[samplesj]

I asked a while back about acceptable THD values, but I titled the thread based on a specific model so I didn't get any responses so I'm trying the more general topic in a more techincal circle.

I'm really most interested in the general effects the THD of a sine wave will have on normal power supplies.  Most "computer" UPS that claim true sine wave state a THD spec of around 5%.  I've found a few that claim 2% or 3%.  However on the other hand we have the audio specific products that claim <1%.  I've also seen power measurements different places on the web that show around 5% out of the wall socket itself (mostly in office buildings though).  If those are realistic than a 2-3% UPS should be step forward.  

One caveat is that I am looking at true online systems so this THD will always be present (not just in a sag/outage).

I don't really care about the battery backup aspect, but I do like the other benefits (line isolation and voltage regulation for example).  Line isolation is a biggy for me.  I have a dedicated line for my good office system, but on my den system I can hear pops from the washing machine.  Of course that line isn't dedicated, but it isn't the same circuit as the washing machine either.

[ctviggen]

you might try one of these:

http://www.smarthome.com/1626-10.HTML

I have a fridge that causes a light to come on for a half second.  I've tried this:

http://www.partsexpress.com/pe/showdetl.cfm?&DID=7&Partnumber=110-452

And that didn't work completely.  So, I'm going to try the smarthome product for all my freezers/fridges (along with the snap on inductor -- heck, it's only $2 and I have four of them).

[samplesj]

Well if I don't go the power regeneration route, I will definately at least use isolation (or balancing) transformers which should isolate the gear from noisy equipment.  Right now I have the HT balanced and I like the results, but regeneration "seems better".

Right now I'm trying to figure out what acceptable THD numbers are.  If those 2-3% numbers are good (or slightly better than average) then a UPS would make the system immune to a lot of issues.  

With a UPS with stable voltage I could even play around with a slightly larger transformer (like Plitron's LoNo Audio 1000VA 45v [not made in the 40/42 safe voltages ]) for my UCD400s since I wouldn't have 10% mains variance to be cautious about.

[ctviggen]

You have HT balanced and can still hear the washing machine?  Yikes!  Even on my system, where my amps are plugged directly into the wall, I cannot hear anything when my fridge, in the same room BUT on a different circuit, turns off, although the fridge causes a light on the same circuit to briefly turn on.  

So, what you're asking is how a THD of 2-3 percent (assuming pure DC to AC conversion, so no noise gets into the AC output) would affect your equipment?  That's a good question.

[samplesj]

No, sorry I wasn't clear about that.  The balanced HT is quiet like it should be.  The den is just a cheap system for background music/noise so it runs straight out of the wall.

For the new 2 channel system I'd love to go balanced, but even DIY the big transformers are really pricy.  

I can get a 2000VA true online UPS with 2%THD sine wave for less than half of what a BPT3.5sig goes for so I got to thinking about if it was better.

[Dan Banquer]

You have asked a tough question that I don't know if anyone has the answer to.  One would initially think that a well constructed power supply would attenuate most of the lower order harmoincs, however I have come to understand that most places that have a lot of computers have very high second order harmoinics on the AC Line, and appears to be cause for concern.
Power Factor correction may be an answer here to reduce the THD, and if I remember correctly Occam has done some experiments in this area that appear to work.  I also think VICOR Corp has a module to reduce this also.
I'll let Occam explain, and I may offer additional comment if necessary.
                    d.b.

[NealH]

I agree that a well designed power supply should attenuate the power frequency harmonics on the input to a point where they should not be much of a concern, assuming attention has also been paid to the signal ciruitry in regards to common mode rejection.  In other words, assuming you have a well designed amplifier like McIntosh, etc.   Unfortunately, cost concerns and insufficient knowledge often stop designers from putting in heavier duty "double L", "pi", "swinging choke", etc., type input filters that were commonplace in yesterday's high quality tube equipment.  The advent of the IC regulator ushered in an era where only a capacitor (not even a single "L") input filter is often employed and, they bank on getting the harmonic attenuation numbers up with the common mode rejection of the PS regulator, amplifier driver and output stages.  

If in doubt there is an MGE power conditioner (model 800 I believe) which is really a modern ferroresonant supply.  It will murder any power frequency harmonics going in, or going out.  The residual THD (3 or 5% max) on the sinewave output is really insignificant.  It's just not going to have an audible effect on an amplifier.   And, common and differential mode noise rejection with this type supply is almost without peer.  Some have complained that ferro transformers emit some audible noise at the oscillator frequency (60Hz) but, this is not the case with modern ferro supplies like this one.

[Occam]

...If in doubt there is an MGE power conditioner (model 800 I believe) which is really a modern ferroresonant supply. It will murder any power frequency harmonics going in, or going out. The residual THD (3 or 5% max) on the sinewave output is really insignificant. It's just not going to have an audible effect on an amplifier. And, common and differential mode noise rejection with this type supply is almost without peer. Some have complained that ferro transformers emit some audible noise at the oscillator frequency (60Hz) but, this is not the case with modern ferro supplies like this one.

Have you actually listened to an audio system powered by a MGE 800 powerconditioner? Could you describe the system and the benefits that the cvt provided. Did you compare it to other conditioners, both industrial and bespoke audiophile types? If so, how would you compare them.
Didn't you find the MGE's stated noise of

Audible Noise: 1 meter A scale - 52dB to 56dB

to be quite intrusive in your listening environment?

TIA,
Occam