Very low noise ~1000VDC power supply.

[Evalon]

"Ouups, can see that I forgot to edit the headline - it's a 600 VDC supply and not a 1000VDC supply."


Hello Forumers,

Hope you can help with this: I am searching for a very low noise ~ 600 VDC power supply. It should have these specifications:

Input: 51VDC +/10 % (if convenient it can be stabilized to e.g. 48 VDC +/- 5 mV)

Output: 600VDC +/- 10% (+/-10% means that it is not crucial whether it is exactly 1000VDC. Also, long-term variations of +/-10% are ok, as long as the period is longer than ~ 5 mins)

Current rating: <10 mA, or, as an alternative 0.1 mA (two different uses)

Start-up time can be < 2 minutes

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.

Must be an oscillator, or the like, and not an SMPS

Should be made from components readily available or which can easily be made (e.g. coils etc.). Light-weight, and preferably small-size, and simple

Any suggestions for diagrams and calculations, or possibly used circuits (??) are appreciated.

Greetings from a very winter-clad Denmark,

Jesper

 

[dougy83]

Greetings Jesper, welcome to the forum

Must be an oscillator, or the like, and not an SMPS

Do you mind if I ask: why not an SMPS? If you're worried about noise & ripple, just shield the converter and put an LC filter on the output. What are your specs for ripple?

 

[Evalon]

Thanks!

Do you mind if I ask: why not an SMPS? If you're worried about noise & ripple, just shield the converter and put an LC filter on the output. What are your specs for ripple?

Thanks - it looks like an interesting and active forum

In reply to your question I can say that it this case I'm working on a hopefully superb audio application and SMPS's just don't do it (in my opinion, that is...). One of the reasons that e.g. batteries or accumulators are sometimes used in audio circuitry is the potentially exceptionally low noise levels (can be >140 dB, static though).

Given the inherent resonances etc. in filtering circuitry - which may remove but also add noise - I prefer to design "quietly" from the beginning, particularly in terms of short terms peaks & spikes that in my experience are the most important.

Hmmm... ripple. Basically as low as practically possible but low levels of sharp peaks and spikes are more important.

Cheers,

Jesper

 

[Nigel Goodwin]

As you're presumably building a valve amplifer (as it's 600V), PSU noise is the least of your problems, as the amplifier itself is inherently noisey.

The obvious answer is not to start from 51V DC (which seems highly bizarre anyway), simply transform the mains as valve amps normally do.

 

[Evalon]

As you're presumably building a valve amplifer (as it's 600V), PSU noise is the least of your problems, as the amplifier itself is inherently noisey.

The obvious answer is not to start from 51V DC (which seems highly bizarre anyway), simply transform the mains as valve amps normally do.

Hi Nigel. Thanks for considering my thread ... however, I'm not designing for tubes, and the 51 volts is the stable voltage from 4 12 VDC lead-acid batteries (actually 51,2 volts) so no mains involved.

Best regards,

Jesper

 

[Nigel Goodwin]

So what is the 600V for?.

 

[Hero999]

An SMPS is the only way to convert a low DC voltage to a high DC voltage.

 

[Nigel Goodwin]

An SMPS is the only way to convert a low DC voltage to a high DC voltage.

Not by any means, but it's the only sensible and efficient way.

 

[Evalon]

Hmmm... this is not a field where I'm very knowledgeable. What I've seen are e.g. small transformer-based circuits that generate 3 KVs or tube based circuitry used for plasma-tweeters that appear to be an oscillator where very high voltages are generated. It's something like the small transformer based circuits - or the like - that I'm looking for - don't know if it's an SMPS ??

Best, Jesper

 

[dougy83]

It's something like the small transformer based circuits - or the like - that I'm looking for - don't know if it's an SMPS ??

That would be a self-oscillating flyback SMPS; the transformer has a feedback winding that controls the driver transistor.

Shield it and filter the input and output; should be enough.

 

[Hero999]

Yes it's a type of SMPS anything which converts the DC voltage to AC and back to DC again is an SMPS.

 

[Nigel Goodwin]

Yes it's a type of SMPS anything which converts the DC voltage to AC and back to DC again is an SMPS.

Again, it's not - I suggest you consider what the initials stand for.

 

[kchriste]

Do you mind if I ask: why not an SMPS? If you're worried about noise & ripple, just shield the converter and put an LC filter on the output. What are your specs for ripple?

I agree. And, if he wants to lower ripple even further, the OP could follow the SMPS with a linear regulator which only drops a few volts.

 

[Hero999]

Again, it's not - I suggest you consider what the initials stand for.

Just to be sure we are talking about a flyback driver aren't we?

It certainly is a switched mode power supply, the transistor oscillator switches the power to the transformer on an of at a high frequency and it is a power supply.

Almost every other way of converting one DC voltage to another relies on switching and is therefore a switched mode power supply, simply by definition.

 

[kchriste]

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

Not by any means, but it's the only sensible and efficient way.

About your comment:

An SMPS is the only way to convert a low DC voltage to a high DC voltage.

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

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.

 

[Nigel Goodwin]

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.
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.

Exactly, it's completely misleading to say that ONLY an SMPSU can do the job.

I imagine that a vibrator based inverter doesn't qualify as 'switch mode' either?, but the two specific cases I was thinking of (and particularly the seciond one, as being practical) definitely aren't.

 

[ronsimpson]

https://www.electro-tech-online.com/custompdfs/2010/01/an118fa.pdf
Linear Technology - Low Noise, High Voltage DC/DC Converters
https://www.electro-tech-online.com/custompdfs/2010/01/dn212f.pdf
Some reading. I have built some of these and found the noise level is very low.
If the links do not work at least read AN118FA

 

[Evalon]

Hello all,

and thanks for your comments and suggestions. Seems that my understanding of what an SMPS includes is not complete ....

I've done a bit of research during the day and the qualities I'm looking for appear to be found in a "Vackar oscillator", where the coil in the oscillator is part of a transformer that generates the higher voltages. It generates a very stable sinus, when simulated.

I've attached an image of an example Vackar oscillator (with coil but not as part of a transformer). I understand that it's a very stable type of oscillator.

Unfortunately, the oscillating frequency of this circuit is about 2 MHz (when simulated) and I would like it to be somewhat lower - in the area of 160 kHz. Any of you know how to calculate this using the circuit I've attached? Or maybe you have a suggestion for an equally good alternative?

Best regards,

Jesper

 

[dougy83]

I agree. And, if he wants to lower ripple even further, the OP could follow the SMPS with a linear regulator which only drops a few volts.

Just to follow on from that, a good regulator this application would be a gyrator type - it's insensitive to supply voltage magnitude (DC part) so you can use a cheap low voltage components (just the cap needs HV rating).

 

[Roff]

Just to follow on from that, a good regulator this application would be a gyrator type - it's insensitive to supply voltage magnitude (DC part) so you can use a cheap low voltage components (just the cap needs HV rating).

Dougy, your attachment doesn't work for me.
EDIT: I reported the problem to the administrator.