Advice please on winding an audio transformer.

[AdrianH]

I am playing again and trying to wind a simple push-pull audio transformer. Only a small unit going for 2 watts output maximum.

Primary 3.250 K Ohm and Secondary 4 Ohm.
core area 279.5mm^2

Guessed at losses in the primary so set efficiency as 0.9 so needing more turns in primary?

Primary 3.250 K Ohm and Secondary 4 Ohm.
Core area 22
My core area implies 8 turns per volt.
Which I get to be 685 turns for the primary and 23 turns for the secondary.

2.83 Volts rms across 4 Ohms for the two watts

For the primary I have assumed with the 0.9 to get the extra turns I have said to myself 2.222 watts input which is 85 Volts on 3250 Ohms so with 8 turns per volt gives 685 turns. I am going to call it 686 turns.

So first question if anyone can check my numbers it is appreciated before I start winding?

Second question can I bifilar wind the primary. As in start two windings stop at 343 windings, add the 23 turns for the secondary and then complete the final 343 turns of the primary?

Many thanks.

Adrian

[retailer]

I've wound a number of PP output transformers lately mainly for guitar, while I know there are those that allow for voltages losses etc I have not ever done this and the results were always as expected, recent Fender Deluxe copy output was 21.5 watts and a Vox ACX15 copy output was 15.25 watts, both as expected.

Your last comment is a bit confusing - bifilar (2 windings together) of 343 turns then secondary then another 2 windings of 343 turns, are these going to be bifilar as well, are you going to have 4 x 343 turns total ? Usually PP transformer primary impedances are assumed to be plate to plate so the primary turns should be around 685 total.

Your the transformer will work but may have a poor low frequency response, I wound a small PP outout transformer for a 'flea power' practice guitar amp project a few years back and while I can't quite remember the full details I recall I used a small core around .75"x.625" so about the same as the one you intend to use and I was quite disappointed with the low end response it started to fall off at around 180Hz where I was hoping for around 80Hz, I ended redoing it with a bigger core. The number of turns on the primary does have a bearing on the low end response if you cant fit the required number of turns on for a reasonable low end response then the only option is to go for a bigger core so you can fit more turns. If you are just doing this as a trial then I'm sure you will get something out of it.

[AdrianH]

Hello retailer, sorry do not have your name.

OK on the ignoring losses and just working on the watts, that would knock a few turns off the primary.

This is really just for a try out, the valve amp is probably flee power really, it is one I did under the thread https://www.vintage-radio.net/forum/...d.php?t=165222

As I am running the outputs in class A (6021 Valves) and I have paralleled the twin triodes I believe I have an ra or 3250 Ohm, so each side of the primary should be 685 turns (686 for ease of split).

What I was trying to avoid was having two different DC resistances on the primaries. Again my thinking is that if I did one primary, then secondary side then the second primary on the top, I will have a longer length of wire, hence more resistance. But thought that if I wound both primaries together, stopped half way, added the 23 turns of the secondary and then continued winding the finish of the primaries, that I could:-
Keep the primary resistances the same and do a small amount of interleaving at the same time.

It may help with balancing the anode currents better in the valves, it may not, but it is all guessing on my side. It's a way to learn.

Adrian

[retailer]

I'm not at all familiar with the miniature dual triode, if you are connecting them in parallel then you only need one primary winding, the anodes are joined together and connected to one primary wire and the other primary wire goes to HT+. If I was doing this I would try a 240v - 6 v mains transformer the turns ratio is close to what you are aiming for and you can see the results without going to the trouble of winding a transformer, you could experiment with different secondaries 240v - 6v, 240v - 3v, 240v - 9v and see which one gives the most output. Normally with a single ended design you would need a gapped core on the transformer but with valves as small as these I doubt an un-gapped core would saturate.

I see you have used the term ra, the valve data gives an anode resistance ra - this is not always the recommended anode load, usually ra is anode resistance and Ra is anode load and they usually differ. As an example of a 'flea power' parallel twin triode, Fender use a 12AT7 and transformer with the triodes in parallel as a reverb driver, I think the 12AT7 anode resistance is around 10k, the transformer that Fender use however has a primary impedance of around 25K and a secondary of 8 ohms.

Anode load is not always given in valve data, from what I have seen it is usually only given if the valve is intended as an output valve, if it not given then you really need to have some knowledge of load line to work out the best anode load, however if you are experimenting then using a variety of small mains transformers might be the best way to do it

[Ed_Dinning]

Hi Adrian, as a rule of thumb you should be looking at running your flux density at 0.6T or less, a little more if you are using GOSS lams. What is your LF 3dB point frequency as this seems a smallish transformer with a lowish turns figure

Ed

Doesn't push pull cancel out the flux?

[Herald1360]

Only the standing dc component. If all the flux was cancelled you wouldn't hear much!

[AdrianH]

Originally Posted by Ed_Dinning

Hi Adrian, as a rule of thumb you should be looking at running your flux density at 0.6T or less, a little more if you are using GOSS lams. What is your LF 3dB point frequency as this seems a smallish transformer with a lowish turns figure

Ed

EH! What!

This is me just following some web page stuff and seeing what result I get, all from a formula of Te=1/4.444 x F x Bm x Ae

Bm set at 1.1 and

Ae set at 279mm^2 but changed to m^2

F set at 100Hz

The small valve amp I am working from is the attached pdf and pic, it can manage about 100mW.

If I do not have enough windings then I can not get more on the bobbin at present. But it has filled a day in.

Let me see if I get sound out of it and it may prompt me to get more into the process and trying to follow some old text that is written in the 50's

Adrian

[AdrianH]

@retailer.

There are two twin tridode's each valve driving one side of the transformer. I think the amp is called Powerman or something very similar. I basically just added some negative feedback and lowered the HT to 150 Volts, apart from the board which I did in Kicad (another learning process) the valves are all out of scrap military gear.

If all the flux was cancelled you wouldn't hear much!

I thought that the "AC" flux was sucked out by the secondary winding, well most of it anyway.

[kalee20]

Hope to add some input later - but I'm a bit confused by the line, "Core area 22" in the initial post.

Also, initial reaction, 685 turns seems a bit low. Without pencil/paper or calculator at hand, I can't comment. But it's the sort of number which would make me recheck everything.

Originally Posted by merlinmaxwell

I thought that the "AC" flux was sucked out by the secondary winding, well most of it anyway.

Nope. AC flux in a transformer core is determined by turns, cross-section area, voltage, and frequency. It's independent of the load applied.

[AdrianH]

Originally Posted by kalee20

Hope to add some input later - but I'm a bit confused by the line, "Core area 22" in the initial post.

Note sure where that came from unless bits left from editing as I quoted the core area as 279.5mm^2 a few lines before, so that is my editing fault sorry. I can no longer change it.

Every day is a school day, I love it.

[Ed_Dinning]

Hi Adrian, the transformer should work, but perhaps not sound as good as it could do.
The B part of the equation is the flux density you wish it to work at and this will vary according to the material and application you wish to work on.

Older material, Stalloy was usually run at 1T for mains use and about 0.6 T for output transformer use.

Unisil, or grain orientated material will run at 1.5 T for mains use and possibly 0.8T for output use.
As ever the are all sorts of caveats in this statement.
The value of F in the equation is broadly used to set the lower frequency 3dB point

Plenty of stuff that can be studied, but transformers are true engineering, in that there are all sorts of compromises to be made for many different rreasons.

Ed

[AdrianH]

Hello Ed,

The main concerns for me were my working out of the formula/turns ratio etc.

Unfortunately I am using old mains transformers, so materials are a guess even to modifying the bobbins to not have a split primary /secondary.

I did find some ferrite transformers from SMPS unit I scrapped, but think it would have an even lower flux rate as designed for 100's KHz.

The first amp is using a Hammond 125B 5 watt rated device and it cost a lot compared to the rest of the kit and even then it was several weeks as it came from the States into France around the houses a few times then eventually to England. So this was an idea I started early this morning, did some reading and then thought I would have a go.

The amp is a very simple and a modified guitar based small practice, I think it was originally done with distortion in mind o a bit of neg feedback certainly helps but is not hi-fi.

In fact just trying it out now, it is working I will sort out the feedback wires and even post it on facebook site, just in case anyone has an interest.

I wish at times there was a Chat room to pick brains in real time!

Adrian.

[Ed_Dinning]

Hi Adrian, as it is an old mains transformer run it at 0.6 T and use 20Hz as the frequency; then re do your calcs. Working in metres, not mm.
Ferrite transformers look nice, but with a max flux density at mains frequency of 0.25 T not very practical !
Note that there are various relationships for calculating the load impedance to be presented to a valve (or 2) that are related to the Ra of the valve (anode impedance).

Push-pull is simple twice the figure for one valve. If you are lucky the R (L) will be given in valve data under various load conditions and you can work on from there.

Note that impedance's are measured in ohms as is the resistance of various windings. These are not interchangeable and a common trap when you are starting out

If you use twice the Ra (of both valves in parallel) of the valve in your calculation it will work, but will not be ideal and may have high levels of distortion.

Ed

[AdrianH]

Just stuck up a basic video on the UK Vintage Radio Repair and Restoration facebook page.

Adrian

[retailer]

Your intended amp is push pull parallel, there was some initial confusion as you seemed to refer to them as one and the same.

A while back I did an emergency repair for my nephew on his Laney 5150 the output transformer had failed he needed the amp for a gig so I stripped down the transformer and rewound it for him over the weekend, I have attached the winding layout, if you are winding transformers for guitar amps then I suggest this is the winding layout you use, it is one of basic push pull layouts and is quite common amongst guitar amps, note that while it looks to have 6 layers it is actually only 5, 3 primary and 2 secondary, I've stripped down a few guitar amp transformers and they all followed this layout more or less.

If you are using an unknown core salvaged from a mains transformer then you should really know something about the core, if you have a variac then wind a wire coil that will fit on the centre leg of your assembled bare core, try for at least 200 turns if you can, set up the circuit in the attachment and starting at 0 volts increase the voltage through the coil in small steps (say 1 volt) and note the volt and amp readings in a list. Initially the current will barely increase with each step in the volts, once the core starts to saturate the amps will start to increase with each step increase in the volts. Stop when you reach a point where the amps rises sharply, the coil may get hot. If you don't have a variac then you may be able use a low voltage transformer with secondary taps from 6 to 30volts in combination with a string 5w resistors to vary the voltage.

Using the amps and volts list draw a graph, of volts vs amps, it will look something like the one in the attachment, a straight line that starts to curve up at the point that the core starts to saturate, you need to pick a point on the graph where you feel the core is starting to saturate, note the volts at that point, then calculate the turns per volt at that point, assuming a coil of 200 turns and volts of 25 then the turns per volt is 8. At this turns per volt the core is starting to satutate but ideally you don't want the core to saturate so to be on the safe side pick a point a bit to the left say it is 22 volts then the turns per volt would be 9. You may want to pick a point even further to the left say 20 volts, then the turns per volts would be 10. Bear in mind this is for 50Hz and for guitar amps this is ok, most guitar amp transformers specs give low end point at around 70 Hz. If you have stripped down a mains transformer to reuse the core and you have taken note of turns and voltages do not use that turns per volt figure, you don't really know how hard the maker is driving the core.

If you have B+ of 150 volts then for instance if using 9 turns per volt the primary winding is 9x150 = 1350 turns, the secondary turns are then an easy calculatiuon. I have used this procedure often and for me it has worked fine for mains transformers as well output transformers in guitar amps, however if doing a mains transformer you need to take voltage drop across the winding into account and compensate for that. If you are wanting something better than a guitar amp output transformer in terms of performance then it gets a bit more complicated and you may need to make tradeoffs.

Once you have settled on the number of turns, then you need to decide if the turns will fit, as a rule of thumb, add up all of the space taken by the wire and insulation and if this figure is more than 75% of the available space then rethink your transformer, ie less turns, thinner insulation, bigger core, you don't want to be getting to the end of the wind and find you need to put the bobbin in a vice to get the laminations to fit, once you become proficient at winding then you may be able to work on 80% or even 85% fill.

[AdrianH]

@retailer

Thank you for your response, I will take note of the variac method, basically finding where the core saturates I guess.

A few things are puzzling me. You call it a guitar amp and yes that is what the powerman amp was for, but in it's simple form without any effects what is the difference between a guitar amp and a standard mono 'audio/sound' amp in construction.

I have read a few articles and most of them tend to go off on topics such as shredding and effects such as tremolo, distortion, volume, bass, mid and treble, and reverb. You can probably tell I am not a musician.

I guess that guitar amps are easy to over drive and produce distortion, where as in audio amps the idea is to keep the signal as linier as possible across the full audio range.

One part that confused me is your comment

Originally Posted by retailer

"If you have B+ of 150 volts then for instance if using 9 turns per volt the primary winding is 9x150 = 1350 turns,"

Is this again in the context of a guitar amp in which you are driving the output stage absolute full swing. I had it in my head that the 150 Volts B+ was just the standing DC supply to the valves and that you operate over the linear region of the valve output which (in my case 70 to 80 Volts pp) hence only going for 2 Watts.

Adding negative feedback helped a lot to clean up the simple amp and I can start to see that a higher feedback voltage from the output transformer could be used to provide better control.

Anyway it is all a learning curve for me and I will not start to understand all the knowledge out there within a few weeks.

I will stick a resistive load on the output and put some the test gear on the amp and see how it does.

I have one last question prompted from you quote. Would the output transformer then have 1350 turns each side of B+ to each Plate/Anode. So a tutal winding of 2700 turns on the primary?

Adrian

So many questions!

[Ed_Dinning]

Hi Adrian, instrument amps are usually designed to allow some distortion, so Stalloy cores are used and may be run up to about 1T. Unisil cores have a less defined and higher saturation point, and are more expensive, so Stalloy is still very often used for instrument amps.
Hi Fi is all about minimising distortion, hence runs at a lower flux density on a more linear material.

Try yous without feedback initially as NFB is a whole lot more brain ache .

Ed