Power conditioning: how did you choose the best approach for your system?

[McNulty]

Over the last couple of weeks I've been reading up on power conditioning, as this seems an area in which I can improve my system. I'm now using a PS Audio Dectet I bought years ago on a separate power line with all audio components (including amplifier) plugged into it.

 

It appears there are several approaches when it comes to power conditioning (e.g. isolation transformers, balanced power, power regenerators, active filtering, passive filtering; also amplifier and other audio components on the same power source vs. separate power source for amp and for components), but with every approach comes the caveat one should try it at home in his/her hifi system before buying, as every situation differs and there seem to be no absolutes in this area. Within each approach one further has the choice between different brands and models of course.  

 

This brings me to the question if people have compared say three or more approaches for power conditioning in their system before choosing and buying their final solution. I would appreciate it if you could share how you came to your decision on what power conditioning works best for your hifi system. Thanks for your input.

 

 

[johndoe21ro]

I'm using an IsoTek Sigmas GII with an Optimum EVO3 cable and I'm very pleased with the results. And I'm using the power amp with it too - absolutely no loss in dynamics.

[Ralf11]

I use properly designed components and had the electric utility test my power

[fas42]

A reasonable approach is to firstly assess how sensitive the individual components are to noise on the mains - if you know where the weakenesses are, then you are far better situated in terms of knowing how to proceed!

 

What I would do is start experimenting with the rig you have now - first step, how much does the SQ degrade if you eliminate the PS Audio entirely, and just plug everything into a conventional power board, on a normal house power spur. Then, a single component at a time gets plugged into the PS Audio conditoned line - how much does this improve things, with just that component and none of the others 'assisted'. Then do two components at a time, etc. ... and further variations on this.

 

What I'm trying to do by this is to precisely nail where the rig is most sensitive to mains quality, and whether having multiple components plugged into the single conditioner possibly makes things worse - trial and error until one has a much better handle on where further conditioning will best benefit.

[One and a half]

@Ralf11 please define improper and how this can make SQ worse.

[gmgraves]

9 hours ago, McNulty said:

Over the last couple of weeks I've been reading up on power conditioning, as this seems an area in which I can improve my system. I'm now using a PS Audio Dectet I bought years ago on a separate power line with all audio components (including amplifier) plugged into it.

 

It appears there are several approaches when it comes to power conditioning (e.g. isolation transformers, balanced power, power regenerators, active filtering, passive filtering; also amplifier and other audio components on the same power source vs. separate power source for amp and for components), but with every approach comes the caveat one should try it at home in his/her hifi system before buying, as every situation differs and there seem to be no absolutes in this area. Within each approach one further has the choice between different brands and models of course.  

 

This brings me to the question if people have compared say three or more approaches for power conditioning in their system before choosing and buying their final solution. I would appreciate it if you could share how you came to your decision on what power conditioning works best for your hifi system. Thanks for your input.

 

 

Well, I figured that a large transformer with lots of iron and lots of copper in the form of transformer primary and secondary windings (for a one to one isolation transformer - 120volts to 120volts) would have so much reactance above 60 Hz that high frequency interference from computers, RF, and switching transients from appliances wouldn’t have a chance of surviving to get out of the Xformer’s secondary.

I had my eye on a 21 Ampere hospital isolation transformer at a local surplus electronics store for years. It was there every time I went in but they wanted entirely too much money for the damn thing. Then one day I approached the front door and saw a big sign over that door announcing “Going Out Of Business! Everything must go. No reasonable offer refused!” I immediately went to the part of the store where the Xformer usually hung out. It was still there. I picked it up, took it to the front of the store and plopped it on the counter. “Fifty Bucks!” I called out. “60”! The counter man retorted. “Done”! said I. 

When I got the Xformer home, being a certified nerd, I took the two end bells off of it to find that in addition to the windings, there was an LRC low-pass filter on both the primary and the secondary sides of the unit. My Tektronix oscilloscope confirmed that the mains was dirty on the primary side and squeaky clean on the secondary side. It immediately went into my main system and I could hear the positive difference at once! Best $60 I ever spent on my audio system.

 

[sandyk]

37 minutes ago, gmgraves said:

It immediately went into my main system and I could hear the positive difference at once! Best $60 I ever spent on my audio system.

 

 I guess that means that it was dead quiet too ?

[shtf]

11 hours ago, Ralf11 said:

I use properly designed components and had the electric utility test my power

 

What utter nonsense.  A $5 AC filter inserted by the mfg'er behind the IEC inlet does not constitute sufficient line conditioning for the noisy AC coming in from the street.  And just because some pimple-faced electric co. technician takes a measurement and tell you your AC is clean doesn't make it so.

 

Do yourself, the forum, and the industry a favor and go audition some superior line conditioners.

[sandyk]

25 minutes ago, shtf said:

A $5 AC filter inserted by the mfg'er behind the IEC inlet does not constitute sufficient line conditioning for the noisy AC coming in from the street.

 

 More likely it is there to meet SMPS RF/EMI specifications from the component  itself.

[gmgraves]

3 hours ago, sandyk said:

 

 I guess that means that it was dead quiet too ?

Yup!

 

[Ralf11]

11 hours ago, shtf said:

 

What utter nonsense.  A $5 AC filter inserted by the mfg'er behind the IEC inlet does not constitute sufficient line conditioning for the noisy AC coming in from the street.  And just because some pimple-faced electric co. technician takes a measurement and tell you your AC is clean doesn't make it so.

 

Do yourself, the forum, and the industry a favor and go audition some superior line conditioners.

 

Let's not be stupid.  One and a Half's question (tho inverted) was a better one.  A discussion of power supply design might be useful, as would a discussion of what add-ons really work.

 

Here are some useful tidbits from Audio Research :

• regulator circuits with wideband response
• Ref6 has a tube-regulated power supply

 

[Ralf11]

I also recently posted a scientific paper relevant here - w/in the last week - IIRC, it is on the thread where the OP was essentially asking for a primer on line noise.

 

Nelson Pass also has some non-technical reading out there on der internet somewhere.

 

Of course, the claim that one needs to add something to deal with line noise pre-supposes that the power supply provided by the multi-thousand dollar component you purchased is somehow inadequate for the amount of noise on your AC line.  And, second, that a home DIYer can improve upon it.

 

Not to say that a Topaz may be worthwhile in some installations - say a NYC apt. (just to pick a likely noise ridden env.)

[PeterSt]

I can imagine that something can be achieved with power conditioning. But people should not underestimate what's all coming through the air and how this is power grid related. PWM pump of your fridge, to name something.

For that reason, power conditioning never worked out for me. The "negative" part for me could be that I can measure it, and that I do. Whatever frequency coming through air (in-band !) is not filtered or -regenerated-out by a power conditioner. Even batteries don't help. Only a new fridge helps. But you won't even know ...

 

The next kind of sad message is that you can be at the recipients of that power, make that immune to whatever it is, and happily be connected to your neighbour's dishwashing grid. The sad part is that again you can not do that. Only the manufacturer of your gear can. And I don't see that happening. It could be done, but it is not too well understood by those manufacturers. Maybe if they first had designed high speed networks (which would make them have the measurement equipment to begin with). But ...

[cjf]

I've tried passive devices and isolation based transformers in my system to date (Torus/Shunyata). In EVERY case, all power conditioners I've tried limited dynamics to my amps in a noticeable way.

 

I'm satisfied with my current arrangement which is a passive Shunyata Denali 6000T on front end gear ONLY and my Amps plugged directly into the wall. The circuit the Amps are plugged into is a dedicated 20A receptacle using 10g Romex to the breaker box. The circuit the Shunyata Denali in plugged into is also a dedicated, separate, 20A receptacle using 10g Romex to the breaker box.

 

I've not tried any "Regen" type devices (ie..PS Audio...etc) as of yet and have no intention of doing so. I'm convinced the outcome will be the same in terms of plugging power amps into them and thus its not worth the time/expense to experiment anymore given what audio gear resells for second hand should my assumption be correct. Been there done that too many times

 

Of course this is all my FWIW, 2cents, opinion

[PeterSt]

39 minutes ago, cjf said:

I've not tried any "Regen" type devices (ie..PS Audio...etc) as of yet and have no intention of doing so. I'm convinced the outcome will be the same in terms of plugging power amps into them

 

People might consider an other line of thinking:

All equipment not on the same house ring - or worse - not on the same earth pin, will imply undesired ground loops. This is similar to the thinking that "a" device on batteries will be for the better because of being on batteries, inherently. No. No, because it implies a different "potential" right away and there you have a first groundloop.

 

Put all on the same ring (no matter computers and amps etc. mixed), probably eliminate PE (protective earth) connections all together (but this is always to be tested) BUT remove all switching supplies from it. Maybe the computer's is allowed to stay (they are not *that* bad) but all what charges phones and feeds monitors should go out. Generally a matter of one separately earthed ring where you plug the whole of the audio in, and make an exception for the monitor which *still* is better to be taken out when music plays. Now it may go a bit beyond comprehension, because it could take a second (separately earthed) ring now, a controlling computer with monitor there, that one over isolated ethernet (RDC) controlling the audio PC.

 

This is infinitely more important than power conditioning, because your individual equipment can not counter attack ground loops. This, while it is already made to some degree to filter mains noise. This goes automatically, via "with some thought" up to very explicitly (ultra low noise internal power regulation).

But groundloops ? n.b.: a groundloop an be considered as undesired current flowing over (digital) interlinks. They eliminate the potential (official ground for the device) difference between the devices. It is a signal. But not carrying any music ... It is NOISE.

 

Edit: The above can't be understood when I forget to tell that interconnecting two mains rings with separate earth (one power regenerator or one isolation transformer among that), will have a large voltage difference for the whole ring to begin with. Thus interconnect those two rings and the voltage difference has to spread over the two and this goes over the interlinks of the devices on the one ring vs those on the other (usually there is one connection point only, but it could be two, like when USB is involved).

[sandyk]

28 minutes ago, PeterSt said:

This is similar to the thinking that "a" device on batteries will be for the better because of being on batteries, inherently. No. No, because it implies a different "potential" right away and there you have a first groundloop.

Peter

While I agree with the rest of what you have posted, this doesn't make much sense to me, and in my experience results in improved performance with USB provided that the lead length from the battery to the powered device is minimised, and it is used in conjunction with a low noise voltage regulator. Even earthing the provided screen of an R-Core transformer causes some degradation with a mains powered  PSU for USB.  I have fitted a toggle switch to such a mains powered supply for USB and this is easy to confirm.

 

Regards

Alex

[semente]

.

[monteverdi]

1 hour ago, PeterSt said:

I can imagine that something can be achieved with power conditioning. But people should not underestimate what's all coming through the air and how this is power grid related. PWM pump of your fridge, to name something.

For that reason, power conditioning never worked out for me. The "negative" part for me could be that I can measure it, and that I do. Whatever frequency coming through air (in-band !) is not filtered or -regenerated-out by a power conditioner. Even batteries don't help. Only a new fridge helps. But you won't even know ...

 

The next kind of sad message is that you can be at the recipients of that power, make that immune to whatever it is, and happily be connected to your neighbour's dishwashing grid. The sad part is that again you can not do that. Only the manufacturer of your gear can. And I don't see that happening. It could be done, but it is not too well understood by those manufacturers. Maybe if they first had designed high speed networks (which would make them have the measurement equipment to begin with). But ...

why not "condition" the offerers? I had a simple filter parallel to my fridge and that of course did not limit the power to my audio system and all my computer stuff had its own cheap filter. In my present (intermediary) rental none of that is happening but I would like to implement that approach again!

[McNulty]

Thanks for all your in-depth reactions so far, much appreciated. 

 

Following the advice of @fas42 , I did some experimenting with plugging components in the PS Audio Dectet and in a regular Brennenstuhl power strip. Note this was not a blind test, so some bias could play a role. Furthermore, in my living room only one wall socket is earthed, creating an additional variable.  

 

The outcome of the experiment was:

- integrated amp better without power conditioning

- DAC: indifferent

- server/streamer: better with power conditioning
- switch: better with power conditioning

 

I intend to try out listening to the system plugged into a quality, non conditioned, power strip over a longer period of time. Later on I might do some further experimenting with specific measures for the server/streamer and the switch.

 

Interesting that @PeterSt and @monteverdi mentioned the fridge. The build-in fridge in my kitchen also creates some problems in my system; as the fridge turns on, sometimes a short tick can be heard over my loudspeakers. The separate power line for my hifi system didn't change this. I also tried to tackle this with an Essential Audio Tools Pulse Protector in the wall socket closest to the fridge, and while this does help a bit, the tick is still not completely gone. So I too have been thinking about buying a new fridge...

 

 

[One and a half]

To expand on  @PeterSt and @sandykposts and inference that a Furman/Panamax/Monster power conditioning type of device, or even a regenerator is effectively zero if the cabling topology is plugged into anything that's convenient.

 

The ideal situation is where all equipment are connected to the one point, to a common ground.

 

Star ground system. Resistors represent ground cable of each device. Center dot is building ground. There are no loops. R1 to R6 are effectively all different, since the length of the cable determines. The situation is not so bad, since the differences are in micro-ohms at the distances we use for audio equipment. This avoids differences in ground voltages between equipment, where ground loops both analog and digital start to form. Here's a good page where this is illustrated further. The concept is quite complex and troubleshooting just as complex, but if a few guidelines are followed, the reduction in noise can and does make a difference.

 

The star grounding system can be replicated with a simple power strip, that has no overvoltage or surge suppression devices, just a plain Jane strip. Especially for sources, where a number of devices feed to a pre or integrated amp. It would follow that to maintain the same ground potential, to use the pre and power amps on the same power strip.

 

There's not much we can do about computers or even music servers, they need to be part of this ground network, but using Ethernet transmission to advantage the music server can be on a long wire (say R1 in the illustration), several metres from the audio system. The ground point connection must still the same, but interference from the music server is well reduced as the longer cable provides difficulties for noise to propagate those distances. 

The traditional approach to keep digital and analog separate with different transformers can be done, the grounding must follow a strict routine where their grounds are brought back to the same star point to the rest of the audio system.

 

So avoid daisy chaining power strips and plugging in different audio equipment, this causes the building ground common point to move, where ground potential voltage increases, use a longer power cord instead. 

 

Once the wiring is complete, it's a another issue to protect the audio equipment against high voltages, perhaps another post. 

[sandyk]

26 minutes ago, One and a half said:

So avoid daisy chaining power strips and plugging in different audio equipment, this causes the building ground common point to move, where ground potential voltage increases, use a longer power cord instead. 

 

Good advice.

[One and a half]

Here's a good page on Digital and Analog mixed grounds and how to avoid drama.

 

 

[PeterSt]

1 hour ago, McNulty said:

as the fridge turns on, sometimes a short tick can be heard over my loudspeakers.

 

This is by almost guarantee caused by an illegal ground somewhere. For example, it would perceivably be allowed to disconnect the ground of au audio signal wire at one end. In an earlier life I did so all the time. Usually the receiving end was disconnected. No problem, you'd say. But since the ground *has* to come from somewhere (actually, needs to  go to somewhere as in "ground return"), it starts to travel illegal paths which start in the receiving device (could be your preamp). Where it goes you don't know. But usually the whole power system is now involved, and switching off/on lights already are audible. Mouse movement may be audible. Disk access. Light dimmers. Led lamps. And certainly the higher powered stuff like fridge motors.

 

Since I started to be "legal" regarding this, all not only went for the better, net, but these anomalies have been gone for 100%. Additionally it now is a measure: if now it happens again, I can be 100% assured that some ground wire (like in a cable connector) went loose. So that too is a reason, obviously. But it is hardly different from doing it on purpose ... 

[sandyk]

5 minutes ago, PeterSt said:

This is by almost guarantee caused by an illegal ground somewhere.

 Peter

 It can also be caused by a poor main earth . I had this problem years ago at a previous address where the main earth connection  was corroded.

Regards

Alex

[PeterSt]

3 hours ago, monteverdi said:

why not "condition" the offerers? I had a simple filter parallel to my fridge

 

Would that limit the high(er) frequency radiation of PWM (pulse width modulation) ?

I gave the example "fridge" to make it easy, but literally I should have said "freezer" (my now dead previous freezer really used that) and otherwise I have a 380V (Grundfos) pump that can be controlled from a distance that has PWM for the explicit purpose (like the PWM controlled fans in your PC - ditch those too !).

 

 

Red is system (DAC) output noise at close to -160db). White is power supply (to the DAC) noise (on average at -165dB). The peak close to 20KHz I don't know at this time (I just forgot). The group you see left of it is that PWM pump of the heating system. Mind you, this goes through air. The higher peak left of that again is ... USB packet noise (16KHz, these days I tamed that to invisible in the DAC output as you can see). The lower peak all in the left is of that freezer.

 

Possibly a next project will be to create explicit antennas to absorb this energy.