DIY Input isolation filter

[guitar89]

I wanted to use an isolation transformer between the input source and amplifier input, but its expensive to buy one (good ones, at least perform good in audible range).

Thus I'm not going to buy one, but build one myself ! (have no clue how to do this)

The first problem is that it must able to perform from 20Hz to 20kHz without problem. (how to achieve that ?)
Second requirement is 1-to-1 ratio turn. (not much problem)
3. Low powered is enough (the source is only about 2V maximum)

I originally was going to throw away the idea of using isolation transformer, but i came across that computer PSU mostly use SMPS(switching mode), which have transformer at high frequency(typically 50 kHz – 1 MHz) so its far above 20kHz, but does the transformer perform bad at low frequency ?
(I have few old unused PSU)

[theAnonymous1]

Originally Posted by guitar89

.....but does the transformer perform bad at low frequency ?

Yes.

There is good reason why audio transformers cost a lot of money and why you will find little info on building one yourself.

[guitar89]

wahaha...... any design/details how an transformer to perform at low frequency ?
Maybe we could do something like crossover to split into 2 transformer group (low and high frequency) ? Or should dump away this idea ?

[theAnonymous1]

Here, grab a bottle of your favorite headache medicine, a stiff drink, and give this a read.....

http://www.jensen-transformers.com/a...%20Chapter.pdf

[chrispenycate]

In my far distant youth I rewound a number of transformers, generally to get a decent six hundred ohm drive from valve (tube) preamp output stages, so involving impedance conversion calculations. I invariably started with a quality audio core, savaged from piece of gear I was scrapping, generally E-I cores. Winding, and impregnating to make certain there were no rattles or buzzes (physical movement of the coil , particularly resonant at one frequency, reflects through as distortion and frequency response variation through the finished product.

Two volts at 20Hz is quite a chunky core (although I tend to think 600Ω, rather than the 10kΩ this will presumably be). What are you driving from? Why do you need the ohmic separation (I'm not saying it's a bad idea, note. But generally your grounds will be linked somewhere, anyway, and inducing enough volts to damage input stages takes power in the order of lightning strikes)? Could you not get by with an actively balanced input, zener diodes and fuses?

The good news is that transformers practically never wear out, so there will be a lot of junk gear around where the input transformers are still good, while the rest has decayed below usability.

All this assumes it is not a guitar or similar instrument you are attempting to isolate (and, considering your user name, this is not a safe assumption). High impedance sources are extremely difficult to build decent transformers for, and if you separate the grounds, to prevent possible electric shock between your amp and PA mics or stage fittings you are going to introduce the most impressive grounding problems you can imagine. Sure, you can redo all the pickup wiring balanced, and connect strings and machine heads to an independent earth spike, but you get enough interference even with standard grounding. I suspect the optimum for that problem would be a radio link.

Or maybe an analogue digital converter in the instrument, and a fibre optic connection…

[guitar89]

haha, the pdf is quite "headaching" me, attempting to understand it later (its late now)

I was intend to use for ground loop prevent (portable solution), so balanced input kinda troublesome connection for this purpose(portable), at least for my understanding.
that was quite some problem..... processing first....

Extra Idea : suddenly think of a new idea, although it increase complexity and cost components.

To make the audio frequency multiple by many times (maybe x 1k), thus making the frequency very high, using the SMPS transformer (50kHz to 1MHz), then de-multiply the frequency, thus increase efficiency and response.
(same theory as Class D amplifier)

[AndrewT]

Hi Anon,
Thanks for the link to that chapter.

I have followed much of Whitlock's papers with keen interest.
He seems to be one you can rely on to tell the truth.

This is the first time I have bumped into a chapter from his book.
Lot's of good info in there. I'll need to do some swatting, even though he claims this as the simplified version..

Thanks again.

[AndrewT]

Bill explains at least part of the difference between output transformers and input transformers.
I have not seen anything yet on the effect of turning an input transformer back to front.

Suppose we had a 1:2 transformer where the two ratio was provided by a pair of series connected near identical secondary windings.

Could this transformer work equally well as a 1:2 (series secondaries), or a 1:1 (paralleled secondaries), or a 1:1 (paralleled primaries, i.e. back to front), or a 2:1 (series primaries)?

[Speedskater]

Originally Posted by AndrewT

Hi Anon,
Thanks for the link to that chapter.

I have followed much of Whitlock's papers with keen interest.
He seems to be one you can rely on to tell the truth.

This is the first time I have bumped into a chapter from his book.
Lot's of good info in there. I'll need to do some swatting, even though he claims this as the simplified version..

Thanks again.

It's not Mr. Whitlock's book, it's a chapter in a very big book (1250 pages) edited by Glen Ballou.

Handbook for Sound Engineers, Third Edition
Glen Ballou, Editor

Some time ago, I read a rumor that Bill Whitlock was writing his own book, but have not seen any evidence.

[counter culture]

@ EI-14 Audio Transformer Customer Made Impedance Ratio | eBay