Power supply concept

[Mosaic]

I was looking at a 5A ,12VDC electronic transformer for a track light and it made me wonder about creating a variable voltage & current linear power supply by using the higher frequency approach to improve ripple numbers and remove low voltage linear regulation heating.

Since I am looking to use 'regular' iron core 60Hz rated transformers at a higher frequency, I guess I'll have to evaluate the balance between increased eddy current losses vs regulation and other efficiency gains. Working between 60Hz and 480Hz (64x eddy currents) is the goal.

The premise is this...
1) Create a variable higher frequency A.C. with a modified sine wave oscillator from rectified A.C. mains.
2) Use voltage & current control feed back opto or hall effect coupled from the low voltage side of the transformer to influence the amplitude of the primary voltage rise by throttling the oscillator frequency. Higher frequency = lower voltage peaks for a given winding inductance. Thus regulation is achieved.
3) This implies that while iron core losses increase at higher frq., this could be offset by the lower power (voltage or current) demand on the secondary side.
4) Low voltage bridge rectification and capacitor ripple handling will benefit from the higher freq. operation.
5) Use a hall effect current sensor to provide isolated feedback to the A.C. oscillator for current limiting. Optoisolator driven by a voltage reference for voltage regulation.

Is this worth a try?

 

[alec_t]

.... to influence the amplitude of the primary voltage rise by throttling the oscillator frequency.Higher frequency = lower voltage peaks for a given winding inductance.

Don't see why/how. Are you confusing current with voltage?

 

[ronsimpson]

So you want to take power from the wall in your house and some how turn it into 480 hz. Then run it through a transformer.
The idea of getting more power through smaller transformers is old. Airplanes use 400hz AC so the size (weight) is smaller. But they have 400hz power lines.

How about this: Rectify and filter some. Use a 400,000 hz oscillator. Use a very small ferrite transformer. (voltage and current feedback, opto isolators etc) Now you only need a small capacitor because you are filtering high frequencies. (oh, wait.....that is a line powered switching power supply) There are every where, in your computer for example.

 

[4pyros]

Airplanes use 400hz AC so the size (weight) is smaller. But they have 400hz power lines.

Even the generators are smaller.

(oh, wait.....that is a line powered switching power supply)

Yep thats what I thought a SMPS, switch mode power supply

 

[Mosaic]

Well the goal was to not use square waves to achieve something like low noise linear regulation with a benchtop PSU for hobby work.

 

[alec_t]

If you don't use square-waves as the high-frequency source then you presumably will have to synthesise high-frequency sine waves. Methinks the semiconductor devices used for that will be dissipating a lot of power .

 

[Mosaic]

I guess it hinges on non harmonic efficient sine wave oscillation which is a challenge.

EDIT:
How about a push pull arrangement for the power sine wave to use the energy without much heating?
As one power transistor ramps down the sine wave on one transformer primary another ramps it up on another transformer's primary. Now we have two transformers sharing the the sine wave energy and not converting it to heat from the mains DC. This will create two independent isolated power supplies on the low voltage side.

 

[Mosaic]

So you want to take power from the wall in your house and some how turn it into 480 hz. Then run it through a transformer.
The idea of getting more power through smaller transformers is old. Airplanes use 400hz AC so the size (weight) is smaller. But they have 400hz power lines.

How about this: Rectify and filter some. Use a 400,000 hz oscillator. Use a very small ferrite transformer. (voltage and current feedback, opto isolators etc) Now you only need a small capacitor because you are filtering high frequencies. (oh, wait.....that is a line powered switching power supply) There are every where, in your computer for example.

You missed a point...variable frequency used to control regulation. Not fixed frequency PWM.
using iron core transformers which are comparatively high inductance will result in the ability to control voltage rise based on frequency. Thus ferrite HF transformers would be pointless here.
Once again the purpose is to use Iron core transformers coupled with higher (not high) frequencies to achieve lower ripple regulation which is better than standard Linear systems that are electrically quiet.

 

[alec_t]

I still don't see how you plan to control transformer output voltage by changing input frequency.
Have a look at this sim, showing secondary voltage unaffected by frequency:

 

[ronsimpson]

I have used variable frequency for regulation, using resonant or quasi-resonant converters.
When you use resonant the (current or voltage or both) can be sign wave or at least will have soft edges on the square wave.

I would not use variable frequency against core loss as regulation. (or transformer inductance)
I do use variable frequency against CL resonance for regulation.

How are you going to make a POWER sign wave oscillator? The reason we use square waves is they are efficient.
With square wave the "switch" (transistor) is on or off. Either zero voltage at high current or zero current at high voltage. Power = voltage X current So power loss is low.
With a sign wave power amplifier there is voltage and current at the same time. So there is power loss.

 

[Mosaic]

alec_t : Here are two Spice sims using a linear transformer with 1mH inductance in both primary & secondary with a step down of 1/2 as an example.
Note the drop in voltage as a result of the change in frequency from 1Khz to 6Khz.( fixed resistive loading)



RON ....I am not trying to make a super energy efficient linear power supply. Better ripple and noise immunity is what I am looking at , which leads to a quieter supply. Square wave switching induce EMI and noise which is my concern.
That said, the possibility of a push pull, dual transformer, sine wave generator seems intriguing from an efficiency point of view.

 

[ronsimpson]

K=0.5??? I see you are relying on massive amounts of leakage inductance. Now vary the load and see what happens.

the possibility of a push pull, dual transformer, sine wave generator seems intriguing from an efficiency point of view.

sign wave is not efficient.

 

[alec_t]

I see that you used k=0.5 in your sim, whereas I used k=1. That explains the difference between the two sim results. Interesting.