Could you help me realise a 110~220VAC to 500VDC @ 0.5A PSU? (PS my first thread!)

[DIYLouis]

Hi guys!

I want to design/build/use a SMPS:

VIN: 110VAC/220VAC
VOUT: 500VDC @ 0.5A , maybe later up it to 1A.
Efficiency : Hell if it's above 70% I will be happy.

I have been browsing and searching the forum and have concluded the following(be it right, wrong or even useful!)

* I can use some kind of quasi resonant current mode flyback converter.
* SMPS Front end PFC feeding a type of Boost converter?
* Not a lot of people build these kind of power supplies and it almost
seems like a secret artform.
* Maybe the opposite is true and everyone builds them I am just to stupid
to realise the obvious.
* This Forum is my friend.

I am assuming isolating the output from input mains is a no-brainer when it comes to question.

Can anyone set me off in the right direction ? If anyone has some experience to contribute regarding topologies to stay away from for this application it would be massively helpful

I by no means want or expect someone to plop down a complete design into my lap. I just do not have any idea where to start with considering designing such a step up converter. Lots of 12VDC or 24VDC to higher voltage step up stuff though.

I am going to be honest, It is for a guitar valve amplifier that I am building. I have xfmr based PSU's, a few, they work fine and have worked a long time. I am not unconfortable around HV and/or high Current. The amplifier is based on the Sloclone(100W output) from the sloclone forums.

The point of the effort in designing a SMPS based PSU for it, is for me to learn and to experiment and ultimately shave a few kg's off of my valve amp projects

I would also like to contribute to the community and post my progress in this thread. Final build details and schematics/PCB's included. It is thanks to communities like this that I have realised many projects and it is time to give something back.

Thanking anyone who can help in advance!

Cheers,
Louis.

[DIYLouis]

NCP1308-D seems like it could have some answers down the quasi resonant road.

The last time this horse has been beaten was 2006:

SMPS for tube amp?

It is now 5 years later and I would like to kow if the suggestion of Quasi-resonant push-pull current mode converter still holds true as a good path to follow?

Thanks!
Louis.

[DIYLouis]

TEA1713 Seems like it might be what I am after.

datasheet: http://www.nxp.com/documents/data_sheet/TEA1713T.pdf

Quoted from Datasheet:

" The TEA1713 integrates a Power Factor Corrector (PFC) controller and a controller for a

Half-Bridge resonant Converter (HBC) in a multi-chip IC. It provides the drive function for
the discrete MOSFET in an up-converter and for the two discrete power MOSFETs in a

resonant half-bridge configuration. "

[DIYLouis]

Also this page : A Kilowatt Switching High Voltage Power Supply might serve useful.

1.5KV output from mains ac for RF Power Amplifier. It has over-current and over-voltage protection and seems to serve it's purpose well.

I think I should start off with the topology that suits my power requirements. And that requires a solid definition of the required input and output parameters unlike my thumbsuck attempt in the first post of this thread. So I shall first do my homework on the exact power requirements before I carry on committing my rants to posts.

I guess it will also serve to help me commit to the project.

Still any help, even if it is "stop wasting our time and search for X" will be greatly appreciated!!

Louis.

[gmarsh]

One thing I'll say: a hard-switched power supply with a 500V output is going to have a very fast dv/dt on the output winding, plus over a KV of voltage swing - you'll have a tough time with output rectifiers. I'd go with a resonant converter of some sort to keep the edges slow, and go for a lower switching frequency.

I'd probably go with a front-end PFC also, as it gives universal mains support without requiring careful engineering of the main switcher to handle a wide input range.

To start, I'd say hit up the Fairchild Semi website and start looking at application notes and evaluation boards for PFC + resonant supplies. Make sure you understand the design procedure, magnetics construction, etc... required for designing these things.

[KatieandDad]

From my Uni days. expect a few failures when chopping HV on the inputs.

Self excited SMPS designs at 240V are a real "Art", best left to the experts or those with limitless, ruthless experimentational drive.

I started to experiment using a much lower starting voltage (ie 24V). Once you get the output to where you want it then you can try switching the input at 115V. Othrerwise you will have so many variables to consider when something goes wrong - AND IT WILL.

[DIYLouis]

@ KatieandDad : You are a 100% correct and I am going forward with caution. I think your suggestion of bumping up a LV design to 115VAC (Or use a variac and try voltages of choice?) is a sound one and I am going to do just that.

I work for a Research and Development company doing a lot of high speed digital stuff and some small signal (and large ) analog stuff. I am actually an embedded developer but the term is more than ambiguous. I work with circuits switching at high speeds on an almost daily basis but the world of SMPS is one that I kind of poke with a stick, and then run off to find the first product suitable for my needs with online calculators haha!

When you are one guy doing everything, you tend to cut corners. and then I realised I am cutting corners where I really should not be... the PSU.

I have built simple single ended step-up converters that steps up the voltage of car/truck batteries to some insane voltages(and less insane) but they were not intended for serious use. They were the result of experimentation.

I think I will start at the PFC part, if I can get one working doing what I want I am basically halfway there. 400V is not THAT far from 500V........ I could even just settle on a lower B+ voltage of say .. 350VDC to 400VDC and maybe the PFC with a 1:1 isolation converter is all I need?

Now I'm left with all the math but hey, who said math isnt fun!?

Thank you for your input and ideas

cheers,
Louis.

[gmarsh]

I wouldn't call using a iron core transformer instead of a SMPS "cutting corners" by any stretch.

You're saving a significant amount of development time and associated funds. You're buying one part instead of many parts, custom magnetics, etc... from potentially a dozen vendors. Your iron core transformer has all the regulatory requirements for safety/isolation so you don't have to worry about it. You also don't have to worry about airflow, cooling, etc. And when all is said and done, unless you're making these things in ridiculously high quantities, the iron transformer is going to be cheaper.

God forbid it weighs a few extra pounds

Edit: Still, SMPS is a fun thing to learn about and understand. Probably a decade ago I wanted to build a massive 12V power supply. Never did, but I still did a fair amount of the design process - selecting a topology, magnetics design, SPICE simulation of the switching components, etc. Gave up on it when I realized I could buy a Mean Well supply for about the same price as ordering the bare PCB All that knowledge paid off a few years later when I ended up getting heavily involved in the design of one at work.

[DIYLouis]

@ gmarsh: I meant cutting corners as in not getting into smps design deep enough to understand every essential around it. All the integrated solutions available with their online calcs makes it easy to design something with a very reasonable chance of success, especially when the online calc gives you layout examples int he same format as the tools you are using for PCB layout

I want to get my hands dirty and understand much much more.

I am going to take your advise and also look at PFC + resonant converter.

Thank you very much for your input!

PS I have had the same experience where some previous fiddling with a technology I did not know/understand well at the time, gave me a significant advantage when I actually had to make it work much later