High Voltage, MCU controled 300Volt power supply

[Ioannis]

Looking for help on High Voltage Power Supply.

Requirements are:
Output 10-300 Volt/400mA
Adjustable voltage and current (CV/CC)
Display status on a LCD by means of a MCU that will control the P.S. as well.
Power from 240Vac line (or lower DC if SMPS is used).

Initially started on a linear power supply but power losses are too much and I reject this idea. So turned to SMPS design but could not find proper parts for the above range of control. What I would like to create is this basically http://en.bjjunyi.com/products_detail/productId=21.html

Any help, ideas appreciated.
Ioannis

[pedja089]

You didn't mention nothing about ripple, resolution, off line or main referenced, earth referenced, isolated...
I would go with offline full bridge phase shifted forward smps with linear output stage.
http://www.ti.com/lit/ug/tidu248/tidu248.pdf

[Ioannis]

Thanks for replying.

You are right about that specs. Not very critical and too many to load the initial thread.

Your post shows a step down (12V/50A) converter. My need is to have an output of 10-300 Volts with whatever input. Either direct from mains or from other source, lower dc maybe.

The design should provide control over the output voltage and the load current (in the range 0-400mA).

That makes for a 120 Watt to the load power supply.

With linear approach the losses are too high in components such as the Pass transistor (10V output at 0,4A drops 130 Watts) and other passive components too, like bias resistor and control circuit.

A linear example is shown in this page: http://sound.whsites.net/project151.htm

A switching of such a wide control (10-300 volt output) is not easy one.

Ioannis

[pedja089]

That is first example that I found to get theory of operation for phase shifted forward smps. You calculate your transformer for maximum voltage that you need on output, with lowest voltage on input.
If your phase is 180deg out of phase, you will have full voltage on output.
If your signals are just little bit off phase you will have very short impulses on transformer, and output filter(this is basically buck converter), will average that short impulse to low voltage. With relatively high ripple, so that is reason I suggested linear output stage. Output voltage of SMPS just need to be few volts higher than output voltage. So linear post regulator will have relatively easy job(not much disipation).
Any other architecture, as far as I know, cant generate very short impulses without causing major loses in mosfets.

For example if you calculate your flyback to give 300V with 80% duty then you need 0,5% of PWM for 10V on output(approximately). If your flyback work around 40KHz, your PWM period is 25uS. Your PWM pulse width for lowest voltage will be 0,125uS.
I wouldn't even try to generate that...
But if you have enough resolution for phase control, you can relatively easy generate that pulse on transformer. You have relatively small output current, so you can omit synchronous rectifying.

[Ioannis]

I wonder what happened to my answer... Just disappeared!

Thank you for the explanation. Looks very interesting in theory and I hope this can be realized for the needs of this power supply.

At 10V I thing it is 2,67% but in any case it is too low. Difficult to obtain such short pulses...

Ioannis

[pedja089]

You are right, I just put low percent from top of my head. Anyway is still sub 1uS.
Just came to my mind, If you have linear output stage, lowest SMPS voltage could be much higher, eg 40V. In that case you will have only 16W of dissipation in linear stage(in worst case).
40-300V is much easier design.
For linear output stage Ill suggest you this design:
https://www.eevblog.com/forum/chat/diy-power-supply-build/
If you carefully read schematic you will notice that all circuit is floating around output voltage.

Also this might help:
Electrophoresis Power Supply Teardown
https://www.youtube.com/watch?v=1j4L4ecjUz0

[Ioannis]

Wow, exactly what I needed! Thanks a lot!

Ioannis

[amgen]

I would..... use triac or power mosfet, phase control or pwm. PiC can do all control, Vfb voltage feed back. current feed back and isolated opto drivers.... smooth the output as desired etc. sorry about messy drawing
8583

[pedja089]

In video above, they use same system as you suggested. Only they have used isolation transformer.
In my opinion that schematic have high ripple and relatively large capacitance on output. So short circuit energy is very large... That means current control would have huge overshoot, until capacitor is discharged.
We don't know spec for PSU, that maybe acceptable, but not very nice solution in my opinion.
I definitely would go with minimal output capacitance, overshoot and ripple.
All that you can get from solution that I suggested.

[Ioannis]

Having a requirement of 400mA means a large output capacitor.

Also the problem of the wide duty cycle variation remains as we need 10-300 volt at the output.

Nice and simple idea but with difficulties in the implementation.

Ioannis