Audio Transformers are like Monster Cables

I'll give this a shot. Old modem cards are virtually free. The transformers aren't shielded though.

Incidentally, it would be the radio driving the transformer.

I have a HP8309B, so I can check the THD.

load.

the

I see Tamura has a mu-metal shield as an option, so this looks like a good candidate. The ultrawide seris is 100-100kHz.

I haven't done transformer coupled design, so here are a few questions:

1) I assume to get the signal from the radio (unbalanced), if I am breaking the ground, then I would use two wire plus shield cable between the radio and the transformer. The radio would not use the shield, but rather the inner conductors. The ground potential of the cable would be the ground of my circuit, which would also be one side of the secondary. 2) Transfomers (from the dark ages of my education) just reflect impedances, so I don't quite get the impedance spec of the transformer. If impedance matching is an issue, I suppose I can use an op amp in inverting configuration to have a known input impedance, but I would have little control over the radio output, other than the impedance is probably low. 3) some radios will have a coupling cap connection. Is this a problem with the transformer?

Maybe I should have started another thread....

How do you know this ?

Graham

If only more ppl understood that !

Graham

In article , Harry Dellamano wrote: [snip]

Yes the data is all there, but encapsulated into an application specific form.

Was it based on this simplified equivalent circuit Harry?

Rg lp Rg = Source Resistance. +---/\\/\\---o o---+---))))---+ | | | Lp = Primary Inductance. | ) \\ Vgen Lp ) Rrp / lp = Leakage Inductance, | ) \\ referred to primary. | | | +----------o o---+----------+ Rrp = Load Resistance, referred to primary.

LF, -3dB = F1 = R/2.pi.Lp, where R = Rg*Rrp/(Rg+Rrp).

HF, -3dB = F2 = R/2.pi.lp, where R = (Rg+Rrp).

Those transformers are all 3dB down at 400Hz and 250KHz, so for any one of them you can calculate Lp and lp.

Factor in the winding resistances when you feel they are becoming significant.

You've already noted that the 50R+50R handles 1Vrms at 400Hz. If you want to run at say 1Vrms at 100Hz then you could choose the one that does 4Vrms at 400Hz, which of course is the 800R+800R transformer..... but the winding resistances then cause trouble. The better answer for a lower frequency response is to choose a larger transformer.

--
Tony Williams.

Hey Tony, I like your "Yes the data is all there, but encapsulated into an application specific form". How true, how true. The 20mW @ 400Hz and 125mW at 1KHz all makes good since because transformers operate on volt*seconds ratings and I^2*R losses but even their Rdc is hard to work with at higher frequencies. Let's say they also told us Lp just to be nice. I guess it's best to buy some and measure the parameters needed for a good design. Regards, Harry

I wouldn't trust those Audio blokes further than I could throw a fat Australien who boot polishes his head to hide his hair loss.

Last time I looked, as a result of the Larkin, all of this shit was made with mu-metal and the idea of some sort of specification of anything meaningfull was just s**te.

Since you do switch mode power supplies you should be able to figure all this stuff out and reach your own conclusions of what might or might not be good... Let's face it, everyone else lurned 4.44 and still can't make a useful approximation.

All you need is a guess at the data. Bloody hell, you might be babble enough to design your own. Ask yourself just how much you care?

DNA

"Harry Dellamano"

** How confused is this cretin ?

Maker's rated load impedance value for an audio band transformer is the

*OPTIMUM* load impedance.

It is optimum firstly in terms of frequency response - when underloaded expect to see a HF response peak, when overloaded expect to see an early HF roll off.

Efficiency ( within the pass band) depends mainly on the ratio of winding resistance to load impedance - typically that ratio is about 10 to 1 but may be higher.

Eg: the F28045 1:1 ( turns ratio) tranny is rated at 100 ohms impedance and has 12 ohms resistance in each winding - so efficiency will be around 80 %.

** No, the correct design procedure is to use a tranny with no better LF performance than needed for the app and operate it at the maker's rated impedance. ** Audio designers know what they are doing.

Been routinely getting high efficiency, low distortion and ruler flat response over a frequency ratio greater than 1000 to 1 for the last 70 years.

...... Phil

I'm not sure if you have twigged it Harry. The data *is* there (well, roughly anyway).

eg, For the transformer designed for 50R+50R source:load.

The 20mW max at 400Hz has already given 1Vrms max at 400Hz, and in turn gives 0.5V/.25V/etc at 200/100/etc Hz.

The 125mW max at 1KHz and up gives a 50mArms maximum current in the windings.

3dB down at 400Hz (F1) gives an Lp of 10mH. 3dB down at 250KHz (F2) gives an lp of 63uH.

Note that the 3dB frequencies are specified at 1mW. This runs the core at a low flux density and keeps the permeability high. This gives the highest Lp and lowest F1 to show on the data sheet. If done at 20mW the Lp would be much lower, possibly in the 3-5mH region.

The only thing the 'encapsulated spec' hasn't given is the transformer's capacitances.

--
Tony Williams.

Tony, I understand what you are saying but to unencapsulate the data for the model you must know how they took the data. What was their source resistance, 0R, 50R or Primary Z? Now primary Rdc must be added to it. I can't believe that with this wide range of matching Zs that all frequencies responses are 400Hz to 250KHz. I can foresee some strange stuff happening with a Hi Z unit at 250KHz. When I do your model in my application, the lowest Z that can support the volt seconds is the best choice. Highest regards, Harry

Usually, it works even better if it's driven from a very low impedance source and loaded with rated or higher impedance. "Rated impedance" is sort of a standard/tradition for specifying transformers, but not the only or often the best way to use them.

Agreed. Years of trial and error will eventually find the solutions to a few simple problems.

John

"John Larkin"

** Fuck off - bloody imbecile.

....... Phil

Getting 3 decades of flat frequency response isn't quite a simple as you might think.

Graham

My transmission-line transformers get 6.

John

I wish you would, Phillis.

--
Service to my country? Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

Darn, you always snip my better lines.

John

John Larkin a écrit :

Yeah, but does it get down to a few Hz, still delivering tens of watts or more, and has a 10:1 or more ratio? And good output transformers did certainly more than 3 decades.

Clearly some people knew what they were doing. I doubt many people could achieve this nowadays.

--
Thanks,
Fred.

might

I don't quite understand your point, but I can sure do stuff with Metglas, modern insulation, and modern softwate that nobody could dream of 30 years ago.

John

Well, they are "Pico!"

John