Explain transformers in the presence of DC flux?

[daviddeakin]

I'm trying to learn about the effects of net DC in transformers (to see what limits their use as cheap smoothing chokes*) and I'm tying myself in knots.

For example, I have a little 12V, 6VA transformer, with a 700R primary resistance.

I put a 10R resistor in series with the primary and measure the AC voltage across the primary (240Vac) and across the resistor: 260mVac, i.e. 26mAac primary magnetising current, or 36mA peak. (No secondary load)

The primary impedance is therefore 240V/26mA = 9.23k ohms

Subtracting the resistance (quadratically) gives an inductive reactance of 9203 ohms, so the gross inductance is 9203/(2 pi 50Hz) = 29 henry. Right?

As I understand it, the transformer is designed to run close to the point of saturation. Therefore 36mAdc is enough to saturate the core, near enough (about 1.5 tesla so Google tells me). Right? So if I (by whatever means) biased the winding with a DC current equal to half this figure, I could still apply 120Vrms and I would still be operating within the limits of the core and I should still acheive roughly the same gross inductance. I have done a ghetto sketch. Is that correct?

*Yes I know chokes have an air gap. No I can't dismantle the core. This is an academic question.

[G8HQP Dave]

I think that is correct, but there are people on here who know far more about transformers than I do.

[G4_Pete]

Hi if you research "Saturable Reactors" and "Magnetic Amplifier" you can get some further info on this.
Pete

[joebog1]

It sounds to me like the transformer has no air gap!! It will work at a few mA only.
Chokes have an air gap, transformers generally don't (there are special "lossy" transformers that DO have an air gap.)
Normally a choke ( there are two "basic" types" ) doesn't get anywhere near saturation unless its at full current rating.
Choke type one is called a swinging choke, i.e. it is normally situated directly after the rectifier, and will "swing" between the maximum current and the minimum current.
Type two is generally called a "smoothing choke" and is situated between the swinging choke and any low current circuits, like the voltage amplifier and phase splitter circuits, or pre driver circuits.
The tap between the two chokes would normally feed the high current circuits, like the output stage of a class B amplifier, or in the case of radio the final output stage.
Between these two chokes could be a number of different capacitors, that also aid the chokes in their duty, which is to limit large current changes, and consequently large voltage changes in the power supply in general.
Chokes "save up" extra power when its not needed, and release it when a large current drain happens. Capacitors do the same thing, but work at different times during the rectifiers cycle and the demand from the circuit it is driving.
There is some very scary math involved to "calculate" the exact values, but this is not normally necessary in most cases, UNLESS you are designing from the word go, and need special requirements.
In essence, a choke stops BIG currents from pulling ALL the power out of a power supply instantaneously, by just allowing a slow decrease over time. It does this by releasing some ( or all) of the flux energy in the iron.
OR should some circuit require a large current, the capacitors involved with the filter circuit will release some extra current, over time, controlled by the inductance of the choke, to the circuit that requires the extra power.
In this way the voltage output is known as a "choke regulated" power supply.

With careful design and good quality chokes and capacitors, it is relatively easy to get a fairly constant voltage output from a power supply, without the need for regulators, or regulating valves or integrated circuits.

You have asked a very simple question, that involves very complex answers .
I hope somebody here can explain far better than I.

best regards

Joe

[kalee20]

I checked the arithmetic. 26mA at 240V is 9231 ohms - yes agree. (We'll differ by 1 ohm due to rounding).

Subtract 700 ohms in quadrature, 9204 ohms, yes agree.

This is equivalent to 29.3H - yes agree.

Now - 36mA (peak of 26mA) being enough to saturate the core - yes you are not far wrong (certainly good enough for a crude estimate on a bog-standard transformer).

So will 18mA take you half way to saturation, leaving you able to apply 120V RMS (which you would then expect to swing the flux from zero to just-saturating? Yes if the core was linear. Unfortunately, the non-linearities of the core will kick in quite severely!

Once you have magnetised the core, removing the current leaves quite a lot of flux remaining as residual magnetism. It's basically become a permanent magnet. So the next application of excitation in the same direction, leaves you with very little flux swing before saturation happens. On the other hand, applying excitation in the reverse direction, gives you with a large flux swing before reverse saturation.

On AC mains, this doesn't matter. But if you switch off at the wrong time, you can leave the transformer magnetised - and next time you switch on, rather from starting from flux zero, you can start from significant flux, saturate the core quickly, and blow a fuse. Toroidal transformers are very prone to this.

Transformer cores don't make very good permanent magnets - if you dismantle them the magnetisation is easily lost. But, while assembled and interleaved, the thing is like a double horseshoe magnet with permanent keepers.

If you had only a very small air gap, you'd find your theory pretty accurate, because just a tiny gap effectively kills the remanant magnetisation. (It also reduces inductance, too - but things get much more linear).

[G8HQP Dave]

Originally Posted by joebog1

Choke type one is called a swinging choke, i.e. it is normally situated directly after the rectifier, and will "swing" between the maximum current and the minimum current.

A swinging choke is one whose inductance changes significantly with current within its normal working range. This does not matter too much, as a choke input supply needs more inductance when current is low (and vice versa) so a cheaper smaller choke can be used. Note that a choke input supply does not require a swinging choke; an ordinary choke can be used provided that it can cope with the voltage variations without getting too noisy.

The basic way a choke input supply works is by low pass filtering the output of a full wave rectifier, so you get a DC output equal to the average voltage in the full wave rectified waveform - which happens to be almost exactly 0.9 times the AC RMS voltage. For this to be true the rectifier must always be on (via one path or the other) which in turn requires that the minimum current through the choke is non-negative, which in turn requires a minimum average current which if not met causes the output voltage to rise because you are now averaging only the higher voltage parts of the input waveform.

If I recall correctly, the 0.9 above is 2 x sqrt(2) / pi. You get this when you do the integral to find the average value of a full-wave rectified sine wave.

[joebog1]

Thanks for a better explanation Dave.