Peak or not peak, that is the question?

[snax]

I have read a lot about power supplies and decide that it was time to experiment.

I have an amplifier that requires 42 volts DC max and I’m assuming that’s RMS.

I designed an unregulated symmetrical supply using 2 rectifiers and a couple of 63v 10,000uf caps. The regulators were connected to 2x30v secondary’s which I calculated to be 42v RMS.

(2x30v)*0.7=42v

So when I measured the DC voltage across the positive and negative rail after the caps and got 84v which I wasn’t expecting.

I’m now thinking that because I’m using 2 secondary’s I should be multiplying it by 1.43 as it’s the sum of the two.

Alternatively should be using 2x17volts AC instead?

What am I doing wrong?

[bobodioulasso]

It seems your amp is single supply? If so, you need a single seconary 30v ac transfo.

You have to multiply AC voltage by 1.4 to get DC voltage.


You can parallel your 30volts secondaries to obtain one single DC 42 volts . Or use one secondary per channel ( double mono)
Or, is your amp +42/ 0/ -42?

[gootee]

Unregulated Power Supply Design

The link above should have everything you need, except your peak maximum current draw, which you will need to know or calculate based on your load's characteristics.

If your amp can handle an absolute maximum of 42 Volts DC, then you will need to consider the worst case combination of AC Mains over-voltage and transformer regulation, and shoot for something lower than 42 V DC, to leave a safety margin. You will need to take into account the worst-case peak value of the ripple voltage, to stay within your 42V maximum.

Aside: "RMS" is usually only discussed for AC or signals. For a pure sine wave AC, the RMS voltage times the square root of 2 (which is approximately 1.414) is the peak voltage. The RMS (Root Mean Square) voltage of a DC voltage is just the same DC voltage.

Cheers,

Tom Gootee

[snax]

Thanks for the speedy reply’s bobodioulasso and Gootee. I will follow your link Gootee and read some more.

The amplifier is a mono block kit that I am using to get a practical understanding of amplifier design and how to build them. Of cause the power supply was not part of the kit, although a design was.

Having been an avid reader of these forums I decided to try a design from another thread.

Nothing ventured nothing gained…

bobodioulasso, so the DC is peak AC then?

By making the supply symmetrical two secondary’s / two bridges / 1 supply, this then doubles (sums) the secondary’s together and therefore the output is seen as (2x30)x(2x0.7)… 84Volts

If I wanted to keep this design and get the 38 volts (with safety margin that gootee mentioned) I should have gone for a 2x15volt AC transformer.

Cheers

Snax

[gootee]

I'm still not sure what power supply inputs your amplifier needs. Often, they need symmetrical + and - voltages. In that case, there would be a ground in the middle, probably where the adjacent ends of the secondaries are tied together. Usually, in that case, you'd have a positive DC output from the top, ground in the middle, and negative output at the bottom.

What voltage(s) does your amplifier need? If you don't need both + and - DC voltages, then you probably don't need to use two secondaries as it sounds like you did.

If you did tie the centers of the secondaries together, then try measuring from each DC output to that (ground) point, or to where the smoothing caps from the two rails are tied together, and you should see half of your 84 V in each case, and hopefully one positive and one negative.

So does your amp want to see +42/-42 VDC max? Or does it want to see a total of 42V, max, between the + and - supply pins, which would be +/-21 V max?

Anyway, if it needs +/-42V max, you could consider adding a regulator circuit to each rail to bring them down to something like +/-38 V.

Going back to your first post for a moment, the "30V" rating is an RMS value, from which you can multiply by 1.414 and get 42.42 peak VAC.

[gootee]

Below is a link to a very good short thread where all of the calculations and considerations needed to design a linear power supply are discussed in detail, with examples. It also includes considerations for regulator drop-out voltage, but you'll get the idea.

Is a 9v transformer large enough?

And here is another good link about linear power supply (PSU) design:

http://www.diyaudio.com/forums/power...r-supplies.htm

Cheers,

Tom Gootee

[snax]

Hi Gootee,

Cheers again for the pointers. You were right about the circuit using ground as the reference 0v and therefore the + to ground is 42 volts.

I have to admit that I understood why the caps rated at 63 volts were not leaking or popping because they only had 42volts through them but I didn’t make the mental connection that the board would only see 42volts with reference to the ground.

I have included diagrams for reference to others who find the tread.

Cheers again.

Snax

[AndrewT]

Snax,
you have posted schematics for a dual polarity Power Supply Unit (PSU) of ~+-42Vdc to +-45Vdc. Yes it will vary as mains voltage varies and as the load current changes.
The amplifier has both the signal input and the signal output referred to audio ground. The amplifier needs a dual polarity supply to power it.
BTW the amplifier is a quasi complementary output stage and somewhat unusually has individual drivers dedicated to each of the four NPN output devices.

[AndrewT]

Gootee,
I had forgotten about that 3year old tutorial.
Your second link doesn't. It's a mix of westhost and diyaudio.

[snax]

Hi AndrewT,

Thanks for the confirmation. I’m now feeling more confident in connecting what I have built to the amplifiers without the fear of a small explosion of resistors and caps.

It’s not paranoia it’s just a healthy fear of high voltage… DC at that!

Cheers

Snax