In a system which consists of equipment with elaborate power supplies featuring high Common-Mode (CM) impedance to the AC line over a wide range of audio frequencies, balanced signal transport between chassis and does not have pin 1 design issues, a power conditioner could potentially (insert pun here) be of minimal value.
The primary component of a proper power conditioner is an isolation transformer. The inductance of a 50 or 60Hz transformer makes it a natural LP filter and a poor transformer for RF, but capacitive coupling between primary and secondary can couple high frequencies. The best isolation transformers feature double or triple shields, the series coupling forming a capacitive divider and allowing for connection of each shield to it's associated AC or signal zero-reference point. In the case of a triple shield, the center shield can be tied separately to technical ground. When correctly done this can cut the I>O capacitance to a fraction of a pF.
However the vast majority of consumer systems feature minimal cost power supplies with high CM AC line to DC output coupling (often >1000pf). I realize this will irritate some traditionalists, but this shortcoming can be successfully addressed using SMPS with their much smaller transformers and potentially lower I>O capacitive coupling, and more easily filtered outputs. The other limitation of consumer gear is the unbalanced signal transport, often coupled with improper grounding of the RCA jacks.
The main function of a "Power Conditioner" in most of these systems if properly implemented is separating signal paths from stray chassis AC leakage paths. A good reference for this subject is Morrison's "Grounding and Shielding Techniques." I was dropped into the warzone of this subject when tasked with building a wideband unbalanced distribution system on a factory production floor with high EMI, hundreds of feet of cable and 80 loads. The signal to be distributed was 80x normal speed audio, so the specs at each load had to be >90dB s/n, and response +/-.5dB from 400Hz to 1.6MHz...which we finally exceeded. Talk about a challenge, it accelerated my aging for sure...controlled impedance cabling and termination, Topaz Ultra-Isolators, massive grounding busses and isolated, chemically enhanced grounding electrodes were key.
The secondary function of an effective power conditioner is actual filtering of the AC power fed to the separate components. Some feature brute-force filters which resonate at the fundamental like the awesome units Ed Simon designed. Most merely feature canned EMI filters with a relatively high low-pass corner frequency, often above the audio band. They can still serve a useful function keeping RF trash out of the AC power system.
Another couple of cents worth...
Howie
p.s. I love this Bob Widlar poster.
The primary component of a proper power conditioner is an isolation transformer. The inductance of a 50 or 60Hz transformer makes it a natural LP filter and a poor transformer for RF, but capacitive coupling between primary and secondary can couple high frequencies. The best isolation transformers feature double or triple shields, the series coupling forming a capacitive divider and allowing for connection of each shield to it's associated AC or signal zero-reference point. In the case of a triple shield, the center shield can be tied separately to technical ground. When correctly done this can cut the I>O capacitance to a fraction of a pF.
However the vast majority of consumer systems feature minimal cost power supplies with high CM AC line to DC output coupling (often >1000pf). I realize this will irritate some traditionalists, but this shortcoming can be successfully addressed using SMPS with their much smaller transformers and potentially lower I>O capacitive coupling, and more easily filtered outputs. The other limitation of consumer gear is the unbalanced signal transport, often coupled with improper grounding of the RCA jacks.
The main function of a "Power Conditioner" in most of these systems if properly implemented is separating signal paths from stray chassis AC leakage paths. A good reference for this subject is Morrison's "Grounding and Shielding Techniques." I was dropped into the warzone of this subject when tasked with building a wideband unbalanced distribution system on a factory production floor with high EMI, hundreds of feet of cable and 80 loads. The signal to be distributed was 80x normal speed audio, so the specs at each load had to be >90dB s/n, and response +/-.5dB from 400Hz to 1.6MHz...which we finally exceeded. Talk about a challenge, it accelerated my aging for sure...controlled impedance cabling and termination, Topaz Ultra-Isolators, massive grounding busses and isolated, chemically enhanced grounding electrodes were key.
The secondary function of an effective power conditioner is actual filtering of the AC power fed to the separate components. Some feature brute-force filters which resonate at the fundamental like the awesome units Ed Simon designed. Most merely feature canned EMI filters with a relatively high low-pass corner frequency, often above the audio band. They can still serve a useful function keeping RF trash out of the AC power system.
Another couple of cents worth...
Howie
p.s. I love this Bob Widlar poster.
+1
Not to forget a faster recovery (when correctly designed) after transients because of HF switching frequency.
I don't understand the bad reputation they have near some audiophiles.
Please do keep in mind the DAC-3 that I put on Richard's prototype power conditioner already does have a switching supply. Conditioner still made an audible difference. And there are already balanced interconnects going from DAC to AHB2, which also uses a switching supply. Just saying.
In other news, my new used Monster 7000 MK II power conditioner should be here in another week. Should be able to power everything in the sound system from that. Interesting to see what happens, I'm curious anyway.
In other news, my new used Monster 7000 MK II power conditioner should be here in another week. Should be able to power everything in the sound system from that. Interesting to see what happens, I'm curious anyway.
That is the model you listened without it and a week later, with it.
I designed the -7000 about 18 years ago.
When the economy fell off the cliff…. thousands of retail outlets closed... where it had been selling. The need still exists but there are no more outlets to sell thru. Its up to the internet only, now. Monster distribution is retired. I will make it myself over in Bangkok where it was originally made.
THx-RNMarsh
Last edited:
Well, with a switching power supply, when the pass element is on, everything on the AC side rides through for free, zero attenuation.
A proper supply for audio would start with a good power transformer that contains an electrostatic shield. That's how I spec my transformers to start with. No matter what happens on the AC side, whatever could possible make a difference must show up on the DC voltages. Do the power supply properly and power conditioners won't make any difference at all.
-Chris
A proper supply for audio would start with a good power transformer that contains an electrostatic shield. That's how I spec my transformers to start with. No matter what happens on the AC side, whatever could possible make a difference must show up on the DC voltages. Do the power supply properly and power conditioners won't make any difference at all.
-Chris
If you are referring to merely the switch and transformer, that is more nearly correct. Indeed the same can be said for a linear supply during the diode conduction cycle. This very issue plagued early direct-conversion receivers powered from the AC line where any stray common-mode RF would be modulated by conduction through the supply and acquire a wicked 60Hz+harmonics humming. However in order to meet FCC Part 15 specs at the minimum all SMPS have stout input filters so this strongly limits what "rides through."
With an SMPS the passband for the period of conduction extends from an octave or so below to several decades past the switching frequency. Frequencies below are subsampled and then averaged by the control loop and output filter which also removes HF switching artifacts (switch freq + ΔI/ΔT spectra). The higher frequency makes both input and output common and transverse-mode chokes practically sized as well as the filter caps. When it comes down to it, the design of a SMPS is largely about magnetics and short switch loop PCB design. I would argue with today's SMPS designs, it is as easy, but smaller, lighter and less expensive to obtain clean output with a SMPS than it is with a 60 Hz PS. There is however something alluring about paying a bunch of money and getting something heavy...
Agreed...although very few consumer gear do it right. When was the last time you saw a preamplifier with a double-shielded power transformer? Without it its a toss up whether there will be inter-chassis noise potentials. They used to be common in professional broadcast gear...now with pricing pressure from junk imported semi-pro gear it is becoming far less common, even in otherwise respectable brands...
Cheers!
Howie
Howie: Can you confirm what you mean by 'double shielded'? I do pay the extra for an electrostatic shield on transformers I buy.
Ref the Benchmark they seem to have done everything right, or at least by the book. Bar a couple of wires that you could ague they should have twisted the conditioner shouldn't make a difference in this case. However stranger things have happened, esp when there is a computer nearby.
Ref the Benchmark they seem to have done everything right, or at least by the book. Bar a couple of wires that you could ague they should have twisted the conditioner shouldn't make a difference in this case. However stranger things have happened, esp when there is a computer nearby.
Thank you. It will be some time before my system has evolved to the point where it would make sense to try it though, so possibly best to wait until the changes are done as everything bar the blu-ray player n the electronics front is changing. Now3 years behind schedule
Don't you forget the output transformer ?
Well, there is very little leakage with a C core transformer and a non conductive C.
I wonder why they are so few in Audio.
Last edited:
Hi Howie,
Yes, I agree with your points as you've made them. I singled out switching supplies because I have heard many comments around that would suggest that they block noise, which we both know isn't true at all. The conduction angle for a switching supply could even be greater than a linear supply with a huge capacitor on it. Either way, the power supply designer has to be smart about it.
-Chris
Yes, I agree with your points as you've made them. I singled out switching supplies because I have heard many comments around that would suggest that they block noise, which we both know isn't true at all. The conduction angle for a switching supply could even be greater than a linear supply with a huge capacitor on it. Either way, the power supply designer has to be smart about it.
I watched that trend with dismay. I used to do service in that market area and it followed consumer goods down the swirly hole. What's the point of having a really good service shop when everything turns into "junk". Compared to what they used to be like anyway. There were some Japanese firms that made really good equipment too.
-Chris
Hi Tryphon,
-Chris
Not in solid state equipment as that was what I was thinking about at the time. I don't for tube products either.
Because, they aren't sexy like a toroid, and they take up more room. Not efficient with space you know. I did see them in some low cost audio products back in 1980. Nothing expensive used them to the best of my knowledge.
-Chris
Correctly designed gear generally doesn't need a line conditioner in a residential setup. Of course the guys who sell them well tell you something else. Show me what you're eliminating from the analog outputs.
Toroids are a good choice because of the low leakage flux. That's one reason why Benchmark is using an SMPS. Easier to filter SMPS output with the huge variety of devices available now than to deal with low frequency magnetic fields.
Toroids are a good choice because of the low leakage flux. That's one reason why Benchmark is using an SMPS. Easier to filter SMPS output with the huge variety of devices available now than to deal with low frequency magnetic fields.
Anatech, I dont understand this sentence.
A SMPS rectify the AC in a big cap. First degree of filtration, maximized when the diodes (of MOSFETS) are not conducting. Then an oscillator transform this DC in AC at high frequency. Don't tell-me the oscillator do not reject anything) then a transformer is used to transform the high voltage AC in the requested output voltage. This transformer is very little, and optimized for the HF of the oscillator: everything below is highly rejected, there is the main isolation between AC outlet and DC output: its capacitive leakage is reduced by its little size.
Then you have an other rectification stage and filtration. That has to be efficient to reject the switching frequency.
Last edited:
- Status
- Not open for further replies.