Quantifying the Transformer Magic?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
3,019
Location
France
I was reading " An approach to audio frequency amplifier design" by the British firm GEC from 1957.

In the appendix, it gave a table to define the properties of transformers that I thought might be of interest to the forum.
2up5zsn.png


This applies to the top end and the amount of Henries will define the bottom end.

Ian's formula is very useful here, as it defines the source resistance required to allow the transformer to operate at -1dB at the frequency of interest:-
First you need to know the impedance of your driving source. Let's say is is a microphone with a source impedance of 150 ohms.

The low frequency response limit of the transformer is determined the primary winding inductance of the transformer at the lowest frequency of interest. For a 1dB drop at the lowest frequency you want the impedance of the transformer to be twice the source resistance at that frequency. So for a 150 ohm source and 1dB down at 20Hz we need:

2*pi*20*L  (transformer impedance at 20Hz) = 300 (twice 150) so L = 300/(40*pi) = 2.4 Henries
https://groupdiy.com/index.php?topic=43026.0  from 2011

Most of the vintage transformer plots I've seen, seem to have less high end response than modern types.  I guess this sweetens up the digital music we have nowadays.

DaveP
 
Aah thanks, but there's much, much more to it than this...

Whoever comes up with a decent transformer model that incorporates "what we like about transformers" is going to end up rich.

And industry is painfully aware of that.

Jakob E.

Your link 404's:  Attachment Not Found
 
The gut tells me this is mainly about what happens with ultr-linear operation and NFB. 

It does strike me that most modern transformers have lower L because they expect NFB, and that also in part helps high end extension. 

These may be thoughts before coffee....
 
It is all horribly inter related. Leakage inductance affects frequency response because it is effectively an inductor in series with the output. It is caused by the primary flux that is not coupled to the secondary. To lower leakage inductance you therefore need to increase coupling. The simplest way to do this is to wind primary and secondary together (bifilar winding) but this is not really possible with very high ratio transformers like tube to loudspeaker ones so instead winding are interleaved. Bifilar winding is great for reducing leakage inductance for example in 1:1 transformers but it does greatly increase inter winding capacitance. This is not really a problem for an output transformer but it is a disaster for an input transformer.

Cheers

Ian
 
I recall back in the 80s talking with Deane Jensen (RIP). He pioneered using modern computer based transfer system measurement technology (IIRC he was a HP guy using a big dog HP ***spectrum analyzer as front end for HP computer system).

My judgement was that Deane was not a "transformer as euphonious effects"  guy, but pursuing (adequate) bandwidth for high fidelity audio with good phase response and linearity. 

JR

*** Coincidentally when I was hired by Peavey in mid '80s they asked my for my equipment wish list. Since they asked, I added the full HP system (a few tens of $k). Rather than ask me about my equipment request they second guessed me and bought only the spectrum analyzer (front end) but scratched the HP computer and software to make it a complete working measurement system. So I ended up with a very expensive doorstop (spectrum analyzer)  that I rarely could use effectively. It was nice but I could have saved them $10k+ if they asked me.
 
There are two approaches to this as I see it, either we say it's all a mystery and we will never know and we must avoid that black cat that crosses our path, ??? or we chip away, test and experiment and find out a bit more about the subject.

At low levels on the straight section, one transformer is going to be like any other if the frequency response matches.
At high levels the particular BH curve will come into play, much like operating a tube into the extremes of the load-line.  So harmonics will be generated.

Science has solved more complicated problems than this IMHO.

DaveP
 
All I can do is measure what lands before me, and note that many of those extended treble modern output transformers will not work in no/low feedback ancient circuits for lack of bass. 
 
DaveP said:
There are two approaches to this as I see it, either we say it's all a mystery and we will never know and we must avoid that black cat that crosses our path, ??? or we chip away, test and experiment and find out a bit more about the subject.

At low levels on the straight section, one transformer is going to be like any other if the frequency response matches.
I am not the transformer guy so forgive me if this is obvious but in the extreme at very low levels it is going to take some finite amount of energy to get the magnetic domains moving. Is this completely qualified by frequency response? 

JR
At high levels the particular BH curve will come into play, much like operating a tube into the extremes of the load-line.  So harmonics will be generated.

Science has solved more complicated problems than this IMHO.

DaveP
 
DaveP said:
That points to lack of Henries in modern transformers, then Doug.

DaveP

Right.  I feel like I'm repeating myself; they are generally designed with NFB as an assumption, not with the assumption a matching condition will ever be encountered.  All the 'old' stuff has the henries for matching condition. 


The attachment still shows up as 'not found' here. 
 
JR> .... Peavey  scratched the HP computer and software to make it a complete working measurement system.

You should have shown them this film. I've linked to the part just before the computer.
https://youtu.be/424oXB6-44Y?t=156
1957 Canada designed and built a sub-chaser. The clip starts with the wind-tunnel then moves over to the ElectroData mainframe. (The ignorant editing butts the digital against an analog machine, but the big Nixies are cool.) Spoiler: the Canadair Argus not only flew, but flew really good.

 
The attachment still shows up as 'not found' here.

Do you mean the GEC attachment, that has been downloaded 32 times already.

The GroupDIY ref is from  Transformer frequency response
posted by sr1200 on 4Feb2011

https://groupdiy.com/index.php?topic=43026.0

DaveP
 
DaveP said:
There are two approaches to this as I see it, either we say it's all a mystery and we will never know and we must avoid that black cat that crosses our path, ??? or we chip away, test and experiment and find out a bit more about the subject.

At low levels on the straight section, one transformer is going to be like any other if the frequency response matches.

At high levels the particular BH curve will come into play, much like operating a tube into the extremes of the load-line.  So harmonics will be generated.

This is not entirely true. At relatively high frequencies it is true but with most core materials  there is an increase in low frequency distortion at both low levels  and high levels.

If you haven't already read it I suggest Bill Whitlock's excellent chapter 11 on audio transformer basics in the Sound Engineers Handbook.

http://www.jensen-transformers.com/wp-content/uploads/2014/09/Audio-Transformers-Chapter.pdf

Cheers

Ian

Edit: link added

 
DaveP said:
The attachment still shows up as 'not found' here.

Do you mean the GEC attachment, that has been downloaded 32 times already.

the GEC attachment is not found
 
EmRR said:
the GEC attachment is not found

nothing here either...... I think it shows up downloaded even if it's unsuccessful.....like my first attempt at spelling unsuccesful.... :-[

Here's an interesting spin on some of this jargon..... I just came across it recently in a Radio Constructor article.... 49 or 50s



 

Attachments

  • choosing an output transformer1.jpg
    choosing an output transformer1.jpg
    227.7 KB · Views: 24
PRR said:
JR> .... Peavey  scratched the HP computer and software to make it a complete working measurement system.

You should have shown them this film.
I repeat it was not a negotiation.. my immediate boss passed along the request for a test bench "wish list", and then passed it back uphill (even before I showed up in Meridian). If he was asked, he wouldn't have a clue about the specific gear. In hindsight I probably should have prepared an actual proposal explaining the sundry parts and pieces.

My real irritation is getting half a system, I hate wasting money, even other people's money (even people with too much money).

JR
 
IIRC Someone that posted here in the past was working on measuring what makes transformers different.
 
Back
Top