Working on a preamp, id like increased headroom and the tonal circuits im using would be easier with a ground reference. ..
What specs do I "need" out of a bipolar supply for it to be appropriately quiet for musical applications?
Alternatively, with a common 9v supply, could i use 3 MAX chips to create +/-9 and subsequently double each 'side' to get +/-18 v?
Don't know about the MAX chips, but for audio use, any power supply needs to be properly filtered. That is usually as simple as adding some R-C sections to take out ripple (manifested as hum). Of course, you can use regulators to do a pretty bang-up job of ...regulating... AND filtering.
So, for "specs" you're looking at (from my point of view) providing your desired voltages, with a little room to drop some in filtering. For +/- 18V, for instance, you might want something more like 24V....regulators take a few volts to operate. It's the way I'd go, YMMV.
Look what they're doing here, might be the way to go: http://www.instructables.com/id/Power-Supply/ (http://www.instructables.com/id/Power-Supply/)
Use one LT1054 (more tolerant to the voltages over 9 volts) to create a bipolar voltage.
As far as headroom.....+9/0/-9 is equal to +18 volts. An 18 volt bipolar supply would be equal to 36 volts.
Keep in mind that some op amps such as the 324 top out at +15/0/-15.
In the world of pedals, it's very unusual to need more than +/- 9V.
If you think you need more, it ought to be based on a specific set of conditions you know you have.
^Alternatively, with a common 9v supply, could i use 3 MAX chips to create +/-9 and subsequently double each 'side' to get +/-18 v?
For reference - you *can* actually get +/- 2*Vin(ish) from a single charge pump. It will tend to sag pretty badly under event quite modest loads, but works at a pinch for a couple of op-amps. As others have said, +/- 9V should be plenty for just about anything you need.
See the 'Bipolar Supply Doubler' on pp11 of the LT1054 datasheet: http://cds.linear.com/docs/en/datasheet/1054lfh.pdf
For a few op-amps, 100r + 220uFd on each rail will clean raw DC pretty good.
IMHO, this goes on the amplifier board. The power supply "should" be clean, but often isn't; it has other jobs to do. The amp designer should defend against the power supply's shortcomings.
I like using two identical wall warts connected together in series using isolated jacks or a single wall wart with ± voltage outputs. You can get wall warts anywhere and they:
1. don't radiate high-frequency noise that can get into inputs and beat with incoming audio to create intermodulation noise or
2. become an issue for a manufacturer who has to qualify anything that generates signals above 9 KHz to FCC radiated emissions standards, a normal cost (~$1500) for a volume manufacturer but a show-stopper for the individual builder of a few units.
Besides, you have to order power converter IC's and their associated components from suppliers whereas every village has a store (including Goodwill) selling wall warts. If you want something more robust, find one of the old Lambda ±12 or ±15 supplies and either turn down the voltage or add regulators inside the stompbox (which are cheap and easy to use).
I am fully in favour of using ± voltage supplies and supplies that are in general greater then the nine volt batteries that came over on the Mayflower. If you look at the spec for the TL07X op amp that most pedals use, it has a headroom problem - the minimum spec for input common mode voltage is 4 volts below the positive supply and 4 volts above the negative supply, leaving 1 volt for the sum of signal and battery degradation. The fact that most designs work only means that the device is typically better than its worst-case ratings, but you should never design something that depends on devices performing better than their spec. I have no problem with a pair of 9 volt batteries in series and battery switching being done by the usual shorting input jack also being used for the output jack for the other polarity and shunt diode protection to prevent possible circuit damage issues with only one supply connected. With enough voltage overhead, polarity protection becomes a 1N400X in series with the input and not the sacrificial shunt diode that sometimes works but may self-destruct before the voltage drops, leaving the circuit unprotected.
I've said it before, and I'll say it again: *AC Wallwart*
Low cost UL/CE approved parts are available as a result of the security industry (typically 16VAC@1A-ish).
Two diodes and a couple of fat capacitors and you have a nice +/- 22V-ish (some hundreds of mA) supply ready for some regulation! *Suggest you might use a plug that nobody will confuse for 2.1mm DC*.
david
Quote from: greaser_au on December 22, 2014, 06:13:31 AM
I've said it before, and I'll say it again: *AC Wallwart*
Low cost UL/CE approved parts are available as a result of the security industry (typically 16VAC@1A-ish).
Two diodes and a couple of fat capacitors and you have a nice +/- 22V-ish (some hundreds of mA) supply ready for some regulation! *Suggest you might use a plug that nobody will confuse for 2.1mm DC*.
david
Old school, like my thoughts on this. Simple and time tested! If you need it cleaner, you can regulate, but it would probably be just fine as you lay it out.
Yup, AC wall warts work great, especially for op amp circuitry that has some power supply ripple rejection.