Hi,

A bit of what I do first, basically I am home taught with regards to electronics. I have a job as a repair engineer, repairing industrial equipment, about 25% of what I do is smps. So I do have good knowledge of how they work.

I am thinking of building my own smps, but there is a big difference between repairing one and designing one.

I was wondering if it is beyond what someone could design themselves at home?

By that I mean a lab power supply smps, one that is reliable, accurate, stable and is not going to blow up within a few years. As I see what happens to poorly designed smps at work on a weekly basis.

I know people are going to say its not worth it, or you wont save any money. But as a electronics enthusiast its a project I want to do as long as its possible of course...

 

It will not be cost effective. But indeed doable. Some parts like the transformers needed may be hard to get. Depending on where in the world you live. But here you may scavenge some parts from your job. So as a educational project just go for it. It is lot of books like this one http://www.amazon.com/Switching-Power-Supply-Design-3rd/dp/0071482725 if you need some literature

 

It will not be cost effective. But indeed doable. Some parts like the transformers needed may be hard to get. Depending on where in the world you live. But here you may scavenge some parts from your job. So as a educational project just go for it. It is lot of books like this one http://www.amazon.com/Switching-Powe.../dp/0071482725 if you need some literature

Thank you very much, The transformer should not be to bad as long as I design it correctly, my old work place makes them so can get a good price. I have a few books, skim read through them to get and idea of what I am letting myself into...

Is there any other issues that might stop me that you can think of?

I don't want to start it unless im sure I have at least a fair chance of getting it working..

* I live in the UK *

 

All SMPS are built around some controller chip(s). And a great help will be the datasheet for this chip. I would select a controller chip. And then design my SMPS around this. Use whatever you can of salvage parts. And ask the forum if you have doubts

 

I always favor do it yourself for learning purposes. Just keep in mind you will likely not save any money. What you end up with will likely be inferior to a commercial product, but then, that is not the point either.

 

Hi,

A bit of what I do first, basically I am home taught with regards to electronics. I have a job as a repair engineer, repairing industrial equipment, about 25% of what I do is smps. So I do have good knowledge of how they work.

I am thinking of building my own smps, but there is a big difference between repairing one and designing one.

I was wondering if it is beyond what someone could design themselves at home?

By that I mean a lab power supply smps, one that is reliable, accurate, stable and is not going to blow up within a few years. As I see what happens to poorly designed smps at work on a weekly basis.

I know people are going to say its not worth it, or you wont save any money. But as a electronics enthusiast its a project I want to do as long as its possible of course...

Why do want to re invent the wheel, convert an atx psu to variable voltage, as these are readly available , thats what i did here>>

http://www.youtube.com/watch?v=EYPhZAnJenA&feature=youtu.be

http://www.youtube.com/watch?v=BuuDoS9NEDY&feature=youtu.be

http://www.youtube.com/watch?v=rpUk9hpFQL4&feature=youtu.be

 

All SMPS are built around some controller chip(s). And a great help will be the datasheet for this chip. I would select a controller chip. And then design my SMPS around this. Use whatever you can of salvage parts. And ask the forum if you have doubts

After looking around, the hardest part to get is the chopper transformer. I have found generic ones from RS, with chips that can be used with it. So I will select the chip depending on the transformer, then build the circuit around that.

I always favor do it yourself for learning purposes. Just keep in mind you will likely not save any money. What you end up with will likely be inferior to a commercial product, but then, that is not the point either.

Indeed, I prob wont use it, it is just a learning experience.

Why do want to re invent the wheel, convert an atx psu to variable voltage, as these are readly available , thats what i did here>>

I already have a variety of power supplies, this project is about learning how to make a smps. Plus atx power supplies have bad regulation on high load, thank you anyway.

 

I always favor do it yourself for learning purposes. Just keep in mind you will likely not save any money. What you end up with will likely be inferior to a commercial product, but then, that is not the point either.

Most of your replies are incorrect.

Depends on what you define as SMPS- if this must include HF isolation transformer.

If you build a buck-mode SMPS, it is possible to archieve about 80%. You need however a suitable storage coil. And cooling can become an issue if you don't know how-to.

TL494 for instance costs 30 cents.

If you want to build your own HF transformer, then some years electronics experience will be of advantage.

Why do want to re invent the wheel, convert an atx psu to variable voltage, as these are readly available , thats what i did here>>

Questionable point of view. Using a SMPS chip as intended = reinventing something? ATX power supplies are rather problematic for instance they don't have good short circuit protection. It is very likely to destroy these and then repairs can be difficult to do (i.e. not worth the labour cost just for one hour).

By the way for my SMPS I have not done one single calculation, and have not drawn any schematics. Preliminary test show a capability of at least 12 Amps (2 to 50 volts). I want to add a 1.8 TFT LCD to display voltages etc., and don't know what should be so inferior about that.

The truth is rather I don't really need adjustable power supply. For 12V that I use, it's cheaper to buy ready-made electronic transformers.

The HF transformer is difficult because they are worked in special modes, so their size can be reduced. High voltage semiconductors aren't easy to work with as well, one mistake and they will become destroyed instantly.

So using a 50Hz transformer is one way that makes it possible to build a SMPS relatively easily.

 

Think again. Maybe you can do as good as professional, and maybe you can advise a total beginner to match your results. But it is unlikely.

It is always important to remember whom your are talking to.

And most people neither have the time or money to match resources you may have. I live in a location where there are about 5 major electronic outlets. Long drives, and gas ain't cheap, but I am blessed given my chosen hobby/avocation.

BTW, I have done the same exact thing as the OP. The results weren't pretty, but I did learn some new things.

Bill's Index

http://forum.allaboutcircuits.com/showthread.php?t=10339

I've done it several times since, and each time it gets better.

 

Any commercial SMPS must conform to strict regulations regarding safety and EMC to mention some. And they have to pass tests before they are put on the marked. As a hobbyist you do not have to think about this. So yes it is most likely that a hobbyist switcher will be inferior to a commercial one. Most commercial SMPS are today primary switched-mode power supplies. That require some(a lot of) knowledge and equipment like a good scope. Strictly speaking I am not sure we are allowed to discuss those here on the site.

 

Yes, we are. DIY is always encouraged. We do not allow unsafe practices, and they are spelled out in the ToS.

 

Think again. Maybe you can do as good as professional, and maybe you can advise a total beginner to match your results. But it is unlikely.

1. For the circuit yes, to some degree approaching professional designs. For mechanical work, like enclosure, no, that's pretty resource intensive.

2. It is unlikely, if 10+ years experience + programming exposure are simply not there. Very unlikely.

It is always important to remember whom your are talking to.

And most people neither have the time or money to match resources you may have. I live in a location where there are about 5 major electronic outlets. Long drives, and gas ain't cheap, but I am blessed given my chosen hobby/avocation.

It is resource intensive, to build some decent stuff from scratch you need a small inventory for electronics components. And datasheets as well, there are some 3000 on my hard drive.

BTW, I have the same exact thing as the OP. The results weren't pretty, but I did learn some new things.

Bill's Index

http://forum.allaboutcircuits.com/showthread.php?t=10339

I've done it several times since, and each time it gets better.

I built my first switcher using LM2576, for the SEGA master system, in the year 2005 or maybe 2006. It was a bit clunky, well overdimensionated, but it worked. No abnormal heat developement anymore.

After building about 30 or 40 switching technology circuits, yes it gets better over time.

Some SMPS stuff can be done at home, with relatively good results, but things like RF shielding, enclosure, and this can become quite expensive.

I would for the most of it, simply buy electronic transformers, and then use a ready-made IC to convert voltage directly on PCB.

Doing high voltage flyback SMPS at home is not really worth it for my part, maybe for experiment/research, but there are other projects that I am busy with. You need expensive equipment as well to test these for instance oscilloscope.

Digital voltage adjustment/ Displays would be easy for me to do, but maybe not so easy for others. Not everyone has some 100 microcontrollers around free to use for whatever purpose.

I think it rather depends what you consider would be a regular SMPS, and what not.

 

Yes, we are. DIY is always encouraged. We do not allow unsafe practices, and they are spelled out in the ToS.

Ok good to know. However a primary switched-mode power supply. Is not a beginner project. It could perhaps be wise to explore more simple switchmode projetcs first. But as I said in my first post. If you feel you are up to the task. Just go for it

 

I'm confused Bill, I just told someone in another thread that discussion of building off-line switching PSUs was banned on this forum. Is that true or not?

I'm sure I've seen discussion of building offline switchers shut down on more than one occasion?

 

To takao21203,

I think we will agree to disagree. I can build many items to very high quality and standards, but a SMPS is simply not one of them. This is what I told the OP. So how are most of my replies incorrect?

BTW, a SMPS is a class of regulators as well as power supplies. HF isolation transformers are not a core requirement, not even close. This is one of those areas where there is lots of room for design. Just a good source of DC voltage regulated with a switching regulator is a SMPS. When you set your definitions too narrow, you tend to leave major things out to the point of getting it wrong.

With as many as you have built, did you document any of the builds? This is what the OP needs, not generalized statements. The designs I have posted suck, but they are documented. How about schematics, or links to designs you know that work, for example?

To Guinness,

State what you want in current and voltage. Ripple can also be a problem, as SMPS will have a minimum ripple spec, this is core to the design, though it can be minimized with attention to detail. An analog regulator is going to be much quieter, always. I have seen a few designs that were hybrids between the two.

End of use application would be nice.

 

I'm confused Bill, I just told someone in another thread that discussion of building off-line switching PSUs was banned on this forum. Is that true or not?

I'm sure I've seen discussion of building offline switchers shut down on more than one occasion?

Show me the post, and I can answer better.

What is not allowed on AAC is transformerless power supplies. We get a constant influx of complete beginners trying to save shekels or whatever by leaving that one absolute component out. It is the same exact thing with the LEDs off mains. The latter of which we should not support but does need taught, since it is a commercial product. It was a standing argument beenthere and I had about the AAC book, one of few.

A HF transformer that provides isolation qualifies as a transformer. The potential path of high voltage to people in possible contact with the low voltage DC is important.

I am human, if I have screwed up I will try to admit to it and apologize. Far as I know this is not the case, but I freely admit it can happen. Far as I know I haven't done it lately.

If the OP builds a conventional bulk power supply with a transformer and follows up with a SMPS regulator this would qualify as a SMPS power supply. On the ones I built I used wall warts, they are cheap (cheaper than building from scratch), quick, and easy.

If he wants to go a more traditional route (I'm thinking of the HF transformer) we can point out the pitfalls. Electronics is not automatically a safe hobby, we set minimum standards and hope for the best. For people we know are rank beginners we go the extra mile.

Knowing some of the crap me and my brother pulled when we were beginning I am aware we have no control over anything anyone does. We just try to prevent it from being AAC's fault when it does.

My job gave me the weekend off, does it show?

 

To many replies to reply to everyone sorry.

If this discussion is banned on this furom, then thats fine and I dont mind thread been deleted.

But if its ok, im going to build a isolated flyback smps. Low power, prob about 5v 2A for my first project.

I have a scope and all other equipment at home needed to design and build any project, and can order whatever components I need through work. Have also been told I can use work equipment, and they spare no expense on us engineers. Very lucky in the fact we all have brand new top of the range equipment on our benches.

I will be adding PFC, short circuit protection and all other safety measures needed to make it safe. No point in building a dangerous device!!!

I fix small smps to 3Ph 600v 50A power supplies at work, and have had a few go short circuit before. So I know more than most people how dangerous smps can be. Thats why I would only build and would only recommend to build low power ones.

That said, thank you all for replying.

Please let me know if this subject is allowed and I will open up a proper thread in the project forum when I start building. If its not then no probs, will solo it

 

Well incorrect towards my own reality. I can certainly build SMPS if I need one, most of the time, this would not be adjustable, and not including any kind of display.

Documentation, PCB layout is another thing, and if I want to do this voluntarily.

But when I replied I was thinking people with some experience should not be told SMPS is too difficult or the results will be disappointing.

I archieved good efficiency even with simple circuits. But yes this can include things like to rewind cores with magnet wire, and to order some range of components to see what gives the best results.

Schematics on the web don't always work and/or are not always suitable for particular purpose.

So a circuit that regulates a DC voltage by inductive means counts as SMPS, I did not know. Not neccessarily including high-voltage flyback transformer.

If asked for a simple reply: No SMPS is not beyond home build- if you are willing to put some effort into it. And- don't put too much effort into it.

 

State what you want in current and voltage. Ripple can also be a problem, as SMPS will have a minimum ripple spec, this is core to the design, though it can be minimized with attention to detail. An analog regulator is going to be much quieter, always. I have seen a few designs that were hybrids between the two.
End of use application would be nice.

Sorry you replied while I was replying, so missed your posts till after I posted mine.

Yea basically, 5v 2A, looking at less than 10% ripple for this first design. Although future designs would want better.

This is just a learning experience at present, it wont be used for anything. When I can build one properly, after prob many failed tries, I would like to build a multiple output smps to go in my home lab for prototyping with.

 

If asked for a simple reply: No SMPS is not beyond home build- if you are willing to put some effort into it. And- don't put too much effort into it.

Thank you for replying, I will start designing now, will open up a thread in the project forum soon, so any help people can give me would be appreciated, thanks again for all the replies.