Where to find PowerSupplies

I'm working on a nice little gain clone. It's going to need 40-0-40 to drive it and mostlikely needing a 400-550VA power capability. I been looking at SMPS and Linear power supplies. If I can with a linear design, where can I find the toroids needed, but at a cheap price (for here in eastern America). I would rather a SMPS just for it's smaller form factor, but does anyone know of a source for them cheap? I'm going to have a supply for each channel, so I'm going to need to of everything. Could anyone point me in the right direction?
 
For linear supplies I have used the transformers from John at www.antekinc.com -- they've just started advertising in AX. Most of John's stuff goes for CNC machines so they may be over-rated for our audio purposes.

For switching supplies -- the telecom 48VDC switchers show up fairly often on EBay. I have used one of these (an old Sorensen Raytheon) on a single ended GC type amplifier. There are Power-One open frame supplies 120/220VAC in 48VDC out on EBay at the moment. (Don't make the mistake of purchasing 48VDC DC-DC converter, however.)
 
MotoMan_Yz400 said:
I'm working on a nice little gain clone. It's going to need 40-0-40 to drive it and mostlikely needing a 400-550VA power capability. I been looking at SMPS and Linear power supplies. If I can with a linear design, where can I find the toroids needed, but at a cheap price (for here in eastern America). I would rather a SMPS just for it's smaller form factor, but does anyone know of a source for them cheap? I'm going to have a supply for each channel, so I'm going to need to of everything. Could anyone point me in the right direction?

That voltage would overheat most of the usual gainclone amps. They do top about 38+38 DC before there's no heatsink large enough. Even at that voltage, its pretty warm. And the usual VA rating need only be 100. In the highest quality center tap transformer (less noise than torroid), that's about $30 or so.

So, what's the application?
 
I'm trying a design for a high output gain clone array. It's going to have a total of 6 clones per channel arrange in a bridged/parallel config. They say more you have in parallel, the more heat you can dissipate. With this design I should get around 300-400RMS into 4 ohms, and be fairly good. I'll have to have some massive heat sinks with fans, but I think I should be able to make it work.

But danielwritesbac, from everything I heard the toroid design produces ALOT less EMF noise. A regular transformer has problems with leaking EMF that can get into the electronics, reasoning behind why they use a toroid. I maybe wrong, but thats what I've heard.
 
Well, to answer the original question, howabout http://www.alliedelec.com There's power suppies aplenty and also bulk heatsink channel.

On the noise part, EI are available in a wide range of qualities. Perhaps considering 10 to 12 amps, a torroid eliminates the guesswork. ;)

Over at newegg.com, you can find Artic Ceramique (most effective thermal paste) and CoolerMaster fans (very quiet). The larger, slower fans are almost inaudiable. Most DC fans run nicely from power resistors to control their speed. ;)

At the local hardware store, you can find a simple bimetallic line thermostat made for attic fans. These are somewhat adjustable, fairly cheap, and easy to mod. I think that you could use one to switch on the fans based on operating temperature.

That's about all the help I can be, because your project is pretty far over my head. ;) I think its great and really fascinating too.
 
One pair 28v can be wired for 56vct, which is 39.5+39.5 volts DC, what you asked for, and still too hot for chip amps.

However, this:
http://www.alliedelec.com/Search/Pr...=&DESC=182U24&R=836-0954&sid=47A8F8803DD9617F
has plenty of strength.

Otherwise, 56vct audio grade EI core can be found at goodwill, thrift, salvation army, ebay, curbside, in old transistor amps that have probably gotten hot enough to melt the solder right off the output transistors.
 
MotoMan_Yz400 said:
I'm trying a design for a high output gain clone array. It's going to have a total of 6 clones per channel arrange in a bridged/parallel config. They say more you have in parallel, the more heat you can dissipate. With this design I should get around 300-400RMS into 4 ohms, and be fairly good. I'll have to have some massive heat sinks with fans, but I think I should be able to make it work.

But danielwritesbac, from everything I heard the toroid design produces ALOT less EMF noise. A regular transformer has problems with leaking EMF that can get into the electronics, reasoning behind why they use a toroid. I maybe wrong, but thats what I've heard.

Toroidals do radiate less. But EIs filter the AC mains better. Take your pick. :)
 
Thank you!

gootee said:
Toroidals do radiate less. But EIs filter the AC mains better. Take your pick. :)

I was sure that I couldn't explain that. ;)
Thank you so much!!

Normally, one can find a 105c temp rated EI core and it will put out much better amperage than advertised, if you don't operate it at really hot temperatures.
However, with his 40+40dc, that's already a furnace, so I didn't know how to select the va versus temp criteria.

Do you think that LM3875T on Mica sheet and foot-long heatsinks, clamp bars, and steel rails, could withstand that voltage "fanless" on 4 ohm loads?
 
So, on AN-1192.PDF from National Semiconductor, it says that the maximum for bridge+parallel mode on 4 ohms is 28+28 volts DC.

But, I believe that they've counted for really hot operating conditions.

Perhaps LM3886T (as recommended by the application document) could operate at 32+32 volts DC?
That's a transformer rated (advertised) 44vct or 22 volt secondaries. . . isn't it?

If it can, then LM3875T can run on this:
http://www.alliedelec.com/Search/Pr...=&DESC=182U24&R=836-0954&sid=47A8F8803DD9617F

I mean "at the most" and not that its the safest config. I'm just asking to find out what the actual limit is. ;)

Thanks guys!

P.S. Just got my new Tripath Trio, up and running. OMG! Concert dynamics and soundfield on 2 ampers? Goosebumps! Can't sleep. ;)
 
Ok I'm making a mistake here... I need to ask another question before I permanently weld this foot in my mouth :xeye:


What is the calculation for find the DC rectified voltage per rail with a given AC input.

Say if I have 40-0-40 VAC from the secondary, after I rectify it, whats my DC voltage out of the rectifier?
 
MotoMan_Yz400 said:
Ok I'm making a mistake here... I need to ask another question before I permanently weld this foot in my mouth :xeye:

What is the calculation for find the DC rectified voltage per rail with a given AC input.

Say if I have 40-0-40 VAC from the secondary, after I rectify it, whats my DC voltage out of the rectifier?

Per each secondary:
40 AC *1.43 = 57 DC (dual 40's is far too much)
28 AC *1.44 = 40 DC (56vct is somewhat unwise)
24 AC *1.45 = 35 DC (48vct is spec for 8 ohm loads)
22 AC *1.46 = 32 DC (44vct is spec for 6 ohm loads)
18 AC *1.48 = 27 DC (36vct is spec for 4 ohm loads)

The "ramp up" in figures above represents diode component selection choices and efficiency erratas. Slower diodes may produce 2 volts less than above.

Demand on input voltage percentage per output also increases with power supply voltage. So, the bridge/parallel is a specific means to get much more power output on much less voltage, according to AN-1192's 200 watt application.

National Semiconductor AN-1192.PDF bridge-parallel application seems to be a much better alternative than cranking up both gain and voltage on a single chip.

You need less voltage than single-chip applications.
On the chart above, it seems that you just "step down" one line to make that difference.

So, I think that a a 44vct (22v dual secondaries doing 32-0-32 DC) transformer will "strike the middle ground" being spec for 6 ohm loads. ;)

EDIT: That's still quite warm. Can someone doublecheck my figures?
 
danielwritesbac said:


Per each secondary:
40 AC *1.43 = 57 DC (dual 40's is far too much)
28 AC *1.44 = 40 DC (56vct is somewhat unwise)
24 AC *1.45 = 35 DC (48vct is spec for 8 ohm loads)
22 AC *1.46 = 32 DC (44vct is spec for 6 ohm loads)
18 AC *1.48 = 27 DC (36vct is spec for 4 ohm loads)

The "ramp up" in figures above represents diode component selection choices and efficiency erratas. Slower diodes may produce 2 volts less than above.

Demand on input voltage percentage per output also increases with power supply voltage. So, the bridge/parallel is a specific means to get much more power output on much less voltage, according to AN-1192's 200 watt application.

National Semiconductor AN-1192.PDF bridge-parallel application seems to be a much better alternative than cranking up both gain and voltage on a single chip.

You need less voltage than single-chip applications.
On the chart above, it seems that you just "step down" one line to make that difference.

So, I think that a a 44vct (22v dual secondaries doing 32-0-32 DC) transformer will "strike the middle ground" being spec for 6 ohm loads. ;)

EDIT: That's still quite warm. Can someone doublecheck my figures?


Thats really wierd.....I'm looking at the Nation Semiconductor BPA-200 documentaion (http://www.national.com/appinfo/audio/files/BPA-200_Application_Note.pdf) and they say they us a 60Vrms toroid. So guessing thats 30Vac per secondary, so thats about 22VDC per rail. So danielwritesbac, your right.


Now I've found those all day long, so finding a dual 30VAC coil is not hard. But my other question is, I need +-15VDC to run my LF412, how would I go about getting that? more secondaries? Regulator? Iontop of the +-15, i would also like 12VDC to power things like LEDs, Fans, controllers, etc. Any ideas?

P.S. the design I'm running off of, is one of the last pages in that PDF linked above. Its the Bridge/Parallel design.



**EDIT** - Oh and it looks as if one could parallel these clones up as many times as on would like to achieve greater wattage outputs, if I were to use say 8 per channel and keep my load at 8ohms (x4 in parallel then have 2 bridges, so total of 8), would that mean I could then up the rail voltage a bit?
 
MotoMan_Yz400 said:

Thats really wierd.....I'm looking at the Nation Semiconductor BPA-200 documentaion (http://www.national.com/appinfo/audio/files/BPA-200_Application_Note.pdf) and they say they us a 60Vrms toroid. So guessing thats 30Vac per secondary, so thats about 22VDC per rail. So danielwritesbac, your right.

Now I've found those all day long, so finding a dual 30VAC coil is not hard. But my other question is, I need +-15VDC to run my LF412, how would I go about getting that? more secondaries? Regulator? Iontop of the +-15, i would also like 12VDC to power things like LEDs, Fans, controllers, etc. Any ideas?

Actually, I have no idea how to calculate the vrms. ;) But, 32v DC per rail is about as hot as you'd like to get that amp. This, is according to my empiric "thumb burn" data. ;)

I am under the impression that the point of bridge+parallel is to do more with amperage, so larger scale versions use even less voltage.

This is only the second time that someone has said I was right. So, I'm quite distracted by that. ;)

Regulator chips should be able to get you 15vdc. Fans and leds operate happily on power resistors (or ordinary resistors). At these voltages, you don't need fans.

Looking at some tripath power designs, a regulator chip was employed for the extra voltages. However, I don't know if they're made at 15v or if you need to use an "adjustable" version.

Hey, zing me an email and ask for AN-1192.pdf, and I'll attach it and send it back. It says that you can use higher voltage with 8 ohm loads, but I say if you shoot for spec voltage at 6 ohm loads then the amp will succeed on most speakers.

Pardon me, but the rest of your questions are over my head. . . quite well. Thanks again for the nice compliment! ;)
 
Well thats what I was thinking. More ICs more current. Well if your load stays fixed, then to up your current, you need to up your voltage. Ohms law.

Ohms Law: V=IR
32VDC = 4amps * 8ohms
40VDC = 5amps * 8ohms

So A higher voltage should pull more current through the load, and with more power ICs, it should be able to handle the higher current. Does anyone object to my way of thinking?