DC current through a transformer, how bad?

[zapta]

I am designing a minimal isolated DTMF dialer for personal use and have a question regarding the transformer.

Is it ok to have DC current through a transformer?  Also, if both primary and secondary have DC current, does the polarity of the currents matter, e.g. to push away the magnetic saturation point?

The transformer I am using is this one http://www.mouser.com/ds/2/449/XC-600132-202474.pdf it is a 600:600 ohm impedance at 1KHz and about 50 ohms resistance. R7 represents a relay switch in the closed position and V1, V2 are the MCU's 0/3.3V square waves at the DTMF low and high frequencies respectively.

Also, the ground at the secondary side of the transformer is for simulation purposes and the line protection is not shown (I am thinking a bidirectional 5V zener across the secondary coil.

Should I worry about the DC current? I would like to avoid coupling capacitors if possible.

[DutchGert]

Well, the datasheet says 75mW maximum so if it has a series resistance of 50R you can do the math

[zapta]

Well, the datasheet says 75mW maximum so if it has a series resistance of 50R you can do the math

Primary current:  1.7V over 100+50 ohms: ~12ma.

Secondary current: ~25ma (phone line characteristics)

15+12 ma over 50 ohms: ~ 70mw.  Good enough for me (the primary current is only for a second or two).

Is power dissipation the only consideration?  Should I worry about magnetic saturation?  Any point in having the DC currents in opposite directions?

https://en.wikipedia.org/wiki/Saturation_(magnetic)

[David Hess]

DC saturation is the real problem and it has nothing to do with the power rating.

[dmills]

There are two types of telephone transformer, one intended to handle the line current and one not, you have one in the not category.

Cap couple the transformer to the line and use a transistor or two as a gyrator to sink the line current (Even a one transistor current sink would probably do).

Regards, Dan.

[zapta]

There are two types of telephone transformer, one intended to handle the line current and one not, you have one in the not category.

1. Should I look for a 'yes' transformer?  What keywords?  Is there a reason why I wouldn't want them (size, price, etc)?

Cap couple the transformer to the line and use a transistor or two as a gyrator to sink the line current (Even a one transistor current sink would probably do).

Regards, Dan.

2. Can I get by by roughly balancing the DC currents in both sides in opposite directions?

3. I am not familiar with gyrators. What do they do?  Will it be powered from the telephone line itself (I want the circuit to be isolated)?

[Kleinstein]

You could avoid saturation of the core by compensation of the DC currents. But than thermal load might become a problem.

Transformers that can handle larger DC current are larger and thus usually more expensive.

I don't think the circuit will word this way anyway, without a relay to disconnect the line, when not in use. Many systems use the DC current to check if the line is in use. So I would consider a completely different circuit, more a an opto-coupler and the transistor circuit on the line side.

[zapta]

Transformers that can handle larger DC current are larger and thus usually more expensive.

I don't think the circuit will word this way anyway, without a relay to disconnect the line, when not in use. Many systems use the DC current to check if the line is in use. So I would consider a completely different circuit, more a an opto-coupler and the transistor circuit on the line side.

Can you draw a quick sketch of the circuit you have in mind?

The current is ~25ma so should be fine power dissipation wise. I do use a relay, the 0 ohm resistor in the schema represents the relay in closed state for simulation purposes.

I did a quick experiment, loading the secondary with 40ma (secondary in a series with a resistor and a DC power supply) and measuring the primary output voltage. As the DC current increases the output signals changes but I couldn't see any difference in the FFT analysis [0 -10Khz, magnitude in DB)


Voltage with no primary DC current:


voltage with ~40ma primary DC current:



Also tried to actually DTMF dialing and it works like a charm, even though the MCU generates a square wave.