Transformer coupled vs Capacitor coupled?


Guys,

I’m sure their are different schools of thoughts on these typologies but are transformer coupled better in terms of sound quality in regard to pre-amps.

Thanks,
Wig
574209a8 b664 45b0 ae5e d4471da41826wig
As with everything it's about implementation, good caps will sound better than bad trannies and vice versa. I've heard both types sound good my current preamp is capacitor coupled.
Transformer-based volume control in a pre-amp is probably the best sounding kind. Implementation is king though. Transformers on inputs/outputs bring system-wide benefits in terms of isolation from common-mode noise that capacitors cannot give.
wig
Transformer coupled vs Capacitor coupled?


So long as the stage before is good in design low output impedance and has more than enough current and all your doing is stopping any dc offset, then a good cap is far better than a good transformer that will always have coloration’s.

But then doing a dc servo instead and having it all dc coupled, is then better than both.

Cheers George
The problem with transformers is bandwidth. Usually in a preamp level situation their bandwidth can be fairly wide, but to prevent phase shift in the audio band the bandwidth of the preamplifier circuit has to be from 2Hz to 200KHz and even in smaller sizes this is tricky for a transformer to do. But its not tricky at all for a coupling capacitor as long as the timing constants have been taken care of around the capacitor.

As George has pointed out, it is also possible to direct-couple. The trick with direct-coupling is the servo circuit has to have good stability.  Many don't and so 'hunting' (oscillating around the correct value) can be a problem. Servo theory is a field of considerable girth bit if a servo is done properly there's no going back to coupling caps or transformers.

Thank you all for your informative posts, I now have a better understanding of those designs and their limiting factors utilized for cost saving measures.

Wig 👍


Tim de Paravicini uses transformers on the balanced/XLR inputs and outputs of his EAR-Yoshino 868 and 912 pre-amps, but of course not on the RCA's.
Tim de Paravicini uses transformers on the balanced/XLR inputs and outputs of his EAR-Yoshino 868 and 912 pre-amps, but of course not on the RCA’s.
I’m not 100% sure (please correct me) but I was told once, that was just to get xlr input and an output connections, but everything inside in the active tube circuit is SE (single ended) in between those input and output transformers.

Cheers George

That is correct George. The EAR-Yoshino circuits themselves are single-ended, the transformers are just to provide balanced connections, as is standard in recording studios, where Tim does a lot of work (Roger Waters’/Pink Floyd’s studio in England being one such).

Roger Modjeski provides a balanced input (only) on his Music Reference RM-200 amplifier, but he told me he did that with discrete components, not transformers. The RM-200 is also a single-ended circuit.

Ralph Karsten’s Atma-Sphere designs are true dual-differential balanced circuits, as far as I know unique amongst currently-available tube amps and pre-amps.

I believe it is Ralph’s contention that balanced circuits themselves provide benefits over single-ended, Tim and Roger that balanced transmission is what they find of significant value, not the circuit itself. About that I may be mistaken ;-) --- Eric.

I believe it is Ralph’s contention that balanced circuits themselves provide benefits over single-ended, Tim and Roger that balanced transmission is what they find of significant value, not the circuit itself. About that I may be mistaken ;-) --- Eric.
I doubt that either Roger or Tim would deny the benefits of balanced differential circuits- in a nutshell, rejection of power supply noise and also rejection of noise induced in the interconnect cables.
A secondary benefit is lower distortion as even ordered harmonics are cancelled with each succeeding stage. This means that the primary distortion product is the 3rd harmonic, which is treated by the ear in a similar manner to the 2nd (contributes to 'warmth' and 'bloom') but its level is at about 1/10th of a circuit that has the 2nd harmonic as its primary distortion component. So distortion components are not compounded from stage to stage as the signal progresses through the signal chain- this results in less coloration.
Basically it comes down to implementation of the design.  All of the designs have their benefits and detriments.  Coupling capacitors have a wide bandwidth but can give up dynamics, details, and transparency to transformer coupled.  With transformer coupled you can get fantastic sound at 20K but designs up to 40K is even better.  They look pretty bad looking at the square wave but you do not need 100 to 200hz to make them sound fantastic.  I have built both (and own both) and the transformer coupled just has more transparency and details.  I use V-Capacitors in the preamplifier as I prefer their sound to the others we have tried.  The Servo design also depends on the implementation.  I have heard some that sounded very poorly but the old Luxman power amp that I have is just incredible.

So is one better sounding than than the other - to me again design implementation is the key but each has its benefits if done well.  Coupling caps is a much easier design to implement IMO.

Happy Listening.

Sound quality
1st Direct coupled using a good dc servo (most transparent/dynamic)
2nd a Good quality Cap (large enough not to roll off the bass too early)
3rd Transformer coupled (using good transformers)

Cost to make, lowest first.
1st Direct coupled ($10-20)
2nd Cap coupled
3rd Transformer coupled.

Least distortion/coloration’s (lowest first)
1st Direct coupled (always lowest, if good design)
2nd Capacitor coupled
3rd Transformer coupled

Lowest output impedance (good to drive next piece of equipment)
1st Direct coupled (lowest from DC to wherever)
2nd Capacitor coupled
3rd Transformer coupled.


No-brainer, direct coupled wins out, cap comes in second.

Cheers George
Guys,

This is awesome and at some point would like to hear a direct coupled tube pre in my system! From my basic research, it's hard to discern if the pre is direct coupled on both in/outputs or if the transformers used are capable of a wide bandwidth; seeing vague terminology...

C or EI cored signal and power transformers are phenomenal but are pricey and would cost almost half of the cost of the pre if they are truly direct coupled.

I know it going to be a costly adventure and knowing that it's solidly designed with quality parts and layout may make the journey worthwhile.

Now the question still remains if more dynamics or transparency will come through as brightness or more forward sounding compared to capacitor coupling? But I guess one can still tube roll in pre or amp to attain the sound they prefer knowing that it's a much quieter pre : )

Wig
wig OP
This is awesome and at some point would like to hear a direct coupled tube pre in my system!
I believe from what he says, "maybe" Atmasphere’s pre "could" be "direct coupled from input to output" ask him precisely that.
I can’t believe there’s no coupling cap/s somewhere between input and output somewhere.
But then it’s a shame if the rest of the system is not true direct coupled also?


C or EI cored signal and power transformers are phenomenal but are pricey and would cost almost half of the cost of the pre if they are truly direct coupled.
Don’t get into the thinking that "coupling transformers" and "coupling caps" are some kind of "magic music making components", their not, they are in many designs a necessary evil to have. Proper direct coupling using a good dc servo control, nails them every time in all avenues of music reproduction except for one "coloration".

Cheers George
@ georgehifi

With direct coupling, are the type of tubes used important? I have owned only capacitor coupled and preferred those with 6SN7s over 12AX7, 6922 small variant tubes; just more body and dimension with 6SN7...

I don't have any direct coupling components, my KT88 Amp and tube Dac is capacitor coupled as well. Do you know of any reasonably priced DC pre-amps I can read up on and possibly add to my system?

Thanks Again,
Wig

Like I said ask Ralph (atmasphere) he the one who says he has a direct coupled tube preamp, whether it's "dc coupled from input to output" only he can say, and what tubes he uses.

Do you know of any reasonably priced DC pre-amps I can read up on and possibly add to my system?
Nearly all "good" solid state preamp should be DC coupled with dc servos, from input to output. It should say it in the features and specs (dc coupled throughout)

PS: Just looked at the Atmasphere preamps, they are "not direct coupled from input to output", there are capacitors in the signal path.
He only says the output is direct coupled. (just the output has no caps or transformers)

Cheers George


@ georgehifi

Thanks again and will do.

Wig 
I can’t believe there’s no coupling cap/s somewhere between input and output somewhere.
But then it’s a shame if the rest of the system is not true direct coupled also?
The problem with direct coupling from input to output is that the circuit goes lower in frequency (to DC) while the power supply (which is supposed to go to DC) in fact has a very low frequency time constant, probably several seconds. If the circuit can go lower than the power supply its possible to modulate the power supply which in turn will cause intermodulation distortion.


So its useful to limit the circuit’s ability to go all the way to DC. This also prevents it from amplifying DC should any be applied to the input of the circuit.


So we have direct coupled inputs and outputs, but coupling caps in between. The advantage here is that since the capacitors are buffered inside the circuit (IOW we’re able to control the impedances which they drive) we are able to keep their values relatively low. This helps with transparency; the larger value a coupling capacitor is, the more coloration it can impose (caused by inductance, ESR and dielectric constant).


When designing a high performance audio circuit, bandwidth is important to prevent phase shift in the audio passband. Generally speaking the circuit has to go to 1/10th the lowest frequency to pass and 10x the highest frequency to pass. So you need 2Hz to 200KHz if you really want zero phase shift. This is challenging if you have an output transformer or coupling cap; in the case of the former getting it to cover that much bandwidth is a real trick! In the case of the latter there is no way to know to what sort of amplifier the preamp will be connected so the coupling cap is often larger on that account, and as a result isn’t going to go all the way to 2Hz in some cases (probably with a solid state amp being driven).

Now in the case of vacuum tubes while it is *possible* to direct couple from input to output, its really impractical and would likely be a reliability issue. So you aren’t going to find a vacuum tube preamp that is direct coupled from input to output. You might find a solid state preamp that is, but even then there are practical issues (most of which I’ve outlined) which make such a thing unwise.


Because coupling caps and coupling transformers are thus a necessary evil there are quite a few boutique products of both types around. FWIW a good line driving transformer is more expensive than a good coupling cap so that is a good portion of the reason that output transformers in preamps are uncommon. But a secondary reason is its also more difficult to get bandwidth and unless one has been quite careful in the design and application of the transformer, higher distortion too.

@ atmasphere

Thanks for your response, the dots are starting to line up in my mind 😁

Outstanding dialogue from all!

Wig