Whenever I have played around with choke input filters, sound quality is ALWAYS inferior!
I have used Telefunken RGN2004, 2504 & 4004, I ave used 5U4G and 5R4GY, I have used even 866A mercury vapor rectifiers. The sonic results were always the same. Without any cap in front, sound is wooly, it is closed in, not shiny and lustrous. Telefunken rectifiers love relatively big capacitances, 6-10 microfarads is a typical value to get started. Even 5R4GY loves C input, even at only 4 uF! The same holds true with 5U4G.
If you plan to use mercury 855 tubes, the upper limit due to known reasons is 2uF, no more! I never liked choke input power supplies. The sketches of Mr. Sakuma never included any choke input filter and that is going on the same side with my preference.
I have used Telefunken RGN2004, 2504 & 4004, I ave used 5U4G and 5R4GY, I have used even 866A mercury vapor rectifiers. The sonic results were always the same. Without any cap in front, sound is wooly, it is closed in, not shiny and lustrous. Telefunken rectifiers love relatively big capacitances, 6-10 microfarads is a typical value to get started. Even 5R4GY loves C input, even at only 4 uF! The same holds true with 5U4G.
If you plan to use mercury 855 tubes, the upper limit due to known reasons is 2uF, no more! I never liked choke input power supplies. The sketches of Mr. Sakuma never included any choke input filter and that is going on the same side with my preference.
I think we're all entitled to a cutoff point where daft ideas are concerned without it inviting suggestions that we take up different hobbies. A degree of scepticism is healthy in some situations.
w
Perhaps, you take my comment -as well as yourself- a bit too seriously.
Last edited:
Forget the power supply that amp layout looks like a dogs dinner, I wonder how prone to parasitic oscillation it is or how much noise it picks up from the front end of the power supply. Some designers really need to look at some quality valve equipment such as communication receivers, telecommunications equipment or electronic test gear to learn layout and shielding techniques before going near an actual project, the schematic only tells half the story.
FWIW I have seen HV power supplies get ripple and noise down to the ppm range using standard design techniques which can be found in any engineering text on the subject.
I took Ken's comment in the spirit in which, I assume, he said it!
My own view, as a sort of EE, is that some of what is said about power supplies is just nonsense. Perceived changes in sound may or may not be improvements, and may or may not be actually caused by the proposed mechanism. I'm afraid my prejudice is to stop listening as soon as someone starts talking too much about energy storage (charge storage is the correct concept), or deliberately making a poor design because he prefers the sound it makes. I'm not a SET fan either.
My own view, as a sort of EE, is that some of what is said about power supplies is just nonsense. Perceived changes in sound may or may not be improvements, and may or may not be actually caused by the proposed mechanism. I'm afraid my prejudice is to stop listening as soon as someone starts talking too much about energy storage (charge storage is the correct concept), or deliberately making a poor design because he prefers the sound it makes. I'm not a SET fan either.
Putting a string of diodes between the parts of a power supply may actually have an effect! Not the one intended nor just dropping a few volts, but actually reducing the ripple! When no current flows through a diode the voltage drop is ZERO volts. At full rated peak current it can be a volt or so. So during the peak of the AC sine wave voltage charging the current will be at a maximum as will the diode drop. After the peak the current will drop and the diodes will effectively stretch out the charge to the following components. The big however is that the voltage drop is an exponent to the passing current so the overall reduction in ripple will be on the order of perhaps tens of millivolts.
But don't forget to consider that any time you unplug and replug your audio gear there will always be some change just due to the wiping action on all the contacts. Of course in this case there is also a clear change in the DC rail voltage also.
But if you want to add half a dozen diodes this is something you should actually be able to measure the change in ripple.
But don't forget to consider that any time you unplug and replug your audio gear there will always be some change just due to the wiping action on all the contacts. Of course in this case there is also a clear change in the DC rail voltage also.
But if you want to add half a dozen diodes this is something you should actually be able to measure the change in ripple.
The effect of a diode could be to sharpen upward ripple and smooth downward ripple. I'm not sure that would be an improvement. The most useful effect might be for Class B P-P amps, as the diode impedance could be highish for normal signals (giving slightly better smoothing) but reducing when current is demanded for a signal peak (giving reduced voltage droop). I think it would be a small effect, and even smaller for Class A SE.
Certainly would be small, less than 100mV and could be as low as a few mv on say a 300 volt 100ma supply. But on a poorly designed power supply with efficient speakers you might notice a difference. And yes there are better ways to get there!
It is just when someone says they hear something that seems nutty it often is, but every so often there actually might be a real difference.
For the question asked on this thread, reasonable results have happened. That is kind of the goal of the website.
So a pat on the back for all of those who do actually usefully contribute instead of just adding noise.
Which should mean that my future comments won't be worthy of your counter commentary - I can only hope.
Last edited:
They'd have to use a bunch of them: taking a 1000kg car to 60 meters/sec (about 150 mph / 240 kph) requires 1,800,000 Joules, and that cap will yield 165,000 Joules at 45 volts (real designers never run at maximum ratings). You'd probably have to exceed the maximum current ratings to set speed records, however.
Two of the power supplies at work have 1.2F 450V capacitor banks (12000uF x 100) Output current is 1000A peak 400A average Voltage adjustable from 0 to 400V.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.