Hi,
Doing the above helps in some areas, however it does not address the root cause, but merely (possibly) mitigates against some of the symptoms, like slowing down the current spikes when the reservoir capacitors are being charged (most extreme in the use of choke input supplies often found in Tube gear but never in solid state gear) and constant current operation. Such power supply work as extra filtering and constant current operation can however help addressing other issues inherent to the circuitry.
To deal with the sensitivity of equipment to power (and other) cable(s) differences we need to employ a different discipline and we must look at the whole interconnected system, including interconnecting cables, common mode and differential mode impedances, balance of the cables carrying power for differential noise and so on.
It is a problem that remains completely invisible on the test bench testing an individual item of equipment and test gear, simply because decent test gear must, by the force of requirements to measure precisely constructed in a way to minimise these effects regardless of the nature of the device under test.
Take the same device into a real system with multiple sources, pre- and poweramp and maybe powered subwoofers and a raft of problems "suddenly appear" that are tough to solve while continuing to comply with electrical codes and that have a lot to do with observed differences between cabled and nothing whatsoever with the direct design of power supplies.
Ciao T
Doing the above helps in some areas, however it does not address the root cause, but merely (possibly) mitigates against some of the symptoms, like slowing down the current spikes when the reservoir capacitors are being charged (most extreme in the use of choke input supplies often found in Tube gear but never in solid state gear) and constant current operation. Such power supply work as extra filtering and constant current operation can however help addressing other issues inherent to the circuitry.
To deal with the sensitivity of equipment to power (and other) cable(s) differences we need to employ a different discipline and we must look at the whole interconnected system, including interconnecting cables, common mode and differential mode impedances, balance of the cables carrying power for differential noise and so on.
It is a problem that remains completely invisible on the test bench testing an individual item of equipment and test gear, simply because decent test gear must, by the force of requirements to measure precisely constructed in a way to minimise these effects regardless of the nature of the device under test.
Take the same device into a real system with multiple sources, pre- and poweramp and maybe powered subwoofers and a raft of problems "suddenly appear" that are tough to solve while continuing to comply with electrical codes and that have a lot to do with observed differences between cabled and nothing whatsoever with the direct design of power supplies.
Ciao T
'Properly designed' means 'fit for purpose' right? So while I agree that from an engineering perspective making an amp immune to power cords is an excellent aim. But the purpose of most amps is to be sold, not just perform according to engineering specs. To get an amp sold it has to be attractive to a dealer and the dealer makes a much bigger margin on cables than he does on amps. ISTM therefore that an amp which is immune to cables won't get sold through any kind of dealer because its less attractive a financial proposition than one which sounds better when partnered with a $1,000 mains lead. Am I missing something?
Vlad, many people here have little experience, and strong opinions as to what SHOULD be important.
Wavebourn, I was designing recording consoles 40 years ago. Please do not think that you are the only person who has faced these problems. How I fixed consoles was to run them from professionally made lab power supplies, and individually buffer each discrete op amp with its own pair of capacitor multipliers. In the old days, I did it with bipolar transistors. Today, I do it with either jfets or mosfets. In any case, even the JC-3 phono preamp, one of my more modest efforts, is double or triple power supply protected, AND for good measure I stuck a 10uf film capacitor across the power supply line where the line cord connects. I try to not leave anything to chance. AND it has worked for me over the years.
For everyone, please empty a little of your 'head' to input new data. Even a PhD does not teach you everything that you need to know. By the way, I hired my first PhD EE in 1974, quite a few years ago. I know when what I need is beyond my educational level, but this does not eliminate EXPERIENCE and feedback from you and your associates' listening impressions. That is critical for successful audio design.
I am surprised that nobody has mentioned circuit breakers, instead of fuses.
Wavebourn, I was designing recording consoles 40 years ago. Please do not think that you are the only person who has faced these problems. How I fixed consoles was to run them from professionally made lab power supplies, and individually buffer each discrete op amp with its own pair of capacitor multipliers. In the old days, I did it with bipolar transistors. Today, I do it with either jfets or mosfets. In any case, even the JC-3 phono preamp, one of my more modest efforts, is double or triple power supply protected, AND for good measure I stuck a 10uf film capacitor across the power supply line where the line cord connects. I try to not leave anything to chance. AND it has worked for me over the years.
For everyone, please empty a little of your 'head' to input new data. Even a PhD does not teach you everything that you need to know. By the way, I hired my first PhD EE in 1974, quite a few years ago. I know when what I need is beyond my educational level, but this does not eliminate EXPERIENCE and feedback from you and your associates' listening impressions. That is critical for successful audio design.
I am surprised that nobody has mentioned circuit breakers, instead of fuses.
By referring to high-power class AB amps, I imply that it is almost impossible to preserve good dynamics of such amps with extensive use of LC and RC filtering in PS. But with standard simple C filter, this C and power cord become included in the signal chain, as well as speaker cables. Even the circuit breaker at house switchboard play a role.
My friend did special reconstruction of power line for audio system, by laying special quality 6mm2 power cable up to transformer substation, and claims it adds to dynamics.
I have choosen another way, low power SE amps with diminishing the role of PS as much as possible, up to 7 RC sections at various places (aimed at HF noise, not at 100Hz ripples). This would immediately kill quality of AB amp, but works excellently for definite kinds of SE schematics.
My friend did special reconstruction of power line for audio system, by laying special quality 6mm2 power cable up to transformer substation, and claims it adds to dynamics.
I have choosen another way, low power SE amps with diminishing the role of PS as much as possible, up to 7 RC sections at various places (aimed at HF noise, not at 100Hz ripples). This would immediately kill quality of AB amp, but works excellently for definite kinds of SE schematics.
Last edited:
John,
I'm afraid this explanation, while reasonable, seems in need of rejection as hypothesis.
In AMR gear we eliminate all fuses except those mandatory for electrical/fire safety by design. Initially we used to have more, but we instead went to self protecting transformers (as we have several in parallel and in theory would need one fuse per transformer).
Now here is the cracker, in the case of the CD-Players all power supplies that feed something variable are CCS fed Shunts and all analogue stages are Class A SE as well (and Tube rectified to boot on the big one).
Incidentally, I use these CCS fed Shunts almost religiously - they do help to keep signal/noise currents contained into controllable loops without having to cut ground planes, offer defined low impedances into the 100's of KHz, keep an absolutely constant load on the power supply transformers and even cure cancer, okay, they don't do the last thing, but they do do all the others...
So there is clearly no variation in current draw whatsoever, no matter what the device does. Yet the one single fuse in the IEC Mains Inlet which must remain has audible effects. I am severely stomped at any explanations and "They are non-linear in operation, and all the current has to go through them." really does not cut it for this case. I am rather puzzled by the why...
Ciao T
I'm afraid this explanation, while reasonable, seems in need of rejection as hypothesis.
In AMR gear we eliminate all fuses except those mandatory for electrical/fire safety by design. Initially we used to have more, but we instead went to self protecting transformers (as we have several in parallel and in theory would need one fuse per transformer).
Now here is the cracker, in the case of the CD-Players all power supplies that feed something variable are CCS fed Shunts and all analogue stages are Class A SE as well (and Tube rectified to boot on the big one).
Incidentally, I use these CCS fed Shunts almost religiously - they do help to keep signal/noise currents contained into controllable loops without having to cut ground planes, offer defined low impedances into the 100's of KHz, keep an absolutely constant load on the power supply transformers and even cure cancer, okay, they don't do the last thing, but they do do all the others...
So there is clearly no variation in current draw whatsoever, no matter what the device does. Yet the one single fuse in the IEC Mains Inlet which must remain has audible effects. I am severely stomped at any explanations and "They are non-linear in operation, and all the current has to go through them." really does not cut it for this case. I am rather puzzled by the why...
Ciao T
Because that's the sort of thing that marketing a five figure hifi item demands. "Special" is important. Otherwise, what's the purchaser buying? This is performance art, not performance.
Dear Vladimir,
Depends, with LC filtering it is most certainly possible. Common "massive" supplies have around 0.5 Ohm DCR (measure by loading the supply, see how much the PSU Voltage drops), so adding 0.1 Ohm like with Manganese/Ferrite cored chokes is not really cramping the Amp's style too much.
It is also possible to use active circuits in the supply to push the power supply impedance into the tens of milliohms way past up to many 100Khz and that without resorting to tons of looped feedback in the power supplies either. In fact, you may find that the output impedance of the Powersupply can be such that it acts as "expander", reducing it's impedance as more current is drawn (one has to watch this effect, it can easily run away on you)...
The role is not direct. There is a rectifier in between and we apply substantial
averaging to the signal before it hits the mains. But yes, with the right amp and speaker and old wiring in the house it is entirely possible to have the lights dim with every big "paukenschlag", a very graphic illustration of this...
Ciao T
Depends, with LC filtering it is most certainly possible. Common "massive" supplies have around 0.5 Ohm DCR (measure by loading the supply, see how much the PSU Voltage drops), so adding 0.1 Ohm like with Manganese/Ferrite cored chokes is not really cramping the Amp's style too much.
It is also possible to use active circuits in the supply to push the power supply impedance into the tens of milliohms way past up to many 100Khz and that without resorting to tons of looped feedback in the power supplies either. In fact, you may find that the output impedance of the Powersupply can be such that it acts as "expander", reducing it's impedance as more current is drawn (one has to watch this effect, it can easily run away on you)...
The role is not direct. There is a rectifier in between and we apply substantial
averaging to the signal before it hits the mains. But yes, with the right amp and speaker and old wiring in the house it is entirely possible to have the lights dim with every big "paukenschlag", a very graphic illustration of this...
Ciao T
Hi,
You are missing that making amplifiers that are immune to cable effects AND can be legally sold, meaning they have to fulfil certain electrical and other safety requirements (which often mandate things that run counter to the requirements for good sound quality) is while not completely impossible (not much is impossible in this universe) is not a trivial matter (to say the least), unless we use transformer coupling and battery power (which in itself is non-trivial)...
Ciao T
You are missing that making amplifiers that are immune to cable effects AND can be legally sold, meaning they have to fulfil certain electrical and other safety requirements (which often mandate things that run counter to the requirements for good sound quality) is while not completely impossible (not much is impossible in this universe) is not a trivial matter (to say the least), unless we use transformer coupling and battery power (which in itself is non-trivial)...
Ciao T
Nope, I'm fully aware that its not trivial. But where's the incentive to try?
Last edited:
I am afraid I disagree with you at this point. So called choke-input power supplies inevitably make amps intended for jazz and vocal, one should forget about punch and drive with rock music. This conclusion is now an axiom among russian audiophiles (equipment users), and from technical point of view, I can not imagine how amp could pump 60A peak current to load during high-level bass notes (maybe many Farads lytics reservoir could help) Active resistance of the choke plays almost no role, but its L affects charging pulses drastically.
Nelson Pass repeated maybe hundred times, that amps are intended for applications. The one best for rock will never be best for jazz.
I am lucky, since any rock is not my kind of music, and I can implement "slow" and high Zout power supplies, but extremly cleaned from any HF noise.
Last edited:
Thanks.
Please send me a PM. I tried to email you and got the same support page. Your PM box is full.
- Status
- Not open for further replies.