in a shield winding between the mains and secondary windings, in addition I'll try to wind the filament windings as bifilar windings which will give a true centre tap which may allow me to eliminate the 'hum bucking' 1k pots - does any one see any issues with my proposed idea.

You are only using the original lams, end bells and mounting brackets?

A few design additions:
Add in a thermal fuse as used in modern transformers
Add a few extra taps to the primary in case of mains voltage fluctuations, eg +/- 2.5% or whatever mains voltage drifts you measure.
Add a couple of secondary taps to get the the HT just right.

What are you using to make the bobbin? Are you replicating the vintage build style eg paper between each layer?

Doug.

[Diabolical Artificer]

I'd agree with adding a few turns/taps on the primary & secondary, you might never use them but they're handy just in case your calculations are a tad off and you get more voltage drop than intended on the secondary. I do this on most of my tfmr winds, that is if you have the room. Often your winding ends half way across a layer anyway and it's no bother to add a couple of taps etc.

Andy.

[Ed_Dinning]

Hi, if you are increasing the core area by 20% (60 to 50Hz) then the turns/ voly will be correct for that stack, so use the same turns/ volt figure and wind for the correct voltages.
Not that sometimes the sec windings will have compensating turns added due to the voltdrop in the windings.
I'll be out in Melbourne in the New Year if you are along there and want to discuss.

Cheers, Ed

[joebog1]

The only thing I would add is,
Try and find true enamel insulation rather than wood varnish AKA polyurethane.
DO find and use onion paper as interwinding insulation.

Somebody said to add a thermal fuse, I would say DONT add a thermal fuse.
Use fuses on primaries and secondary HT ( you shouldnt need heater fuses ) .

When actually testing the amp will be the time to start with say 250 mA on the primary and go up from there. If you go crazy mad and add 100's of uF on the secondary, you will have to determine the correct fuse by trial and error.

Joe

[dougietamson]

From the assembly and operation manual:
https://www.thehistoryofrecording.co...kit_SCA_35.pdf

Ventilation
When using the SCA-35 it is essential that it have adequate ventilation. This unit dissipates 110 watts of heat,
and you can see that there is the same heating effect as
with a 110 watt incandescent lamp in a small metal container-the case will get quite watm to the touch, and
should have air space above to permit the heat to dispell.
There should always be several inches of air space above the
unit and behind it. Never place anything directly on top
of the cover when the amplifier is operating.

I've read these amps run warm due to the full enclosure hence the thermal fuse as a design option.
They are a pain but you can plan for some degree of 'expert user' replacement access.
Not sure if you run AC in the room. last time I was in Adelaide, April 2018, it was unexpectedly hot during the day.

The Power Supply
The SCA-35 is powered by a power transformer and full
wave rectifiers using silicon diodes. Separate heater windings are included in the power transformer to provide means
for individual hum adjustments on each stereo channel.
The power transformer is impregnated with an epoxy
resin which serves the dual purpose of heat dissipation
from the core and prevention of noise and vibration.

Can you check how much epoxy they used, might make it a pain to get apart with minimal damage.

Good idea from Joe (resident transformer SME) about adding a fuse on the secondary, an in-line fuse holder would be a good option.

[joebog1]

I suspect the Epoxy they used will be Glyptal, and if the transfomer was baked after it was impregnated it will be a proper ^%(*&% to pull apart.
Dougie's quote doesnt say if it was vacuum impregnated or just dipped, but after all these years I suspect the varnish will have set like concrete. If the transfomer was properly vacuum varnished I wouldnt even try.
I would be tempted to buy an 240 - 120 stepdown transformer and leave it at that.
As far as running at 50 Hz rather than 60 the transformer will handle that OK, albeit a little hotter than it did on 60 Hz. You could add some vetillation additions to the case to remove the heat. Even a small fan similar to that used on a PC video card will be more than enough, and if you run it at half voltage it will be very quiet indeed.

Joe

[retailer]

After taking some voltage measurements I immersed the transformer in solvent in a large coffee tin and closed the lid, after a 4 or 5 hours I was able to separate the lams one at a time - it was no harder than any other transformer I've stripped down, I now have the lams soaking in a closed container of solvent so I can clean off the varnish, this is a bit of an unpleasant job, the solvent MEK doesn't remove the varnish but rather turns it into a jelly consistency so it can be scrubbed off with a rough cloth one lamination at a time, not a pleasant job at the best of times, but needs to be done as any varnish left on lams will affect the stacking factor.

With the windings free of the laminations, I unwound and counted turns of each winding, the original windings had been put on the winding tube at an angle which resulted in almost zero margin at one corner - I wonder if this could have caused transformer failure down the track. The core is 25mm x 50mm, wire sizes are - primary 0.6mm, secondary 0.22mm and each filament is 1.10mm, filament winding is 16 turns for 6.70 volts - HT windings are 792 turns each and primary is 287turns so turns per volt is close enough to 2.39, plugging this into the universal transformer equation shows the original transformer was run at a Bmax of approx 1.25T, I'm going to increase the lamination stack to 65mm (a 30% increase) so I should be able to reduce the Bmax and possibly increase wire sizes, while increasing the core area is good, adding laminations does not increase winding window size, I'll have to take some time to play around with turns/per volt and wire sizes, like others I've read that some users report hot running transformers so cool running should be a priority

Adding a few taps to the primary is not such a bad idea if it can be accomodated, while I've wound some transformers on a tube with insulation between each layer, they do take longer, a little bit of extra care is needed so I'm going the easy route and I'm going to make a bobbin - at least that is the plan at this stage - by cutting the parts from some 1.8mm glass fiber sheet on my cnc mill, I've just about finished the design in Fusion360, once the parts are cut out they should clip together to make a nice bobbin. I've made bobbins by hand this way in the past but found one needs to needs to have very good marking out and cutting skills so the finished product is square and stays together with out glue

Ed, thanks for the kind offer of assistance your, transformer winding skills are highly regarded, but I'm a bit of distance (720Km) from Melbourne so not very practical.

[Ed_Dinning]

Hi, Melbourne was only a suggestion if you get over that way at all, I doubt we will be out in the Adelaide direction this year, too much DIY at son's.

Primary taps often give problems and take up space, so I personally try to avoid them unless necessary.
Try and get your primary as smooth and even as possible and use insulation between layers. If you really need to pack the turns in it is possible to insulate 1/2 layers (between wires with the greatest voltage stress), but be certain to insulate at the ends of a layer where a turn could slip down.

Double insulated wire is now used a lot in Europe, without interlayer insulation
(Grade 2) especially on secondaries (it is a moot point if the extra wire insulation is thicker or thinner than the additional paper thickness)

Note that there are a variety of insulation coatings available, only some of which are solderable and different types will have different thickness. Not much, but which can cause an upset in a "tight" transformer with many layers.

Some transformers use a safety screen/ EMC screen, this is a thin layer of copper foil in a single turn, but with ends insulated. It can provide safety from a pri/ sec short (foil thickness important here) or to act as an earthed screen for interference (usually thinner or can be a single layer of, say, 0.063 dia or similar, connected at one end only)

If you can only increase the wire dia on one winding, make it the pri, as being the innermost it heats the most.

If the core is Stalloy it is being worked hard at 1.25T and will have a highish mag current, adding to heating. If Unisil then well within its capabilities (1.5T). Possibly worth doing a mag curve on the completed transformer to see how close it is to the knee.

LV secs can be single insulated wire

T/V figure for the pri will be what is required to give the core flux density. T/V for the secs are sometimes higher to give compensating turns to allow for the I * R drop on load.
Especially on LV secs, a thinner wire is sometimes used than standards dictate, for economics as well as window area. The load is constant so the extra drop is easily compensated for

Bobbin winding is probably your best route, it is easier to do and gives added safety for wires slipping down at the ends.

Your figures for 50Hz and increased stack should compensate for the frequency difference, you will also win out slightly with the thinner primary as it has a better packing factor.

The nominal standard for winding is 3A/mm^2

Hope that helps

Ed

[joebog1]

Not much I can add to what Ed has already said. Apart from the flux density. I ALWAYS try and run it lower rather than higher. Its probably a bit over the top, but you dont know what iron you are using. Dont forget, although you have increased the stack height, you havent increased the window size so I would definately be running your calculations at least a couple of times to see that it will fit. My own transformers, even when using 5%SiGO laminations is ONE Tesla, or in my language, 64500 lines per square inch.

All the best with your rewind!!

Joe

[retailer]

The bobbin if finished - while the end result is better than those I have previously made by hand it was not quite as straight forward as I thought it would be, I still needed to do a bit of fettling, a rotating cutter can't cut square internal corners so I needed to finish off all internal corners by hand, I did learn about clearances and where I should reduce dimensions slightly to reduce the amount of fettling needed, not counting the hours spent on Fusion360 the actual cutting operation only took about 90min, still as with all things new, the more of these I make the easier it will get.

With a completed bobbin I now know the size of my winding window and can start playing around with wire sizes etc. I have an Excel spreadsheet that allows me to plug in wire size and turns for each winding and it returns a % fill factor, I have more or less settled on 2.5 turns per volt which gives a Bmax of approx 1.1T and should just fit nicely.

Ed: I don't have the luxury of choosing between suppliers when it comes to wire, buying from overseas or even interstate is not cost effective due to shipping costs, my supplier is local, his main customers are industrial motor rewinders.
Not sure what you mean by a mag curve, or the test procedure involved - I did do some rough and ready tests on the original core comparing to a known ordinary standard grade core - 50 turns of wire on the assembled core - connect a variable ac supply (variac and isolating transformer) voltage and current through the coil is monitored - increase voltage until current is 1 amp - voltage for the ordinary standard core at 1 amp was 17V and for the original Dynaco core it was 19V so it would appear that the original laminations are a better grade.

[joebog1]

VERY nicely made bobbin. What I do when I have to do a rewind using old iron is to use .75mm thick fibreglass PCB material that I leave in the ferric chloride for a day or so. Its excellent insulation, its quite thin enough to make it easy to cut and file, although you really need to spin the drill up to get clean drill holes. I glue with superglue.

I have never had a breakdown to core using this method.

Joe

[Ed_Dinning]

Yes we are somewhat spoilt for suppliers in Europe.

Mag curve is simply a plot of volts/ amps taken usually on the transformer primary with all other windings unloaded. It is then easily seen where the curve starts ti ben and saturation is starting to occur

Commercial winders usually run Stalloy cores at 1T; Unisil (GO) ones up to 1.5T.
GO usually has a smoother transition into saturation.

I did do a transformer design foe a Mil TRU where we used Supermendur alloy run at 2T and 400Hz. It was mounted on a wing spar and in the airstream, so cooling was not a problem. It squealed like a pig due to magnetostriction, but next to a jet engine it was no prublem.

Ed

[retailer]

I had to order some wire for the primary and can't pick it up until Monday so while waiting I wound a temporary primary with scrap wire - 600 turns 17 ohms - when connected to my mains supply the current draw was 65mA I left it powered up for a few hours and it was slightly warm to the touch, for comparison I hooked up a commercial A&R transformer with a similar core Xsection it drew 64mA so I'm assuming the 65mA is in the ballpark. Out of curiosity I unwound around 5 or 6 turns and found the current draw went up to 74mA a bit of a surprise that this few turns would result in such an increase. Assembling the core did give me a chance to check the clearance of the bobbin within the bell ends, I found I need to have a few mm extra clearance to allow the lead out wires to cross over to the middle and down through the bottom hole in the bell end - in addition I need to provide a small amount over overlap on the holes drilled in the bobbin end cheeks, the temporary winding showed me the wire lead out holes don't coincide with the end of the winding so it's back to Fusion 360 to modify bobbin end cheek and machine a new pair.

[Ed_Dinning]

Hi, Sounds like your calculations were spot on. It would also appear you are close to the saturation knee with the rapid change of current with voltage.

Ed

[joebog1]

A&R transformers were mostly very well made. They were specified for almost all the kits designed by R,TV & H , Electronics Australia etc.

With regards to the current going up by removing only a few turns, I would in fact add a few turns.
Its extremely rare that windings fit perfectly without using up the full layer width. So if six turns increased the current by that much, theoretically adding six turns would take it down and away from the knee that Ed mentioned.
If you get a high mains voltage area, your transformer starts to get a bit dodgy in design, so adding a few turns might even make it very easy to add a tap for, say, 250 or in Western Australia 260 volts.

As far as the amplifier goes, it wont make a huge lot of difference if your HT is a few volts low.
But seeing as the 6BQ5's are running close to maximum already, and are known to run stinking hot anyway, a few volts lower HT wont hurt either.

Just my wanderings thru my brain.
with thx and credits to the Bonzos.

Joe

[Station X]

Two posts deleted. Can we keep posts respectful and clean please.

[retailer]

I finally found some time to spend on the Dynaco SCA35 mains transformer rewind 110V to 240V, I needed to increase the core area by at least 20% to account for the 60Hz/50Hz mains frequency difference between Aust. and USA, I have a small selection of salvaged laminations and found some that are the same size and pattern, the elongated Dynaco chassis transformer mount holes will allow a core stack increase from the original 50mm to 65mm - this more than the needed 20%, however I believe the original Dynaco laminations are a better than standard grade and the laminations I planned to add are an unknown and more than likely a standard grade so I hoped that the extra would make up for the standard laminations.
A turns per volt figure of 2.55 would give a Bmax of approx 1.1Tesla which should ok for the core I was using. I had previously assembled the amp and at normal listening levels I measured the HT current draw at approx 155-165mA adding the total filament current of 4.5A gives a total secondary power of approx 90W - allow for losses and the primary needs to supply 100W, a proper calculation is probably not quite this simple but it's close enough to the mark.
100W on the primary equates to approx 420mA - the primary wire then needs to be approx 0.42mm diam, this I don't have - I have 0.4 and 0.45 so I decided to go with 0.45, with a secondary current of approx 155mA the HT winding needs to be just over 0.25mm diam - I have 0.3mm and 0.25mm, I'll try 0.3mm - the 2 filament windings will be 1mm diam each. In theory the rectified dc voltage obtained from a bridge rectifier is the ac volts x 1.414, in practice I have found this is not the case and a more accurate approximation is the ac volts x 1.3 - for a final HT of 380V the secondary needs to be ac 295V - to allow for voltage drop across the resistance of the secondary winding (190 meters) I add 10 volts and settle on a secondary of 305V - primary 240V/612 turns and HT secondary 305V/780 turns.
I plug the wire sizes, turns, insulation etc into my excel worksheet that calculates the fit of the windings and decide that 0.3mm secondary wire won't fit and is out of the question, I decide to go with 0.25mm for the secondary - the original Dynaco secondary used 0.2mm wire so 0.25mm wire should be ok, the windings fit comes out to be 91.5% ie 91.5% of the available bobbin space is taken up with wire and insulation, I know for me at least this will be a tough ask, a fill factor of 80-85% for me is good, with care everything fits comfortably, I'm not confident with the 91.5%. When doing these calculations I use the measured wire size to allow for the varnish thickness.

I go ahead anyway and wind the primary and secondary and when these are done I measure the remaining room and see I have approx 2.3mm for the last 2 windings which are 1mm each for the filaments I can see things won't fit. Again I go ahead anyway thinking that at least I can test the finish windings and see how my calculations worked out, as I feared I had to use a small G clamp to squash the windings down flat or I couldn't fit the laminations over the bobbin where the windings bulged out the side. The secondary voltage comes out at 304V so I feel I'm close to the mark, I do a temporary connection to the amp and run a test, final HT is 370V slightly down from the Dynaco specified 380, but I feel this is ok. I run the amp with music at medium volume for 1/2 an hour or so but can't really gauge temp rise as it is quite hot today - 38deg C, everything in the workshop feels hot to touch. So while the transformer seems to be working ok I don't really want to leave it as is, it would have been nicer if I managed to get the windings fill at 80-85% - I may have to rethink things.

[joebog1]

Seems the weather over there is similar to here!!. You bought new wire so it's usually pretty good to 110 C so I wouldn'tt worry about temperature rise too much.
Squashing the windings to fit the lams I have done myself, it's not recommended though! I still have boxes of new John Sankey 5% SiGO that you can have but it's much larger than you require.
I also have gate laminations which have twice as long centre leg so your window space dramatically improves. Once again it's much larger than you need at 38mm 1 1/2 wide centre limb. For filament windings, have you done a calculation using twin wire of half the gauge ?.
It sometimes ends up as a single full layer rather than 3/4 width layer.

You could also "push " the wire a little using a smaller slightly smaller gauge.

I have wire tables that go anywhere from 500 mA/CM to 1.5 A/CM ( Milliamps per circular mil. ) if you need them.

Sounds like you are learning real fast!! excellent work.

Joe

[retailer]

I got a chance to run the amp for around 2 hours, and used my non contact thermometer to check temperature rise, the ambient temp in the workshop was around 31C and the transformer made it to 49-50C after 2 hours. I did a few voltage checks and found one of my filament windings is under voltage at around 5.99V - I had to wind these by hand and some how messed one of them up, this is not so bad as I wasn't 100% happy with it, and in the back of my mind I was thinking of trying something different, gives me a good excuse to unwind the filament windings.