Anyone add an isolation transformer to a solid state amp?

Discussion in 'Amps/Cabs Tech Corner: Amplifier, Cab & Speakers' started by doc, Jul 11, 2016.


  1. doc

    doc Member

    Messages:
    5,187
    Joined:
    Jan 26, 2005
    Location:
    Middle Tennessee
    Just curious if anyone has tried this. Apparently some of the "tube tone" of guitar amps we're so fond of is due to output transformer saturation, and my thought is that you could just add an isolation 1:1 transformer to the output of a transistor amp between the transistors and speaker and add "instant tube tone". I'm probably missing something here. Anyone try it or know some problem with that approach?
     
  2. Malcolm Irving

    Malcolm Irving Member

    Messages:
    748
    Joined:
    Aug 27, 2014
    Location:
    Derbyshire, UK
    Great idea for an experiment. I've got my own thoughts about whether it would make it sound 'tubey', but I'll wait and see. (Been wrong too many times before! ;)) I can't see any problem - in fact I can see a couple of advantages:
    Some faults in SS amps can put DC through the speaker and blow the speaker (this would prevent that).
    Secondly, if you had an SS amp which produces say 100W into 4 ohms, but only 25W into 16 ohms, you could use a 1:2 transformer (like the ones used to convert 110V to 230V) to extract maximum power to a 16 ohm speaker.
     
  3. xtian

    xtian Member

    Messages:
    1,522
    Joined:
    Dec 15, 2010
    Location:
    NorCal
    You're suggesting the 1:1 trans might allow for "sag" as you approach it's max available current?

    I was just repairing a Peavey Series 260C bass combo, and I had a 60 watt bulb limiter in-line. I never bothered to remove the bulb limiter during testing, because all functions were working fine! My "kill-a-watt" was showing about 15 watts, and the bulb was glowing dimly, except when I cranked up the amp, then the bulb would glow brighter.

    Perhaps this would allow for some sag at high volumes, or you could adjust the wattage of the bulb as needed.
     
  4. doc

    doc Member

    Messages:
    5,187
    Joined:
    Jan 26, 2005
    Location:
    Middle Tennessee
    It should allow for some compression and reduced transient response, maybe a little sag. Several of the things we tend to like about tubes over transistors are contributed by the transformer rather than the tubes themselves.
     
  5. Jeff Gehring

    Jeff Gehring Silver Supporting Member

    Messages:
    3,297
    Joined:
    Jan 17, 2008
    Location:
    Kansas City area
  6. doc

    doc Member

    Messages:
    5,187
    Joined:
    Jan 26, 2005
    Location:
    Middle Tennessee
    I was going to try polishing it, but glitter might be better.
     
    Steppin' Wolfe and Jeff Gehring like this.
  7. darkfenriz

    darkfenriz Member

    Messages:
    214
    Joined:
    Sep 4, 2015
    Location:
    Warsaw, Poland
    How exactly do you guys associate the presence of a transformer with 'sag', presumably meaning DC supplies sag under load?
    I don't see a connection here...
     
  8. wall_of_sleep

    wall_of_sleep Member

    Messages:
    693
    Joined:
    Jul 31, 2011
    Location:
    Omaha, NE
    Most solid state amps I've opened already have a transformer in the supply.
     
  9. pdf64

    pdf64 Member

    Messages:
    4,509
    Joined:
    Aug 9, 2008
    Location:
    Staffordshire, UK.
    Yes, I think that the use of the term 'saturation' here is one of those situations where the colloquial and technical meaning of a word are rather different, or at least the colloquial use is generic covering a multitude of effects, whereas the technical meaning is specific.
    No physical component is ideal, all have non linearities and parasitic effects to consider, but for transformers (electromagnetic components with soft iron cores), these effects are way more significant than with most other passive components, eg resistors, film caps.

    My understanding is that a transformer whose core is under continuous saturation would be pretty much useless, with massive losses / little signal output.

    I suspect that actual core saturation is unusual in a guitar amp OT being used for guitar, but that hysteresis and other core and copper losses act to add distortion and 'colour' the output of a transformer, at levels below saturation. In addition, the non-ideal aspects of the circuits connected to the transformer may interact with the transformer's non linearities.
    The above effects (and more!) will combine and and result in the particular sound of the amp's circuit implementation.
    Here's an interesting article on the general topic http://sound.westhost.com/articles/audio-xfmrs.htm

    I think that when trying to analyse an amp's performance at a non-trivial level, it's beneficial to consider both the schematic, which shows the circuit with 'ideal' components, and it's implementation, and so should attempt to factor in the more significant parasitic effects of assembling that schematic in the real world.
     
  10. darkfenriz

    darkfenriz Member

    Messages:
    214
    Joined:
    Sep 4, 2015
    Location:
    Warsaw, Poland
    Guys, please keep in mind that the transformer is very often just a component in a negative feedback loop system.
    It is very often the case that the loop manages to correct both non-linearities and bandwidth limitations of the transformer, until it doesn't and makes the transformer characteristic visible again.

    In a slightly over-simplified model one could assume the transformer effects are not there until the whole power amp clips and then the transformer attributes suddenly start to appear. This can be a more vivid effect than the static transformer distortions themselves.

    So - I would say it makes all the difference whether you consider a feedbackless or feedback amplifier with a signal transformer.
     
    Ben Furman likes this.
  11. doc

    doc Member

    Messages:
    5,187
    Joined:
    Jan 26, 2005
    Location:
    Middle Tennessee
    I suspect that Peavey, Carvin, et al. have had the use of an OPT occur to them, the problem is that transformers cost considerably more than a handful of resistors, caps, etc. It may be that they tried it and found it wasn't a good solution sonically, but I suspect that it was more they realized on the front end that financially it was a non starter so they didn't really ever start with that approach. Anybody know of a good cheap source of a 1:1 (or close to that) transformer of the appropriate power handling ability for a 30 to 50 watt solid state amp? I'm not curious enough to shell out $100 for an experiment, but if I can get it down to the $10 - $25 range I might give it a try unless someone sees a reason I'd be at high risk of breaking something. The plan would be to simply hook the output of the amp to the primary and the speaker to the secondary and A/B the setup at different volume levels. Nothing like real world experience.
     
  12. Silent Sound

    Silent Sound Member

    Messages:
    4,170
    Joined:
    Jun 8, 2012
    At that price, you're going to have to find something used. And it probably won't be an audio transformer either.

    Still, I've experimented with this in the past, though not for speaker outputs, but rather for line level outputs on recording equipment. Here's my discovery:

    A 1:1 transformer won't sound much different than no transformer, except you'll likely lose a little bass and treble. Transformers tend to roll off at the upper and lower extremes of their frequency range. It serves a purpose if you're trying to balance an output, match impedance, or isolate DC, but otherwise doesn't do much good.

    Transformers can make a difference in tone, but only when the circuit is setup to take advantage of the transformer. An isolation transformer thrown on to the end doesn't really do that. You need to have some pretty significant step up or step down for the transformer to actually start to work it's magic.

    If the transformer is playing a significant role in the sound, then you'll likely want to use a nice one. Cheap transformers will sound cheap (usually harsh and metallic). Good sounding transformers are expensive. And most newly made audio transformers these days are made to sound as transparent as possible.

    Transformers don't really sound as great as you think. There's a reason that amp manufacturers use tubes instead of transformers for the phase inverter, even though a transformer would also work (and with far less complexity). There's also a reason they use capacitors on solid state outputs instead of transformers, and cost isn't as big of a deal as you would think.

    What I'm saying is, experiment with the idea if you like. It's a great way to learn how stuff works. But I wouldn't waste a whole lot of time and money on it. I've played with SS amps a good bit in the past and found many different ways to improve the sound of most of them. But that usually involves carefully analyzing the schematic and identifying and beefing up the weak points. And that can require a bit of working knowledge of SS amplifiers, which isn't a weekend project. Just inserting components at random is much more likely to make things worse than better. But if you really want to play around with this, it might be a better idea to put a 1:1 transformer on the input, rather than the output of the amp. That way you wouldn't need anywhere near as big of an transformer and could save yourself a ton of money. At least that would give you a vague idea of how much of a role a transformer plays in the tone and might help you to decide if it's worth spending more money on.
     
    UsableThought likes this.
  13. doc

    doc Member

    Messages:
    5,187
    Joined:
    Jan 26, 2005
    Location:
    Middle Tennessee
    Thanks for the input. I'm all for this not costing much, but if you put the transformer at the input, I don't think you'd drive it into saturation and get the same effects. I have thought about using a smaller value transformer somewhere after the VAS, but my knowledge of SS amps is limited and I'm worried I'll fry something in the process.
     
  14. J M Fahey

    J M Fahey Member

    Messages:
    2,163
    Joined:
    May 22, 2009
    OTs do not saturate over the audio range at guitar frequencies, so what does not happen canĀ“t have an audible effect.

    Magnetic saturation is an inverse function to frequency, so any saturation will appear at the lowest of the low frequencies, or at subsonics.

    The effect correctly described by Peavey is something else, and once you reread their explanation, it "clicks" in your mind.
    We are talking guitar amps of course:

    MANY use cheap undersized transformers, the emphasis being on **cheap** .

    As such, they are made as small as possible (save iron) , with as few turns as possible (saves copper) , both combine to reduce primary (parallel) inductance so we lose bass.

    Another problem is that simple wound transformers have a lot of (series) parasitic or leakage inductance; also relatively large layer to layer capacitance, both combine to cut highs.

    Solution for the latter 2 problems is to interleave, meaning both primary (plates) and secondary (speaker) windings are split and literally interleaved (think mixing a stack of cards) where you wind a couple layers primary, then some secondary, then some primary again and so on.

    Since every time you switch wire size (primary thin, secondary thick) you waste a lot of time resetting the winding machine many times (I know, I wind my own transformers), so again cheap transformers are wound like glorified power transformers, or with very simplified interleving, say one plate primary, single secondary, second plate primary.
    Now you are losing highs.

    So Peavey properly calls cheap transformers a bandpass circuit (good mids, not so good lows and highs), unexpected result is that it removes some harshness and mud.

    Put that transformer in a tube amp, add some feedback, and it will work better than expected, good for clean sounds, but overdrive that power amp, NFB goes down the drain, and bandwith narrows to its "real" value.
    It definitely is part of "guitar tube sound" .

    Now: adding it in a passive way, between SS power amp and speaker?

    Well, if cheap enough, it will cut some bass and treble and even maybe cut some harshness, why not?
    But in that case it will do so all the time, so sound in general will be somewhat poor.

    Given that, you might just use cheap speakers and call it a day ;)
     
  15. Adub

    Adub Member

    Messages:
    1
    Joined:
    Dec 31, 2017
    No, it won't make your amp more tubey. It will make it sound fuller and louder. It will isolate the power amp from the frequency dependent impedance fluctuations of the speaker. These fluctuations are what make solid state amps sound brittle and shrill when over driven as the highs are being accentuated and the mid lows and lows are less powerful. If you want tube sound, buy a tube amp. If you want to see what your ss amp really sounds like, slap an output transformer on it, or better yet, for the price of a resistor and some solder, convert it to current drive. Then the power will follow the speaker impedance fluctuations in a positive way and you will have a ss amp that is watt for watt louder than tube.
     
    darkfenriz likes this.
  16. doc

    doc Member

    Messages:
    5,187
    Joined:
    Jan 26, 2005
    Location:
    Middle Tennessee
    Hey, thanks Adub. That's interesting stuff. Any advice on more info on converting to current drive? I googled briefly and came up with enough info to get the basic idea, but not good leads on where to find practical tips for implementing this on a guitar amp.
     
    darkfenriz likes this.
  17. teemuk

    teemuk Member

    Messages:
    2,823
    Joined:
    Apr 21, 2009
    Unfortunately, transformers - as is - contribute very little to any "tubey-ness". As was mentioned, in practice saturation is a rare occurence and basically sounds horrible when happening.

    Last time someone actually cared to test actual effects of transformer saturation (this was with a tube amp with moderate output power and decent output transformer) the results were that it happens at frequencies below usual guitar input (e.g. 30 Hz), and introduces plenty of very high order harmonics because of the clipping mechanism where out-of-phase current and voltage basically just randomly "chop" pieces away from the waveform. Believe me, it will sound way more nasty and shrill than any "hard clipping" of waveform peaks.

    But usually saturation will not happen under typical operating conditions, which is nice.

    Yes, a little cheap output transformer might saturate much easier (though the aforementioned band-pass -effect is probably way more prominent) but then again the most common modification for small tube amps is an upgrade to decent output transformer. Wonder why..

    And speaking of solid-state and output transformers,,,, Yes, they were very common at a time. A typical PA amp from 1970's usually featured an impedance matching (auto) transformer - especially if it was designed for 70V or 100V speaker systems. There are also examples of SS guitar amps with transformer coupling: Univox, Marshall, Trace Elliot, Peavey... They all made SS guitar amps with output transfromers. Nope, no one regarded it as a big deal or heard them introducing any apparent "tubey-ness". In fact, I'm pretty sure that most people were completely unwarare that some of those amps even featured transformer coupling. In practice, any decent output transformer is virtually "transparent" one could say.

    It just matches impedance. That's all.

    There is a difference in designs where a transformer is used as a plate, collector or drain load vs. using it as cathode, source or emiyyer load, and it is principally the difference between current "follower" amps and voltage amp topologies. This circuit architecture difference is substantial. The push-pull vs. single-ended architectures also contributes a lot to circuit behaviour, but even in that using transformer is not mandatory for any of those effects.

    Today, with solid-state stuff, we can easily do without them.
     
    Last edited: Jan 2, 2018
    HotBluePlates and darkfenriz like this.
  18. teemuk

    teemuk Member

    Messages:
    2,823
    Joined:
    Apr 21, 2009
    Current amplification is much more typical to solid-state -guitar- amp than to any other solid-state amp. For aforementioned reasons. AND it has been that since late 1960's and very, very prominently since early 1980's. In fact, I wager a guess that about 90% of SS guitar amps would require no additional implementation whatsoever.

    And, the effects actually have very little to do with using an output transformer and all in the world with effective output impedance. Why this relates to transformers is that winding resistance (slightly) increases output Z (the effect is even slighter with transformers matching low "SS impedances"), and phase shifts introduced by the transformer limit the magnitude of NFB one can apply, which in turn limits how low effective output Z can be brought with negative voltage feedback.

    That's all. There's nothing magical in those components.

    All the transformer can do is reflecting impedances of the windings. It will not "isolate" these fluctuations at all, it will simply reflect them as usual. All in all, effects of output Z have basically nthing to do with transformers...

    And for "tubey" effect we actually want moderately high-ish effective output Z so that the "fluctuating" impedance actually causes the amplifier to introduce something else than "flat" frequency response to load in question. If anything, adding a coupling transformer can only enhance these effects due to its contributions to increase effective output impedance.
     
    Last edited: Jan 2, 2018
    HotBluePlates likes this.
  19. teemuk

    teemuk Member

    Messages:
    2,823
    Joined:
    Apr 21, 2009
    This was the main application of output, impedance matching, transformers in SS. But... DO NOTE that for anything other than lowest of "lo-fi" one needs a transformer specifically designed for audio and for entire band of audio frequencies. Power transformers are nice for 50 - 60 Hz but more or less useless for audio.

    You can still buy such transformers from some suitable vendor. Peavey, for example, had such product few decades ago.

    The "modern" (as in, employed for few decades already) approach is lighter and more cost-effective: Simply alter the supply voltage that powers the power amp section: Lower supply voltage for lower impedance loads, higher supply voltage for the higher. Do the math. It works.

    One only needs filter caps and output transistors rated for higher voltage (no problem) and in return a very, very expensive and bulky component can be eliminated as redundant.
     
    Last edited: Jan 2, 2018
    HotBluePlates likes this.
  20. pdf64

    pdf64 Member

    Messages:
    4,509
    Joined:
    Aug 9, 2008
    Location:
    Staffordshire, UK.
    Modern solid state 100V line amps seem quite common, they're intended for sound distribution systems in shops, pubs etc, and have output transformers.
    The one I had a nose around in had multiple taps on the OT secondary (100V, 60V etc) the lowest of which was 8 ohms.
    This type of thing http://avslgroup.com/assets/manuals/9/5/953115UK.pdf
     

Share This Page