Very low noise ~1000VDC power supply.

[dougy83]

Dougy, your attachment doesn't work for me.

Doesn't do it for me either. I removed it cause I decided it was rubbish.

EDIT: This should work though (it's a bog standard gyrator + zeners to protect against fast-rising VIN)

 

[Roff]

Doesn't do it for me either. I removed it cause I decided it was rubbish.

Well, I guess that explains it.

 

[Hero999]

Hero999,
I think you have to go back to Nigel's original comment:

About your comment:


Nigel is correct that there are other, albeit less practical and efficient, ways:

Motor --> Generator.

All right, that's not an SMPS, you could also use solar cells and an LED.

In the old days they used to use motor generator sets to convert one voltage to another. A vibrator was another method but I consider it to be a type of SMPS.

SinewaveOsc --> Amplifier --> Transformer --> rectifier and filter.

I guess you could say that the commutator in a DC motor does the "switching" but no switching is done using a Sine-wave oscillator.

Yes it is, I would classify a sinewave oscillator and a transformer as an SMPS and some SMPSes indeed use sinewave oscillators e.g. a resonant converter does.

In fact it's generally not possible to convert one DC voltage to another without converting it to AC first, whether this be by a motor generator and LED to photodiode; at first you might dismiss the latter but electromagnetic radiation is alternating electric and magnetic fields even through it is not an SMPS.

 

[Roff]

A sinewave oscillator can be a completely linear circuit. It can also be class C, which I suppose could be classified as switch mode. Conversion to DC with a rectifier is nonlinear, but I don't believe it is considered switch mode.

 

[tcmtech]

As far as I have ever understood power sources the method of getting one DC voltage up to another typically requires some form of switching method to produce the voltage change. If a pure sine wave source is made there is still the normal problem of every part of the wave below the output voltage, in this case 600 volts, is not used however every part of it above the 600 volt point is being rectified and dumped into the capacitor and creates the DC output.

Depending on how high the open circuit sine wave peak voltage is above the 600 volts its going to create the same output power characteristics of basically any PWM switching power supply and the same resulting harmonics and other ripples associated with any rectified source of power that oscillates or switches above and below the output DC voltage level for any length of cyclic time period.

Unless the OP uses a 600 Volt battery source there will always be some degree of switching effect harmonics or distortions in the DC output that will require some form of filtering above and beyond a simple capacitor.

 

[kchriste]

Yes it is, I would classify a sinewave oscillator and a transformer as an SMPS and some SMPSes indeed use sinewave oscillators e.g. a resonant converter does.

I think we are splitting hairs here. Any line operated DC power supply could be considered part of a SMPS then.
In my mind a SMPS is a supply where the active device switches fully off during some part of the cycle and is fully on in another. Where the active devices are always operating in "class A" mode, the supply cannot be considered as a Switched Mode Power Supply, IMHO. I wouldn't consider a class A/B stage as Switched Mode either since neither transistor is in saturation under normal conditions.

 

[Oznog]

Motor --> Generator.
SinewaveOsc --> Amplifier --> Transformer --> rectifier and filter.

I guess you could say that the commutator in a DC motor does the "switching" but no switching is done using a Sine-wave oscillator.

An oscillator does indeed use transistors as a switch, they might be in a linear mode instead of a straight on/off switch, but it's kinda beside the point. It makes AC and puts it through a transformer, obviously, so the result is not inherently different. There's AC noise that needs to be filtered.

Evalon, your job could be serviced just fine with an SMPS. In fact it should be. Just drop that unnecessary, arbitrary specification and the question can answered better.

The SMPS output imperfections is probably better described as "ripple" than "noise". And if you can clear the audio range of ripple/noise you're pretty good to go. I mean it's not going to be very sensitive to 100KHz ripple, for example. There is some potential to affect the output but it's not "sensitive" so basic filtering to keep the amplitude down would easily prevent any distortion that might create.

 

[eblc1388]

It must have very, very low levels of short-term voltage variations such as peaks and spikes. What is important is that there are no "sudden" variations in the voltage. Whether it varies in the longer term is less important. Basically short term ripple should be as low as practically possible.

Produces a higher voltage of say 700V using SMPS and follows that with a linear voltage regulator to get 600V.

At 10mA maximum loading, the regulator only dissipate 100*0.01 = 1W and is easily doable.

 

[Nigel Goodwin]

Yes it is, I would classify a sinewave oscillator and a transformer as an SMPS and some SMPSes indeed use sinewave oscillators e.g. a resonant converter does.

And you would be completely wrong!

What does the 'S' stand for - SWITCHING - no switching in a sinewave, it's identical to using a standard mains transformer.

 

[Roff]

An oscillator does indeed use transistors as a switch, they might be in a linear mode instead of a straight on/off switch,<snip>

Come again? That's like saying an op amp uses transistors as a switch.

 

[Hero999]

And you would be completely wrong!

What does the 'S' stand for - SWITCHING - no switching in a sinewave, it's identical to using a standard mains transformer.

But the DC to AC conversion involves switching so it is an SMPS. It's true that the part after the inverter is no different to a normal transformer power supply but the fact that there's an inverter to convert DC to AC makes it an SMPS in my opinion.


Anyway I don't see the point about arguing about terminology as long as we agree that the only sensible way of converting a low DC voltage to a high DC voltage is to use a traditional SMPS. Old fashioned methods such as a motor generator sets and vibrators are less efficient and more noisy than modern SMPSes.

I would recommend using a flyback converter with a transformer for this application.

What are the voltage ripple requirements?

Bear in mind that 600V with a ripple of say 3V isn't going to be easy to achieve: higher voltage normally means higher ripple.

 

[Roff]

Bear in mind that 600V with a ripple of say 3V isn't going to be easy to achieve: higher voltage normally means higher ripple.

The load is only 10mA. How difficult can that be to filter?

 

[Hero999]

That's true, I forgot the current was only 10mA.

 

[Nigel Goodwin]

But the DC to AC conversion involves switching so it is an SMPS. It's true that the part after the inverter is no different to a normal transformer power supply but the fact that there's an inverter to convert DC to AC makes it an SMPS in my opinion.

No switching involved - and your opinion is completely wrong.

Anyway I don't see the point about arguing about terminology as long as we agree that the only sensible way of converting a low DC voltage to a high DC voltage is to use a traditional SMPS. Old fashioned methods such as a motor generator sets and vibrators are less efficient and more noisy than modern SMPSes.

The reason for pointing out your error is that this is an historical record, people will read this and think "what a load of muppets, they don't know what they are on about on these forums".

SMPSU's use SWITCHING - it's in the title. A sine converter doesn't use switching, it's a linear system - do you consider a conventional mains transformer 'switching', because that's exactly the same thing.

 

[Hero999]

Just how do you convert the DC voltage to AC, to power the transformer without doing any switching?

 

[Nigel Goodwin]

Just how do you convert the DC voltage to AC, to power the transformer without doing any switching?

Exactly as a conventional mains transformer is fed, you feed it an AC sinewave - do you not understand what 'switching' means?.

 

[Hero999]

But the input isn't AC it's DC remember? As you know a transformer will not work with DC so it needs to be converted to AC which involves switching as I said a couple of posts ago.

 

[Nigel Goodwin]

But the input isn't AC it's DC remember? As you know a transformer will not work with DC so it needs to be converted to AC which involves switching as I said a couple of posts ago.

Not at all, no switching involved, or required. Does your stereo at home use switching?, does your car radio?, switching is only used for specific purposes and reasons. Using 'switching' is only one way to convert DC to AC, and a specific type of AC at that.

 

[kchriste]

The transistor driving the transformer will always be turned "on". Just how turned "on" the transistor is depends on where in the sine cycle we are. The transistor will never be turned "off" or switched off in class A. The current in the transformer reverses polarity during the cycle when the magnetic flux in the transformer core collapses and begins to "source current" into the collector. That is why, with a transformer coupled class A amplifier, the voltage on the collector of the transistor can reach a theoretical value of x2 Vcc. But the key here is that the collector current NEVER drops to zero in a class A amplifier/osc unless it is being driven into clipping.
Circuit for reference:

 

[Roff]

The transistor driving the transformer will always be turned "on". Just how turned "on" the transistor is depends on where in the sine cycle we are. The transistor will never be turned "off" or switched off in class A. The current in the transformer reverses polarity during the cycle when the magnetic flux in the transformer core collapses. That is why, with a transformer coupled class A amplifier, the voltage on the collector of the transistor can reach a theoretical value of x2 Vcc. But the key here is that the collector current NEVER drops to zero in a class A amplifier/osc unless it is being driven into clipping.
Circuit for reference:

View attachment 38400

The current in the primary never reverses direction. If it did, the transistor would be cut off through part of the cycle.