[warnerwh]

make? I am trying to understand what the difference between one amp with 100wpc has with 40,000 to one with 80,000pfd's of capacitance is.  Does the higher capacitance mean a higher dynamic headroom? Also transformers seem to vary in size quite alot.  How do these two features affect the performance?  Thanks

[amplifierguru]

Hi warnerwh,

For a given size of amplifier, say your 100Wrms, doubling capacitance will halve the 100Hz ripple superimposed on the DC supply voltage. It could also halve supply impedance but this may be limited by wiring.

Neither of these things will necessarily improve the sound. A well designed amplifier has a high enough supply rejection built in to the amp cct proper to render it insensitive to these, such that PS becomes little more than an adequate resource.

My Simple Killer Amp (SKA) is a prime example of this where it requires a lesser, adequate, power supply resource to deliver it's renowned sonics, while others depend on a very overbuilt PS to help their amp design reach acceptability.

Dynamic headroom is more related to an amp that has a less 'stiff' (often smaller) power supply where the supplt DC sags under continuous rms power testing but delivers considerably more power under the transient, low duty, demands of music, which does not cause the DC to sag so much. Such an amplifier will deliver higher transient power from a usually smaller PS resource (read transformer mostly).

Hope that helps.

Cheers,
Greg

[marvda1]

isn't there a school of thought that smaller caps sound better than the large soda can size caps with their high values?

[amplifierguru]

Hi marvda,

Yes there is and I subscribe. The power supply module I make to pair with my SKA module is an example -

http://members.dodo.com.au/~gregball/guru_009.htm

The idea behind this is that smaller C's have their (lowest) impedance sustained to a higher frequency and 3 (for example) in parallel reduces impedance 3 times for all frequencies. They also have size and cost advantages as well as distributed heating in the ESR so possibly longer life.

To take this heating to a logical conclusion, for best lifetime (an inverse function of temp) the capacitors should be stripped of their plastic sheath rendering them nude for better heat dissipation.  

Cheers,
greg

[_scotty_]

Of course it helps to remember that the metal shell is live and if it touches the chassis you have a potential short to ground.  amplifierguru, if I understand you correctly your amplifier's circuit speed is not particularly dependent on
how low the impedance of the power supply is. This is a good trick if I have
understood you correctly.
Scotty

[amplifierguru]

Hi scotty,

Yes that's correct. Once the amplifier becomes highly independant of PS quality, it can perform well with almost any PS and the PS simply becomes a resource that is either adequate or not.

For examply, my GB150D is capable of 150W into 8 ohms if the PS can maintain +/-52V at 100W draw. This can be a +/-53V huge overbuilt PS or it can be a 160VA +/-60V supply that sags considerably under draw.
The latter will have dynamic headroom to over 200Wrms on music as average demand will not drag down the PS, allowing higher peaks. So the latter with the underbuilt PS resource is the more powerful amplifier, on audition.

At lower cost. Win,win.

Cheers,
Greg

[warnerwh]

amlifierguru: Thank you very much. This has been the answer I've been looking for.

This leads me to another question.  Why not build amplifiers with the slightly weaker power supply if they will deliver more power under real world loads? I've read who knows how many times that a stiff power supply is the most important part of an amp. Now I am wondering why.  Maybe I should take an electronics class.

Part of the reason I was wondering these questions is because of the Blue Circle amp I have. It's rated at 80wpc into 8 ohms and only 100wpc into 4 ohms with 3db dynamic headroom.

When I put in a fresh pair of tubes I noticed that there's 6 15,000pfd caps in there.  That's as much as many amps with 200wpc.  The Rotel 991 I have running my woofers only has 60,000.  

When it comes to sound the Blue Circle amp is easily the best sounding amp I've ever owned so I really don't care what is what other than curiosity.  It probably doesn't measure as well as the Rotel but if I only could have one it would be the little BC 24.

[NealH]

An amplifier not affected by  power supply droop will likely employ gobs of  feedback in the output stage.  With gobs of feedback transient modulation distortion rears its ugly head.  As well there can be more of the high order distortion artifacts which are not desirable.   Dynamic headroom will be short lived too due to the lower input capacitance.  

I would pass on this type amp.


An amplifier which provides a stable and consistent power supply rail (lots of capacitance or LC) allows the designer to employ less and, more basic componentry especially in the critical output stage.  This generally leads to a more organic and fluid sound character.  Also, it can maintain it's dynamic headroom for a longer period of time so, it won't whimp out when the going gets tough.  

I would pursue this type amp.

[Dan Banquer]

Hi warnerwh,

For a given size of amplifier, say your 100Wrms, doubling capacitance will halve the 100Hz ripple superimposed on the DC supply voltage. It could also halve supply impedance but this may be limited by wiring.

Neither of these things will necessarily improve the sound. A well designed amplifier has a high enough supply rejection built in to the amp cct proper to render it insensitive to these, such that PS becomes little more than an adequate resource.

My Simple Killer Amp (SKA) is a  ...

I think there is another issue to be considered here, and that is with more capacitance you will better regulation of the supply. As with most of audio and the post quoted above the concept of better regulation is dismissed.
I beg to differ.
When considering better regulation one must look at the gain bandwidth of the transistors in the amp and how that relates to supply voltage. If one reads the transistor data books carefully you will find a spec for gain bandwidth at a certain supply voltage. That spec holds for that supply voltage but not other supply voltages within x amount of volts.
Now that initially sounds like a lot of techno babble but consider the following: An unregulated supply's voltage will vary under load so by definition at higher loads the voltage across the amp will sag. We also notice that when we do distortion measurements we find that the distortion increases as the load increases for amps with unregulated supplies. Now lets add some real world conditions and put the amp under reactive load (like a loudspeaker) and we are now faced with voltage and current that are out of phase at the output transistors.
The distortion can go even higher under these circumstances because the feedback that is applied to the amp was based on a relatively stable supply voltage. If that supply voltage goes outside the parameters set by the feedback loop then the gain bandwidth has changed due to supply voltage changing and the feedback loop is no longer "correct".
In short: Change that supply voltage too much and your feedback loop is not working for the initial parameters it was set for and you can get anything from instability to upper order distortion.
There is a penalty for the lack of regulation, and it goes right back to classic analog design rules.
Hope this helps:
         d.b.

[amplifierguru]

Hi rnhood,

An amplifier with good power supply rejection (PSRR) does not need to have high feedback or have significant output stage gain. In fact, the best place to have gain is in the input stage, before the signal is alternately referenced to the supplies where intrusion begins. It is considerably compromised by the common high gain intermediate voltage amplifier stage (Vas) of conventional design topology, as it simply amplifies the intrusion at it's input due to the non-infinite Zout of the input stage. This is why I dispensed with the Vas stage completely in my SKA topology putting most gain into the first stage (80dB) and a small amount  (30dB) in the output stage -

http://members.dodo.com.au/~gregball/guru_003.htm

TIM has been thoroughly debunked by Robert Cordell and is simply Slewing Induced Distortion (SID) and does not occur in amps with adequate slewing rate. I published an article in Wireless World in Dec 1990 entitled "Distorting Power Supplies" for more on the subject of PSRR.

Hi Dan,

My,and most, feedback networks are not supply sensitive. FB works as advertised largely independant of supply bounce.

Warnerwb's statement -

..."Part of the reason I was wondering these questions is because of the Blue Circle amp I have. It's rated at 80wpc into 8 ohms and only 100wpc into 4 ohms with 3db dynamic headroom....When it comes to sound the Blue Circle amp is easily the best sounding amp I've ever owned..."

Says heaps.

cheers,
Greg

[Greg Erskine]

hi warnerwh,

Can I make a few novice observations. I have been building quite a few 50-150w DIY amps and I try to build them to the designers specifications, including the PSU. So far, in the 50-100W range the usual recommended capacitance is 20,000uF (some say a max of 40,000uF). So far, all my discrete amps have 20,000uF except for the SKA 150 which has 56,000uF. Ironic that the SKA has nearly 3 times tha capacitance of my other amps.  

It is also my understanding that as you increase capacitance, you increase the initial current through your bridge when you turn on the amp. So if you significantly increase you capacitance, ensure the bridge is up to the job.

BTW, a lot of Gainclones somehow get away with only 2,000uF or 3,000uF.

regards

[Dan Banquer]

"My,and most, feedback networks are not supply sensitive. FB works as advertised largely independant of supply bounce. "

Then your not testing very thoroughly, because if you did you would come to the realization that the classic rules of analog still apply and J.L. Linsley Hood was far more knowledgeable than most people realize.
                           d.b.

[amplifierguru]

Pray tell, Dan. I'm all Ears. On any level.

IMO stiff and particularly regulated power supplies are generally overkill, often used by manufacturers as a cover up of an otherwise poor design, which, for a smaller outlay and better, informed, conceptual design of the amplifier proper,  could be eliminated.

*************edited*********.

Cheers,
greg

[Dan Banquer]

"Pray tell, Dan. I'm all Ears. On any level.

IMO stiff and particularly regulated power supplies are generally overkill, often used by manufacturers as a cover up of an otherwise poor design, which, for a smaller outlay and better, informed, conceptual design of the amplifier proper, could be eliminated.

I have proven this time and again with a variety of giant killer products. "

You obviously have not read any of  the articles and books from Mr. Hood. I consider that rather interesting since, if I remember correctly, you claim to have wrote for Electronics and Wireless World where Mr. Hoods' article were featured extensively.  His amp designs were not exactly poor designs by any stretch of the imagination, the only exception being....................
               d.b.

[amplifierguru]

Hi Dan,

I read Mr Hood's early articles in the '70's when I was teaching at Uni. era. He lost me when he came up with a (from memory) 270W amp using one pair of darlingtons! No SOA to speak of.

Hi Greg #1,

So you've finished your SKA then?

Cheers,
Greg

[Dan Banquer]

Thank you for your reply Greg; I've learned what I needed to learn.
                            d.b.
P.S. You should have read his articles in the 80's and 90's.