Battery-powered supplies

[Steve Eddy]

Originally posted by ALW
'Cos they are very noisy, have very limited bandwidth, and rubbish impedance.

How much/what kind of noise, what's the bandwidth and what's the impedance?

se

[ALW]

How much/what kind of noise, what's the bandwidth and what's the impedance?

Dynamic load-related noise, and significant chemical noise from the process going on inside the battery.

Just try looking with a scope - the output is horrible under load. The noise lasts long beyond the initial stimulus too, as the reaction carries on.

Bandwidth is very narrow.

Impedance from the data sheets - usually 10's of milliohms, much higher than a well-designed conventional supply, leading to some of the noise problems above...

Couple the above with the impracticality of battery PSU's (which usually need mains power somewhere for charging) I fail to understand the appeal. When I listen, I feel the appeal drops further

Andy.

[Elso Kwak]

Originally posted by ALW


Dynamic load-related noise, and significant chemical noise from the process going on inside the battery.

Just try looking with a scope - the output is horrible under load. The noise lasts long beyond the initial stimulus too, as the reaction carries on.

Bandwidth is very narrow.

Impedance from the data sheets - usually 10's of milliohms, much higher than a well-designed conventional supply, leading to some of the noise problems above...

Couple the above with the impracticality of battery PSU's (which usually need mains power somewhere for charging) I fail to understand the appeal. When I listen, I feel the appeal drops further

Andy.

Hi Andy, Very interesting observation!
I noted the Mark-Levinson JC1-DC cartidge preamplifier had some sort of stabilisation.(battery powered )
The JC1-AC (AC-powered) had similar circuit and sounded a lot better but was plagued by some hum problems.

[Steve Eddy]

Originally posted by ALW
Dynamic load-related noise, and significant chemical noise from the process going on inside the battery.

Just try looking with a scope - the output is horrible under load. The noise lasts long beyond the initial stimulus too, as the reaction carries on.

I guess it's safe to assume you wouldn't touch a tube amp with a 10-foot pole.

Fortuantely, I'm not an oscilloscope. And the only thing that matters to me is the end result.

But in any case, the noise is quite benign, being random in nature as opposed to the periodic nature of power supply ripple and its assocated harmonics. Also, the bulk of the energy is in the low frequency range and falls off at a rate greater than 1/f. It's effectively eliminated from the output by virtue of the power supply rejection of the circuit.

Bandwidth is very narrow.

You said that already.

I asked you what the bandwidth is.

Impedance from the data sheets - usually 10's of milliohms, much higher than a well-designed conventional supply, leading to some of the noise problems above...

I haven't found the noise to be any problem. Perhaps it's only a problem to those looking at it on an oscilloscope.

Couple the above with the impracticality of battery PSU's (which usually need mains power somewhere for charging) I fail to understand the appeal.

Well, battery supplies can be rather impractical if you're using speakers that are inefficient and require large amounts of power to achieve sufficient output levels. But this doesn't describe all loudspeakers.

And I could care less if the mains were used for charging. I just don't want it powering my amps.

When I listen, I feel the appeal drops further

Listening is where I find the most appeal. It's the only thing that matters for me. Again, I'm not an oscilloscope.

se

[A 8]

Originally posted by Steve Eddy


Ok.



Why? If it's not reproducing anything up to 100kHz and above, why would you care what the ESR is up there?

Since when don't we want low and linear esr within the power supply for the full bandwith of the power amp?

Besides the obvious issue of feedback, ringing etc there are for example harminics created by the delivery chain and/or if you run a dac with little or no analog filtering....

Also, by the time you get up to 100kHz, ESL begins to swamp out ESR anyway.

eh, Why? 100kHz is not a big problem from an esl perspective.



That was reproducing 100kHz and above?

Don't think I need to answer that. My point was that there is a big difference running a op amp with 10Kohms or so load or running a pair of 3 ohm speakers....ie a differense of whatever in esr will be a much bigger part of 3 ohm then 10kohms.


I've never tried it with a gob of capacitance. Only with and without about 22uF. Didn't seem to make any difference but I kept it in so I don't have to worry about lead inductance if I ever need to lengthen the power feed lines.

Try film caps!



It's an important metric on caps largely because it determines how much heating you'll get and ultimately sets the maximum ripple current the capacitor can handle safely.

Really....don't think that is the only reason.
In many applications esr over frequency is way more important then handling current.
se

[ALW]

Listening is where I find the most appeal. It's the only thing that matters for me. Again, I'm not an oscilloscope.

You must have heard some truly awful PSU's - batteries sound terrible to me...

Andy.

[Steve Eddy]

Originally posted by ALW
You must have heard some truly awful PSU's - batteries sound terrible to me...

Key phrase being "to me." I could care less how they sound to anyone else. I don't listen for anyone else's pleasure but my own.

If you don't care for battery power supplies, it's no skin off my nose. So what exactly is your point?

se

[Steve Eddy]

Why do you write your replies within the quote box? The whole point of the quote box is to separate quotes from responses to those quotes.

And when I reply to your message, the system strips out everything in the quote boxes so I have to go keep scrolling down to your post and copy/paste your responses in here.

Since when don't we want low and linear esr within the power supply for the full bandwith of the power amp?

When the amp isn't being called upon to reproduce anything at 100kHz. If the amp's not drawing current at those frequencies, then the ESR at those frequencies isn't going to have any relevance.

Besides the obvious issue of feedback, ringing etc there are for example harminics created by the delivery chain and/or if you run a dac with little or no analog filtering....

In which case the amp will be drawing current at those frequencies and now we're into a completely different context than originally.

eh, Why? 100kHz is not a big problem from an esl perspective.

It can be more of a problem than ESR because it's ESL which causes the cap's impedance to start rising at high frequencies. Depending on the cap, ESL can begin to overtake ESR well in the audio band.

Don't think I need to answer that. My point was that there is a big difference running a op amp with 10Kohms or so load or running a pair of 3 ohm speakers....ie a differense of whatever in esr will be a much bigger part of 3 ohm then 10kohms.

And my point is that if you're not drawing current at 100kHz, the ESR of the cap at 100kHz isn't going to matter to the amplifier.

Try film caps!

I have. But currently I prefer the 22uF OS-CONS that I've been using.

[QUOTE]Really....don't think that is the only reason.[QUOTE]

Where did I say it was the only reason?

se

[ALW]

I asked you what the bandwidth is.

Sorry missed that one - it matters not one jot what the bandwidth is specifically, as in the case of batteries it's so spectacularly low, relative to the audio band, that's the dominant factor.

In fact it's also difficult to measure, since the other appaling performance factors of the battery get in the way and dominate once current is drawn. Your conclusions re: the 1/f status of the noise profile is unfortunately not true except for the steady state case - I have some FFT plots somewhere should you be interested.

I cannot, unfortunately, tell you the real reason why battery supplies are totally unsuitable for audio use, due to some current work I'm doing - but all will become clear in the fullness of time.

Since though I have tried all sorts of battery supplies and all manner of linear ones my experiences are based solely on the sonic effects, not the measured ones.

What topologies have you compared to battery PSU's?

What do you listen for as an improvement?

The one thing I can assure you is that, apart from basic functional measurements, I listen first, then measure later in order to try an analyse what I'm hearing, in order to be able to progress the designs further.

Key phrase being "to me." I could care less how they sound to anyone else. I don't listen for anyone else's pleasure but my own.
If you don't care for battery power supplies, it's no skin off my nose. So what exactly is your point

In the interests of this shared community, where I thought the whole idea was to pass on experience to others so that we can all learn something, I was trying to get across the fact that to my ears, and everyone else to whom I've demonstrated the effects, battery PSU's suck big time, compared to a less costly, far quieter, wider bandwidth, easier to use, linear PSU. Not dogma, just an experience.

If you like them, that's no problem to me either, but as Iasked above, what have you compared them to - that might tell us a lot about why you didn't like the mains-based approach.

What's important to you in an audio supply?

Yours in audio harmony,

Andy.

[Steve Eddy]

Originally posted by ALW
Sorry missed that one - it matters not one jot what the bandwidth is specifically, as in the case of batteries it's so spectacularly low, relative to the audio band, that's the dominant factor.

Wow. Can't even make it out to 20kHz, huh?

In fact it's also difficult to measure, since the other appaling performance factors of the battery get in the way and dominate once current is drawn. Your conclusions re: the 1/f status of the noise profile is unfortunately not true except for the steady state case - I have some FFT plots somewhere should you be interested.

Yes, let's see those FFT plots.

I cannot, unfortunately, tell you the real reason why battery supplies are totally unsuitable for audio use, due to some current work I'm doing - but all will become clear in the fullness of time.

Obviously battery supplies are not totally unsuitable for audio use as many are in fact using them.

Sounds like you've got some ulterior motives going here.

Since though I have tried all sorts of battery supplies and all manner of linear ones my experiences are based solely on the sonic effects, not the measured ones.

"Sonic effects" are in the mind of the beholder. Or are you attempting to assert your subjective preferences as somehow being inherently superior those of others?

What topologies have you compared to battery PSU's?

Cap input (both regulated and unregulated), choke regulated as well as PWM switching supplies back when I was dabbling in car audio.

What do you listen for as an improvement?

I don't listen for anything in particular. I just go by the gestalt experience.

The one thing I can assure you is that, apart from basic functional measurements, I listen first, then measure later in order to try an analyse what I'm hearing, in order to be able to progress the designs further.

How exactly do you go about determining that your subjective perceptions are in fact due to actual audible stimulus? Or are you working from the erroneous assumption that your subjective perceptions are always the unerring result of some actual phyiscal reality produced by the equipment?

In the interests of this shared community, where I thought the whole idea was to pass on experience to others so that we can all learn something, I was trying to get across the fact that to my ears, and everyone else to whom I've demonstrated the effects, battery PSU's suck big time, compared to a less costly, far quieter, wider bandwidth, easier to use, linear PSU. Not dogma, just an experience.

Sharing one's experiences are fine. To denegrate the experiences of others, not so fine. At least with respect to subjective issues.

If you like them, that's no problem to me either, but as Iasked above, what have you compared them to - that might tell us a lot about why you didn't like the mains-based approach.

What exactly would it tell us?

What's important to you in an audio supply?

The same thing that's important to me in all the rest of the system. The enjoyment of music.

se