Isolation transformers - where do I start?

I did this the easy way by purchasing an Audio Power Industries PowerWedge Utlimate 116, which features 6 isolation transformers for sources components (it isolates components from one another as well as from the line). It's been a long time since I compared using this to plugging straight into the wall (I do remember an improvement, otherwise I wouldn't have kept the unit), and it also puts the power through HF filters as well as transformers. But I can handily compare another power aspect, that of running balanced AC (+60v/-60v as opposed to +120v/0v), which is selectable/defeatable from tranny taps switched individually on the rear panel. Based on my experiences in this area, I would say if you're going to use isolation transformers, you should go the balanced route if you can.

Some general suggestions here.

"Old school" iron core transformers are superior to toroidals in terms of noise rejection. Toroidal designs do work, but are not as efficient in terms of total noise rejection.

Look for transformers that use the most iron for their cores and are rated the highest in terms of KVA. "Good" isolation transformers are phenomenally heavy and expensive. My bare 3KVA isolation transformers weigh over 100 lbs apiece. The cores are actually good for about 5KVA. As such, i might saturate the windings but i'd never saturate the core. The 1.8 KVA bare transformers that i have weigh 42 lbs apiece. I used to think that these had "healthy" cores until i picked up the 3 KVA units : ) The bare "little" .25 KVA transformers i have weigh 12 lbs apiece. Just the sheer shipping weight and cost of raw materials for these units might tell you why many manufacturers opt for lighter and less costly toroidal designs.

If you must share transformers between components, keep the digital separate from analogue and vice-versa. If using a transport and dac, further gains can be made by having an individual transformer for each unit. If possible, take that a step further and use individual isolation transformers for each component.

If you are going to use an isolation transformer on a power amp, the transformer should be rated quite a bit higher than what the amplifier would draw at the point of clipping. Otherwise, you might run into core saturation on the isolation transformer during momentary peaks or sustained low frequency passages.

Bare in mind that transformers produce a magnetic field around them. You will have to take steps to shield them if placed near the gear. Keep signal cabling FAR away from ANY type of power transformer.

Start with your digital gear and work your way down the line.

I've been told to stay away from Sola units as they supposedly do "bad things" sonically. Besides that, 3% distortion is pretty high. According to what i've read, the power coming off the grid is "supposed to be" less than 5% at any given time. If it is higher than that, your pole transformer is probably defective.

Short of purchasing a TRUE power regenerator ( expensive and in-efficient ), there is nothing that you can do to remove ALL of the noise coming in off of the line. Using good quality iron core isolation transformers will give you the most benefits with the least drawbacks, but they are sloppier and far less convenient to work with than commercial "audiophile" PLC's.

Keep in mind that this is just my take on the situation. Others will have varying opinions that are probably just as valid, maybe more-so. Sean
>

Sean's advice is very good. I would recommend looking at EquiTech's wall mounted units. These are very large isolation transformers that take 220 in and then drop it down to 110 with two balanced poles (thus further rejecting the noise). I made one similar to these with a transformer I had gotten from a medical device company that was no longer used. It was about a 6.5 KVA unit (yes it was heavy). The only downside to the EquiTechs is that they are expensive, but I think they have really engineered a superior product (it's the brute force method--but works quite well). It's used by many recording studios, they're just not as well known in the home market.

Cdc, I was in a very similar situation this past winter. I looked at all the pieces recommended above and found the Hubbell to be one of the best out there. Rick at Virtual Dynamics has recently informed me that he is modifying Hubbell Iso Transformers with Cryogenic treatment, internal wiring tweaks and some other stuff you'd have to ask him about. My power needs have been met lately so I am not in the market for one of these... If you end up getting one please let me know what it does in your system.

would dedicated outlets take care of most of the "problems" that transformers fix? what is it exactly that they fix anyway? fluctuations in current? noise?

I would start by buying a couple of smaller isolation transformers from eBay ($15-$35). Play around with them. Open them up and clean all contact points w/ contact cleaner--you'll be surprised...

I have several isolation transformers and I really like my C core ones. Some are "Chicago Transformer Co." 3.5 amp isolation transformers and they have three output taps: 105, 115, 125V. I plug my DAC to the 125V tap and besides lowering the noise floor it increases slam and authority in the music. Very similar to when I use my 3,000W 220V ONEAC to plug my power amp. Sean is correct when he says C cores are superior. My big ONEAC is also C core.

However, don't turn down a great deal just because it's toroidal--right, Sean?

Dennis: Dedicated lines simply separate the power lines for your audio gear from all of the other wiring in the house. This in itself can reduce noise within the AC feeding your A/V system, but you are still susceptible to the noise that is coming in from the outside line feeding your AC mains. As such, an "optimum" installation would be one with dedicated lines and either a very high current power regenerator or extensive isolation via low impedance filtering / isolation. If you wanted to get "really crazy", especially if you had BIG high powered amps, would be to bypass all of the AC completely and feed your system off of a bank of batteries. If i had a simple, low powered system, that is probably what i would do. Since i am not in that boat though, i've had to go the first route.

An isolation transformer "uncouples" or "isolates" the A/V components from the outside lines. It is basically a 1 to 1 transformer. In other words, if you put 120 volts into it, you get 120 volts out of it. Due to the indirect coupling via the transformers windings though, you end up losing the mass majority of noise that may have been riding on the line. This is a good thing in every aspect. As stated above, "old school" iron core transformers are more efficient at "electrically isolating" the AC signal from the noise, but a toroidal design can also be quite effective. As Francisco stated, i would not pass up a good deal on a legit isolation transformer just because it was toroidal based. While it may not work quite as well as an equivalent E-I or C core "old school" transformer, it would none the less reduce the noise on the line very drastically. The drawbacks to the reduced efficiency ( in terms of reducing noise ) that the toriodals may suffer from can more than be made up for by compact size and measurably less weight than if one were to make comparisons with standard transformers.

The drawbacks to using isolation transformers is that they can introduce losses into the AC system and play games with the phase of the signal being fed to components. In both cases, this is typically a side effect of trying to use too small of a transformer for too big of a job. If the core of an isolation transformer is getting physically hot, you can bet that it is not big enough for what you are using it for and "bad" side effects will result. It is not abnormal for any device that is passing a reasonable to large quantity of power to undergo some type of temperature rise, but a transformer should NOT get physically hot to the touch. If it is, the transformer is simply not up to the task or being pushed too close to its' limits on a steady state basis.

As such, that is why i recommended using transformers that were rated for well above the amount of power that you intend to pull through them. This minimizes the potential for thermal losses / phase shifts while allowing the system to obtain all of the benefits of such a design. The use of higher current rated transformers also helps to keep the line impedance down, which is always a good thing. The "filters" used in many PLC's are of somewhat higher impedance and this can introduce problems of a different nature.

Hope this helps and explains a few things. There are variations to what i've mentioned, but this should cover the basics. Sean
>

Thanks for the replies. The Equipower is impressive but their T1000 is $989 and the 1.5Q (15 amp) is $2,189. I would like to start cheaper as $1,000+ is beyond my budget.

Hopefully the basic ISO transformer gets pretty close. I will have to check on the weights and type of core before I buy.

I'm still not sure whether I should go with a big ISO transformer for my entire system ($250 for 2kVA) or a .14 kVA power conditioning "constant-voltage-transformer" for just the source and/or preamp. Maybe I'll just buy both and return what does not work as well.

Thanks Sean for the in depth reply and warning about the Sola.

Thanks Buckingham, I will look into the Hubbel. I couldn't find Virtual Dynamics Company or Hubbel searching here at Audiogon. Do you have a link?

thanks sean... so what does "noise" sound like (the kind of noise that is eliminated from an iso tran)?

I'll say it again: buy one or two cheap ones and play with them. You'll get a good feel on their use and they can always end up in your computer or bedroom system. After you use them for a while clean all the contacts inside and you will be surprised...

Using a good quality isolation transformer or "less intrusive" PLC typically results in the following results. That is, so long as the transformers or filters are properly sized and do not restrict current.

1) Lower noise floor ( "blacker backgrounds" )

2) Less grain and glare ( digital sounds less digital )

3) Increased liquidity i.e. voices and instruments sound more natural. ( Many PLC's don't do well here as they they can tend to sound more "sterile". Many people are willing to live with this trade-off due to having some form of spike / surge suppression )

4) Increased separation of instruments and notes ( there is less noise to fill in the gaps that would normally be quiet )

5) Increased dynamic range ( due to lower noise floor )

6) More stable soundstage and imaging ( less noise "blurring" the actual spatial information / musical content of the recording )

I think that you get the idea. I must state that results will vary as AC varies from location to location quite a bit. Both the effectiveness and the results observed will vary with the types of filtration / isolation devices used, the quality of AC to begin with, the amount of filtration designed into each individual audio component, etc... My experience is that all systems will benefit from isolation transformers, so long as they capable of passing more than enough current than any individual component would need at any given time. As mentioned, it is best to isolate each component if possible or at least separate analogue from digital.

The most "bang for the buck" installation would make use of an isolation transformer for digital, one for analogue and one for the amp. Obviously, one need not use mega-watt devices for the digital and analogue gear unless those devices pull a lot of power. If all of your digital gear adds up to pulling 100 watts, shoot for an isolation transformer that can sustain 250 watts or .25 KVA. If all of your analogue gear pulls 400 watts, shoot for an isolation transformer of 1 - 1.2 KVA. In essence, you want to keep the load passing through the isolation transformer at around 40 - 60 percent of capacity. This pretty much guarantees that you will not saturate the core under any circumstances ( the one potential major drawback to this approach ) AND gives you some headroom in case you swap gear for something that is a little more "thirsty" for current in the future. As far as power amps go, especially big brutes, you can't have a big enough isolation transformer. Use the biggest one that you can find with the highest KVA and most massive iron core.

Someone that was shooting for the "ultimate" in noise reduction would use a properly sized isolation transformer for each individual component. This is obviously easier to do if you have a smaller system with less componentry.

I do not recommend using voltage stabilizers or regulators unless you have a problem with voltage in your area. Most of these devices produce / add distortion of their own. You may now have more stable voltage on the average, but you've substituted one problem for another. Obviously, there are a LOT of different products and ways to do this, so there are bound to be exceptions to what i've stated. Running dedicated heavy gauge lines directly back to the mains with clean connections and no breaks in the line will typically reduce voltage drops or "sagging" to a minimum. If you've got an over-voltage condition, that could be a problem that only voltage regulation could fix if the AC provider is unresponsive. To me, AC regulation would be a last resort but at that stage of the game, it might be a necessary one. Sean
>

Just FYI, the API PowerWedge Ultra 116 (sorry, not 'Ultimate' like I typo-ed above) incorporates 6 iron-cored iso trannys, 5 of which are 120w rated, one 150w rated, and all are capable of running in *balanced AC* mode or having their grounds lifted at the flick of a switch. It also provides 4 current-unlimited outlets without iso trannys for power amps. All the outlets employ line filtering as well, and there are MOV's and circuit breakers for protection. Everything in one 65lb. nice looking, amp-sized component (smaller versions with fewer outlets are available). The only quirk in this product is that the AC mains connector is a special Neutrik twist-lock variety, which cramps your ability to play with cords going from the wall to the unit - you basically have to use one of API's or have a custom one made up. One of the nice things about this approach though, is that ground paths are all more nearly the same length for all components, reducing ground loop potential. Since none of my source components is rated to draw above around 5w-30w nominal or so, I can't see needing a more complicated setup than this unless I want to try running my amps from isolation transformers, but this seems like it might be more trouble than it would be worth for relatively high-powered tube mono's. The unit has been most beneficial for my digital front end components and FET preamp; it didn't do all that much for my tube preamp for some reason (and my phonostage is battery powered). No sonic downside that I've heard, but the PW Ultra 116 costs over $1,500 new, and usually 1/2 to 2/3 that much used when you can find one. A PS Audio Power Plant might do all this and more, but would cost more for the same number of outlets, draws a lot of power, and produces a lot of heat, and as Sean alludes to, I for one have never measured worrisome voltage conditions on my lines. (Interesting side note: The Ultra was designed in conjunction with Kevin Halverson of Muse fame.)

Z: That sounds like a very nice product. However, i would still bypass this device and run the amp direct to the wall. Otherwise, you're trying to pull enough current through one power cord to feed the entire system. As such, the amp WILL pull hard enough to create sag on the line for the other components being fed if hitting the volume control. Other than that, it sounds like it would be a very convenient and clutter free device. Sean
>

Luke Manley has suggested the same thing concerning the VTL's, but at the moment I am not set up conveniently near enough outlets to do this (I assume only a different duplex could make any difference in this regard provided the single main AC cord is comparable in guage to the in-wall wire). At some point I intend to test this proposition however, especially since API maintains that their filtered outlets will improve the wall power, and direct comparisions using my smaller stereo amp tended to bear this out. I am contemplating getting the listening room rewired for dedicated lines.

Doing more research, not all isolation transformers are the same.
1) "Multi-shielded IT with interwinding capacitance .005 pF max and common mode noise attentuation 126dB. 120 volt / 4.17 amp. $360.00"
2) "Low leakage (hospital type): leakage current primary to secondary rated less than 50 micro-amps and is measured at less than 10 micro-amps. Power: 250 VA $133.00"
3) "Isolation transformer with static shield 150 VA $36.80."
4) "VA 150 capacity / 7.0 pounds / $45.34. The case and core and electrostatic shield are internally connected to a third conductor of NEMA standard cord and receptacle for safety"

So far #1 and #2 hospital grade looks best. Any thoughts?
I checked on ebay but not really sure what I would be getting / not returnable / get what you pay for / etc.

You are right about transformers being "different". Not only are we talking about the design of the unit, but also the quantity and quality of materials used. The 3 KVA transformers that i have use as much iron as other 5 KVA units that i've seen. Obviously, the smaller one is built a LOT "beefier" than the other even though it is rated for only 60% as much power. Retail price on the 3KVA transformers when new was up around $3000 and i've seen them selling for appr $900 in electronic surplus outlets. The only reason that i can afford the things that i have is because i am a careful shopper and know what to look for. I would imagine than many of you here are in the same position, hence the perusal of Audiogon's used gear listings. As such, I picked up the two above mentioned transformers for $210 : ) Then again, shipping set me back another $130 due to the high ( 230 lbs ) weight : (

When it comes to looking at electrical specs with these types of transformers, the models with the lowest stray capacitance is typically going to do a better job of isolation. Increased capacitance will couple ( you've heard of "capacitively coupled" ??? ) the main side of the transformer to the secondary side in a more direct fashion. In effect, the higher the capacitance per winding, the more of a "leaky" isolation transformer. The end result is only partial isolation if using a transformer high in capacitance.

Think of the amount of leakage as being compared to a signal to noise ratio ( S/N ratio ). The greater the isolation, the less leakage and the less noise gets by. Just as a higher s/n ratio is rated at a higher number in terms of dB's, so is the isolation factor. An isolation transformer with 146 dB's of isolation is "better" or more isolated / less leaky than a transformer that is rated at 126 dB's. At the same time, a transformer with a bigger iron core can pass more current without saturation or distortion. If you can achieve these two things in a transformer, you are most of the way there.

One should also take into account that any transformer generates a magnetic field. If you have multiple transformers within the same chassis, it is possible for the magnetic field of one transformer to "modulate" or "super-impose" its' signal onto another transformer. This occurs because the fields inter-act with each other due to close proximity. This is the very same reason that we do not want signal cables near power cords, etc...

As such, transformers that have "end caps" or shields over the windings are normally preferred over those that are expose the windings. While it is true that toroidals do produce a smaller field around them, they are also not quite as efficient at reducing / isolating noise to begin with. As such, one must pick and choose their trade-offs accordingly. If you have limited space and want convenience while feeding several different components, one can shoot for one chassis with several toroidals or fully shielded yet smaller sized iron core transformers in it. While one stands the potential for greater inter-action between them, the benefits would typically far outweigh the drawbacks in terms of having to use several different chassis, taking up tons of space with the associated power cord nightmare, etc... This is especially true if pulling minimal power from the transformers i.e. to feed line level sources, a preamp, etc... It would be a different story if you had multiple large transformers with their bigger magnetic fields trying to feed power amps, etc.. Sean
>

I use Equitech and www.b-p-t.com products

Storytime: Interesting update on the issue Sean raised above concerning my use of the API PW Ultra 116 PLC (although it admittedly doesn't have to do with isolation transformers). To recap briefly, I have everything plugged into this one unit, the source components into the individual isolation transformers, the VTL MB-185 tube monoblocks into the filtered but non-isolated high-power outlets. The 116 Ultra is then plugged into a single wall outlet via its own cord (and a grounded 3-prong adaptor). Sean mentioned that running everything from this one AC cord could be choking off the whole system's power somewhat, and suggested that I run the power amps to the wall separately. The thought had occurred to me before, but as I explained I don't have the outlets near the system to accomplish this, in addition to which API recommends plugging everything into their unit (VTL doesn't agree, FWIW).

Tonight, I was doing some lowish-volume comparisions between my old C-J MV-55 amp and the VTL's, as part of some auditioning I've been doing over the past few days just to get a handle on where I am with the new amps, now that their re-tube complement is fully burned in and I've installed a pair of Shunyata Sidewinders PC's on them, and before I decide to either send the VTL's away for a while to be upgraded and/or sell the C-J after 3 1/2 great years. Both the VTL's and the C-J were plugged into the Ultra's amp outlets (it has four), but of course were not all powered up at the same time, since one or the other was without a speaker load until I switched the cables over.

Listening to several audition cuts while A/B'ing the amps had me feeling pretty good about my upgrade once again, after not having heard the old amp for a while. The C-J obviously doesn't have the power to go loud like the VTL's (which was the main instigator for my upgrading, because I moved to a larger listening room), but it held its own on transient articulation and freedom from sounding mechanical or hard. Overall though, even at these late-night levels, it could not match the new amps in terms of frequency extension both up and down the spectrum, transparency to fine detail, resolution of images' spatial context, image separation, relative lack of imposed timbral colorations, bass authority and tautness, dynamic independence of images from one another throughout the soundstage, depth of field, and the smooth cleanliness that indicates relative freedom from noise and distortion artifacts. Next to the VTL's (at about 2 1/2 times its price when new), the C-J sounded smaller, flatter, coarser, more colored, more veiled, more congested and prone to glare, looser, more sonically changeable in the face of its input signal and load demands, less extended spectrally at both ends of the scale, and also dynamically at the small-scale end (didn't get into high level dynamics during this session, but of course the VTL's rule there as well). Don't get me wrong, the MV-55 was still enjoyable in its own right (and represented a huge improvement over my previous Classe Seventy SS amp at the time that I got it), but in short it was confirmed once again as sounding less natural and imposing more of its own character on the sound than the much more powerful mono's that are replacing it.

About the only possible negative on the side of the VTL's that I could find in these direct comparisions was a slight suspicion that they might be just a tad lean tonally through the midrange, but I couldn't be sure whether this might seem to be so simply because in A/B'ing the two amps, they were more extended and present at both frequency extremes, with less thickening congestion through the center. When I had last done comparitive auditioning on the VTL's, it was with the Sidewinders vs. the stock cords, and though I generally found the Sidewinders were a pretty universal improvement in terms of laying bare the soundstage and improving dynamic liveliness, I did wonder whether they could be just a little too bright-sounding in my setup. I chose not to get overly concerned about this for the time being however, since the VTL 'sound' had struck me as having been just a smidge dark and lacking in sparkle before replacing the cords. But now, I was wondering about the cords again, when Sean's admonition came back to me. Maybe if the amps were plugged into their own outlets, they might sound slightly heartier in balance.

That's when I suddenly remembered something I had forgotten before - I still had my old original, non-'Ultra' Power Wedge downstairs somewhere. This smaller model didn't have the balanced AC option on the isolated source outlets like the current version, but I could use its filtered amp outlets. I went and found it, and with its power cord I could stretch far enough to reach the AC duplex on the adjacent wall, and then use it for plugging in the monoblocks alone. The mono's would still have to share one main AC cord to the wall, but not in addition to the source components. The outlet in question was also on a different circuit breaker, something confirmed when the amps' turn-on current in-rush dimmed the lights on that side of the room much more than when they're plugged into the Ultra along with everything else as usual.

So, with the mono's on their own cord and circuit, I powered them back up, expecting to hear at the very least sound that was merely equally good as before, and hopefully better in some way. But when I played the last cut over again, I was startled to discover that it sounded as if something had broken. The VTL's were now rendered as sounding much closer in almost all ways to the old amp, losing most of their considerable advantage at the flick of a switch. I couldn't believe it. How could amps that should by rights be so easily superior suddenly sound not all that much different in a head-to-head, no matter where I plugged them in? In some ways, like smaller image and stage sizes, and sounding pinched and glarey on unision horn charts, they now sounded worse if anything. Feeling freshly vulnerable to the whims of fate, I started the track again from the beginning, but could only get halfway through it before I needed to jump up and put things back the way they had been.

With the MB-185's restored to drawing juice from the 116 Ultra with the rest of the system, I again heard the sound they ought to have, no problem. Whew! I checked both outlets with a voltmeter, both read 120v. I'm not gonna dwell on this episode too extensively right now, either concerning its possible causes, or to extrapolate it to when (or if) I decide to get new multiple dedicated lines and outlets installed. Although I believe the old Power Wedge isn't supposed to have quite as good an AC filter as the Ultra does, I really don't think that this could account for everything I heard. I tend to think that either it's because of something having to do with the way the outlets are wired, or because API is right about the advantages of plugging everything in centrally, but I can't know for sure. All I know is that for now, I'm sticking to the one main AC cord for the system, even if Sean is technically correct (and he always is), and am not going to lose any sleep over it. Good night. :-)

Zaikesman: I'm human and make mistakes too. Wanna see some threads and posts to prove it ? : )

As such, most of my comments are based on either first hand experience, common electrical / electronic theory, logical deductions and / or any combo of the above. With that in mind, nothing in audio is set in stone and one REALLY needs to try things out for themselves within the confines of their own system to see what works best. I have seen / heard components and cables sound like hell in one system and really compliment the performance of another system. You really don't know until you try it.

Having said that, i would be curious to see if the outlets that you plugged the amp into were on the same breaker and circuit that the line level gear was plugged into. Can you verify this and let us know ? It is possible that one set of outlets was on one leg of the circuit and the other outlet was on the other leg of the circuit. The fact that you have also introduced a completely new device and source of filtration into the system could also introduce another variable into the equation.

Is there any way to plug the amps directly into the same outlet as the line level gear is plugged into using one heavy duty extension cord with some type of outlet splitter ? I would be curious to see what happened in such a situation. Sean
>

Sean: Your first two paragraphs are a given of course - I was jes' funnin' (and am not one to talk when it comes to freely dispensing 'authoritative advice' that amounts to little more than my take on common sense, bountious caveats included gratis :-). You may still well be correct in your basic presumption here, I just may not be able to take advantage of it.

As I say, the possibility of the second PLC tossing a spanner in the works does exist, but I could test this by plugging it in to same wall duplex as my regular one and trying this hook-up again, and then doing what you suggest in the last paragraph. The outlets in question are confirmed as being on separate breakers, but on the same side of the box.

I did notice something else today though - the two outlets are oriented 180deg. upside-down from one another, so that the polarized wide plug-blade slots are on the left with one duplex, on the right with the other. Since the adaptors I'm using are also polarized, and I am grounding them at the center screws in each case, this means that by necessity, both the 3-prong PLC wall plugs were plugged in oriented in opposite directions. I don't know how wiring behind the outlets looks, so I don't know for sure if this really resulted in the two PLC's being plugged in with reversed respective polarities, but the evidence is suggestive.

Reverse AC polarity within a system can play a lot of games. One may not realize these "problems" exist until the situation is corrected. Once that takes place, the "sonic haze" has been lifted and you can now hear things that were previously not noticed.

I'm sure that you've seen this before, but take a look at this thread about Noise, Hum and AC Polarity to really sort through what's going on with your system. Don't forget that you have to remove ALL interconnects from the system as you can't have any of the component's chassis' tied together. Sean
>

Now that I think about it though, wouldn't the fact that everything else in the system was switched to balanced AC via the PW's iso-trannys render the AC polarity of the power amps moot? Anyway, I realized that it's more difficult than I thought to do the above tests using just the main system AC duplex, due to the fact that I can only install one polarized 3-prong adaptor so that it can be grounded at the center screw (unless, I guess, I run a little piece of wire around from the second adaptor's ground connection - if I get that motivated, I'll see what I can do. I suppose I really should at least replace that one duplex with a proper 3-prong receptacle variety. :-).