200va smps topologies

[derick007]

Hi

I have just joined the forum as I have an interest (as a hobby) in smps. I have a little knowledge, absolutely no experience and deceptively slow on the uptake.

I would like to design and construct a smps with 8 dc output rails :

15v @ 3A x2 = 90VA
-15v @ -3A x 2 = 90VA
6.5v @ 1A x2 = 13VA
-6.5v @ -1A x2 = 13VA

TOTAL = 206 VA.

I believe either a flyback or forward topology would be most suitable ?

Any help would be gratefully received.

Kind regards, Derick007

 

[derick007]

Apologies should have mentioned, the 6.v rails power ic's and therefore are held constant.

The 15v rails are to be used for testing purposes and therefore these rails will be variable.

 

[blasphemy000]

Your 206VA rating is assuming a 100% efficiency rating which isn't going to happen.

Another thing. If you only have limited knowledge and no experience with Switch-Mode Power Supplies, I would recommend that you do a lot of studying and reading on the subject and also become very familiar with High-Voltage saftey measures. The voltages and currents commonly found inside a SMPS can kill you instantly 1000x over.

That being said. You could build a 500W half-bridge supply quite easily to get the power levels you are looking for. This could also be done with a 2-switch forward converter topology as well, although I think the half-bridge would be easier to configure for a beginner. I would also recommend a "hard-switched" power supply for this power level and because they're easier to get working for a beginner.

Here is a link to a website showing the different topologies of SMPSs and it also offers some calculation help as well, although I've never used the calculations on this website before.
http://schmidt-walter.eit.h-da.de/smps_e/smps_e.html

Also, for figuring out the number of turns and other design factors for your switching transformer, I've found the program in this link to be very helpful. I think this program was created by a fellow user here at diy-smps.
http://www.diysmps.com/forums/showthread.php?275-SMPS-transformer-design-tool

Good luck in all your adventures into SMPS design and above all...BE CAREFUL when building and testing these things. DEATH IS PERMANENT!

Also, if you don't have an oscilloscope, don't even try to build a SMPS. I've managed to get a couple working w/o a scope, but it was nearly impossible, and I blew up many components on the way. All I had was a frequency counter and duty-cycle monitor and it took me forever to get one of these PSUs working. Once I got my scope I was easily able to diagnose problems in the supply. Also, if you have a scope, become very familiar with the proper way to take HV measurements and be very careful where you place the ground lead of your probe. Also, NEVER try to connect/disconnect a probe while your "device-under-test" is running. If you accidentally touch the probe in the wrong spot, very bad things could happen, like blowing up your scope, your power supply, or death. I would suggest, even insist, that you either buy or build a differential scope probe so that you can take floating measurements between two points above ground. Here is a link to one I built myself from scratch because I couldn't afford one. They are very expensive for a real one, but they may save your scope and a lot of headaches.
http://www.diysmps.com/forums/showthread.php?511-Built-a-Differential-O-Scope-Amp

 

[derick007]

Thanks for all your advice, particular regarding the dangers of HV and the websites.

I have PROTEUS simulation software, which I am hoping to use and an old second hand Tektronix scope.

I have also found some interesting literature on 220w smps for LCD TV from ON-Semiconductor. They also sell the board, which I am thinking of purchasing to gain some knowledge/experience on working with them.

 

[blasphemy000]

You're welcome. I always preach safety first when it comes to building these supplies. The input stage typically produces 340VDC which is extremely dangerous if you're not careful, and the storage caps can stay charged to this voltage for a very long time if the circuit is not equipped with appropriate bleeder resistors. Purchasing a pre-made board with assembly instructions or a DIY SMPS kit would be a good way to get started and there are several companies that offer these. On-Semi is a great company, I use their stuff a lot, and I do a lot of prototyping with their stuff since they offer free samples on a lot of it. Good luck in your endeavor into the world of switching power supplies and above all, be careful!

 

[derick007]

I would be interested to find out about the several other companies re. DIY SMPS kit

 

[KX36]

I'd aim for a single switch forward converter for this application, prefereably with current mode control. Half bridge or 2 switch forward converters reduce the voltage stress on the transistors but require a way of driving the high side transistor such as a gate drive transformer. Half bridge can handle more power than forward but you have to use voltage mode control which doesn't perform as well and is more difficult to compensate than CMC (CMC is possible but it's a lot more effort on HB), so if you don't need the extra power it's probably not worth the bother.

UC3844 is one of the standard CMC controller ICs with the ubiquitous TL431/optoisolator working as the voltage error amp. There are probably better performing ICs than this now though. Optos suck but there's not a lot of choice.

I agree with blasphemy000 that it should be hard-switching. Quasi-resonant/ZVS/ZCS design is complex and imposes limits on load current etc. It's use is best saved for much higher power supplies.

To add to the safety bulletin, if this is an off-line supply, you should power it from an isolation transformer if you're going to probe the primary side.

I'm curious as to why you have duplicates of all the rails.

 

[derick007]

Thanks KX36.

I hope to design and build a lab./bench power supply - some commercial ones have duplicate rails which are variable. I guess flexibility - I can connect them in series and have +,- 15v and +,- 30v.

 

[KX36]

If it's flexibility you're after, you could just build 2 separate units and then they can be separately controlled and outputs separately regulated. The rails of a power supply follow each other so they'd each need to be separate modules if you want them to be separately variable. I have found that having a widely variable output voltage adds significant challenges especially if you want a wide range of input voltages. I've done something similar. If you only care about the end product though, you might be better off with a linear supply with a big power transformer with various relay controlled taps. Most commercial supplies are still done this way at this power level, there's less noise on the output and it may be cheaper to just buy one rather than make one. SMPS bench supplies tend to be single output and high enough power that linear isn't an option.

 

[derick007]

I did start with the linear method, but I would like to learn about smps as they seem to be in every electronic device there is.

Apart from learning about them I have half a notion to have a go at repairing LCD/LED TV's. Having some knowledge of smps would be a big help.

 

[KX36]

Re: LCD TVs:

If the power supply is at fault, it's usually because the output cap has gone bad and replacement is quick and easy

 

[derick007]

do you know of another forums/websites re. repair of LCD/LED TV's ?

I recently came across a guy Jestine Yong ?

Also do you know of any websites where I could download service manuals for LCD/LED TV's - preferably for free ?

 

[KX36]

Jestine Yong has a power supply repair book and I believe is the person behind the Blue ESR Meter and Blue Ring Tester (hence the plugs in the book). It's no textbook but it's an OK read for an absolute beginner and if it has the solution to your problem it's worth having. Someone here is advertising that they'll give you a copy for free if you PM them.

Unfortunately I can't really help you more than Google can at this point. It's not something I have much experience of. There are some videos of people repairing computer monitors or TVs on youtube which might help you. I don't have a good source for service manuals (beyond sometimes getting lucky on google) unfortunately.