30-32volt 4A regulated dual power supply...

Thread Starter

excuseme

Joined May 19, 2008
6
I need a scheme or a simple suggestion for a +/-31V 4A regulated dual power supply...

I wish to build an integrated power ampli based on ta2022 tripath's chip (2x60W 8ohm)

I've just build a lower power ampli based on ta2024, and I am really enthusiast of the result in terms of high quality sound (so that I abandoned my expensive onkyo integra for this:D).

With that ampli, I verified that the power supply was critical for audio performances, and, after lots of trials, I chosed a regulated power supply with a simple 317 adj. voltage regulator.

So I desire to use the same type of supply also in this new project, but the ta2022 needs more power (2x60W 8ohm) and a higher supply voltage (+-31 volt is the best)

I am studying how to add a transistor to 317/337 regulators to achieve higher current, but I lost myself in calculations.:confused:

I am a pharmacist, and not an electronic engeneer!
If you help me... maybe I can help you if you have doubts about medicaments.....:D

Thank you very muuch
 

thingmaker3

Joined May 16, 2005
5,083
The following won't be elegant or highly efficient, but it will be easy and it is a time-proven technique.

You'll want an LM317: http://cache.national.com/ds/LM/LM117.pdf

and an LM337: http://cache.national.com/ds/LM/LM137.pdf

You'll also need a bridge rectifier, and a transformer with a center-tapped 48Vac secondary. The transformer would be better if rated to supply 2 Amps or more, but a 1.5 Amp output will work. Find yourself some capacitors as well - working voltage should be not less than 60V.

Go from the following instructions, but use the bigger transformer and higher voltage caps, and adjust your regulators to your desired voltage: http://www.electronics-lab.com/projects/power/015/
 

beenthere

Joined Apr 20, 2004
15,819
Most power amps are happy with unregulated rails. Get a 48 volt center tapped transformer, or a pair of 24 volt 4 amp transformers and a couple of diodes rated at over 6 amps, plus a couple of big filter caps - about 4700 uF @ 50 volts.
 

dileepchacko

Joined May 13, 2008
109
Hi
you can use LM338 5A voltage regulator. LM338 is a +tve voltage regulator
It can handle up to 5A without reducing the output voltage. This device will be an ideal for this type of application .For negative adjustable voltage regulator, you can go for UA79HGSC from Fairchild . Download data sheet from the respective companies. you will get lot of application hints from the data sheet.
 

Thread Starter

excuseme

Joined May 19, 2008
6
thankyou! thankyou!

we are very near to the solution:

the lm338 is good for the +31v, but ua79hgsc doesn't go below 24v!

the problem is the -31v:

it's strange: datasheet of lm317 explains how to add a transistor to achieve higher current, while I don't find nothing about high currents in the datasheet of lm337 (the negative one).

I think the options are:
1)paralleling more lm337, but I can't find any scheme,
2)add a boost transistor, but... which?!

anyway thank you very much for your precious suggestion!
is incredible: so many answers in a few hours... you are great!
 

Thread Starter

excuseme

Joined May 19, 2008
6
Most power amps are happy with unregulated rails. Get a 48 volt center tapped transformer, or a pair of 24 volt 4 amp transformers and a couple of diodes rated at over 6 amps, plus a couple of big filter caps - about 4700 uF @ 50 volts.


well... I tried this solution with the first ampli I build (based on a similar but less powered chip: ta2022)

I used an oversized toroidal shielded transformer, big big caps, lots of money but... the sound of the amp was of high quality but not exellent.

I tried the lm317 and... my little creature became a very high end audio ampli!

I think (but is a personal opinion based on my ears) that tripath's cips work better with regulated power supplies.

I invite everone to try those ampli... they are really amazing!
 

mik3

Joined Feb 4, 2008
4,843
[QUOTE

I think the options are:
1)paralleling more lm337, but I can't find any scheme,
[/QUOTE]

Paralleling the regulators is not a good idea because they dont output exactly the same voltage so there will be current flowing between them which may cause them overheat due to their low internal resistance.
 

thingmaker3

Joined May 16, 2005
5,083
If you use the lm337 in the TO220 package, with appropriate heat sink, it will source enough current without a pass transistor.

If you want a pass device anyway, reverse the polarity on your electrolytics and the diode, substitute a 2n2219 for the 2n2905, and a suitable PNP in place of the LM195s. Note that NPN choices superior to the LM195 have become available in the eight years since the data sheet was published.
 

Thread Starter

excuseme

Joined May 19, 2008
6
If you use the lm337 in the TO220 package, with appropriate heat sink, it will source enough current without a pass transistor.

This would be the simplest solution... at 30V if I reach 2,5A it means 75W(x 2).
The ampli is 60Wx2 with an efficiency of 90%.
Do you think I can reach 2,5A 30V? (with big dissipator, I don't have space limitations)

If you want a pass device anyway, reverse the polarity on your electrolytics and the diode, substitute a 2n2219 for the 2n2905, and a suitable PNP in place of the LM195s. Note that NPN choices superior to the LM195 have become available in the eight years since the data sheet was published.
More amperes!! yes!!

Look at this scheme: http://img181.imageshack.us/img181/2339/reg1vh0.jpg

is taken from the 317 datasheet.
I don't undrerstand the value of R?, can you hepl me?

now there is my project, where I added the "negative part":
http://img527.imageshack.us/img527/5391/regcomplqi7.jpg
please tell me if the polarities of Q1, Q2 and D2 are right.

Thank you very, very, very much!
 

thingmaker3

Joined May 16, 2005
5,083
"R?" is appropriately labeled "RL" and is called "load." It is a simplified schematic of your amplifier, or of anything else the supply might power.

For "??Q1," try a 2n2219.

Hopefully, other members will have suggestions for the pass transistors.
 

Thread Starter

excuseme

Joined May 19, 2008
6
solution!!!!

I found a negative voltage reg capable of 3 Ampere, from 2 to 32Volts: the LM333.

On the LM333 datasheet there is a scheme of a dual adjustable voltage regulator using LM333/LM350.

Thank you all!:D:D:D
 

Thread Starter

excuseme

Joined May 19, 2008
6
"R?" is appropriately labeled "RL" and is called "load." It is a simplified schematic of your amplifier, or of anything else the supply might power.

For "??Q1," try a 2n2219.

Hopefully, other members will have suggestions for the pass transistors.
Thingmaker, you are great!
 

Junkman

Joined Jan 7, 2009
2
One thing to remember: The watts out of the audio amp is (Are?) not continuous... That's a maximum output rating. At any instant in time, it may actually be only outputting a fraction of the total possible. Therefore, the power supply need NOT be rated for the actual total output. And in fact, audio amps rarely output their maximums. At least, not for very long. Just on the peaks.
That's what those big, massive caps are for, to handle the surges, or peaks. The big final caps 'fill in' the momentary surges, and after that musical peak has passed, the power supply 'refills' the big caps. (The exact same idea as DC ripple, but on a somewhat grander scale) So, keep that in mind when sizing your power supply. (You do know that speakers are not really so much voltage driven, as much as they are actually current driven devices, right? Makes a big difference in how you 'feed' them. The voltage is actually quite low, it's the current that makes the magnetic field, {In the voice coil} that reacts to the permanent magnet's field, that moves the diaphragm, aka, speaker cone, and hence, makes , um, music?)

So, your power supply need not be rated for the maximum instantaneous output power of the audio amp on a continuous basis.
Only the average, with a bit to spare. The bigger the 'bit to spare', the better, of course. Depends on what you can manage to get, build, buy, tolerate, fit in, etc. Personally, I would not be too worried about making the power supply capable of sourcing the maximum possible peak output of the audio amp on a continuous basis, but that's me.

Junkman
 

SgtWookie

Joined Jul 17, 2007
22,230
Top