# Output Audio Transformer Impedance

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

#### kawasaki99

Is there a standard frequency at which the impedance of an audio output transformers primary winding is stated? I have a Hammond output transformer whose primary winding measures 3.3k @ 100cps, 20k @ 1000cps and 40k @ 10000cps.

Regards, Richard

#### bob91343

The primary impedance is NOT the open circuit impedance. It is the loaded impedance, with the secondary having the rated load.

Having said that, the 'standard' frequency of measurement is 1 kHz.

#### SemperFi

Is there a standard frequency at which the impedance of an audio output transformers primary winding is stated? I have a Hammond output transformer whose primary winding measures 3.3k @ 100cps, 20k @ 1000cps and 40k @ 10000cps.

Regards, Richard

As mentioned u must test OTs under load. I would think u measure inductance of the primary when it is unloaded.
So since Z of inductance = 2*pi*L*f, u get an inductance of L = Z /(2*pi*f) which gives u 5.3H, 3.2H, and 0.6H at those frequencies. The capacitive part of the OT is obviously taking charge as frequency goes up. Tho I'd think it shouldn't vary as much as this below 1kHz?

#### Yvesm

. . . Tho I'd think it shouldn't vary as much as this below 1kHz?
The level is also preponderent, most inductance testers (even highly respected and respectable bridge meters) usually give very lo numbers because the iron is unsuficiently "excited".
Above approx 5Khz the iron vanishes

Yves.

#### kawasaki99

Thanks for the reply; since a voltage drop measured across a component is, in fact, a measure of inductance, placing a known inductor in the test circuit along side the test component provides a comparsion of the two individual voltage drops providing a good measure of inductance at any particular frequency. Obviously the test circuit must, as closely as possible, reflect the working circuit. The original question still stands: is there a standard test frequency for stated transformer inductance ?
Regards, Richard

#### THD+N

Thanks for the reply; since a voltage drop measured across a component is, in fact, a measure of inductance, placing a known inductor in the test circuit along side the test component provides a comparsion of the two individual voltage drops providing a good measure of inductance at any particular frequency. Obviously the test circuit must, as closely as possible, reflect the working circuit. The original question still stands: is there a standard test frequency for stated transformer inductance ?
Regards, Richard

I use 1kHz to measure unknown transformers to find the impedance ratio.

#### Steve Eddy

Transformers don't have any particular "impedance" per se.

What they do is reflect impedances as a square of their turns ratio.

Of course the reflected impedance appears in parallel with the primary inductance so you don't want to do your measurements at too low a frequency. 1kHz is about as good a "standard" as any as if it's a decent transformer the reflected impedance shouldn't be too drastically effected by primary inductance.

se

#### kawasaki99

Thanks again, I think I got it.

Regards, Richard

#### BudP

Just a couple of details to make things harder to comprehend... of course!

The impedance formula given is correct, for -3 db at the frequency of interest. To achieve -0.1 db at this frequency multiply the derived inductance by 2.76. If you measure with 120Hz @ 1 volt, you will be measuring the core at rest, which is not zero flux but remnant flux, what it will provide with minimum excitation. To understand what happens with more power applied you must have permeability curves and an inductance formula. There are two permeability curves, one for AC signal and one for any core with DC on it and a gap.

The impedance of a primary is equal to the inductance of the primary. To match impedance to a tube you need to add the transconductance and turns forced load line in parallel, then add the DCR of the primary in series. This is the number to use to decide if your transformer will match with a tube at the lowest frequency you care about.

For E/I commercial grade core the core is all done with power transform, except as a ferrous bounding box for the coil antenna event, from 250 Hz to 400 Hz, core grades M50 to M3 respectively. The capacitive coupling and the direct EMF flux transform from primary to secondary provides everything else.Other metals work at higher frequencies, but there are trade offs to be made when using them so don't assume that the very best core will provide you with what you are looking for, once you decide to move beyond Hammond and EDCOR levels of performance. Not that either of these two have anything to apologize for.

Bud

#### Steve Eddy

Uh... what was that middle part again?

se

#### SemperFi

The original question still stands: is there a standard test frequency for stated transformer inductance ?
Regards, Richard

Wasn't that answered clear enuff in post 2?

#### artosalo

I have a Hammond output transformer whose primary winding measures 3.3k @ 100cps, 20k @ 1000cps and 40k @ 10000cps.

In general I prefer to see the inductance of the primary. Then the min. usable frequency of the amplifier can be easily calculated.

For example Hammond specifies the impedance ratios / reflected impedances of their lower cost 125 -series transformers, but not give the value of primary inductance (which are too low for HIFI-purposes).

Therefore many beginner have dissapointed about the actual results of their amplifiers. The reason is the the impedance values have been sufficient, but due to low primary inductance the real performance is limited to 100 Hz and above.

#### SemperFi

True, but when they claim a given OT has a response from 20 to 20kHz, the inductance must actually be enuff for those specs and we shouldn't need to measure the inductance. Tho I agree I also like to measure the inductance so I know what I got.

Those of u with low-L OTs simply aren't using the right tubes Or build a guitar amp.

#### Yvesm

As long as the source impedance (as set by tube's Rp, amount of feed back and all those kind of things) is not specified, the published frequency response is meaningless.

With a true "zero" impedance source any transformer will pass any frequency range at any power . . . or die

Yves

#### SemperFi

Isn't it published? Isn't that why they spec load impedance? So a 5kohm primary should be flat from 20Hz when driven by a tube suitable for that load?
In said example the L must be at least 5000/(2pi*20) = 40H
Wow, that's a lot of turns! Any one wonder why I stick to low Rp tubes?

#### Yvesm

Yeap, 40H for 20Hz at -3dB . . . with a source having large impedance relative to 5K as for a pentode for exemple.
But with a 1K one (as typical for a 300B) L should only need to be 1000/(2pi*20) = 8H
Of course this inductance is to be measured at 20 Hz and for the level you plan to obtain.
A light unit with little iron core could saturate with few milliwats at this frequency not to tell about the effect of DC component.

"In doubt, use more iron"

Yves.

#### SY

Yves, Bud, while I have you guys here... All thing being equal, when you increase the core size, the leakage inductance also tends to go up? If we ignore the bass performance, is there an advantage to be gained by making a physically smaller transformer?

#### Yvesm

Yves, Bud, while I have you guys here... All thing being equal, when you increase the core size, the leakage inductance also tends to go up? If we ignore the bass performance, is there an advantage to be gained by making a physically smaller transformer?

Hi SY,

Almost all formulae that I've found here and there to compute leakage inductance have a common structure like that:

LL = turn number squared * total winding thickness / link surface / K

K being a mysterious coupling factor involving number of sections and how they are disposed and interleaved (but you have seen that in Linear Audio #0)

With a larger iron, one can reduce turn number (for equal primary inductance), reduce thickness and increase link surface so I firmly beleive that the leakage inductance is lowered.
Unfortunatly, various parasitic caps increase and this may be worst because unwanted and sometimes deep resonances start to appear.

Yves.
We drank "Rasteau" without you

#### SY

We drank "Rasteau" without you

I forgive you, since that evening, we drank Batard-Montrachet without you.

If you were going to design a transformer that completely ignored the bass (let's say below 100 Hz) in favor of optimizing treble, how would you do it?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.