• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

switching power supplies

So no one has ever tried it? They sound pretty simple; there's no need for a huge expensive transformer, just a toroid you can wind yourself. Anyway, just thought I'd check.....
 
I've worked with them the last 5 yrs. I've not designed them however, I have helped prototype a few. The ones I work with are simple buck and boost switchers from Nat semi but only have a 5A continuous output, 7A peak. This is not enough for your application I believe. A few Things to consider when laying out your design that might be problematic: Heat, emi, and grounding. I stress the last one over the others... especially in such a high power design. It is essential that there be large amounts of ground plane in order to stablize your circuit. Oh, always buy a few extra mosfets incase you see smoke!!

P.S. Looking at the schematic from the website you linked I'm assuming this is for your car. If its not, and can be powered from mains, you wouldn't need a switching power supply!
 
Well, truth be told, I would power the amp from whatever I could at this stage. It just seems cheaper to build a switching supply than to try to get dual rail power from the wall as that seems like it would require a huge transformer. Is there any cheap way to get a split power supply from a single AC source?

About the ground plane: that's the main thing I'm really worried about is how to actually put the thing together. I realize it will need heavy current-carrying capabilities, but I don't know if it'll end up on a PCB or not as that would be another chunk of money to get, especially a high-current board with huge traces.
 
Do you want to build a well-designed one, or one that just works?

There's a lot fo theory involved in switching power supplies, and I'm not talking about just conversion ratios, duty cycles, and efficiencies.

Looks like the one you want to build is a push-pull type with the transformer center-tapped to provide +- rails. You can only control one side of the rail with this topology, but for the application you are using it in, you can assume that the two rails draw the same amount of current.


How much voltage and current are you looking at? I can probably whip something up for you real quick.
 
Hey bobsmith, I just did the same project as you. Gave up and bought a car battery, charge it every night and use it to power them when I wanna listen. Not the smoothest but it works. My application included two amps, and at peak they use 200 amps of power! So check your manufacturer specs to see how much it uses.
 
If you've never built an SMPS before, then don't start with a 200W version.

Try something simple like a 10W version. Saturating your output inductors on a 10W supply is somewhat less expensive and destructive than doing the same of a 200 W supply.

Presumably, you've got the basic test equipment you'll need - DMM, 'scope, and ideally an LCR meter, and are familiar with designing simple PCB layouts and have some way of getting your boards (ideally double layer) produced.

Why not take a look at some of the simple regulator chips from Nat Semi or ST - there are a wide variety of near monolithic devices, as well as some very good app notes with helpful hints and tips - like how to calculate the poles and zeros of the feedback transfer function!

Of course, you could just wuss out at the last minute - like I did. I'd gone to a lot of effort to design a constant current regulator for some high power LEDs - but realised, that trying to build and debug it without a scope was going to be virtually impossible.
 
Originally posted by: crazySOB297
Hey bobsmith, I just did the same project as you. Gave up and bought a car battery, charge it every night and use it to power them when I wanna listen. Not the smoothest but it works. My application included two amps, and at peak they use 200 amps of power! So check your manufacturer specs to see how much it uses.

No, you see I don't have a car amp I'm going to power; I'm building an amp and this seems like the most reasonable way to get the required +/- ~30-70VDC required - this thing WILL be run from a car battery.
 
Originally posted by: blahblah99
Do you want to build a well-designed one, or one that just works?

There's a lot of theory involved in switching power supplies, and I'm not talking about just conversion ratios, duty cycles, and efficiencies.

How much voltage and current are you looking at? I can probably whip something up for you real quick.

What would you "whip up?" I've got some plans for it already (link at the top of the page.) I was hoping for maybe ~300 watts MAX output at probably +/- 45VDC, but more likely 200W.

I've got all the parts coming now, and I'm planning on building it. Like I said, construction might be just on perfboard or something, as I don't want to spend megabucks on a single (or a couple) PCB boards.

Looks like the one you want to build is a push-pull type with the transformer center-tapped to provide +- rails. You can only control one side of the rail with this topology, but for the application you are using it in, you can assume that the two rails draw the same amount of current.

Is that alright though? The actual voltage output isn't that big a deal as long as it stays below like 70 volts per rail.
 
Originally posted by: Torched
I've worked with them the last 5 yrs. I've not designed them however, I have helped prototype a few. The ones I work with are simple buck and boost switchers from Nat semi but only have a 5A continuous output, 7A peak. This is not enough for your application I believe. A few Things to consider when laying out your design that might be problematic: Heat, emi, and grounding. I stress the last one over the others... especially in such a high power design. It is essential that there be large amounts of ground plane in order to stablize your circuit. Oh, always buy a few extra mosfets incase you see smoke!!

P.S. Looking at the schematic from the website you linked I'm assuming this is for your car. If its not, and can be powered from mains, you wouldn't need a switching power supply!

Heh, I know about the smoke thing... at work they're trying to get this huge power supply for a 3-phase DC brushless motor up and running; it's blown up a few times so far, but that's to be expected seeing as it needs to provide 3 lines at 500VDC, 50A.... if anything goes wrong..... BOOM! 😛
 
Originally posted by: bobsmith1492
Well, truth be told, I would power the amp from whatever I could at this stage. It just seems cheaper to build a switching supply than to try to get dual rail power from the wall as that seems like it would require a huge transformer. Is there any cheap way to get a split power supply from a single AC source?

Yes there is and its the second part of the circuit design that you've linked. Essentially your design is inverting the 12VDC to 60VAC then using a transformer to step the voltage back down...
There is a better way to rectify the voltage after the transformer and that can be found here Look for the dual bridge design(i think it's the third one down).
 
That bridge design is exactly the same as the one at the back end of the transformer in the switcher; it still requires a center-tapped secondary, which implies needing a large, expensive transformer. Although, it would be a good idea to use two bridge rectifiers, even though high speed diodes are expensive....
 
http://www.alliedelec.com/Search/Produc...-0864&SEARCH=%09182F30&ID=&DESC=182F30

This would be the transformer you need although its too expensive.. Or you could buy an old power amp on ebay for cheap and just rip the xfrmr out of it...

The second part of the design you have has that circuit. It's whats used to generate the negative voltage not the switching power supply! Take into consideration that most of the components needed for audio have to be high quality, low noise components, i.e. low ripple/ low e.s.r. Diodesand the bridges(both if you go with the 2 bridge design) need to have the fastest ratings you can find.

For prototyping you could just use prepunched copper clad boards from here and for running the ground plane use 16AWG or larger hookup wire(also lots and lots of solder where you can)
 
I am a bench tech for a company that makes Salt Chlorine Generators, for your pool. The basic idea is to add a bit of rock salt to your pool (like 200 lbs), and put a "Cell" inline with your pool filter. The cell has titanium plates that the water passes through, and the voltage from the power supply is applied to the plates. The Na and Cl are split from the salt, the Cl, does its job of reducing or oxidizing the impurities, and then it naturally goes back to plain old NaCl. The power supplies I work on (fixing, and solving production problems) are 22-29V 8+ amp jobs based on a PWM (SG2525). Essentially its a big transformer and giant rectifier to get 45V DC, the PWM turns on and off 4 IRFP250 FET's in parallel to handle the power. Put a big ole inductor (with protection diode! Always remember the protection diode! Schottky's a must!) and a fairly large cap on the end of it, and viola, 29V @ 8amps. I've thought of doing this myself, I'm fairly certain that I could use it directly from the output to power the PA's. Rated at 250W, Can actually handle a hundred more before failure. Its a bit costly doing it this way, but very efficient and comparitively little noise.

EDIT: Just looked at your schematic, it seems to be based on a SG3525, and similar in design to the ones I work on. BAH, I understand your problem now! You want 12V input..... Gonna need to use another PWM, or maybe even an oscillator, and get your 12V into 12V AC, then into a voltage doubler or tripler. Then rectify, then put through the output FET's, THEN into your amplifier.

No easy way to make 12V from your car into the 45V+ you need for a 300W amp.
 
The SG3525 as used in that circuit does contain an oscillator; it puts out a a pair of square waves 180 degrees out of phase with eachother. These outputs drive directly power mosfets (I'm using BUZ11s) to power the coil. According to that website, that circuit should work....
 
Back
Top