Blocking DC offset with a transformer okay?

[mkiijam]

I have a small-ish (about 2"x2" x .5" lams) output transformer that is dealing with about 200-500mV of DC offset coming from the previous stage. Of course it's blocking it nicely... But will it get saturated? Permanently magnetized or something? I know I can solve it by putting caps before the transformer, but find myself curious about trade offs caps vs DC offset?

[Kleinstein]

The normal transformer material can work with some superimposed DC. There are no long term effect expected - so if it works for a 1 second it will also work for longer.
The details depend on the construction of the transformer (e.g. ring core or EI)  and material (e.g. isotropic or grain oriented or amorphpous or nano--crystalline). The DC bias will reduce the useful magentization (lower max amplitude) and can add some distortion (especially 2nd harmonic). Tube amplifier in part used DC current through the output transformer, but higher power ones tend to avoid / compensate with 2 tubes as push / pull.

The DC voltage is not relevant, the point is the current relative to the magentizing current. So we can't tell if 200 mV may be too much of insignificant depending on the turns and transformer.

[TimFox]

In general, it is advisable to block DC offset from a transformer winding.
Even below saturation, it can add even-order distortion, since the hysteresis in the transformer will be different for positive and negative swings about a DC value.
In single-ended vacuum-tube amplifiers (which have a cult following), the substantial DC current through the output transformer primary winding from the plate to the power supply is tolerated by using a substantial gap in the iron core magnetic path, which has the side-effect of reducing the magnetizing inductance.
If there is no need for DC current through the transformer, it is good practice to add a capacitor in series with the winding.
Past that, you need to give more information about the transformer characteristics.

[mkiijam]

EI. Grain oriented. Primary is a split design totaling 1000 turns of 36 Gauge. Secondary is 1200 turns of same sandwiched in the primary

[TimFox]

How much DC current will there be from your roughly 500 mV offset voltage?
What is the core flux density (B field) from that current?
Is there a circuit problem from an appropriate series capacitor?

[jonpaul]

any DC bias will cause audio distortion odd order.

Use a series electro cap as required.

NEVER DC couple an audio trsf

Jon

[TimFox]

Actually, the distortion caused by DC bias is even-order, since there is an asymmetry between negative-going and positive-going excursions from the quiescent point.
(Some people like that distortion in single-ended tube designs, but I prefer accuracy.)
Odd-order distortion distorts positive and negative excursions in the same way, such as the current waveform in a non-linear resistor excited by a sinusoidal voltage.

[mkiijam]

Okay, so how do I calculate the current. Sorry to ask a 101 type question (maybe it's more 090) But I find myself actually asking that.

The transformer is being driven by a discrete style op amp Jensen 990 / API 2520 etc. And is then married to whatever as a load. But for example could be a DAC with a 10k input impedance or higher. If one side of the primary is grounded then me thinks it would some sort of ohms law calculation with the primaries DCR or impedance and the voltage that the amp is putting out?

[TimFox]

For DC current from a DC voltage, you only need to know the resistance of the primary coil, which is much less than the operating impedance level seen there for audio signals.
The flux density B is a trickier question, since you need to know the complete description, including the magnetic material.
Normally, the manufacturer specifies a maximum DC current for good operation, which should be far below the smoke level.

[mkiijam]

Yes, the DCR is about 35 Ohms.

These transformers are home grown...

[TimFox]

500 mV / 35\$\Omega\$ = 14 mA is not trivial:  it might even load the driver excessively.
I need to review my magnetics before I give you a formula for the B field in the iron, but I believe you need to know the inductance and the number of turns in the winding.

[mkiijam]

Actual tests come in as low as about 15mV

[TimFox]

Calculation of B field in an iron-core inductor:
For ferromagnetic material, the relation between B, the flux density in the iron, and H, which is the product of current and turns of wire, is not linear.
Therefore, to get an estimate, you want a measured value of the self-inductance L near the level you are driving into the inductor, along with the cross-sectional area A and the number of turns N.

The flux linking the windings is  N x B x A = L x I, since the flux linking N turns is N times the flux in the iron.
Knowing N, A, and the measured L, one obtains the flux density (in Tesla)

B = (L x I) / (N x A)

where L is in H, A is in m² and I is in A

[mkiijam]

Wow. I guess I have my homework cut out for me!

[jonpaul]

Mkiijam:

No homework needed.

1/ ANY DC bias in audio trsf, speakers is unacceptable.

Audio amps are designed to have ZERO or perhaps a few MV DC!

2/ Calculate the cap so Zcap <<Zload at lowest Freq Easy to find them in any desired capacitance. Perhaps $1

3/ Get electrolytic , place in series. Voila no DC bias.

SOURCE>>cap>>Trsf

DONE

[TimFox]

Not quite:  plate output transformers for single-ended tubes use transformers with a gap in the core to tolerate a finite amount of DC.

For the series electrolytic capacitor, it needs to be even larger than that required for the high-pass filter formed with the primary impedance.
To minimize distortion in the (non-linear) capacitor itself due to the AC current through it, the voltage across the capacitor should be very small at the lowest frequency in the band.