Many people like the sound of a tube amp where the tube rectifier signature is left in the signal path. That is personal taste, however the reputation for poor bass response is due to using PTX with too low a VA rating and rectifier tubes that can't deliver enough current for the design. Now that we have that cleared up, wouldn't one argument in favor of tube rectification be less filtering required? Some SS rectified designs have so much filtering they get sluggish by the time the switching noise is filtered out? I'm just throwing this out there, I have built/designed tube amps with either or, but do tend to prefer the tube rectifier with its slow startup which I think is easier on the amp tubes.
Last edited:
Quote:
So, let me try and understand you… you think that if a resistor sounds ‘lush’ then playing Motorhead through that resistor will make Motorhead sound lush????
Components do not have a ‘sonic signature’; they have electrical characteristics which must be understood, selected and blended to produce the desired circuit and system response.
---
If a resistor as has an impedance at AC frequency due to inductance then as you put this is an electrical characteristic.
It is the electrical characteristic that creates the signature of the component!
If you say you cannot hear the difference between polypropylene and polyester caps then yes I agree "in reverse" the circuit is a bad design!
You say that you are going to produce what is on the source CD Etc.
I say yes that is the aim, however what you get from your system and what another will get with the same amp with different components will be presented differently!
This could be detail, imaging, tighter base, improved treble or the reverse where yours system is better.
I have listened to Ongaku and money aside I was not a fan, many are! I prefered the PP circuits they had on display. Different harmonics.
Any way you are probably correct. Electrical characteristics do not effect sound only the Dielectric sorry is that a characteristic!
Regards
M. Gregg
So, let me try and understand you… you think that if a resistor sounds ‘lush’ then playing Motorhead through that resistor will make Motorhead sound lush????
Components do not have a ‘sonic signature’; they have electrical characteristics which must be understood, selected and blended to produce the desired circuit and system response.
---
If a resistor as has an impedance at AC frequency due to inductance then as you put this is an electrical characteristic.
It is the electrical characteristic that creates the signature of the component!
If you say you cannot hear the difference between polypropylene and polyester caps then yes I agree "in reverse" the circuit is a bad design!
You say that you are going to produce what is on the source CD Etc.
I say yes that is the aim, however what you get from your system and what another will get with the same amp with different components will be presented differently!
This could be detail, imaging, tighter base, improved treble or the reverse where yours system is better.
I have listened to Ongaku and money aside I was not a fan, many are! I prefered the PP circuits they had on display. Different harmonics.
Any way you are probably correct. Electrical characteristics do not effect sound only the Dielectric sorry is that a characteristic!
Regards
M. Gregg
like anything else - it depends. I've had good luck with both SS and tube rectifiers. My friends prefer tube rectification, but I personally think that is the "tube only" bias. I'm in the weird camp of thinking that SS is good for some things (MC carts and big power amplification) while tube is good for others (MM carts, linestages, low power for high efficiency speakers).
Back to rectification - for a SE (or Class A) amp, I usually stick with a tube rectifier like the 5AR4. For Class AB/high powered amplifiers - I like the "speed"/dynamics of solid-state like the lowly UF4007.
Back to rectification - for a SE (or Class A) amp, I usually stick with a tube rectifier like the 5AR4. For Class AB/high powered amplifiers - I like the "speed"/dynamics of solid-state like the lowly UF4007.
I think it comes down to R/C time constants.
If you go too big you will have problems, if you go too small you are going to have problems.
That is why power supplies are designed for the amp at hand.
Too low of filtering and the caps won't be able to keep up with the bass and sagging will occur.
Too much filtering and it can make it sound sterile and lifeless.
Get it just right and mmmmmm, that's it.
Makes me wonder about the SS amps, with their gazillion microfarads of filtering.
If you go too big you will have problems, if you go too small you are going to have problems.
That is why power supplies are designed for the amp at hand.
Too low of filtering and the caps won't be able to keep up with the bass and sagging will occur.
Too much filtering and it can make it sound sterile and lifeless.
Get it just right and mmmmmm, that's it.
Makes me wonder about the SS amps, with their gazillion microfarads of filtering.
I agree with all those that have said that components do not have a sonic signature of their own, but have electrical characteristics which must be understood before a reasonable design is achieved. I would add that the farther you get from the signal (power supply, line cord, house wiring, substation, generating plant) the less impact its characteristics will have on the amplified audio.
This goes for rectifiers too- once the DC supply is quiet enough to meet the demands of the design, it makes little difference whether the design is thermionic or solid state.
I also agree with (who WAS it, now) that said it's not always about JUST the electrical design. There is a certain amount of aesthetic to it too- otherwise we wouldn't be designing with tubes at all (oops- NOW I've said it!).
All that being said, I usually prefer solid state diodes since they're far more efficient than thermionic devices- higher output voltage from a given transformer and far less heat- we usually have plenty of heat in a tubed design as it is.
And there's usually enough nice, big, warm, glowing glass bulbs visible to satisfy the desired result. -I really don't miss seeing the rectifier tube enough to justify the additional heat. I admit 5U4s are visually appealing but personally I have a hard time justifying their use.
Cathode stripping is also really not an issue below several thousand volts so there's really no need to delay B+ while the heaters get up to temperature.
This goes for rectifiers too- once the DC supply is quiet enough to meet the demands of the design, it makes little difference whether the design is thermionic or solid state.
I also agree with (who WAS it, now) that said it's not always about JUST the electrical design. There is a certain amount of aesthetic to it too- otherwise we wouldn't be designing with tubes at all (oops- NOW I've said it!).
All that being said, I usually prefer solid state diodes since they're far more efficient than thermionic devices- higher output voltage from a given transformer and far less heat- we usually have plenty of heat in a tubed design as it is.
And there's usually enough nice, big, warm, glowing glass bulbs visible to satisfy the desired result. -I really don't miss seeing the rectifier tube enough to justify the additional heat. I admit 5U4s are visually appealing but personally I have a hard time justifying their use.
Cathode stripping is also really not an issue below several thousand volts so there's really no need to delay B+ while the heaters get up to temperature.
I had no problem justifying the use of 5U4GBs in one project. The design nominal Vpp was 350Vdc. The power xfmr I had in the junk box was a 650Vct @ 150mA device. With SS diodes, it overvolted badly. With a 5U4GB, it produced the design nominal DC (352Vdc to be exact). Why not use the 5U4 since it's obvious this power xfmr was designed for it, and the design would either require a different power xfmr for SS diodes or some circuit to dump nearly 100V anyway. Besides, I have essssss-loads of 5U4GBs and its compactron version on hand anyway.
Cathode stripping may not be an issue, but over voltage with DC coupled designs definitely is. That was another problem: the use of SS diodes to provide a negative rail. That comes up to some 450Vdc within a couple of seconds, way before cathodes can warm up. That exceeded the Vhk rating of the 6FQ7s I was using as grid drivers. Forch, I could float the entire heater supply without causing objectionable hum. Otherwise, some sort of arrangement with Zeners (and their nonlinear capacitances) would have been required to prevent poofing the HK insulation.
If you DC couple, best to preheat when using SS power supplies.
I use the same filtering regardless. Yes, you can filter more heavily with SS diodes, as these have much higher Isurge ratings. However, vintage power xfmrs weren't designed with this in mind, and may not appreciate Isurge way in excess of what designers could reasonably expect of hollow state diodes. Same goes for modern power xfmrs like those "Hammond Classic" power xfmrs which are recycled designs from the days before Si diodes were ever available.
Yes!
Yes, but the level and it's effect surely depend upon the competence of the overall psu design.
Not necessarily. If the designer has dumbly just bunged in SS diodes then he / she deserves the noise. But if the designer has chosen SiC Schottky diodes, or correctly applied RC snubbers, then used an inductor, then a voltage regulator (and perhaps isolated the current loop by a shunt regulator) then the power supply is likely to be a quiet and competent foundation upon which to build the amp.
So the key to it is competent design for the specific application (in my humble opinion).
J.
Jen-B,
Of course the circuit must be well designed, however that is not the be all and end all of it!
Quote:
But if the designer has chosen SiC Schottky diodes,
------
Please show us all one with 1000V 1A+ we would all like to use it!
The PSU you describe above adds components to remove problems with components, these "add" more problems even though they remove others.
How many valve amps do you see with complex power supplies?
Many "users" not builders will say they don't like the stiffled sound from regulated power supplies!
Then again some like them!
This is also true with DC or AC heaters they "sound" different. I do not mean that the music has some how changed Motor Head into a classical piece. The presentation has changed.
This is subjective to the listener. I watched a "blind guy" listening to an audio note kit amp on demo, we had accidently stood in the sound stage of the music. we were not talking, we were casualy listening. I looked up to see a frustrated person franticaly moving his head about. I walked over and apologised he said he could not "see" the music had something happened to the amp! I spent some time talking to him its interesting from a different perspective!
This can be even more of an issue with guitar amps. When the sustain and harmonic content change the overdrive sound. Yes guitarists have their idea of perfection and are just as crazy as audiofools.
You can fit components (a diode is a diode) no its not. A cap is a cap no its not. A resistor is just a resistor no its not! The design of a circuit is very very important, However a component catalogue does not say (non audiophile) this cap is smooth and lush. Then again it does not say this resistor has a high inductive value at AC frequency etc. Or the dielectric causes HF roll off at AC ?Hz. in this capacitor.
I managed to find in a TV repair shop some "Old new" paper plessey caps still in there plastic sealed bag "paper type". I substituted them in an amp, guess what they made it sound like "old valve" equipment. Sine / square wave looked exactly the same!
Regards
M. Gregg
Last edited:
So you believe that once the circuit is optimised and proven in it’s intended application (i.e. it is well designed) then that is not the end of it? Oh well, good luck with that philosophy.
Last year I used some with 600V rating, so two in series should work… or go for the new Cree C2D05120A (RS components 702-7487) with a 1200V, 5A rating.
No, and this is a big point, they do not add problems IF DONE COMPETENTLY !
I note with interest that many Valve amp builders adopt a self-defeating philosophy of ‘simple at all costs’ because it seems to satisfy their psychological desire for involvement with their equipment… it allows them to harp on about changing capacitor A and getting result B, and changing resistor X and getting result Y, etc., etc.. They seem to want to identify with a given philosophy so as to be “in the gang”. This seems quite common in hi-fi circles. The problem is they then spout nonsense which misleads those that are innocently following-on and trying to learn the facts. (Facts such as a rock solid power supply is more suitable for accurate amplification than a flimsy supply which allows you to “hear” the rectifier!)
This is another major point. Users (not builders) have no idea what aspect of a system is causing a particular quality (for even if only one part has been changed you always listen to the interaction of parts and not just the single part), yet time after time they draw pathetic and shallow conclusions which they then present as facts. This leads to misinformation, and after a while sad myths take hold, such as “all feedback is bad”, “all solid state sounds harsh”, “all silver cable is bright”, “all tubes sound warm”, etc., etc..
I guess that you and I will not agree on these things, so I’ll leave it to our readers to decide. My final comment is simply this: if you can hear the rectifier there is something wrong with your design. And that’s a fact.
Kind regards,
J.
Just yesterday I put the finishing touches on a single tube pre-amp that easily drives 1k loads, makes 0.01% predominantly second harmonic distortion at 1 volt rms and a tenth that at normal levels, is within a couple tenths of a dB across the entire audio band, and has a noise floor beneath my Audiophile Firewire at 24-bit, but you keep right on believing that was done to impress people instead of being better engineering than cookbook band-aid complexity.
Strange path this forum has taking.
"Facts such as a rock solid power supply is more suitable for accurate amplification than a flimsy supply which allows you to “hear” the rectifier!"
Power supply 1. GZ34 rectifier, choke, 600uF of large computer grade electrolytic caps. Difficult to hear any difference with a GZ32
Power supply 2. GZ33 rectifier, choke, 120uF of motor run caps. Difference with a 5U4G audible.
One supply sounded distinctly better than the other. Which, I leave to your imagination.
Andy
Power supply 1. GZ34 rectifier, choke, 600uF of large computer grade electrolytic caps. Difficult to hear any difference with a GZ32
Power supply 2. GZ33 rectifier, choke, 120uF of motor run caps. Difference with a 5U4G audible.
One supply sounded distinctly better than the other. Which, I leave to your imagination.
Andy
What is meant by "better"? More accurate as a reproducer, or more in accordance with my particular tastes in sound?
Having five times as much capacitance in a choke input supply is bound to make the rectifier characteristics less obvious. To a first approximation, the charging current pulse is set by the choke but also by the rectifier. The resultant voltage output ripple is set by the capacitance value. Less capacitance means more ripple, so more intermodulation in any less-than-ideal amplifier (especially SE!). Less capacitance therefore means rectifier characteristics more noticeable - exactly as you found. So 120uF is not enough capacitance for that particular amplifier - the 600uF PSU is better.
Audio is one of the few branches of engineering in which something which is demonstrably better can be considered to be actually worse! I am glad motor engineers don't work in the same way: the best car suspension would be the one which lets you feel every bump in the road (this shows that it is "transparent", rather than the "sluggish" ones which give a smooth ride).
Actually a real driver does like a little bit of road feel but that is a topic for another forum. There is a place for technical measurements and specifications but at the end of the day it is about the music that we hear. If it measures perfectly but the life of the music is lost then maybe we are looking at the measurements in the wrong way. If we improve the measurements and the music improves also then I suppose we have evidence that we are on the right track.
I don't have a dog in this show as I haven't enough design experience to warrant an opinion, but I think we need to be a little bit humble about our ability to describe everything 100% numerically.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.