Can a single transformer be used as an RF mixer?

[neazoi]

Transformers if saturated pass to their non-linear region.
Assuming a single small toroidal core is saturated, can it be used as a simple RF mixer?

1. How should the signals applied? All in one winding? Separate windings for RF, IF and LO?

2. Would that be better to use square loop toroids?

3. Can I DC bias one winding so that I can saturate the transformer with the DC?

 

[BradtheRad]

3. Can I DC bias one winding so that I can saturate the transformer with the DC?

This is what's known as a magnetic amplifier.

https://en.wikipedia.org/wiki/Magnetic_amplifier

Sorry, can't answer your other questions.

 

[neazoi]

This is what's known as a magnetic amplifier.

https://en.wikipedia.org/wiki/Magnetic_amplifier

Sorry, can't answer your other questions.

So the DC bias can be used in conjunction with the input AC applied to the core (RF and LO signals). What the bias will do, it will shift the operational range of the core from it's linear region more towards it's saturation region, am I right?

 

[BradtheRad]

Yes. From the Wikipedia article:

"In saturation, the AC winding on the saturated core will go from a high-impedance state ("off") into a very low-impedance state ("on")."

This method of using transformers is no longer as common as it once was. It does look intriguing. It might find new applications today. Technological concepts have a way of re-emerging, in new forms.

 

[FvM]

I'm not sure what you want to achieve, just demonstrate that the transformer can basically work as a mixer (similar to it's possible to talk through a tin can telephone) or want to reach specific performance parameters.

Regarding DC bias, without it you have a symmetrical I/V characteristic (with only odd terms ax + bx³ + cx^5...) and you'll respectively miss some mixer products associated with even characteristic terms, like f1 +/- f2, you only get f1 +/- 2*f2 and 2*f1 +/- f2.

 

[Dan Mills]

An interesting old school use was in the linearity circuit of a crt telly where a saturating inductor was sometimes used to control pincushion distortion, sometimes even with a magnet glued onto it to offset the BH curve.

The 'DC' in this case was the supply current to the vertical deflection amplifier with the signal being controlled being the horizontal deflection current (Or the other way around, it has been a long time).

Regards, Dan.

 

[neazoi]

I'm not sure what you want to achieve, just demonstrate that the transformer can basically work as a mixer (similar to it's possible to talk through a tin can telephone) or want to reach specific performance parameters.

Regarding DC bias, without it you have a symmetrical I/V characteristic (with only odd terms ax + bx³ + cx^5...) and you'll respectively miss some mixer products associated with even characteristic terms, like f1 +/- f2, you only get f1 +/- 2*f2 and 2*f1 +/- f2.

I have thought that a saturated transformer that works in it's non-linear region, could be used as a mixer like any non-linear device. I am interested in it as a demo only. In fact I have seen a document (https://www.google.gr/url?sa=t&rct=...nCF6IiqHOqRhG2aWfnKX7JQ&bvm=bv.75774317,d.bGQ) that used a saturated coil (single winding) as a means of producing harmonics of a single frequency sine input signal. And guess what, wou are right, it produced odd harmonics. To produce even harmonics the author fed the output to a bridge rectifier.

- - - Updated - - -

An interesting old school use was in the linearity circuit of a crt telly where a saturating inductor was sometimes used to control pincushion distortion, sometimes even with a magnet glued onto it to offset the BH curve.

The 'DC' in this case was the supply current to the vertical deflection amplifier with the signal being controlled being the horizontal deflection current (Or the other way around, it has been a long time).

Regards, Dan.

Yes the magnet to saturate the coil could be a good solution to avoid bias.
I remember I have found some coils in an old TV that had a permanent magnet glued onto them and I did not know what was that for. They were for that reason, thanks!

- - - Updated - - -

The paper states that the non-linear coil is a un-polarized inductor.
What does it mean? are inductors polarized?

 

[FvM]

The paper states that the non-linear coil is a un-polarized inductor.
What does it mean? are inductors polarized?

Magnetic components with permanent magnet polarization are used in RF, e.g. circulators.

 

[neazoi]

Magnetic components with permanent magnet polarization are used in RF, e.g. circulators.

Oh, so it means square B-H loop inductor that has been already pre-magnetized in the upper or lower saturated region?

 

[vfone]

Even if mixers use a nonlinear device, by definition a mixer is a LINEAR translation device, which theoretically should convert a input signal with frequency f1, to an output signal having frequency f2.

Against frequency multipliers which also change the frequency of the input signal f1, mixers theoretically preserve the amplitude and phase without affecting modulation properties of the input signal.

Using a saturated core transformer might be used as a frequency multiplier, but not as a mixer.

 

[Borber]

Single balanced mixer uses only one transfomer. DC saturation of ferrite core will only change inductance of windings. Mixed signals are producing much less change in magnetic density of core than DC current in separate winding can do. Signals are practically always in linear part of magnetic curve although it's slope is changed.

 

[FvM]

Signals are practically always in linear part of magnetic curve although it's slope is changed.

The practical question related to the topic of this thread is, which RF level is required to run through a sufficient curved part of the B/H characteristic, even with DC or permanent magnet bias. We need to know about frequency.

 

[neazoi]

The practical question related to the topic of this thread is, which RF level is required to run through a sufficient curved part of the B/H characteristic, even with DC or permanent magnet bias. We need to know about frequency.

I am basically interested in LF/HF 100KHz-30MHz but it may be too optimistic.