Power Supply Preference - Page 6 - diyAudio
Go Back   Home > Forums > Amplifiers > Power Supplies

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 7th December 2012, 12:08 PM   #51
Mooly is online now Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
I'm curious to see how this works out and its good to go over the theory once in a while. And while we all agree music and speaker impedance are anything but constant we have to have some "standard" method of arriving at realistic figures. So here's what I came up with

100 watts into 4 ohms. That means the required amplifier output voltage is V= √W*Ω that calculates as 20 Volts RMS.

What rail voltage do we need. 20 volts RMS is 28.28 volts peak and 56.56 volts peak to peak. Knowing this enables us to "guestimate" the likely supply voltage. No amplifier can put its supply "across the load", there are losses. These occur in the output devices themselves, in the emitter resistors, in the wiring, in fuses etc etc so we have to make assumptions now. The figure could be anywhere in the 5 to 15 volts range. Lateral FET's in particular have a high Rds value and are consequently "lossy". Lets say 8 volts is a "reasonable" overhead. That implies we need our 28.28 volts peak PLUS another 8 volts.

So we get 36.28 volts as the supply voltage (that's plus and minus 36.28 volts don't forget) because the amplifier can only "use" one rail at a time.

So moving on to the transformer we get 36.28/√2 which is 25 volts. Thats the AC voltage needed to give 36.28 volts after rectification and smoothing. Or is it ? The bridge rectifier has losses and these are higher at high currents. Higher than you might think. We probably need add another two or three volts or more to overcome these losses.

So now we are at around lets say 28 volts AC. For the dual rail supply that means a 28-0-28 (or higher) Vac transformer for our 100 watts into 4 ohm amplifier.

What about current though ? This is where it all gets more complicated and I had to refresh my knowledge on some of the more obscure formulas (college was a long time ago )

The peak current in the load is easy though. Ohms law. 28.28 volts peak across 4 ohms is 7 amps. The average current is Ipk/pi giving 2.2 amps. (Thats for one channel remember and can be used as a guide for transformer current rating. For 2 channel we are doubling the current rating)

What size reservoir cap. That depends what ripple voltage you consider acceptable. Also remember that huge capacitor banks draw hugh peak currents when supplying a steady steady state current into a load. Why ? Because the large cap discharges "little" in between mains cycles yet that energy taken out has to be put back in. This means the bridge and wiring and transformer supply huge peak currents but over ever shorter time intervals. Why so. Because the capacitor peak voltage drops little between cycles so that means that no charge is put back until the transformer voltage reaches at least this value (plus losses) and so the charging occurs only near the "tips" or peaks of the AC waveform.

Ripple voltage across the caps is the average current drawn * discharge time. I found conflicting info for this discharge time value with some formulas just quoting half the mains cycle time. i.e. 10ms for 50 Hz. That doesn't sound logical. Another reference quoted around 7ms on the basis that this is the average discharge time with around 3ms as the charge time. That sounds reasonable so we'll go with that. So ripple voltage is (Iaverage/Tdischarge)/capacitance.

Putting some numbers in and using our 2.2 amps average current and selecting say an 8200uf reservoir cap and we get around 1.9 volts pk/pk of ripple. That ripple voltage "reduces" our steady state DC rail value because the minimum rail voltage is now the lower part of that ripple waveform. So thats another thing to factor in the guestimation. That minimum dip has to be equal or greater than the steady state DC rail voltage needed.

And again remember. The calculations were for one channel not two.
__________________
-------------------------------------------------------
Installing and using LTspice. From beginner to advanced.
  Reply With Quote
Old 7th December 2012, 12:31 PM   #52
MarianB is offline MarianB  Romania
diyAudio Member
 
MarianB's Avatar
 
Join Date: Jul 2011
Location: Romania
Send a message via Yahoo to MarianB
Nice marathon

All is verry usefull info, and i hope everyone will take it into account. Still i would think that 8V drop for BJT output devices seams like a bit of a stretch, 3-4V would seam more likely, FET's would eat about 8 or even more so, that depends on theyr Rds-on at stated current and also on Vth for the gate witch is usually arround 4V.

Regards,
Marian.
__________________
Trance Session Live by Dj Marik
  Reply With Quote
Old 7th December 2012, 01:01 PM   #53
jlind54 is offline jlind54  United States
diyAudio Member
 
Join Date: Nov 2012
Location: Kansas
Thanks for getting me on the right track and explaining things so well. Now that I have a bit better understanding I would guess that a standard mains transformer capable of the output described above would have the same sound as a smps capable of outputing the same voltage/current. Ideally there wouldn't be a corralation between the amps psu and the sound produced correct?
  Reply With Quote
Old 7th December 2012, 01:10 PM   #54
diyAudio Member
 
Join Date: Jun 2011
Reality check:

And the fets will never be full on either so rds-on has quite a little to do in the reality with the losses and power consumption of the amplifier.
In reality the amplifier biasing will require the most of the losses, then losses to the heat and some to the rectifiers and very little to the cabling (hopefully) and so on...

You guys are also using quite complex methods to arrive to your conclusions, when you can make it really really simple. Im a simple guy and liking simple things.

To have 100w of music power (mono), you need a 100w powersupply! It is that simple - in reality...

First off all - music is not a constant sinusoidal signal and the most power consuming parts in the music signal are the basses. Everything else is recorded with 'lower volume'.
So how much there is bass in the signal, will in reality determine your power supply requirement.

And now to the fun part!
If you want 100w power, you will need the 100w to the BASSES! :-)
Remember everything else is recorded with lower volume...
Because 100W 2kHz 90dB/W will make you deaf or at least give you permanent hearing disability.

How much bass is there?
Depends on your liking. Lets assume you like some ridiculous tecno beat - is there 50% bass?

So are we already down to 50% of our powersupply power requirement - in reality?
:-)

Voltage calculation has been precented.
For current: I=P/U
That is the average.
And for Andrews transients we need a capasitor bank - a big one.

Many of you will now disagree and i accept that, because this is only my opinion based on my limited knowledge of the basic electro physics.
  Reply With Quote
Old 7th December 2012, 01:20 PM   #55
MarianB is offline MarianB  Romania
diyAudio Member
 
MarianB's Avatar
 
Join Date: Jul 2011
Location: Romania
Send a message via Yahoo to MarianB
I like simple things too but sometimes it's just not that simple, you cannot espect to get 100W of audio power from an 100W PSU, i think not even Class D power amp does not have anywhere near 100% eff. So you really need to take into account the losses when calculating the power of the supply, and @Mooly presented us with a verry usefull way to go ( cind of more detaliled/complete version of what i have presented ), it is not that complicated at all if you think about it, but it is the correct way to go.
__________________
Trance Session Live by Dj Marik
  Reply With Quote
Old 7th December 2012, 01:31 PM   #56
diyAudio Member
 
Join Date: Jun 2011
Quote:
Originally Posted by MarianB View Post
you really need to take into account the losses when calculating the power of the supply
You really need to take in to account the losses when you calculate the VOLTAGE of the supply - this was my point.
Therefore 100w music = 100w power supply.
  Reply With Quote
Old 7th December 2012, 02:08 PM   #57
MarianB is offline MarianB  Romania
diyAudio Member
 
MarianB's Avatar
 
Join Date: Jul 2011
Location: Romania
Send a message via Yahoo to MarianB
Well it is not like that ( and you do not seam to understand why ), maybe you are confusing switching with linear, well with linear power amp ( class B for example ) it is not enough to get the right voltage level, you need the correct amount of current as well, and it has already been presented a detailed way to do it, you cannot drop the current just because you have enough volts to get that 100W, cus the less curent you supply, the more volts will drop on full load, and you do have to think about tha too if going for the correct way to design the supply, look at @Mooly way, +/-36Vcc @2,2A ( witch ofcourse i would round it up to 3A ) gives about 150W of power from the supply, for 100W of power to the load, so again you are wrong and i hope u will understand why, sooner rather than later, it will save you poor designes.
__________________
Trance Session Live by Dj Marik

Last edited by MarianB; 7th December 2012 at 02:13 PM.
  Reply With Quote
Old 7th December 2012, 03:48 PM   #58
diyAudio Member
 
Join Date: Jun 2011
Yes. But:
P=U*I and real music in real life is not constant sinusoidal signal at full power CONSTANTLY, that is why 100w power supply for 100w music is more than you will ever need - in real life situations.

I.e. in this case 100w/(2*36v)=1.4A is what is needed...

You can measure the real life power consumption of the amplifier really easy...
You will be surprised how little it needs when the volume is fully cranked - it needs little... You should know this - because you are the man!

I find it complete waste of money to design for double the max power for a constant sinusoidal (test)signal - complete waste of money...

But this is diy - we can do what we each want. And I do find Andrews transients an important aspect in the design procedure - therefore money saved in the transformer goes to the bank - capasitor bank that is - many many many many of them.......... :-)

The one on my drawing board right now the bank holds 2500W of capasitor power!!!!
Amplifier is only 2*100W.
  Reply With Quote
Old 7th December 2012, 04:23 PM   #59
MarianB is offline MarianB  Romania
diyAudio Member
 
MarianB's Avatar
 
Join Date: Jul 2011
Location: Romania
Send a message via Yahoo to MarianB
Quote:
Originally Posted by palstanturhin View Post
I.e. in this case 100w/(2*36v)=1.4A is what is needed...
Whatever!... i do not have either the time nor the will to get on the right track every misguided, self convinced technician, who ever listens guys like you may verry well pay the price on weak designs, i cannot care less. on the other hand whoever really wants my help, and accepts my imput i will be more than glad to give it any time i can.
__________________
Trance Session Live by Dj Marik
  Reply With Quote
Old 7th December 2012, 04:47 PM   #60
Mooly is online now Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
Quote:
Originally Posted by MarianB View Post
Nice marathon
Thanks...
__________________
-------------------------------------------------------
Installing and using LTspice. From beginner to advanced.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Making Symmetric Power Supply from Single with 4x ATX power Supply? lupetalo Power Supplies 11 20th November 2012 11:51 AM
Listening preference.... Harlan Tubes / Valves 34 5th November 2011 05:58 AM
preference between 5842 and 6c45pi dsavitsk Tubes / Valves 0 8th December 2006 09:43 PM
Matlab or Mathematica -- which is your preference? rif Everything Else 7 11th May 2004 05:34 AM


New To Site? Need Help?

All times are GMT. The time now is 06:29 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2