tmbg
Well-known member
How much does noise on the phantom power supply affect mic noise? Is it important to have a near-noiseless phantom supply, or does the balancing cause noise to cancel?
phantom power noise
- Thread starter tmbg
- Start date
Help Support GroupDIY Audio Forum:
SSLtech
Well-known member
Noise on the phantom supply will be cancelled to the degree that the mic preamp rejects common mode signals. CMRR at 100Hz is usually very good, and ripple is usually in the 100Hz-120Hz region.
That having been said, there is no real reason that significant amounts of noise should be present on the phantom rail in the first place, and a 60dB CMRR is not that fantastic if you're operating the mic pre at 60dB gain... 1Volt of ripple will come out of the mic pre at 1Volt under these circumstances.
-If you're wondering -for example- whether or not you have to shield 48V wires carrying signal to the phantom supply, you certainly don't have to be obsessive, but the occasional well-placed local capacitor to ground will cover the chance of noise ingress by bleeding it to ground.
Using regulated supplies and local decoupling I've never had a problem with phantom noise getting into a mic signal, but if I ever encountered more than 110mV of noise of any description, I'd fix it rather than simply expect it to not get into the signal.
Even at a more reasonable 20dB of gain, a 60dB of CMRR performance is reduced to 40dB rejection... that's not going to necessarily be absolutely silent in the quiet passages...
So to answer your question, yes and no. My own view is that if a phantom supply is reasonably clean, then the natural CMRR of the preamp will make any residual noise pretty much disappear. I've run a volt of AC at 60Hz through a preamp with decent CMRR figures and been pretty much unable to hear it at normal operating gains with moderate signal levels, but change any of those consitions and you'll get different results.
I've never had to obsess over it, just start from reasonable numbers and the rest takes care of itself.
Keith
That having been said, there is no real reason that significant amounts of noise should be present on the phantom rail in the first place, and a 60dB CMRR is not that fantastic if you're operating the mic pre at 60dB gain... 1Volt of ripple will come out of the mic pre at 1Volt under these circumstances.
-If you're wondering -for example- whether or not you have to shield 48V wires carrying signal to the phantom supply, you certainly don't have to be obsessive, but the occasional well-placed local capacitor to ground will cover the chance of noise ingress by bleeding it to ground.
Using regulated supplies and local decoupling I've never had a problem with phantom noise getting into a mic signal, but if I ever encountered more than 110mV of noise of any description, I'd fix it rather than simply expect it to not get into the signal.
Even at a more reasonable 20dB of gain, a 60dB of CMRR performance is reduced to 40dB rejection... that's not going to necessarily be absolutely silent in the quiet passages...
So to answer your question, yes and no. My own view is that if a phantom supply is reasonably clean, then the natural CMRR of the preamp will make any residual noise pretty much disappear. I've run a volt of AC at 60Hz through a preamp with decent CMRR figures and been pretty much unable to hear it at normal operating gains with moderate signal levels, but change any of those consitions and you'll get different results.
I've never had to obsess over it, just start from reasonable numbers and the rest takes care of itself.
Keith
DrFrankencopter
Well-known member
It's all dependant on the CMRR of your input stage, and how well matched your 6.8K current limiting resistors are. The more mismatched they are, the more noise will leak through, and it's no longer common mode.
When I've built phantom power supplies, I shot for 0.1 mili volt noise.
Cheers,
Kris
When I've built phantom power supplies, I shot for 0.1 mili volt noise.
Cheers,
Kris
tmbg
Well-known member
thanks guys... reason I ask is because I've been thinking about ways of deriving the 48V from a +/-15vdc supply, and thought about a dc/dc boost converter, but they are switching and tend to be noisy...
DrFrankencopter
Well-known member
I wouldn't go with a DC-DC converter....that 100kHz switching noise is almost impossible to tame.
If you're making the +/-15V supply yourself you could use a voltage trippler approach...they're a pretty common way of deriving the phantom voltage.
Cheers,
Kris
If you're making the +/-15V supply yourself you could use a voltage trippler approach...they're a pretty common way of deriving the phantom voltage.
Cheers,
Kris
tmbg
Well-known member
yeah, the voltage tripler I've seen a couple places, but I don't fully understand how/why it works.
SSLtech
Well-known member
...or there's my KPS-1 all-in one solution...
Keith
Keith
tmbg
Well-known member
[quote author="SSLtech"]...or there's my KPS-1 all-in one solution...
Keith[/quote]
Yes, but that doesn't come with the side-benefit of me understanding how it works
Keith[/quote]
Yes, but that doesn't come with the side-benefit of me understanding how it works
SSLtech
Well-known member
Ah, it's simple enough... a standard bridge rectifier and smoothing cap arrangement, into a pair of positive and negative voltage regulators. This is the absolute most basic, straighforward regulated ±18V supply that you can build.
On to the above, the thing that makes the KPS-1 unique is that it uses a diode/capacitor arrangement rather like a voltage multiplier, but it draws from both legs of the AC supply, not just one or the other. Since the positive side alternates from one leg to the other, this method works much better than the usual single-leg voltage multiplier, since it doesn't spend 50% of the time doing nothing at all... (the AC voltage spends 50% of it's time in the "wrong" polarity, so nothing happens, requiring larger capacitances to keep the charge from decaying)
It's no more difficult than most voltage multipliers, plus it's compact... I like it (but then I'm biased! :wink
Keith
On to the above, the thing that makes the KPS-1 unique is that it uses a diode/capacitor arrangement rather like a voltage multiplier, but it draws from both legs of the AC supply, not just one or the other. Since the positive side alternates from one leg to the other, this method works much better than the usual single-leg voltage multiplier, since it doesn't spend 50% of the time doing nothing at all... (the AC voltage spends 50% of it's time in the "wrong" polarity, so nothing happens, requiring larger capacitances to keep the charge from decaying)
It's no more difficult than most voltage multipliers, plus it's compact... I like it (but then I'm biased! :wink
Keith
tmbg
Well-known member
[quote author="SSLtech"]Ah, it's simple enough... a standard bridge rectifier and smoothing cap arrangement, into a pair of positive and negative voltage regulators. This is the absolute most basic, straighforward regulated ±18V supply that you can build.
On to the above, the thing that makes the KPS-1 unique is that it uses a diode/capacitor arrangement rather like a voltage multiplier, but it draws from both legs of the AC supply, not just one or the other. Since the positive side alternates from one leg to the other, this method works much better than the usual single-leg voltage multiplier, since it doesn't spend 50% of the time doing nothing at all... (the AC voltage spends 50% of it's time in the "wrong" polarity, so nothing happens, requiring larger capacitances to keep the charge from decaying)
It's no more difficult than most voltage multipliers, plus it's compact... I like it (but then I'm biased! :wink
Keith[/quote]
Ahh, well, the +/-18v part I'm perfectly comfortable with... I whipped up a +/-16.9V unregulated supply out of stuff I found in the junk box the other night. It's only unregulated because I dont have any 7815/7915s on hand, but that'll change soon. The bit I'm not all that comfortable with is the voltage multiplier part. Is your design open enough that I could peek at the scheme?
On to the above, the thing that makes the KPS-1 unique is that it uses a diode/capacitor arrangement rather like a voltage multiplier, but it draws from both legs of the AC supply, not just one or the other. Since the positive side alternates from one leg to the other, this method works much better than the usual single-leg voltage multiplier, since it doesn't spend 50% of the time doing nothing at all... (the AC voltage spends 50% of it's time in the "wrong" polarity, so nothing happens, requiring larger capacitances to keep the charge from decaying)
It's no more difficult than most voltage multipliers, plus it's compact... I like it (but then I'm biased! :wink
Keith[/quote]
Ahh, well, the +/-18v part I'm perfectly comfortable with... I whipped up a +/-16.9V unregulated supply out of stuff I found in the junk box the other night. It's only unregulated because I dont have any 7815/7915s on hand, but that'll change soon. The bit I'm not all that comfortable with is the voltage multiplier part. Is your design open enough that I could peek at the scheme?
SSLtech
Well-known member
Absolutely! -It was conceived as a group DIY project, open for all to view, build, evolve and develop as they see fit. There's even a PCB layout and everything!
http://www.beatbazar.com/guests/ssltech/kps-1/index.htm
It's in there with the other projects, the 9k mic preamp and the OB-1 output balancer (with 'accidental phantom' protection!!!)
-Enjoy!
Keith
http://www.beatbazar.com/guests/ssltech/kps-1/index.htm
It's in there with the other projects, the 9k mic preamp and the OB-1 output balancer (with 'accidental phantom' protection!!!)
-Enjoy!
Keith
CJ
Well-known member
They have these new devices that just came out called transformers, and they work great for changing ac voltages to anything you want so you won't have to get silly ideas in your head on how take something rediculous easy like buliding a power supply and turning it into a phd project involving doulblers, triplers, dc to dc converters, regu;lators,,... :razz:
sorry for the rant, just a little tired of all the 48 volt threads.
sorry for the rant, just a little tired of all the 48 volt threads.
SSLtech
Well-known member
CJ,
When are you going to give up being so friggin' negative, stop pî$$ìng on everyone else's pickle-patch, and actually start contributing for a change?
Honestly, with some people, it's just "take, take, take and never give"!!!
:wink: (-juss' kiddin!!!- luv ya!)
Seriously though, transformers with seperate 48V-able windings are much less common than 20-0-20 or similar... that's why I came up with the KPS-1, to help people out and try to steer the many "how do I...phantom..." threads to a happy outcome.
Keith
When are you going to give up being so friggin' negative, stop pî$$ìng on everyone else's pickle-patch, and actually start contributing for a change?
Honestly, with some people, it's just "take, take, take and never give"!!!
:wink: (-juss' kiddin!!!- luv ya!)
Seriously though, transformers with seperate 48V-able windings are much less common than 20-0-20 or similar... that's why I came up with the KPS-1, to help people out and try to steer the many "how do I...phantom..." threads to a happy outcome.
Keith
tmbg
Well-known member
Thanks SSL...
I'm really glad I found this forum, I'm more concerned with learning the whys and less the hows, and so far everything I've seen has been great.
Rest assured I've been frantically studying my horowitz & hill book, so I'll try to keep the dumb questions at a minimum I'm going to try to whip up a debalancing stage tonight, just using a dynamic mic to test, and I'll order some 7815/7915 and LM317s to finish up my PSU.
SSL - do you think that a xformer putting out 17/0/17 will be enough voltage to drive your phantom circuit up to 48V?
I'm really glad I found this forum, I'm more concerned with learning the whys and less the hows, and so far everything I've seen has been great.
Rest assured I've been frantically studying my horowitz & hill book, so I'll try to keep the dumb questions at a minimum
I'm going to try to whip up a debalancing stage tonight, just using a dynamic mic to test, and I'll order some 7815/7915 and LM317s to finish up my PSU.SSL - do you think that a xformer putting out 17/0/17 will be enough voltage to drive your phantom circuit up to 48V?
SSLtech
Well-known member
17-0-17 might be a little low... but you can just change the voltage-setting resistor on the LM317 and aim for 44V or something... remember that 48V phantom mics will usually happily work down much much lower... an AKG C414 for example, will usually work just fine all the way down to about 10Volts...
Keith
Keith
tmbg
Well-known member
Well the mic I'm most concerned about working properly is my 30-year-old U87... although I'm a bit leery of hooking anything I built myself up to it without first using my Oktavas or something as a guinea pig.
CJ
Well-known member
I nomiate Keef to do a power supply Meta, including Phantom!
Yes, people need options for the phantom circuits. Glad to see that project at Kev's place. I wonder what else is hidden over there?
:grin:
Yes, people need options for the phantom circuits. Glad to see that project at Kev's place. I wonder what else is hidden over there?
:grin:
Kev
Well-known member
[quote author="cjenrick"] Glad to see that project at Kev's place. I wonder what else is hidden over there?[/quote]
what
where
who
I didn't do it
no one saw a thing
carry on
nothing to see here
diy diy
happy happy
... ...
:razz:
it's all good
what
where
who
I didn't do it
no one saw a thing
carry on
nothing to see here
diy diy
happy happy
... ...
:razz:
it's all good
G
Guest
Guest
[quote author="SSLtech"]Noise on the phantom supply will be cancelled to the degree that the mic preamp rejects common mode signals. [/quote]
Dear SSL and the others.
Somewhat of you answer can become unclear and to obey thermodynamic law.
There are two kinds of phantom - circuit noise:
1) The noise of phantom power supply unit
2) The noise of phantom circuit (resistors, etc).
The first (PSU) is generally common mode systhem noise (50/60 Hz
and/or harmonics) This noise CAN be suppressed by CMRR of input amp,
but some noise affect microphone s power supply rejection.
The second noise is thermal and flicker noise of phantom resistive
network. this noise is uncorrelated in each wire (hot, cold) and CAN NOT BE SUPRESSED by any CMRR or
other trivial circuits. You must use Maxwell daemons or perpetuum mobile.
xvlk
Dear SSL and the others.
Somewhat of you answer can become unclear and to obey thermodynamic law.
There are two kinds of phantom - circuit noise:
1) The noise of phantom power supply unit
2) The noise of phantom circuit (resistors, etc).
The first (PSU) is generally common mode systhem noise (50/60 Hz
and/or harmonics) This noise CAN be suppressed by CMRR of input amp,
but some noise affect microphone s power supply rejection.
The second noise is thermal and flicker noise of phantom resistive
network. this noise is uncorrelated in each wire (hot, cold) and CAN NOT BE SUPRESSED by any CMRR or
other trivial circuits. You must use Maxwell daemons or perpetuum mobile.
xvlk
> The second noise is thermal and flicker noise of phantom resistive network. this noise is uncorrelated in each wire (hot, cold) and CAN NOT BE SUPRESSED by any CMRR or other trivial circuits.
True, assuming two 6K8 resistors. (One 3K4 resistor to a transformer center-tap should not induce any thermal noise.)
But it IS supresssed by the low resistance of the mike. We are adding 2*6K8= 13K6 in parallel with 200Ω. Signal is reduced 0.127dB, noise power is actually DEcreased 0.063dB. S/N is reduced 0.06dB, which is quite acceptable. (Because of the random nature of noise, it would take an hour just to measure a 0.06dB shift in noise level.)
> You must use ....perpetuum mobile.
In english: "perpetual motion".
Did you see that someone found a math "proof" that Maxwell's little demon won't work for free? Maybe you are too young to be amazed, but I've thought about that little guy for decades and could not find a flaw in the argument that he could make hot/cold for free. I don't really understand the "proof" because it goes deeper into sub-atomic space than I ever learned. But it goes along with the idea that, even if we could stop building tubes/transistors out of resistors, there would still be an energy-cost of computing, Ergs per Bit (more like femto-ergs per megabit of thought).
True, assuming two 6K8 resistors. (One 3K4 resistor to a transformer center-tap should not induce any thermal noise.)
But it IS supresssed by the low resistance of the mike. We are adding 2*6K8= 13K6 in parallel with 200Ω. Signal is reduced 0.127dB, noise power is actually DEcreased 0.063dB. S/N is reduced 0.06dB, which is quite acceptable. (Because of the random nature of noise, it would take an hour just to measure a 0.06dB shift in noise level.)
> You must use ....perpetuum mobile.
In english: "perpetual motion".
Did you see that someone found a math "proof" that Maxwell's little demon won't work for free? Maybe you are too young to be amazed, but I've thought about that little guy for decades and could not find a flaw in the argument that he could make hot/cold for free. I don't really understand the "proof" because it goes deeper into sub-atomic space than I ever learned. But it goes along with the idea that, even if we could stop building tubes/transistors out of resistors, there would still be an energy-cost of computing, Ergs per Bit (more like femto-ergs per megabit of thought).
