Gyrators to replace PSU chokes?

[Merlinb]

Originally Posted by ray_moth

Thanks for the information. Can anyone explain how to calculate the simulated inductance for a MOSFET gyrator? Would an IRF820 be OK? I'm thinking of using it in a stereo EL34 PP amp.

Here:
Active Choke for Valve Amplifiers

Gyrators can indeed take the place of smoothing chokes. The circuit shown in the link is current limited to about 50mA, but you can increase the zener to 12V, and R3 to 10 ohms to get up to 200mA dc continuous, with a clip-on heatsink.

(You can even make a gyrator-input rectifier if you isolate the gyrator with diodes, although I can't see much point in such a topology.)

[revintage]

Can anyone explain how to calculate the simulated inductance for a MOSFET gyrator?

Ray,
Forgot your question about simulation. It´s quite easy if you do it by trial and error:

Do two models with the same capacitors.
Do first wih a choke and the other substituting the choke with a gyrator.

Then apply the same DC-voltage with added ripple.

Try out what choke-value gives equal ripple as the gyrator one.
You can use the .step directive to find the right value.

[Merlinb]

Originally Posted by revintage

Ray,
Forgot your question about simulation.

A more accurate method is to do a frequency response plot of the gyrator and then work out the inductance from the resonant frequency:

L = (1/2*pi*f)^2 / C

[revintage]

A more accurate method is to do a frequency response plot of the gyrator and then work out the inductance from the resonant frequency:

MerlinB,

You´re correct but why doing it the hard way.

I´m sure you have a Spice to find out. That is the easy way and accurate enough.

Besides, asked Ray for how to simulate. I know he also uses LTspice and can get file the from me.

[Merlinb]

Originally Posted by revintage

MerlinB,
I´m sure you have a Spice to find out. That is the easy way and accurate enough.

Actually I was talking about SPICE!

[Michael Koster]

Originally Posted by Merlinb

A more accurate method is to do a frequency response plot of the gyrator and then work out the inductance from the resonant frequency:

L = (1/2*pi*f)^2 / C

You mean put it in say a series resonant circuit with a known C value and do a frequency vs impedance plot?

Wouldn't it be easier to plot the gyrator impedance vs frequency and calculate L based on the -3db cutoff frequency?

With the MOSFET based gyrator circuits I build there is a simple RC that determines the Fc. I usually calculate the -3db point in circuit and don't even worry about the equivalent inductance.

If it's a power supply filter, you can use a constant-voltage-referenced follower and get both ripple reduction and regulation with the same pass element.

Cheers,

Michael

[TheGimp]

I looked at Gyrators a while back and A couple of things come to mind with regards to their use

(1) A Gyrator does not store energy like an inductor.
(2) because they don't store energy, they will dissipate more heat to achieve the same voltage drop.

With these in mind, wouldn't one achieve better PSRR (assuming the goal is to reduce ripple for lower IM and hum) using the transistor as a series regulator rather than as a Gyrator?

The simulations I've done with the gyrator show higher ripple than the simple pass regulator with not many more components. The pass regulator does require a larger input filter cap as a trade off.

An additional advantage of a true inductor is the result of their storing energy. It is possible to take a transformer whose ratings are with a capacitive input filter, change it to an inductive input filter and then draw a greater amount of current at a lower voltage.

[Merlinb]

Originally Posted by Michael Koster

You mean put it in say a series resonant circuit with a known C value and do a frequency vs impedance plot?

Something like that, yes. I was thinking of a Gain V. Frequency plot really, which will show a resonant peak just like any LC filter stage. Impedance v. frequency would work too though.

With these in mind, wouldn't one achieve better PSRR (assuming the goal is to reduce ripple for lower IM and hum) using the transistor as a series regulator rather than as a Gyrator?
The pass regulator does require a larger input filter cap as a trade off.

Exactly. A series stabiliser can give better performance, but needs a larger / high-voltage capacitor.

[Tubologic]

Tubes can be used as filter reactor (chokes) in power supplies,the concept is not new and there even was a type especially designed for that application: the CBS 6216.
This small 9-pin miniature high perveance beam power pentode can actually replace a 12H/100mA/350DCR filter choke in a power supply,as illustrated in this 1953 CBS ad below:

6216CBS.1.jpg

More common (high perveance) tubes could be substituted in this circuit with similar results. For many reasons, the 6216 and associated filter reactor circuit remained obscure and was seldom used. My guess is that the good old choke was cheaper,more reliable and easier to implement...

[ray_moth]

Thanks very much, everyone, for the interesting reponses.