Inadequate power supplies. .

Art,

Wall warts are ok for stomp boxs eqs and even compressors. The Speck ASC EQ uses a wall wart and it's a very good equalizer!I only have issue with them in mic pre applications where you need to juice up a mic level signal to line level and power condenser mics as well ...



A few people will disagree and that's ok. If they like how a pre amp driven by a wall wart or line lump sound, I don't care.

I have an early POD and the newer rack unit (XT Pro). The differences in their sound are not linked to the power supply design differences.

Kurt is right when he states that the units that are most affected negatively are mic pre's trying to deliver good headroom and phantom power. There is too much junk out on the market these days that doesn't deliver REAL power to the mic. I am still waiting on the verdict over the Portico debate...

Kurt Foster wrote: A few people will disagree and that's ok. If they like how a pre amp driven by a wall wart or line lump sound, I don't care.

I'm not against external power supplies as much Kurt is and in theory I see nothing wrong with switch-mode supplies.

History does tell us that quality sizable analog supplies in quality gear works the best.



as with many subjects in audio ... a little knowledge is dangerous ...

but

here goes anyway



Lets not talk about pre-amps or stomp boxes to start with. People have re-conceived ideas so it might be better to start with something less emotional and more comic so everyone gets the point.



Computer Speakers

:lol:

We like computer speakers cos they are so big and strong.



Everyone would laugh at a computer speaker that has 600watts PMP or what ever they want to call it. I've exaggerated but so what.

We all know these things don't have the power that is advertised and it is crap.



Things you do know about is the power supply or the wall wart and perhaps the fuse , if it has one and the voltage and current it is capable of delivering.



For example

... 10Volts DC and 1 Amp (1000mA) = 10x1 = 10W

Power = Volts x Amps

... 10V DC and 500mA = 5W



more typical for a large wall-wart might be

... 12V DC and 850mA = 10.2W



For a stereo computer speaker that's about 5W per side and that is only if the system were 100% efficient

which they are NOT

at about 60% efficiency we end up with about 3W per side if we are lucky.



To move this on to mic-pres or stomp boxes we use the same maths .... but before I do can anyone tell me ...

How many watts is required to drive two channels of 600 ohms up to 24dB ??



is this worth continuing ?

moonbaby wrote: There is too much junk out on the market these days that doesn't deliver REAL power to the mic. I am still waiting on the verdict over the Portico debate...

delivering power TO the mic ????



Phantom Volts required by mics may be 48 Volts but the current required is usually very VERY small

and so the POWER is also very small .... perhaps less than the power consumed by one LED.

This is cool Kev.

Please continue my education. This is the kind of stuff I always look for, If someone can show where I am misunderstanding the concepts I am happy to pay attention.



I still have never heard any pre powered with a wart or lump that sounded good though.

Kurt Foster wrote: I still have never heard any pre powered with a wart or lump that sounded good though.

yes

and this may be a true generalisation ... but might be more about cost and compromise than just the fact that it is an external supply.



we could look at dBu, dBV, dBm and so on

but it may just confuse

have a look at

http://www.rane.com/par-d.html#decibel

for some definitions



Lets just look at some max levels to be found in typical op-amp circuits.

Generally we see +/-15 volt rails and some times 20 and 24 volts with the classy op-amps.



using the dB calculator at Analog Devices

http://www.analog.com/Analog_Root/static/techSupport/designTools/interactiveTools/dbconvert/dbconvert.html

we can get some raw data.



first we check a couple of favourite alignment levels at 600ohms termination.

0.775 Vrms = 0dBu = 1mW

1.23 Vrms = +4dBu = 2.5mW

and the dB calculator does seem to agree ... this is good.



with 15 volt rail an output signal is not likely to peak much above 14 volts, allowing for a little loss at the top through the op-amps.

using V PEAK this time

14Vpeak = 22.13dBu = 163.3mW

for 20 volt rail

16Vpeak = 24.78dBu = 300.8mW

and 24

23Vpeak = 26.44dBu = 440.8mW



Now some of the above dBu figure may look familiar to people with modern interfaces from the usual suspects.

It should also be noted that some of the favourite op-amps may deliver the required voltages into 10K inputs BUT suffer at these levels into 600 ohms. There will begin to current limit and self protect.



The big CRUNCH comes when we double terminate as you might if you used a Y cord into an old 600ohm mixing desk, for hardware monitoring and then an LA2 feeding the interface.

Double terminated = 300 ohms.

Use the calculator to see the new current requirements. Only the units with transistor outputs are likely to be able to do this ... and larger tub'ed & transformer'ed outputs .... 1176, LA2, 1272 etc ...



BUT for interfaces that are 10k and above .. AND only require 22dBv to attain 0dBFS

you may never have a problem.



SO at this stage one output stage, double terminated into 300 ohms with 15 volts peak we need 650mW of power.

Lets assume 50% efficiency so now we need 1300mW

:roll:

Dual channel unit

so now we need 2600mW

:roll:

Add some LEDs and a Meter light and some phantom volts and we will be over 3000mW in no time.

this is 3 W

some meter lamps may suck much more and we do seem to see LEDs lighting meters these days ... less power and cooler.



A 15 V power unit at 500mA to 850mA ... even 1 AMP seems like it should do a simple Dual Mic-pre.

BUT does it ? ... and if not ... what is going wrong.



:roll:

anyone see any mistakes above ?

lets not get picky as it was all ball park figures.

actually there is a mistake so if the geeks are concerntrating, they can jump in an tell me.

8)

I want to be Kev when I grow up.



:cool:

kev you forgot about all thouse lights behind the tubes :)



im sorry i wish i had something helpfull to add here.

Kev wrote: anyone see any mistakes above ?

actually there is a mistake so if the geeks are concerntrating, they can jump in and tell me.

Guess that would make me a geek :roll:

The mistake I see is the 16Vpeak = 24.78dbu = 300.8mW

The correct equazation is...

16Vpeak = 23.29dbu = 213.3mW



IMHO 1 Amp would be plenty for the pre-amp as long as the peak was not the constant.

You know the whole RMS deal. I use amprobes all day long at work, checking on motors and control circuits. I used my Amprobe on a mic once, with my meter on several settings, to see what the output on a microphone was and how high the dail would go (singing, screaming, whispering). Just because I'm a geek.

Do I win anything? Give me give me! :)

so if you got the 15 volt rail and lets say 16Vpeak = 24.78dBu = 300.8mW, with the 600ohms being double terminated OH who am i kidding i have only a small clue of what i am takling about.



I am definately learing from you though Kev.



thanks,

-Blaumph

its that guy again wrote: Guess that would make me a geek :roll:

The mistake I see is the 16Vpeak = 24.78dbu = 300.8mW

it was a mistake but it was supposed to be 19V ... my dislexia at work agian ... 8)

I was trying to show 15 20 and 24 volt systems



the geeky bit I was looking for was more about +/-15 volts is 30 Peak ....

or is it ?

therecordingart wrote: I wish I understood a single thing that was typed!

darn it

:roll:



perhaps I need to think and then present a web page that better explains this.



Why is it that people seem to understand loading when it comes to 8 ohm 4 ohm and 2 ohm speakers

BUT then totally forget about current and power when it is a 600ohm line level vs 10k+ of a brifdging input ???

Or

Headphone levels ... Impeadance of headphone are all over the place.

the geeky bit I was looking for was more about +/-15 volts is 30 Peak ....

or is it ?

Yeah that's it Kev.

Split rails so add those rails for the full swing.



The specs don't look too bad after that.

28.75dbm , but that's at full efficiency as you say.... into 600 ohm ...



For 18 V +/- ( or 36 V full swing ) that's around 30 dbm@600 ohms

I think.



WOT's it sound like though?

For 18 V +/- ( or 36 V full swing ) that's around 30 dbm@600 ohms

I think.

Actually that 36 volts is pure rail not p to p , so it should be even better than that.In theory but there are always substantial losses.



All this maths doesn't go far in the real world.

The unit should be evaluated by it's performance versus retail cost.

The H.P.F is a good move and an external power supply seriously takes off $200 or more in manufacturing costs ( smaller cabinet ,, less internal shielding... ) ......



What's the frequency of the internal oscillator for the PSU?

What's the bandwidth of the unit?

Is it true class A or is he just biasing the opamps halfway?

Sebatron wrote: All this maths doesn't go far in the real world.

The unit should be evaluated by it's performance versus retail cost.

Hi Seb



yeah and as you said before

Sebatron wrote: WOT's it sound like though?

You have raced to the end zone and brough a bunch of things too fast.

Yes there is lots to be considered and as I said before it can be a real compromise when you are trying to earn a living.

In DIY you can just build what you want and not worry about expense or spec sheet ... or even trying to explain things to a client.



All I was trying to do was set up a little bit of maths to help see through the crap.



If the joib seems to require a 10 W supply and you only see 5W then you have good reason to look again.

If the power supply is getting HOT ... same as above



BUT

if you seem to need 1 W and you have 5W and the thing sounds good ... then do look deeper to the other aspects of the design and use ALL infortmation to help your purchase choices.



Living in a bridging input world of 10K and above ... the driver currents are much lower and so to are the power supply requirements.

On the other hand 8 channels of Mic-pre ALL with phantom and ALL with a peak light to flash often

AND

light meters

AND

double terminated at 300 ohms .... and so on





The other points Seb brings up about

oscillator frequency

bandwidth

true class A or the bias trick

All make a difference



complicated isnt it ?

and a little knowledge is dangerous



Those average analog Plug Pack / Wall Warts can be very flakey with voltage and current capability .... and sometimes even unpredictable,

C'mon they're CHEAP.



There can be lots of reasons that a unit doesn't deliver big levels into a given load.



is this worth discussing ? ... if so we need to take one point at a time or we end up in circles

Phleeez continue guys. This is good content. '8-)' :!:

ok Kurt

but , point me in a direction cos I seem to have lost my way

what was I saying ?

:roll:

Sebatron wrote: Actually that 36 volts is pure rail not p to p , so it should be even better than that.In theory but there are always substantial losses.

yes there are always some losses but for ball part stuff we can use the rail as a MAX voltage source and work from there



" ... is pure rail is NOT p to p ... should be even better "

?

:?

err I must be reading this wrongly

I've been on night shift for the Cricket ... my brain may need a re-boot ...



:roll:

perhaps we sould pick a simple well know circuit and just explain some of the requirement in different situations ?

This is great reading Kev (& Seb) please enlighten us more.

perhaps we sould pick a simple well know circuit and just explain some of the requirement in different situations ?

Sounds great, I'm all ears.

I started to knock up a chart to compare levels and rail voltage and current requirements



(Dead Link Removed)