Transformer design.

[bikerhifinut]

Split from this thread

Originally Posted by joebog1

Amps AC multiplied by .62 is the DC capability of the winding.
Although If you are conservative it wont matter too much.
The reduction of the current capability is because of the charging current peaks into the filter capacitor after the rectifier.

So does this also apply to HT windings?
For instance these mains transformers are specified as capable of 150mA 410-0-410V. I have always taken this as the total current from both halves of the winding which is just enough for a single 5-20 2xEL34 design (145mA draw). The transformers are marketed as being suitable for a Mullard 5-20 design. I think I wouldn't be happy with LESS than 200 mA now.
And on another tack, I have generally taken the VA rating of a lower voltage transformer as being exactly that whether rectified or not. This suggests that a transformer for a transistor amplifier must be 1.6 times the minimum VA required to maintain maximum output to me.

Puzzled.

A.

[glowinganode]

Indeed it does and is referred to as "utilisation factor", W/VA. Rectifier and filter design affect this factor, bridge rectifier being better than full-wave and choke input better than capacitor input.
Typically fwr / choke input = 0.63, capacitor input could be as low as 0.4 - 0.5
Citing the example above, input 820 V x 0.15 A = 123 VA, assuming zero voltage drop across the rectifier, output 580 V x 0.145 A = 84 W. This implies a UF of around 0.67 which sounds unrealistically high. I'd agree that a 200 mA rated secondary might be more appropriate.

Rob.

[joebog1]

Gday to the fellow "nut".
In Australia ( yes yes, we are all criminals, WITH education now, transported from Blighty ) Transformers in the 50,s and 60,s were advertised as ( quote) DC output, and were wound with wire guages to suit. Rectifiers ( GZ34/5AR4, 5U4, 5AS4, etc etc )were always run with AC on the filament/cathode, so no derating was nessesary.
A VERY general rule of thumb for Hi-Fi stuff is to run a wire rating of "about" 700 circular mils per amp!. BUT ratings from 500 circular mils/amp ( ****** hot) to 1500 circular mils/amp ( very cool and conservative, The bowler hat and umbrella transformer) were common. Today insulation is far better than it used to be!! I remember "black enamel" which was nothing more than shellac painted onto the wire, and the "varnish" could be stripped off using a thumbnail .
Then comes " classes" of transformers. A whole box of mixed fish in itself, as the world doesnt use common "denominators". ( take a yank pint of beer ) Its only half full, but demands full price. Take their "wattage rating" its peak watts or music power, NOT R.M.S. as those of us, educated with the full value being the CORRECT and mathematical answer.
In the 1950's, transformers were "sort of" in the class B class, i.e. ~ 60 degrees C temperature rise, ABOVE ambient!! Today its quite common to use class F transformers that are ( I think, correct me if im wrong) 150 C

One other point I would like to make to you !!! and I have been designing and winding transformers for 50 years, 5% Si Grain Oriented M35 laminations are ALL I CAN BUY !!!!! ( 5% is the silicon content)
They are made AND stamped in India, to a very high quality !!! They are ALL thats available in Australia for ( ordinary) transformers, so I think they have either lengthened ONE of your legs, or perhaps shortened the other .

Err Hmm, with respect, go read up ( online if you wish) the Radiotron Designers Handbook on magnetics and transformers and chokes ( chapters 4 & 5 from memory)
The maths is my school ( inches)and the design criteria are STILL used in the finest military equipment today ( NO sand pit in a modern fighter aircraft, to defeat Atomic Pulse damage) I learned ( well nearly all ) ) my traffo stuff from that book!!

I hope I havent offended anyone !

Joe

[bikerhifinut]

I get you Joe.
Firstly, Majestic are a long established and respected manufacturer here in the UK.
I don't think they deal in "Leg pulling".
As I understand it, output transformers use a different thickness of lamination in general and after that my head spins but I understand it allows them to squeeze a bit more out at the expense of a bit of extra heat. This won't be an issue with mine as the whole point of overspecifying was to gain a "cool runner".
I had a long chat with Ed about my other transformers and i am sure he told me that there were a number of different lamination steels available. Also a tale of a CHinese transformer from a supposedly respectable amp that had very inferior lams.
But I thank you for your concern Joe, and your experience and enthusiasm is very appreciated here.
To reiterate, Majestic were at great pains to understand EXACTLY what I was doing, which Valves were going in the amp and what sorts of HT current draw I was expecting/getting. So I am confident that I am getting something that is "fit for purpose" and I think the price was reasonable, all things considered.

A.

[joebog1]

You are correct re : thickness of laminations. .35 mm = 14 thou, and is most commonly used on small transformers that most of us use/ play with. .175 mm or 7 thou laminations are sometimes used in the top class output transformers. They are also very common in chokes.
Using .35mm lams is fine, especially as ALL the laminations I and I suspect many others use come from the same manufacturer. We used to have Australian made lamination ( John Sankey ) but like everything else, we import rather than make.

Williamson designed the first really fine transformers to be used in closed loop feedback circuits. Since then we have modified his designs to make them easier and cheaper to make with little if any difference in performance.
Another thing that effects the performance is the insulation used, and the method of insulating the windings. Williamsons origional design is on the net if you wish to have a look. I have wound them exactly as specified and it takes days!!!
Paper insulation ( known as onion paper) is still the best for audio. I can no longer get paper !!! so Im forced to use mylar tape. A better insulator perhaps but not as neat and precise as paper, with its consequent degradation in performance.
Flux density within the iron is another thing effecting performance, with audio it amounts to bottom end performance, with power transformers it amounts to heating.


Anyway, your specs look pretty good so I know you wont have any trouble

Joe

[Ed_Dinning]

Hi Joe, the standard here is M6 35 Unisil, so grain orientated.
I have seen some very shiny lams but probably .7mm thick and believe quite lossy/ a straight Stalloy.
Stalloy was normally run to about 1Tesla, Unisil to 1.5 Tesla. I havev seen Superpermendur run at 2T in a mil app, but it screamed with magnetostriction.
The above figures are for mains transfrmers; output transformers are normally run at a much lower flux density depending on how big your pockets are.

There is a debate among some of the Dutch amp builders about the "bottom end" of the iron curve and noise/ distortion added by magnetic domain movement at low signal levels. A dodge that was used with current transformers was to use some silicon iron and some nickel alloy lams to get fidelity up but keep cost reasonable.

Nickel alloys (mumetal, rho-metal, etc) work well but are susceptible to mechanical damage and are very expensive; they were often used in the better intervalve and driver transformers as well as on power grid current transformers with heavy fault currents.

Modern wires are normally grade 2, with 2 layers of enamel so no need of inter-layer insulation. This certainly saves space on the smaller wire sizes. Polyester tape, with its higher permittivity my increase the self capacitance of a winding and upset the feedback if it was at a critical level.
Where pri/sec are interleaved as in the Williamson type of transformer, it may be advantageous to add an earth screen layer. This takes the interwinding capacity to earth, not into the feedback loop.
It is also important to look at the layout of the windings and how low potential and high potential winding ends coincide; in may mean reversing the direction of wind to reduce coupling.

Langford -Smith is a very good basis for transformer design, but is from some time back; there are also a few other transformer books from that era "Electronic Transformers and Circuits" ~ Lee, "Transformers" ~ Connelley . These go pretty deeply into the theory but do not give much practical advice. The practical "how to" books of the era lacked most of the theory.
Modern books, since the introduction of ferrites are much better but still lack some of the basics on laminated transformers.

A transformer is a true engineering problem in that there is never a "correct" answer to the design as it needs much forethought and compromise. That is why I enjoy it.

Typical current density in modern mains transformers is 3A/mm sq for the metricated.

But bear in mind, if the resistance of the winding is important in the greater circuit, copper has a tempco of about 0.3% per deg C

Perhaps we should start another thread on transformer design.

Ed

[bikerhifinut]

Originally Posted by Ed_Dinning

Perhaps we should start another thread on transformer design.

I would be keen to read that Ed.
I think this thread has got stuck on the thorny issue of my noisy mains transformers rather than the other issues arising during the build and testing of the amps.
Thanks for all the education Ed, as you probably realise, only so much sticks in my bonce, an awful lot is above me for now, but the more I understand the more I get out of this hobby.
I would hope to rescue these transformers for another less demanding application in the future.

Andy.

[joebog1]

A nice idea Ed, and thanks for your comments.
I guess its up to the mods/sysops to decide on a transformer education thread
( Comments ?? Mods ?? )
I know I can still learn lots myself, just by banging around a few ideas.
For one how about vertical sectioning rather than horizontal as is "normal".
Im talking audio output and driver transformers of course. Williamson did both I guess with his twin bobbin idea.

I can no longer obtain C core iron, which was also very good ( much better than standard EI stampings anyway).

With insulation Ed, I still use interlayer winding with tape, and of course I still use Empire tape between sections, especially with OPT,s. With mains transformers I ALWAYS use an electrostatic screen, either wire wound ( usually with the primary gauge of wire so I dont need reset my old Glow winding machine) or copper shim, depending on quality.
You seem to have picked up on metrics, which I havent so I still use line/sq"
About all I do know is that 1T = 64500 lines/sq". small transformers ( say up to 75 watts ) I run at maybe 80,000 lines, with transformers of 100 watts to say 250 watts I use 64500. over that ( I rarely make anything bigger) I use regular design rules.
As far as output transformers go, I use a maximum of 64500 lines at maximum rated output and 30Hz, so my output transformers always perform rather well. As an example, my 50 watt traffos are 1 1/2 X 2 1/4 core, so about 250 watts as a mains transformer core. Thats also the easiest tongue width I can buy bobbins for, and iron as well.
I have been sitting on 20 kilos of 2" tongue John Sankey iron for some years,
so maybe I will do a design to match the modern sand amps with their 200 watts RMS output ( push pull parr KT88's ?? )as it is the wifey chases me with a broom when I use my little 6BW6 8 watter. I also have a stack of left over "gate" laminations ( square rather than the "scrapless" variety )

Anyway Im off topic somewhat.

best to all
Joe

[julie_m]

My number one tip for designing transformers is, learn a foreign language! Textbooks in any language besides English will use the proper units in all calculations. That one simple step makes electromagnetics actually comprehensible! CGS and American units are just a hindrance.

There's also this excellent article.

[Radio Wrangler]

The whole education industry conspires against magnetics.

Whether it's school A levels or a degree course, they teach electric fields and capacitance things first, and then magnetics and inductors, but oops, they're only a third of the way in when it's time to start revising for the exams. They all do it in the same order, they all run out of time, so all people get educated on e-fields and are all at sea on H-fields.

I thoroughly agree about using SI units. Trying to use the old FPS and CGS stuff... lines per square cubit anyone?

David

[bikerhifinut]

ALL things Magnetic were a form of black magic to me............

I dug out some of my old study notes the other day from the 80's and it made no sense at all.............

I do remember squirming in front of a DTp examiner (now the MCA) trying to explain the workings and adjustment of a standard magnetic compass.
"now then Mr Miller, tell me what you know about the magnetic compass.....") Oh yes the questions I think were deliberately obtuse and all encompassing (ooops).
Mind you I must have known enough because I passed.

I did enjoy "Split B" calculations though for a perverse reason.

Digression but I thought a quick look through some old theory might help but it hasn't.

A.

[bikerhifinut]

Looking at these buzzy transformers, and wondering........... Could the Steel Chassis have anything to do with making things worse?
I have done as much as I can to isolate them from mechanical and electrical and magnetic influences but I wondered if using a "gasket" made of single sided copper clad Fibreglass PCB is having an effect? It was there because the opening was too large to use the fixing holes on the transformers even though the weight is mostly borne on the main chassis via the oversized grommets i fitted.
I still think the transformers are mechanically noisy, but maybe i am making matters worse?
They other thing I noted was how the vibrations transmit from the chassis and the gel feet i stuck on, through the glass shelves of my equipment stand, which is nothing special just a repurposed widescreen TV table.

A.

[joebog1]

I think I mentioned it before, but I thought that these transformers are not varnished at all. If any part of the windings or laminations are not tied down securely, they will vibrate.
Thats why I varnish every transformer and choke I make, because my transformers will "rattle" as well. I think Ed will agree with me. Varnishing is cheap, so those transformers must be real "economy" units. Rattle will also be caused by running the iron at a high flux density too.

Maybe the cheapest thing would be to approach a transformer manufacturer and see if they will vacuum varnish them for you. Alternately, you could try varnishing them yourself with a spray can. Use one of those long thin tubes that come with WD40 type products and inject as much as you can into every crevice you can see, especially where the wires come out of the bobbin. Do up the screws and spray the outside of the transformer as the last thing. Its definitely NOT a proper job, but it may cure the problem. I dont know what spray cans you can get in Britain, so I looked at what I can get here and you will have to try and chase them down I have included the part numbers.
CRC Clear urethane seal CRC2049
Electrolube Clear protective lacquer CPL200H
RF Circuit board lacquer RF61

Thats about all I can think of, as you have already tried everything else.

Joe

[bikerhifinut]

I have a notion to approach the manufacturer as although i didn't buy from their official outlet, they may appreciate a "right of reply" and the opportunity to inspect these units in case they were a "rogue batch".
I need to be careful due to the laws of libel amongst other things.
The odd thing is, the output transformers are stonkingly good.
Measure and sound well.
The worst that can happen is they would say no...............

A.

[joebog1]

I agree Andy!!
As far as the output transformers go when there is no sound the transformers have no flux, or near enough to zero. When you play music/speech LOUD then the flux density is at its maximum, especially with heavy bass. Try a little reggae at full tilt, if the OPT's are going to rattle, that will be the time. ( 1812 Overture if you have a decent recording, I think its the Kaytel 200 gram vinyl that has the cannon shots recorded so loudly that 99% of turntable arms, jump completely out of the groove, the actual groove tracing is like 1/4 inch apart the modulation is so heavy)

With a push pull circuit and no sound the transformer doesnt do anything, its just idling.
At worst if the output valves are really badly matched, there will be a few milliamps of offset current, but its DC, there is no AC component to "rattle" any loose bits.


Regards
Joe

[bikerhifinut]

Oh its had a good test.
Apart from buzzy mains trannies it performs amazingly well.
The way it dealt with some Metallica tonight leaves me in no doubt as to the output transformers ability to handle fast transients and drums and bass etc.
Ok slight veer and back on topic Joe.

A.

[joebog1]

Originally Posted by julie_m

My number one tip for designing transformers is, learn a foreign language! Textbooks in any language besides English will use the proper units in all calculations. That one simple step makes electromagnetics actually comprehensible! CGS and American units are just a hindrance.

There's also this excellent article.

You are absolutely correct Julie!!
I have spent this morning looking and reading his website!
I recommend anybody having trouble with transformers ( and LOADS of other very well written articles) read and understand what he has written.
Its superbly put together with enough dry humour to keep one interested, and instilling the required knowledge without any pain.
Thanks for that link
Regards
Joe

[Ed_Dinning]

Hi Folks, yes Julie's article is a good one that has been available for many years.
Problem was that school teachers generally did not understand magnetics themselves and although physics was a metric subject it was taught in CGS, not MKS which gives much more meaningful numbers in all sectors of science.

Try soaking the transformers in molten paraffin wax as an impregnant. This will also act to protect the transformer from short term overloads due to the latent heat of the wax.

Ed

[Radio Wrangler]

In my early years at HP we had an in-house transformer design manual. It covered all the theory needed as most people needed a remedial course due to it getting skipped in most forms of education. Being American in origin, everything was in 'quaint' units but it produced good transformers if you followed the guidelines. It also contained data for all the standard wires, bobbins and lams used in our in-plant transformer shops. It would be great to have a copy today, but I think I'd translate it into SI. The transformer shops it covered have all long been demolished and the ground salted...

David

[bikerhifinut]

Hi Ed, Joe, David and all.
I phoned the manufacturer yesterday.
I didn't expect them to be too helpful as I didnt buy the transformers from their approved retail outlet, rather from a contact on another forum, both of us in good faith as they were unused still in their shipment boxes.
All That was suggested to me was that they should be supported fully around their base in case of a loose lamination on the end, which made not a lot of sense to me. All my experience with drop through transformers so far has been with isolation washers or grommetts or pukka tophat spacers which, by definition, only support the transformers around the bolt holes at each corner.
I will try an experiment with one though in the interests of thoroughness. and drop it onto the insulating gasket spacer minus fibre washers or rubber grommits. If that works then fine, and I can use them in next years SE project which will only demand 100mA maximum from the output valve.

Joe, I had a good look at the transformers again and I think they have been varnish impregnated after construction as the laminations look well coated in some form of lacquer/varnish.
Ed's suggestion of wax impregnation is interesting if messy sounding. I guess a box or 4 of cheap "Prices Candles" would donate the wax. I think wax impregnation was a technique used by Hinchley on the transformers used by Leak as they are notorious for leaking wax when hot and in warm climates (Australia).

If all else fails, I am hoping the lams and bobbins will serve as the basis for a pair of lower rated transformers for the future project. (rewind).


Andy.