Diodes...
I have been using ON MSRF ultra soft recovery diodes lately. The specs show these to be about as quiet as they come for a given rating, and I like the resulting sound, so have not felt a need to try anything else.
I think one should consider that diode switching noise definitely broadcasts into the enclosure, and as such could be picked up elsewhere in the circuit, despite, perhaps, being filtered by the reg(s). Best to keep diode switching noise to a minimum when possible.
Anyone else try these diodes?
I have been using ON MSRF ultra soft recovery diodes lately. The specs show these to be about as quiet as they come for a given rating, and I like the resulting sound, so have not felt a need to try anything else.
I think one should consider that diode switching noise definitely broadcasts into the enclosure, and as such could be picked up elsewhere in the circuit, despite, perhaps, being filtered by the reg(s). Best to keep diode switching noise to a minimum when possible.
Anyone else try these diodes?
I have been talking about my experience with the 11DQ06. At 0.15 cents each, I hardly think it is hard to obtain, exotic, expensive, or "high end".
Lets get that misconception out of the way.
Seems like there are many people on this thread who don't want to know why, and would rather spend time arguing the point rather than trying it.
What is a "signal line"?
Have you read in a book that rectifiers make no difference to sound or have you actually experimented?
In such discussions it's always useful to establish what one can actually hear on their system. Do you hear capacitors, resistors, chokes, termination? Can you hear difference between amps at all? MP3/WAV? Phono cartridges? Tonearm alignment? Absolute phase?
If you hear most of the above but not rectifiers your opinion will be useful. Otherwise...
It's has no place with the signal purely power. I don't need to read books to tell me that they wouldn't make a difference, what is there to make a difference? They don't emit noise, DC is DC after all.
How did you prove to yourself that you could really hear all the differences you claim? Did you just use your 'auditory memory' aka 'golden ears' or did you do a proper test to ensure you were not deluding yourself?
After you post details of your testing procedure, your opinion will be useful. Otherwise........
Boscoe:
if rectifiers were "perfect" I believe you would be correct. But nothing is perfect, and rectifiers do emit noise, and there is more than "perfect" DC on the output of a rectifier bridge.
Additionally, the power supply is directly in the signal path as far as I am concerned; the output signal is just a power supply modulated by the input signal.
Also, consider the fact that rectifiers also broadcast noise internally in the components case-there is the distinct possibility that this broadcast noise will be picked up through antenna effects in the output signal.
if rectifiers were "perfect" I believe you would be correct. But nothing is perfect, and rectifiers do emit noise, and there is more than "perfect" DC on the output of a rectifier bridge.
Additionally, the power supply is directly in the signal path as far as I am concerned; the output signal is just a power supply modulated by the input signal.
Also, consider the fact that rectifiers also broadcast noise internally in the components case-there is the distinct possibility that this broadcast noise will be picked up through antenna effects in the output signal.
I have used FRED and other non-standard rectifier diodes when the circuit designer specified them.
I've not tested whether they 'made a difference', but if it is a difference of a few cents (or even a couple of dollars), there is little to lose...and I'm assuming that the designer has a good reason for the spec. I doubt that I could hear a difference, but as I say, I haven't tested that hypothesis. I have tested coupling caps, and could not hear a difference, so I use the ones that look the best in my amps - USSR PIOs with lots of Cold War aura!
If George Anderson (Tubelab) calls for a specific component, I try to listen. Other than that, it's off-the-shelf 1N4007 or bridge rectifier modules for me.
I've not tested whether they 'made a difference', but if it is a difference of a few cents (or even a couple of dollars), there is little to lose...and I'm assuming that the designer has a good reason for the spec. I doubt that I could hear a difference, but as I say, I haven't tested that hypothesis. I have tested coupling caps, and could not hear a difference, so I use the ones that look the best in my amps - USSR PIOs with lots of Cold War aura!
If George Anderson (Tubelab) calls for a specific component, I try to listen. Other than that, it's off-the-shelf 1N4007 or bridge rectifier modules for me.
What type of noise is this? Not in a signal path, signal paths carry signals. More of a discussion although still pointless would be about capacitor smoothing and PS sagging due to transformers or rail regulation. A discussion on whether a rectifier affects sonics is completely absurd.
I just gave three
solid reasons why such a discussion is not absurd. If you still insist that such a discussion is absurd, clearly your time would better spent elsewhere.
solid reasons why such a discussion is not absurd. If you still insist that such a discussion is absurd, clearly your time would better spent elsewhere.
Just a bit of history. Digikey some years back introduced a new manufacturer of diodes. Until then no one much really noticed problems with diodes. Well these diodes made so much noise in switching off even some digital circuits had problems. The quick answer was that the diodes were good and switching transients were not specified for any diodes.
Some place I have a few of them the noise you see from them in a simple bridge circuit can be twice the power supply voltage for a MILLISECOND or so.
Now that prompted some folks to look at all diodes and some noted improvements.
The best kept secret is the 1N4148 it is good for 100 PIV and 250 ma. Costs pennies and has nice clean switch off, at least from the brands I have tried.
Some place I have a few of them the noise you see from them in a simple bridge circuit can be twice the power supply voltage for a MILLISECOND or so.
Now that prompted some folks to look at all diodes and some noted improvements.
The best kept secret is the 1N4148 it is good for 100 PIV and 250 ma. Costs pennies and has nice clean switch off, at least from the brands I have tried.
With what I've seen the last year or two here, about using active shunt regulator designs for lower noise (there was someone selling PC boards for these), it wouldn't surprise me too much about what some people find audible (perhaps even in a DBT!).
As possible "research" for my post yesterday I did a Google search I didn't use anything yesterday, but now after reading this post I can't help but post one of the first hits:
Diamp
As far as the 1N4148, I wonder what peak current it can handle - I can imagine someone reading that and making a low-current power supply using a 25-volt transformer that happens to be sitting around (capable of several amps short-circuit) with a couple 50,000uF capacitors for "good filtering," and the current at power-on being enough to make the thing light up like a light bulb during the first half cycle. "But I was only going to pull 100mA from it. "
Well, I have experimented a lot with diodes over the past months, and it certainly is audible in a standard CLC regulated low voltage power supply!
1N4148: cheap but nasty compared to Sic diodes, an upgrade in cheap equipment, but sounds 'hyped' IMHO. A bit too bright for me.
Vishay HEXFRED: excellent midrange, but loses some highs and little less tight in the bass. My choice for guitar amplifiers.
Cree Sic: best diode i have tried so far, unfortunately also the most expensive...still the difference is big enough (to me) to justify the price.
Diodes do make a difference in 99.9% of the appliances, because all commercial designs will cut costs at the PSU!!! (no snubbers, no elco bypass caps, single stage filtering before regulator)
Why do tube freaks (being one of them) still swear by tube rectifiers?
Why do most people prefer a linear psu over a smps?
all for the same reasons you can't hear?
just my (tried and tested) two cents...
Marc
1N4148: cheap but nasty compared to Sic diodes, an upgrade in cheap equipment, but sounds 'hyped' IMHO. A bit too bright for me.
Vishay HEXFRED: excellent midrange, but loses some highs and little less tight in the bass. My choice for guitar amplifiers.
Cree Sic: best diode i have tried so far, unfortunately also the most expensive...still the difference is big enough (to me) to justify the price.
Diodes do make a difference in 99.9% of the appliances, because all commercial designs will cut costs at the PSU!!! (no snubbers, no elco bypass caps, single stage filtering before regulator)
Why do tube freaks (being one of them) still swear by tube rectifiers?
Why do most people prefer a linear psu over a smps?
all for the same reasons you can't hear?
just my (tried and tested) two cents...
Marc
If you can hear the difference with different diodes (!) then your PSU is modulating the output, so there should be observable data, and the PSU design is rubbish, IMO
Slightly disagree. In many tube amps, the power transformer has secondary windings for both the high voltage and for heaters (typically 6 or 12V). If you use conventional diodes for HV, the HV rail can look perfectly clean, but the sound compromised. No magic involved- the switching transients of the diodes can couple via the power transformer into the heater supply and thence to the cathodes of the tubes.
You could certainly argue that this means the power supply is rubbish (I might not argue back!), and it's a trivially easy problem to fix if one is aware of it. But it does involve a subtlety and illustrates that having a clean rail from a power supply is necessary, but not sufficient. To your point, it will still result in an observable and measurable difference in the amplifier's output.
In my experience the power supply for any electronics system is the heart of that system, get the PSU right and a lot more down the line will fall into place, not always an easy task.
My own take on a lot of this is that we souround what is quite often good engineering with esoterica, with minimal data, etc wgere as I would like to see more empirical data and scientific based explanation for these differences so that an engineered solution can be found. Unlike many others I will not trust my ears as an arbitor for a change as I know they can be affected by so many variables, so in my case I need the data so I can form an opinion and see where the cahnges will be benefitial. We need data as well as listening impressions in my view.
My own take on a lot of this is that we souround what is quite often good engineering with esoterica, with minimal data, etc wgere as I would like to see more empirical data and scientific based explanation for these differences so that an engineered solution can be found. Unlike many others I will not trust my ears as an arbitor for a change as I know they can be affected by so many variables, so in my case I need the data so I can form an opinion and see where the cahnges will be benefitial. We need data as well as listening impressions in my view.