Tube amplifiers and interstage transformers


I have seen that a lot of very expensive tube amplifiers use interstage transformers to handle impedance between stages. There are other two methods to handle impedance between stages: capacitor coupling and RC. I would like to know if somebody with more knowledge can explain the difference in the sound of these 3 styles.

Somehow, it looks like some of the most expensive SET amplifiers have interstage transformers like Audio Note or Shindo.

I also read about the highly regarded amity and karna amplifiers that D Olsen built and they have also interstage transformers.

So, what are the benefits of using interstage transformers?


I am sure that one can find a lot of people in either camps -- those that hate "iron" (limited bandwidth, phase shifts, etc) and those that hate caps. I have not personally experimented with design, so I cannot speak to theoretical advantages. It just happens to be the case that the amps I like are in the "lot of iron" camp. I have a pair of Audionote (uk) Kageki amps. These are interstage coupled.

I also own a pushpull amp based on an old Western Electric 133-A design, and a linestage that is also based on a Western Electric design. The linestage has input transformers and output transformers (no caps other than in the power supply) and the amp has input transformers, and again, no caps other than in the power supply.

My phonostage has a step up transformer in lieu of an active gain stage for low output moving coils.
Have to agree with Larryi, definitely two opposing camps.In theory the
limited bandwidth argument against transformer use makes some sense,
until you put theory aside and actually listen.In my experience the best
sounding amps and preamps more often than not take the high quality iron
route. Tube amps with interstage transformers sould exceedingly natural
and realistic compared to other topologies.They are utterly transparent and
have superb tone and body. My 2 cents worth.One of the
very best power amps I've ever heard, the Absolare Passion PSET 845
tube has the IT and is stunning! Both my amp and preamp use them and
wonderful to me.
Out of curiosity, it appears as though the interstage tranny is a big help with impedance matching. What I don't understand is why bandwidth is limited within the realm of human hearing i.e,why they even have this effect. Seems like you should get cleaner more powerful signal from the stage that needs to be matched to the next stage when impedance is better matched.
In reality you do get the advantages you describe despite what opponents
of the IT may claim.As alan pointed out, Many of the very custom and cost
no object designs use them .For good reason, when implemented
appropriately they produce excellent sound. I do recognize that other
designs can provide excellent sound also.Interstage transformers are a
time tested and proven pathway for superb sonic results.
I agree with Charles1dad and Larryi, most of the best tube amplifiers I have heard have interstage transformers. And like I said before, some of the highly regarded ones also have interstage transformers.

It looks like if you want your interstage transformer not to reduce the bandwidth or affect negatively the sound you need to buy a really good one.

Still, a lot of amplifier designers and people with high knowledge of audio will go for no interstage transformer approach.

The issue is that besides the impedance matching Interstage transformers bring a kind of naturalness to the sound, question would be how?

I have a 45 SET amp built by Bob Danielak and it sounds amazing. It does not have interstage transformers at all.

Some of the Audio Note kits have interstage transformers and people really talk great things about them.
As is usually the case the key is execution. A top quality IT used by a talented builder and placed in a good circuit will turn out fine.I certainly can't offer an explanation but this proper use of IT yields such naturalness and striking realism.My amp's builder Israel Blume much prefers the IT rather than a capacitor in the signal path. Of course other builder/designers will feel the opposite and have their own approach.I heard my amp paired with the Atma-Sphere MP-1 and the sound was also fabulous. My line stage has an IT and is sublime with this amplifier.There's always different paths to great results.
Regardless of the technical hows, that is the bottom line: it depends on the implementation.

Maybe I could ask Bob Danielak to build a SET IT amplifier for me so I can compare for myself at home. Maybe one with the darling 1626, which by all accounts looks like a "magical" tube in the presence and soundstage area.
The use of the term "impedance matching" isn't really appropriate for the majority of interstage transformer usage.

Typically output transformers are used to "match" the high impedance of a tube to the low impedance of a speaker so they usually have rather large stepdown ratios.

In the case interstage coupling, tranformers with ratios around unity are typically used. A 1:1 doesn't help with impedance matching, Small step up ratios make the impedance matching worse and it is only the transformers with a stepdown ratio that help in the impedance matching situation and you typically see them used for the drive of an A2 grid.

Thanks for jumping in Dave with a real technical explanation. Could you help us here giving us an idea why interstage transformers when properly implemented add a kind of natural and openness to the sound?

This might not be an appropriate response but system synergy is the only thing that can give you what you want.

The problem is, you need to know what you want prior to asking the question or you risk a 1000 post thread from others telling you what they think you need.

Intactaudio (Dave) is brilliant and balanced. Read his posts a couple of times to make sure your getting what he is saying.

from what you have experienced with the Duelund caps, what do you think a "cap coupled" amp using a cap that was that good do for a challenge to the same amp with an IT? Theorize of course.
I've had the same hypothetical thoughts.I believe either one can be done with a fine result.Given my deep satisfaction with the use of IT in my Coincident components, I'd probably favor the transformer approach.It's so wonderfully natural .For all I know it could be as good with CAST caps in their place.It would be fun to build both examples and listen.It's interesting to note that many cost no object designs choose the IT when they could use the best capacitors money can buy.
Dave, thats a bigger truth, system energy affects everything and is the ultimate sound shaper.

And yes, it would be nice to know what exactly are you looking for, but in reality (at least me) I know mostly what I do not like and what I could like. For sure, while I listen to more and more different speakers and different kinds of amplifiers i can narrow my options, but still is not so easy.

I've not found that transformers contribute to transparency. IME they always detract from it. Bandwidth aside that has been the main reason I avoid them if possible.

Distortion is however a main reason people will often prefer transformers, as these days of inter-stage transformer use it is mostly in single-ended circuits. The type of distortion I am referring to is the lower ordered harmonics, the 2nd, 3rd and 4th.

We have built a number of OTL guitar amps, and one of the problems we ran into was the fact that the amp was transparent and not a lot of distortion. This is great for a high end audio amp but bad for guitar. So we built an OTL that used a single-ended driver circuit and an inter-stage transformer to go from single-ended to push-pull. This got us a lot more 2nd ordered harmonics, which made the amp sound richer.

Sunn used this technique in their early transistor guitar amps, which have a reputation amongst guitar players as one of the few transistor amps you can use with electric guitar, due to the richer sound of their single-ended circuits (which were also mostly zero feedback).

When you are using a transformer for audio purposes, loading of the transformer is a big deal. If not loaded at all, the transformer will merely express the inter-winding capacitances rather than the turns ratio. This will result in poor bandwidth and distortion. As you load the transformer more and more, approaching a value known as 'critical damping' that bandwidth will widen and the distortion will drop. Because there is not a lot of power involved, the bandwidth of the transformer can be quite decent, easily exceeding 100KHz, so the real issue is how well it plays bass.

In SET circuits this is usually not much of a problem because once you get above about 7 watts or so you aren't going to see much low frequency bandwidth out of the amplifier anyway. Plus, to take advantage of the SET's ability to have unmeasurable distortion (in a lot of SETs distortion decreases linearly to unmeasurable as power is decreased towards zero) you need a really efficient speaker (103 db or more) and you are not going to get much bass out of most of those speakers either.

In order to get a condition of no phase shift in the audio passband (20-20KHz), the amplifier must have bandwidth from 2Hz to 200KHz, IOW 1/10th to 10X of the frequencies to be amplified. There are very few tube amps with bandwidth to 2Hz, and not many speakers that go to 20Hz, so this bass thing is where most of the fudging occurs in tube amplifier design, which is exacerbated when interstage transformers are employed.

Phase shift in the bass region translates to a loss of impact to a frequency up to 10X the frequency of the cutoff. So if the amp rolls off at 15Hz its a problem up to about 150Hz or so. This is the other main objection I have, as the speakers I play at home go to 20Hz. I played bass in the orchestra for a long time and as a consequence I have found that I am hard to please in this regard, hence I stay away from transformers in general.
This is why I hope you`re able to hear the components that use an IT and listen for yourself.I appreciate explanations from the design/engineering perspective but nothing beats actually hearing the end result.Andy Grove an Audio Note engineer has written very compelling arguements for the advantages of interstage transformers. Both sides make good cases. Listening is the best solution.
Thanks a Atmasphere for his technical explanation of his experiences with transformers in amplifier design.

And yes Charles1dad, I will have to listen for myself and compare.

I have listened to some amplifiers that have ITs and I liked their sound. Mostly, I listened to them not in a proper listening session but in a demo room.

Best way to get out of doubt will be to get me one and live with it for sometime.
That`s good to know and the right direction to take.Based on theory,measurements and the viewpoint of designers of solid state components,tubes should`nt even be a consideration for best sound aspirations.They`re too colored,too much distortion,outdated and measure poorly. Heaven forbid those comptemplating the SET amplifiers(worse of the tube lot).Funny thing is with all this logic, SET are to people like me the most realistic and natural of any type of topology I`ve heard.The very best sounding system I heard in Las Vegas a few weeks ago(CES and T.H.E. SHOW) was the Absolare-Rockport at the Flamingo Hotel.

The Absolare Passion was a PSET with 845 tubes(and an interstage transformer!), it was just stunning good and utterly realistic and sublime. Other 'exceptional' systems in comparison came off as only very good "Hifi". You just have to listen and trust your ears and spontaneous reactions.I`m sure no way, no how this amplifier has bandwidth or measurements that would compete with many other amps I heard(it`s SET, how could it?). I believe what matters most for excellent sound are qualities we don`t/can`t measure.Alan, you say your SET sounds amazing, I believe you.

Based on theory,measurements and the viewpoint of designers of solid state components,tubes should`nt even be a consideration for best sound aspirations.They`re too colored,too much distortion,outdated and measure poorly. Heaven forbid those comptemplating the SET amplifiers(worse of the tube lot).

Charles1dad, I can't agree with that statement, unless you exclude the 'theory,measurements' part of the first sentence.

The human ear is more sensitive to odd-ordered harmonics than just about anything else, as it uses them to determine the volume or loudness of a sound. This is why transistor amps sound bright even though the odd ordered distortion coming out of them is very low. Its just that the odd ordered harmonic distortion coming out of a tube amp is usually a lot lower, which is why they sound smoother.

SETs in particular have an additional quality which is that all forms of distortion they make decreases to unmeasurable at lower power levels (our amps do that too FWIW), something that transistors don't do (below a certain power level, distortion rises with them) and also something that most push-pull tube amps don't do (the traditional thinking being that this is caused by the phase splitter circuit).

IOW, the **theory** says that an SET (or other amplifier with similar abilities) should sound better to the ear. It is unfortunate that so much weight is given to specs on paper that don't mean anything to the human ear, which is the final arbiter. It is possible to create and measure specs that *do* mean something to the ear (IOW have a direct correlation to what we hear) but the industry has zero motivation to do anything like that. So for now, SETs will have to 'measure poorly' as you put it :)
I understand and agree with your post.My point is the common argument against tubes is their generally higher distortion and less than impressive measurements. It was my way to encourage alan to rely on his own hearing rather than what some pre conceived prediction or claim what the result will be.Alan says his SET amp is amazing and he reached that conclusion based on his ears.He did it the right way.
My experience is similar to Charles1dad's, in that my favorites are SET amps, especially 300B SETs. I've heard superb SETS with interstage transformer, and superb SETs with interstage cap.
Wavelength offers(I think,could be wrong) both IT and non IT versions of their excellent SET
amplifiers.I'd love to hear Gordon Rankin's opinion on how the two options
vary and why.
Charles1dad, something that has puzzled me for a long time is the also 'common argument' that triodes are the most linear form of amplification known to man, yet they have more distortion??

The fact is that they don't. Try putting a transistor in a single-ended zero feedback configuration and see what happens :)

The 'common argument' is apparently false, and should be addressed where ever it appears.

We found that by adding feedback to our OTLs we could get the distortion down well below that of conventional transistor amps, but in doing so the amp also took on some of the sterile qualities (read: brighter, less musical) that are hallmarks of amps with feedback. We chose at that point (about 1980) to eschew feedback; if the speaker thus did not allow the amp to sound right on that speaker, the conclusion is that the speaker would never sound right with *any* amplifier! This has limited our market, but the upside is that if a person sets up the amp right its hard to beat.

SETs have a similar 'weakness' (although I regard it as a musical strength) on the same account.

Transistors 'measure better' on account of the fact that the 'measurement' has almost nothing to do with human hearing rules and are simple a method of making the amp look good on paper.
Thanks Atmasphere, I realized some time ago that the present measurement standards and the sound quality have hardly any correlation. In some ways to a degree it's almost an inverse relationship. Stereophile's JA measurements compare to Art Dudley's listening impressions is an example.Measurements do have a relevant use, they just aren't useful as a predictor of sound quality. I don't know why the laws of human hearing are'nt appreciated more and respected.

Measurements do have a relevant use, they just aren't useful as a predictor of sound quality.

bingo! Measurements are objective, sound quality is subjective and there is a word that describes ambiguous blending of the two in order to manipulate.


I don't know why the laws of human hearing are'nt appreciated more and respected.

A lot of what we know about human hearing rules has only been worked out in the last 35- 40 years. But by 50 years ago, most of the understanding of how to analyze audio products was already in place. To get the really low THD specs you have to use feedback; when it was discovered that the human ear uses odd orders to detect volume (and also that trace amounts of odd ordered harmonic distortion is really irritating to the human ear), the fact that negative feedback was contributing to that became an 'inconvenient truth'; essentially the audio industry paid no attention and may have even swept the discovery under the carpet.

So for the last 45 years or so the spec sheet has told us nothing about how an amp sounds- in fact as you say may even show an inverse relationship to the sound of the product. But changing the way the audio industry does things is hard- there are a lot of companies that simply will not accept human hearing rules as important, even though they pay lip service.

The result is that audiophiles, no matter what they have read or heard from others (and no matter who those others are) have to audition things for themselves. You would think that the spec sheet would help in that purpose- what else could it be for??

Of course the issue is pursuit of the almighty dollar. Tube amplifier power is more expensive than transistor power and high efficiency speakers are more expensive than low efficiency speakers. But the hype is that the cheaper transistors and lower efficiency speakers are somehow better because of the 'specs'! Whatta world!
The thread has become really interesting!

I also believe that specs are just a way to measure audio, but those measurements do not have a direct relation with how we humans perceive and process sounds.

"So for the last 45 years or so the spec sheet has told us nothing about how an amp sounds"

Great Line!!

The mechanism that works inside our ears is not the same mechanism that works in the machines that produce those amp specs. We are a much complicated machine with lots of nuances to take into account.

As Atmasphere said, most of the companies need something to show or rumble about and most of the time is something related to specs, although some of them play the review game very well...

We just need to listen and that is always a pleasure.

Personally, I regard that the pleasure of listening to music in a good hi end system that is able to conjure the emotions of the musicians to our living room is a kind of spiritual quest that tries to find in sound a connection with the invisible, with something that lies beyond the organized noise, that somehow make us aware of some intuitive knowledge we have lost along the years as a civilization...

How can we measure that?
I found a website that produces ITs that has a documents explaining the advantages of this approach.