DIY phono preamplifier part 1: low noise power supply

[@sc0or]

An ultimate PSU for vinyl would be a charging station in an iron case distant from a riia corrector, charging one battery bank while another one feeds the corrector. I used to calculate that 6x18650 are enough for 100-200 hours (depends on a schematic, but 2000mAh and 10-20mA).

[@yeoldestuff]

This is correct, however such a PSU would be way beyond the point of diminishing returns. The power supply does not need to be any better than the Johnson noise generated by the loading resistor and the inductance of the cartridge itself.

[@sc0or]

@@yeoldestuff That seems to be a quite low noise. If I'm not mistaken -88dB is about 0.5mV on 12V rail in a case of a zero PSRR (which is not a case in most schematics). That is far worse than modern LDO can provide. May be you are right.

[@gino3286]

Hi thank you very much for your great video I have a question about traditional power supply point 1:45 in the video You mention an inductor to be placed after the regulator To suppress what ? i mean what kind of AC noise could come out from the regulator not suppressed Have you tried also the classic regulator circuit ? Thanks a lot again Kind regards, gino

[@yeoldestuff]

Hi gino, traditional monolithic regulator ICs such as LM7815 and LM317 can be quite noisy depending on the load capacitance, with noise peaking anywhere between 1 and 20 kHz depending on the load and other parameters of the circuit. Placing an LC or an RC filter after the regulator should help to suppress the aforementioned noise peaks.

[@gino3286]

@@yeoldestuff Hi thank you very much for your kind and extremely valuable advice I clearly have to study more the topic and i am about to do it Could these filters be places indifferently after or before the regulators ? and what would be your best option RC or LC filter ? thanks a lot again

[@yeoldestuff]

@@gino3286 You’re welcome. The filter should be placed after the regulator. It is easier to get a lower cut-off frequency using an LC filter but frankly an RC filter should be just as efficient. For example, using a 1000uF capacitor and a 2.7Ohm resistor would effectively kill everything above 60Hz.

[@gino3286]

@@yeoldestuff thank you very much again I am starting now to do something I will start trying to measure noise from power supplies in the audio band using a low noise sound card and Arta software power supplies quality is very important in audio equipment Kindest regards gino

[@simplereef4854]

Very nice video, how much it cost you in total to make the PSU?

[@yeoldestuff]

Hi, the two most expensive parts were the PSU board itself (50 euro) and the transformer (17 euro). The case I bought on Ebay for 14 euro, and the rest of the components I had lying around but they were really cheap (e.g. you can buy a bag of 10 IEC sockets for a couple of euro, hook-up wire is typically about 5-7 euro for a spool etc.)

[@ckngmad1357]

Why no ground? Why not using metal case as shielding?

[@SylwerDragon]

I know it is a bit late to comment. but in general there are few problems with your design 1. Plastic box is not good as there is no shielding that is needed to decrease influence of noise. 2. problem with transformer isn't trivial. I was also surprised that is actually matter what kind of transformer is used. Toroidal transformer is actually worst transformer used. The best is specially shielded transformer that is shielding primary from secondary side. Another option is to use coils on different lets say Legs so that means primary coil is alone and there isn't another on on itself as it is on most transformers. even these separate transformers are fine they needs additional shielding. 3. pcb design isn't bad problem is it isn't so good. I would suggest to use low noise circuits..it ofc depend what kind of Amps you need. but in general up to 1Amp is ok..then i would suggest LT3015 if lower current is possible then i would suggest : LT 3042 up to 0.2A or LT 3045 up to 0.5A. 4. It isn't simple to reduce noise as low as possible it is pretty complex problem. These solutions and circuits are used for advanced Voltage standards to reduce noise as much as possible.

[@yeoldestuff]

1) a metal case would be completely useless against low frequency EMI unless it's made of ferromagnetic materials, e.g. cast iron 2) you can get a toroidal transformer with electrostatic shielding as well, although the effectiveness of electrostatic shielding at audio frequencies is disputable. The advantage of toroidal in this particular case is that it has the weakest external electromagnetic field of any transformer type, which helps reduce induced hum in downstream circuits 3) I am highly skeptical that any monolithic regulator would be half as good as Sulzer's or Jung's regulator

[@SylwerDragon]

@@yeoldestuff Nice response. As they say prove me wrong :). To point 1. Shielding isn't useless as such if there isn't any material or only plastic it works both way high frequency could go in and also out.. if you would go super op you could use double shielded cables but that might not help you in general at all in simple project as power supply. If shielding wasn't needed we would have all laboratory equipment from plastic instead of metal. To point 2 even properly shielded toroidal transformer is too noisy and as such not used for low noise applications. i hardly doubt any regulator is as good as monolithic one such as LT3042 prove me wrong. I 'm glad to learn something new and improve my pcb design.