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.
My apologies, there may have been  security setting in its properties, but I've unblocked that now so we'll try again.

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? 
This does not seem to be a problem for microphone transformers which may have only microvolts at times.

DaveP
 

Attachments

  • GEC_Approach_to_Amplifier_Design.pdf
    1.8 MB · Views: 15
There is much to it beyond frequency response. Look into hysteresis and barkhausen effects. Many non-linearities, quite a few of them interacting. This why a convolution of a transformer never sounds like it, not even on dedicated multi-sampling systems like the Nebula..

Jakob E.
 
EmRR said:
It does strike me that most modern transformers have lower L because they expect NFB, and that also in part helps high end extension. 
They don't necessarily have a lower inductance, but they certainly have less leakage inductance.
That's a necessity. Stability of an amp under NFB demands minimum phase-shift in the usable BW, which puts constraints on the leakage inductance and stray capacitance of the xfmr.
The most significant benefit of NFB in a power amp is not, as many believe, reduction of THD, but it's reducing the output impedance of the amp, which in turns gives a better control over the speaker movement in its resonance region, helping reduce the boominess. However, in order to achieve this, the NFB ratio must be about 20-26 dB, so the output impedance for an 8 ohm tap would be about 0.5-1 ohm. Reduced THD is a bonus.
This puts a serious constraint on the xfmr's phase response, which requires reducing the leakage impedance.
So the need for less leakage come from the constraints of NFB, not because NFB allows to be lax about it.
Indeed, some manufacturers took the easy path and chose to decrease the nominal impedance at the same time; these are not the designs that have stood the test of time.
 
There is much to it beyond frequency response. Look into hysteresis and barkhausen effects. Many non-linearities, quite a few of them interacting. This why a convolution of a transformer never sounds like it, not even on dedicated multi-sampling systems like the Nebula..
The Barkhausen effect will generate multiple very high harmonics due to its "square wave", a kind of digitization of the curve, however, the capacitance of most transformers will leak these away.  The capacitance will act like a DAC on the discrete steps and smooth them out IMHO.

DaveP
 
Hysteresis may have been a problem for very old transformers, but modern  grain oriented steel laminations have reduced this effect due to their low coercivity.  i.e."soft" iron.  Hysteresis is employed for hard drive magnetic materials with high coercivity.

Maybe CJ can give us more info on very old laminations?

A clue as to compressors that used transformers with poor top end response can found from the schematics.  For example: The Western Electric 1126A and the Collins 26C, which are non feedback designs, incorporated some form of treble boost.  Later, mainly post war designs, incorporated feedback between the output stage and the intermediate stage which cured the problem as EMRR has already pointed out.

DaveP
 
DaveP said:
The Barkhausen effect will generate multiple very high harmonics due to its "square wave", a kind of digitization of the curve, however, the capacitance of most transformers will leak these away.  The capacitance will act like a DAC on the discrete steps and smooth them out IMHO.

DaveP
I remember having seen an experiment on Youtube, where a winding was submitted to a DC ramp and another winding was connected to an amp. The sound of the domains reorienting was definitely midrange.
 
abbey road d enfer said:
They don't necessarily have a lower inductance,

True, but the majority do, and spec far closer to bridging drive Z, almost never for matching condition.  Almost any comparison of new transformers versus old is apples to oranges with that being one of many parts. 
 
if you could somehow read all the files on transformers, Western Electric, Peerless, UTC, Triad,  Telefunken, Partridge, Magnetic Materials, National Arnold, i bet you would find a ton of interesting material.

i wish i knew what dumpster they went into,



 
Where I used to work, at BBC External Services, they had bay upon bay of valve line amps and rep coils. The rep coils were about the size of a small Hovis loaf (that's a small loaf for those on the other side of the pond) and were designed to pass 17 Hz ringing tone from remote stations so presumably had an awesome frequency response

Fixed lines between sites were equalised with a fixed equaliser and attenuator followed by a line amp. The lines were copper pairs leased from BT

I wonder what happened to all those transformers? There were hundreds of them. Sold for scrap, I imagine, although a lot of the other hardware turned up in auctions when External Services moved from Bush House

Nick Froome
 
One of my most major experiences later in life,  is getting hold of some few bunches of  local  vintage traffos  ... I'm sure many, many were recycled from the day  ... the fifties and sixties.

By golly, the construction and performance can be an eye opener, in today's age where stuff is landfilled after only a short period of use.

These guys I have, early to mid 50s, Australian made in Melbourne ...  (local subsidiary of  Swedish giant  Ericsson)  are works of art. They recall the glory ways of the US and UK greats, and the not too much worse manufacturings from the colonies.

I do recall, working for STC late 80s-early 90s  ...  local subsidiary of UK giant STC which became part of the huge Alcatel of France  ..    a lot of those 'line' traffos  ..  used in telephony  . Like many, many .. many!

...

It's mostly the reason I like to test  tubes  .... just to experience these historical  great  combos  ... but in a measuring environment and only after I've really honed in on the basics.
 

Attachments

  • Trimax 30K-600 +27dBm traffo pic5.jpg
    Trimax 30K-600 +27dBm traffo pic5.jpg
    262.2 KB · Views: 20

Latest posts

Back
Top