Preamplifier power supply

I am building a PS that woud power preamps of various types. Some of them are using single ended supply and some are opamp based that need +- rails. Is it wise to make power transformer secondary center tapped to create standard dual rail supply and then unequally load rails: positive rail would draw around 1.5A of current and negative around 0.5A ( because positive rail will supply both types of preamps ) ? Or I should make separate secondary winding for single rail ? Or this does not matter at all as long there is enough current in the rail ?

Best
Robert

I don't see a particular disadvantage of using just one secondary for both positive rails, except that you have to buy a transformer whos capacity is not used for the negative rail. You could use more smoothing capacity for the positive rail (for 1.5 vs. 0.5 A 3x that of the negative rail) for about equal ripple on the positive and negative rail.

Samuel

what voltages are you after?

regulators would help get even voltages for the dual supply,

DCR in sec transformer winding will create a larger voltage drop across the winding which might translate into a slightly lower voltage on the top rail depending on regulation specs (wire size) of the transformer,

don't forget the phantom supply also,

Thank you, you also answered my next question about applying asymmetric smoothing capacitor value

Samuel Groner said:

except that you have to buy a transformer whos capacity is not used for the negative rail.

Click to expand...

Do not understand what you meant by this.

My plan was to make 56V center tapped secondary and use LM317/LM337 regulators to deliver +-24V. With LM317 I would drive base of 2N3055 to get more current. LM337 can deliver 0.5A by itself. Transformer used would be 250VA ( oversized ), from which i would dedicate 24VA to phantom power ( separate 44V secondary winding ) and rest would be allocated to 56V secondary ( plenty of power in transformer )

CJ said:

what voltages are you after?

Click to expand...

+-24V

CJ said:

don't forget the phantom supply also,

Click to expand...

it is included

Best
Robert

I've assumed you're using a off-the-shelf transformer, where typically both secondaries are rated for the same current. So you'd pay for one over-rated secondary. If you get a custom one that's no issue (and for a one-off project it's not much of a problem anyway).

Samuel

In case of interest,  I have custom transformers with up to 5 separate secondaries in that VA range which will serve your needs perfectly. I also have single, dual and triple output regulator boards.

Samuel Groner said:

I've assumed you're using a off-the-shelf transformer, where typically both secondaries are rated for the same current. So you'd pay for one over-rated secondary. If you get a custom one that's no issue (and for a one-off project it's not much of a problem anyway).

Samuel

Click to expand...

Sam, your brains must still be in vacation. 
Using the two secondaries in series, with the resulting center tap in a bridge rectifier makes the currents equal in both.

Right, thanks! Somehow I was thinking about fully independent rectifiers (and indeed I'm just back from a few days in the alps).

Samuel

What do you think about idea of using dedicated voltage regulator on each preamp board which would be fed from LM317/337 output?

This way I could go with higher voltage from the first regulator ( lets say +-34V ) so I could power also those JLMs high voltage op-amps. The challenge in this is to determin proper secondary voltage since everything is getting to close to the LM317 max input ( which is 40V, I think ).

PS would power 8 preamps that use only single-end +24V supply and consume around 110mA per preamp, in total 880mA. It would also power 8 various op-amp based pres which consume max. 60mA per rail. That would load my positive rail by 880mA + 480mA = 1.4A ( I did not calculated phantom here because it has its own winding which would be able to deliver 500mA ).

Negative rail will be loaded with 480mA in full working scenario.

If current per positive rail is max 1.4A and I have 39V at the input of the first regulator , it means I would have 5V of voltage drop * 1.4A which would give 7W of heat for the first positive regulator to dissipate. Pres that need either single end +24V or dual rails would have dedicated regulator on board locally, which would need to drop voltage by 10V. Since it is only feeding one pre that is consuming around 110mA of current, heat that local regulator needs to dissipate is 1.1W ( for single end ) and 0.6W per rail for dual-end.

Are my numbers approximately correct and what do you think about this kind of construction ?

Best
Robert

Your numbers are correct but why would you want to boost the +V with 2N3055 when you only need 1.4A?

If you want to have additional power use LM338.

If I were you I would keep the single +24V separate from the dual.

But the option to also power high voltage JLMs changes the set up.

In any case if you are going to generate almost 20W of heat I'd recommend to spend it in the power supply case not to heat up the preamp case.

sahib said:

Your numbers are correct but why would you want to boost the +V with 2N3055 when you only need 1.4A?

If you want to have additional power use LM338.

Click to expand...

I fully agree. Also remember that the wonderful overcurrent/overtemperature protection built in the LM317 does not work with external transistors. 

If I were you I would keep the single +24V separate from the dual.

In any case if you are going to generate almost 20W of heat I'd recommend to spend it in the power supply case not to heat up the preamp case.

Click to expand...

+1

sahib said:

If I were you I would keep the single +24V separate from the dual.

Click to expand...

You are right, separate winding for 24V would reduce amount of wasted heat somewhat

Best
Robert

> Transformer used would be 250VA ( oversized ),

That is one HELL of an over-size!

Neves pull say 24V 0.1A or under 3VA.

Two 990 opamps pull up-to 48V 0.050A, again 3VA.

Lots-n-lots of preamps pull under 10mA, more like 0.3VA.

Even double or triple allowing for regulation loss and AC/DC conversion stress, you could feed 25 to 250 preamps from this one large lump.

Yes, my numbers represent worst scenarios, power wise is an overkill, but there should be some sonic advantages that comes with big power transformers. This subject has been covered already and I am interested to try it myself.

Only disadvantage is the weight, but I am use to lift heavy things

No promises, but if I could get a smaller tranny ( like 50VA ) , and if I find time,  I could prepare some audio samples running through same pre but with different power tranny ( unless this test has been done already ). If this happens, it will not happen in soon future ( probably months )

I decided to size smoothing capacitance to allow 1V P-P of ripple voltage ( 10000uF/1A ), do you think this is ok for this purpose

Best

I just realized this was my first topic to GroupDIY, so hello to everyone 

Robert

> sonic advantages that comes with big power transformers.

Some people think there are sonic advantages to having a separate power supply for each major unit. Look at "Dual Mono" power amplifiers.

I think you are using a 100 HorsePower excavator to dig holes for tulips. But they are your tulips, do what you want.

There are also sonic disadvantages that come with using oversized transformers:

i) Larger EMI field. May or may not be an issue, depending on implementation

ii) Greater propensity to generate mechanical noise (maybe it's just more noticeable once you go into hundreds of VA). This can potentially get worse over time as well (t-fs of this size can benefit from a soft-start circuit, which will hopefully decrease mechanical stress to the laminations on power up).

Some people think there are sonic advantages to having a separate power supply for each major unit. Look at "Dual Mono" power amplifiers.

Click to expand...

According to an interview I read with Tim d P, humans only have around 14dB of separation between L/R. But I'm not advocating that we all settle for 14dB x-talk in our preamps  ;D Nor am I averse to a little specmanship; market forces and all that...

edit - before anyone jumps on me, I realise that measured x-talk is a potential indicator for PSR - so it's good to have a healthy figure.

thermionic said:

(t-fs of this size can benefit from a soft-start circuit, which will hopefully decrease mechanical stress to the laminations on power up).

Click to expand...

thanks, I`look into it

According to Rod Elliott no need for soft-start circuitry for transformers below 300VA

I would like to attach leds on each rail to indicate "power good". Can led be connected in series with resistor accros rail or some more complicated approach is needed?