AND: Using the engineering equation: dV/dt = I/C Let's see how long it can maintain its voltage: Let's say, dV is a change of 1V. Let's say, C=58F, then 1X58 =IXT(in seconds) Then you will droop 1V after 58 seconds with a 1A draw. Not too long. See the problem Let's say instead that we use a battery with 1A-hour capacity. Then it will be equivalent to about 3600F. or about 22 times better.
Jan, I have played around with a couple of different EDL caps and have some left which you may have after returning from your trip to the sun. ESR at DC is pretty low and ESL I don't expect to be much different from Elco's, looking at their construction. I decided not to use them because they self discharge quite quickly.
On batteries in general: regulators like the Jung/Didden or the Sjostrom are already pretty noise free, and if noise were a problem, there is a trick to get it even further down. Also, output impedance of this kind of regulator is pretty good in the whole audio band. The limiting factor here is the wiring between the PS and the current consumers. The impedance of this wiring pushes up the ripple on the supply rails under fluctuating load. There is no way an even better PS or batteries could help out here. The only solution is to put the regulation as close to the current consumers as possible, or, of course, use decoupling caps with small leads near the point of consumption.
On batteries in general: regulators like the Jung/Didden or the Sjostrom are already pretty noise free, and if noise were a problem, there is a trick to get it even further down. Also, output impedance of this kind of regulator is pretty good in the whole audio band. The limiting factor here is the wiring between the PS and the current consumers. The impedance of this wiring pushes up the ripple on the supply rails under fluctuating load. There is no way an even better PS or batteries could help out here. The only solution is to put the regulation as close to the current consumers as possible, or, of course, use decoupling caps with small leads near the point of consumption.
Last edited:
going by the datasheet 1000000 cycles.
17.8Wh at fully charged
10years life at rated voltage at room temp (25C)
max peak current 2000A for 1sec
max continuous current at 15C 100A
max cont. current at 40C 160A
ESR ~2.1mΩ
^^ for the 500F/16V
given that I dont think in audio applications we would expect much less than the 10 year life as long as the chassis was built with that in mind.
you would never use batteries on noise alone, floating is a bonus for some circuits and again, why does everyone automatically compare a battery, completely by itself with no filtering to a super regulator? wouldnt a fairer comparison be a battery to a rectifier or at best a raw DC supply?
17.8Wh at fully charged
10years life at rated voltage at room temp (25C)
max peak current 2000A for 1sec
max continuous current at 15C 100A
max cont. current at 40C 160A
ESR ~2.1mΩ
^^ for the 500F/16V
given that I dont think in audio applications we would expect much less than the 10 year life as long as the chassis was built with that in mind.
you would never use batteries on noise alone, floating is a bonus for some circuits and again, why does everyone automatically compare a battery, completely by itself with no filtering to a super regulator? wouldnt a fairer comparison be a battery to a rectifier or at best a raw DC supply?
Last edited:
Study i did before the invention of super caps was that for fast transients, standard batteries (carbon, mercury, NiCad etc) didnt do well. Must be cap bypassed for speed (and to maintain low Z after discharge time). DC esr was great fully charged.
Last edited:
Reason why care have to be kept inside the amp itself with good fast current sources and good voltage filtering near the low power stages, where it is easy and more sensible, specially in current feedback topologies. That is often done in OPAs, where good CFA devices are not so bad rated, regarding PSRR. After that, and with some precautions, it is not so difficult to provide BIG rails with good enough noises. That, for the numbers.
But little changes of signatures are still audible depending of various power supplies topologies, mostly concerning dynamic behaviors during transients. I tend to prefer simple emitter follower stabilization VS feedback regulations. May-be some conflicts between the power supply servo delays and the amps feedback ones. Anyway, care had to be taken, at any steps, to get high speed regulations/filtrations.
For preamp, no real issues, you can find integrated circuits able to provide <15µv rms of noise, like TPS7A30/ TPS7A49. (~ -120dB)
Last edited:
Phased array loud speakers have been around for decades.
See the text "Sound system engineering by Don & Carolyn Davis"
( 4th edition soon to be released) http://www.amazon.com/Sound-System-...4501&sr=8-6&keywords=sound+system+engineering
See the text "Sound system engineering by Don & Carolyn Davis"
( 4th edition soon to be released) http://www.amazon.com/Sound-System-...4501&sr=8-6&keywords=sound+system+engineering
Last edited:
Exactly! The only way to make those regs 'better' for the load is either using separate regs at the load point, or remote sensing at the sensitive load point.
The Zout of these babies was about the equivalent of an inch of wiring.
Remote sensing was used in the regs we worked on in 1994, see Home.
jan
or just use sufficiently thick wire and decoupling for the load current in the first place...
nothing wrong with regs, I use mostly regulated supplies, but I just find most of the performance stuff mentioned with regs vs batteries is kinda irrelevant for most cases and if it isnt, regulate the battery. why does AC get a free super regulator for the purpose of comparison?
nothing wrong with regs, I use mostly regulated supplies, but I just find most of the performance stuff mentioned with regs vs batteries is kinda irrelevant for most cases and if it isnt, regulate the battery. why does AC get a free super regulator for the purpose of comparison?
Last edited:
Jan, i know you're involved in this fabulous 1.5µv regulator design. May-i ask-you an objective answer ? Did-you find any audible advantage in a preamplifier with the complexity involved for any PSU noise improvement under -120db ?
Of course, for the +5V analog reference of a DAC, it is an other issue.
It is a real question because i never tested that, and i suppose you did.
Of course, for the +5V analog reference of a DAC, it is an other issue.
It is a real question because i never tested that, and i suppose you did.
Please share how to lower the noise of those regulators.
I've seen one such suggestion by having 3 reference (zener) diodes in parallel, though I doubt how effective it may be, since the noise of the reference diode(s) is filtered by an electrolytic cap. According to Jung, the noise of the regulator is being governed by the noise of the error OpAmp, so I wonder how the noise can be lowered, except by choosing a quieter OpAmp, like OPA211 or AD8597 or LME49990.
It is not so much the resistance but rather the inductance of the wiring that limits the output resistance of the supply as seen by the load, thickening the wires doesn't help. Good layout and local decoupling is more important.
I didn't really do any controlled tests for that. To do that you also would have to define with what you should compare it?
The design exercise was to get the best we could do with a reasonable cost.
If you want to know audible results, John Walton and the New Jersey Audio Society did test 13 different regs in Linear Audio Vol 4, measurements as well as controlled listening tests. He used as 'load' a Borbely-type preamp known for not-so-hot PSRR. He reports clear audible differences, with the Jung/Didden (ahem) one of the best.
So, you could say that almost 20 years after, the soundness of the design was confirmed.
But, if you have a preamp with 120dB PSRR the effect might be much less or even non-existent. Everything is relative (at least since 1915).
jan
Thanks for your answer, Jan. Was the transformer the same ?
I ask that because AC leakage between grounds which pollute non symmetrical signals seems to me more a problem than little noises on rails.
Any integrated circuit they found among the best ones ? Are their preferences directly correlated with noise levels ?
Last edited:
L'Audiophile n°29, septembre 1983 Super capacitors in French of course
"The internal resistance varied widely and measured a low 155 mOhm for nickel-cadmium, a high 778 mOhm for nickel-metal-hydride and a moderate 320 mOhm for lithium-ion. These internal resistance readings are typical of aging batteries with these chemistries."
how does internal resistance affect performance
"The internal resistance varied widely and measured a low 155 mOhm for nickel-cadmium, a high 778 mOhm for nickel-metal-hydride and a moderate 320 mOhm for lithium-ion. These internal resistance readings are typical of aging batteries with these chemistries."
how does internal resistance affect performance
Attachments
Here are some of the measurement results.
The general ranking they found is attached. But you should really get the whole article 😉
jan
The general ranking they found is attached. But you should really get the whole article 😉
jan
Attachments
Christophe,
When you are stating the -120db noise rating is that an "A" weighted average or is this unweighted?
When you are stating the -120db noise rating is that an "A" weighted average or is this unweighted?
Is it really such a mystery as to a possible mechanism for sound changing via equipment support? A conductor inside a magnetic field, say a circuit trace in a pre amp moves in relation to the field, what do you think happens?
Audibilty of this may be an issue, but I have found changing equipment support changes the sound, sometimes not for the better, but sometimes in important positive ways.
- Status
- Not open for further replies.