• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

advice needed on designing power supply

Thread starter #3
I think I've worked out the right question to ask. I'll start by saying hello to everybody and thanking electromaster for getting this board together. It's difficult to explain the problem I'm having designing this power supply so I've put the schematics and my questions on a web page (link below)




New Member

I've just been looking at your site and can't see any reason for keeping the two +12v rails seperate - just common 'em up and use a single supply.
Your way of duplicating the circuit for stereo is correct.

The only current used by this circuit will (unless you try driving headphones directly) be the two JFET op-amps. I can't find any consumption info for these but it wouldn't surprise me if 10mA would be enough for the whole mixer; 100mA is bound to be plenty. Just build the top half of the power supply you show on your site.
The transformer's lower half-winding can be dumped (possible single 15v winding-type transformer) if you use a full wave rectifier, or two of the diodes on your diagram (D1 & D4) can be dumped if you keep this transformer.

As you mention 'studio' I wonder if you are aware of 'Hum loops' :twisted: - multiple earth points on an audio system causing a constant audio humming noise. Try leaving the earth connection off this supply and rely on the audio leads connecting to an earthed amplifier/tape deck/whatever. Earthing in complex audio set-ups is rarely straightforward :( .

Play - the only way to learn.

ps. when you blew up the capacitors you will probably have cooked the associated op-amp. at the same time :oops:
Expensive smoke - there's no substitute :!:
Half supply..

If you dump half the supply wont the sound sound bad? Considering it wont swing below 0 V? Wont you need coupling caps all the way through the circuit otherwise?

Maybe I misunderstand you....


New Member
'Single Rail' Amplifiers

This amplifier is capacitor- coupled at each input (fig. 2.20 C5 to C8.) and output (C10) :) .
The amplifier is a standard 'inverting' design. A bias is provided by R5 and R6 to the non-inverting input - this will be half the supply rail (therefore 6v). The feedback resistor R7 causes the inverting input to match this voltage in order to balance the output (this inverting input is referred to as a 'virtual earth point' - you will never see its voltage change #unless the output goes full-scale 0v or 12v#).
At this point you can imagine the supply as -6v and +6v if measured from the virtual earth point.
Just to complicate things though, neither inputs nor the output use this virtual earth directly - the capacitors allow for a DC voltage shift so that the power supply negative rail can be used as 'real world earth'.


New Member
Thread starter #7
Mechie, thanks for the reply. It's been nearly 2 months since I posted this question and I'd given up on getting an answer. By trial and error I worked out how to build a +12v 0v +12v power supply, only to discover (as you pointed out) that a single +12v rail was all that was needed, but I learnt a lot in the process. I just checked the specs for those opamps, they appear to need a supply voltage of 1.8 mA so I could have saved a few quid on the power supply. Who cares I learnt loads doing it :) . So... if I do want to use this mixer with headphones will the 100mA from the power supply be OK to drive them?


New Member
Driving Headphones from a mixer

The 10uF output capacitors on your mixer will protect headphones from the 6v offset the op amp generates so there's no danger there (but you will get a hefty 'thump' in the headphones when you first switch on the mixer :shock: ):-

Going back to your website diagram Fig 2.20.
The amplifier is unity-gain (input resistors R1 to R4 = 100k and feedback resistor R7 = 100k) with a potential divider on the input for volume control (the pots VR1 to VR4). The best this amp can do then is give out the same voltage as is applied to the input (sum of all four inputs). If this is a standard 1v p-p sort of studio then you won't get much power into the (assume 8 ohm) headphones - P = v squared / R = (1^2)/8 = 120mW (and this assumes 100% efficiency), but it would work.
At this power you would be taking I=V/R 120mA from the power supply PER CHANNEL; but this is PEAK current at MAX output :-
If you listen to lots of square waves then :-
... This is 50% Max current and 50% Min current (0mA)
... this must then be only half the power and current calculated above.
... you must be weird :wink:
Real audio will be very few peaks and so even lower current.
A 100mA power supply will just about cope.
The downside is that you will 'load' the amp output and reduce the signal at JK5 whenever the 'phones are connected - does it matter?

If you want a headphone output why not add a second 'mixer' amp with a larger feedback resistor (say 330k), still a 100k input resistor (only one this time) - gain now about 3x, take it's input from your existing mixer output (which still also gives the JK5 output signal) and use this to drive headphones? (via the output capacitor of course).
The same type of op amp (LF351) or a 741 would do - they are both very tolerant of output overloads so experiment (with cheap 'phones to start with!).


New Member
Ahh.. didn't find this thread before but this is all I'm looking for, excellent!

Thanks for pinpointing me mechie :D

//Albert "thec" Sandberg
Thread starter #11
Ok. I just found some time to get back to this project. I've simplified it to use as a stereo volume control, between my satellite TV receiver and the hi-fi. The problem is, the damn thing is whistling at me. It's ok until the powers switched on. Then this whistle that changes pitch as I move my hand over the circuit starts. If I remove IC1, the IC2 side works fine and it's the same if I replace it, and then remove IC2. Is this something to do with component placement on the PCB/breadboard i.e. certain components to close or too far away from each other?


Oscillating Amplifier

As a rule audio amplifiers don't care about component layouts- apart from hum pick-up from power lines there should be no problems.
I think it highly unlikely that you have enough capacitive coupling from amp output back to its input to form an oscillator.

You appear to have (re)invented the theremin :shock:

My guess is a wrong component connection somewhere >>>

Double-check the op-amps wiring -- could you have swapped inverting and non-inverting inputs for each other?
Could one of the feedback resistors (R8 and R9) be open/missing/connected to non-inverting input?
Connect a 100uF or bigger capacitor across the power lines (careful of polarity!) just in case it is some weird supply being overloaded type of problem (most likely with switched-mode supplies so unlikely to fix the problem :?: )
Thread starter #14
Thanks again for your input mechie. Someone else suggested putting a 100uF cap across the supply rails. I tried that but it didn't fix it. However, I have discovered that a 0.01uF ceramic disk across the outputs ( after C7 and C8 ) cured it. I'd love to know why.

Your amplifiers are used in inverting mode and so have 180degree phase change. In addition you have several RC networks in the signal feed path, each giving a phase change which varies with frequency. If the overall phase change at any frequency exceeds 180 degrees and there is feedback from o/p to i/p due to the physical layout, then the circuit will oscillate. That may be your problem.

You would have done better to have wired the amps as followers with unity gain. It would have been more stable - and simpler.

Latest threads

EE World Online Articles