A question about double secondary windings transformers

[mushroom]

Hello,

I have no double secondary transformer, so I cannot test for now. I created a LTSpice model, but I am not familiar with this software, I just downoaded and installed it one hour ago ! Usually, I simulate with an old version of Multisim...

I'd like to know what happens to voltage on secondary winding #2 (L3) when current is drawn from secondary winding #1 (L2) by a rectifier bridge and a capacitor, without magnetic saturation (ideal transformer).

Spice oscope shows me that L2 current has no effect on L3 voltage. The circuit seems to be equivalent to a pair of independants transformers. I was thinking that mutual inductance should cause some perturbation.

I am interested in winding #2 voltage only, as I plan to use it as a "voltage sensor".

Is this true, or am I missing something ?

I understand that "K1 L1 L2 L3 1" creates an ideal air core transformer

[EDIT] I had a double secondary transformer ! A 100 VA industrial 2x24V. Pulling current from one secondary (1 to 2 amps) significantly modify voltage on the other secondary : +- 100 mV or so.

Now, the question is : what's wrong with the model ? The magnetic circuit ? How to modelize a realistic transformer ?

[bugs]

If you use an ideal transformer you will see no voltage drop. You can change the coupling factor K to less than 1, for instance 0.95 to add leakage, but this makes it dependant on the magnetization inductance. You have selected 1H which seems a little too high. you can either add leakage inductance and parasitic serires resistance as separate components, or select proper values for K, and the magnetization inductance, and add some series resistance to the inductors.

[T3sl4co1l]

For k=1, you will indeed observe no difference.  Real transformers might be 0.95 to 0.9999, and you will observe a change, due to equivalent series primary impedance, and whatever paths through the coupling matrix you may have.

Note that, in general, the coupling between N windings is a matrix of mutual couplings, and the total must be conserved (namely, the matrix is symmetrical, and I think the determinant is normalized and nonzero, i.e. the matrix is nonsingular).  The shorthand of "K1 L1 L2 ... Ln 1.0" means the whole matrix is 1, so there is no difference between any windings.  But in general, you can have windings in some arbitrary order where k12 != k23 and so on, in which case you will observe different voltages from different loadings of those windings.

Tim

[Zero999]

I'd even say 1H is too low for a mains transformer, as that would make the magnetising current >730mA which is far too high. 20H is probably more reasonable.

You should add some resistance too.

If you're simulating the mains as an input voltage then it needs to be the peak value which is 325V.

In a real transformer. The turns ratio will also be a bit lower than ideal to compensate for the losses under load, so a 240V to 24V transformer would be 10:1.2, rather than the expected 10:1, assuming the voltage drops by 20% under load.

[Yansi]

For somewhat realistic simulation of a mains xformer, you need to know the primary magnetising inductance, coupling factors or leakage inductances for all windings, primary and all secondary series resistance and loss resistance in parallel with the magnetizing inductance. (represents losses in the core).

[mushroom]

Thanks for your answers.

This is far too complicated for me !!! And when I read "matrix math", this recalls me sigmas, normalizations, reductions and other huge developpements I always hated (and was happy to forget as fast as possible) !

I attempted to understand models with the help of this document : http://cds.linear.com/docs/en/lt-journal/LTMag-V16N3-23-LTspice_Transformers-MikeEngelhardt.pdf

I will probably not go further, because it is too difficult, and because I probably found a better way to solve my problems.

I need a mains voltage sensor (power meter, oscope V & I probes). I discovered high isolation high sensivity current transformers (1000:1000 or so ratio) ; to use them as voltage sensors, they just need a resistor in series with the primary winding, and a burden in parallel with the secondary. It is far more linear at the same expense. (it is somewhat disapointing, as I was (successfully) coding routines to compensate distorsions and phase shifts ; but it's just a hobby !)

I found these components as I was looking on eBay for small double secondary transformers (2x9V, 2VA) in order to experiment what cannot be modelized using my knowledge.

Here is a TV19 transformer, it can outpout 2.5Vpp with a 500 ohm burden, a very low distosion, less than half a degree phase shift, and gives a 2 KV isolation. I did'nt find european source for such components. Will have to order from HK... A current transformer for voltage measurements.

http://www.yhdc.com/en/product/521/

http://www.yhdc.com/en/product/518/