The Pearl 3 needs +/-15V power. Hmmm, perhaps there is a wall wart solution which would save space and money. An old HP Printer of mine bit the dust and it has a +32V/5A supply. Maybe a filter and rail splitter. Here's the problem:
Attachments
I'm not so thrilled about the idea of connecting a synthesized ground (a/k/a rail splitter) to a phono stage. Including the possibility of shooting a third of an ampere through the windings of a moving coil cartridge, if the rail splitter circuit fails.
The famous Victor very low distortion oscillator uses a 35V supply to synthesize +/-12V (or 15?) with a quite simple circuit. It still manages -150dB THD. Worth a look.
Jan
Jan
Doesn't the synthesized ground have to source and sink current? Each channel's 4xJFET first stage squirts approximately 12 mA into the ground network (schematic in post #1 of the Pearl 3 thread). Are you confident the LM317 called "U2" is able to sink 24 mA?
You may wish to pursue the idea of shunt regulators rather than series; shunt regulators can source current and also sink current (within limits).
You may wish to pursue the idea of shunt regulators rather than series; shunt regulators can source current and also sink current (within limits).
Last edited:
How about an AC wall-wart and a simple linear power supply like Rod Elliott's ESP-05 Mini? AC wall warts are not common but I have picked up a couple (16V AC) at car boot sales for a song. Creative Labs used to use them for a PC speaker system.
Here on diyAudio there's a project with schematic and Gerber files called "VRDN". It takes an AC-to-AC wall wart input, and produces +/- VOUT bipolar supply outputs at 1.5 amperes each. You get to choose anywhere in the range (11V <= VOUT <= 20V). A suitable wall wart for Pearl 3 applications might be the Jameco ReliaPro ADU150100RV0360 (link). Of course 1.5 amperes output capability exceeds Pearl 3's requirements considerably, so you could consider cheaping-out on the heatsinks and the total bulk filter capacitance. If you know what you're doing.
I used a 12V / 500mA linear wart for a TL072 based phono pre amp, with a split rail, as the load was low it ran about 17V between rails, which was enough.
They sag to 12V on load, and SMPS gives 12V at all times, but I wanted it higher...and above that wall warts are difficult to get.
Those are gradually going out of production due to the high cost of Copper, SMPS are cheaper.
But still about $1.50 or so here, and about $3 for a custom made one with electrical steel laminations and overall good build, the cheap ones are a bit iffy about that.
You could mount a transformer inside the pre amp body, better choice and magnetic shielding, if it is metal.
Printer power supplies may be noisy, use proper filters.
Also, some printer supplies are two voltages with a common ground, and a feedback loop between one and ground, and sometimes a loop from device, which can be a headache to use, please check that out before using it, as the unused volts may be part of a loop....and it wil try to keep correcting that.
They sag to 12V on load, and SMPS gives 12V at all times, but I wanted it higher...and above that wall warts are difficult to get.
Those are gradually going out of production due to the high cost of Copper, SMPS are cheaper.
But still about $1.50 or so here, and about $3 for a custom made one with electrical steel laminations and overall good build, the cheap ones are a bit iffy about that.
You could mount a transformer inside the pre amp body, better choice and magnetic shielding, if it is metal.
Printer power supplies may be noisy, use proper filters.
Also, some printer supplies are two voltages with a common ground, and a feedback loop between one and ground, and sometimes a loop from device, which can be a headache to use, please check that out before using it, as the unused volts may be part of a loop....and it wil try to keep correcting that.
Last edited:
The various laptop supplies need to hand-shake. (Over the decades I have owned HP, IBM, Lenovo, Toshiba and Dell laptops). This HP Printer supply is not a problem in that respect.
I have used the Murata DC-DC +/- 15VDC converters. They are a bit noisy for this application.
Of course, the lowest noise DC-DC converter design from Jan is no longer available.
I have used the Murata DC-DC +/- 15VDC converters. They are a bit noisy for this application.
Of course, the lowest noise DC-DC converter design from Jan is no longer available.
I got some photos from a member who is experimenting with a MeanWell bipolar AC-to-DC converter SMPS, to power his moving coil phonostage. Apparently the MW model PD-2515, which delivers ±15V at 0.8 amperes per rail, is both cheap ($15 at Jameco) , ($16.79 at Mouser), ($16.92 at Amazon), and quiet. He seems very pleased with the complete lack of hum and noise at the output of his phonostage, when driven by the MeanWell and a pair of diyAudio "AmyAlice" noise filter boards.
He was so enthusiastic that I bought a MeanWell for myself, and fooled around with it to learn about what the SMPS datasheet calls Hiccup Mode. That's its self-protection mechanism which (a) shuts off the output when the load current is too high; (b) does nothing for a while; and then (c) attempts to drive the load again, hoping the load current is now within spec limits.
Not surprisingly, I found that when you ask the SMPS to drive a giant electrolytic capacitor, the output current at switch-on ("inrush" current into the load cap) is waaaaay bigger than 0.8 amperes, and the MeanWell hiccups. On my unit, 6800 microfarads on each of +15 and -15, with no other load, was enough to make it hiccup forever and never stabilize to +15VDC , -15VDC.
Also not surprisingly, I found that the more DC load current you draw from the MeanWell, the less capacitance it can drive without hiccupping at switch-on. Which makes sense: output current at switch-on is just the sum of load current plus capacitor inrush charging current. More load current means less inrush charging current, which means less capacitance. Do you suppose there is a graph on the MeanWell datasheet, with LoadCurrent on the horizontal axis, and LoadCapacitance on the vertical axis, showing the region where smooth switch-on is guaranteed to occur (no hiccupping)? If you were MeanWell, would you bother to characterize this, guarantee it over 100.00% of all production units, print it on the datasheet, and answer customer inquiries about it?
edit- if anyone is thinking about buying one of these MeanWell bipolar supplies and playing with it, let me suggest that you ALSO purchase an assortment of 10 watt wirewound resistors to load down the ±15V rails by different amounts. They're pretty cheap, less than 75 cents apiece, so buy a bunch and be ready for anything. Plus an assortment of 25V electrolytic capacitors of course, to explore what is and is not possible to drive at switch-on. At minimum, 470 microfarads {the caps in the diyAudio Store "PO89ZB" filter} and at maximum, about 6800 microfarads {which made my MW unit hiccup forever}.
He was so enthusiastic that I bought a MeanWell for myself, and fooled around with it to learn about what the SMPS datasheet calls Hiccup Mode. That's its self-protection mechanism which (a) shuts off the output when the load current is too high; (b) does nothing for a while; and then (c) attempts to drive the load again, hoping the load current is now within spec limits.
Not surprisingly, I found that when you ask the SMPS to drive a giant electrolytic capacitor, the output current at switch-on ("inrush" current into the load cap) is waaaaay bigger than 0.8 amperes, and the MeanWell hiccups. On my unit, 6800 microfarads on each of +15 and -15, with no other load, was enough to make it hiccup forever and never stabilize to +15VDC , -15VDC.
Also not surprisingly, I found that the more DC load current you draw from the MeanWell, the less capacitance it can drive without hiccupping at switch-on. Which makes sense: output current at switch-on is just the sum of load current plus capacitor inrush charging current. More load current means less inrush charging current, which means less capacitance. Do you suppose there is a graph on the MeanWell datasheet, with LoadCurrent on the horizontal axis, and LoadCapacitance on the vertical axis, showing the region where smooth switch-on is guaranteed to occur (no hiccupping)? If you were MeanWell, would you bother to characterize this, guarantee it over 100.00% of all production units, print it on the datasheet, and answer customer inquiries about it?
edit- if anyone is thinking about buying one of these MeanWell bipolar supplies and playing with it, let me suggest that you ALSO purchase an assortment of 10 watt wirewound resistors to load down the ±15V rails by different amounts. They're pretty cheap, less than 75 cents apiece, so buy a bunch and be ready for anything. Plus an assortment of 25V electrolytic capacitors of course, to explore what is and is not possible to drive at switch-on. At minimum, 470 microfarads {the caps in the diyAudio Store "PO89ZB" filter} and at maximum, about 6800 microfarads {which made my MW unit hiccup forever}.
Last edited:
Patrick (EUVL) has some very high performance drop-in 3-terminal regulators that would be just the ticket.
Jan
I will check. It's somewhere in the Group Buy section I think.
Edit: https://www.diyaudio.com/community/threads/possible-gb-for-lt3045-3094-pcb-ref-smd-fc-cen-iv.402297/
He has stock I believe.
Jan
Edit: https://www.diyaudio.com/community/threads/possible-gb-for-lt3045-3094-pcb-ref-smd-fc-cen-iv.402297/
He has stock I believe.
Jan
Last edited:
I'm not sure what current you were drawing from the wall-wart in your original pictures,
but most SMPS start to chirp ( become discontinous ) at loads 100's of times less then they were designed for.
Drawing milliamps out of a 5A SMPS at low noise and low currents might be difficult.
but most SMPS start to chirp ( become discontinous ) at loads 100's of times less then they were designed for.
Drawing milliamps out of a 5A SMPS at low noise and low currents might be difficult.
Last edited:
I keep wanting to say this, bus as I've said it before here's a link:
https://www.diyaudio.com/community/...voltage-regulators.359652/page-3#post-6330067
This requires two identical secondaries, but a lot of 50/60Hz transformers thesedays have that.
https://www.diyaudio.com/community/...voltage-regulators.359652/page-3#post-6330067
This requires two identical secondaries, but a lot of 50/60Hz transformers thesedays have that.
What's the current requirement? For reasonably low current, there's the idea of a single 15-20V AC secondary winding (from an old-fashioned 50-60Hz walwart) powering two half-wave supplies. Switching supplies for audio seem worriesome with the noise generated by fast switching and such.
Yes, and if you did a board for it it (with implementation of your insights added) might very well fly off the shelves !