Hi soundcheck,
Yes, I "invented" it, unless someone already came up with this idea before I did. For me, most important thing is that I finally seem to have solved the mains power supply issues that significantly degrade sound quality.
The charge-transfer power supply performs like a very low noise battery power supply. So the power supply you are now using might perform similar to the charge-transfer power supply.
Other advantage of the charge-transfer power supply is that it can drive higher voltage loads like (tube) power amplifiers. In that case, the tube rectifier can be replaced by hybrid HV Schottky diodes, and the large chokes could be replaced by capacitance multipliers, these will also emulate the slow rise in anode voltage during power-up. This is possible because the mains interference path is interrupted / blocked, so the whole power supply becomes far less critical.
You are always welcome to come over and listen to the DI4T, you could bring your DACs for direct comparison. You will have to wait a bit for the modules, as the ones on the photograph are fully handcrafted prototypes.
Yes, I "invented" it, unless someone already came up with this idea before I did. For me, most important thing is that I finally seem to have solved the mains power supply issues that significantly degrade sound quality.
The charge-transfer power supply performs like a very low noise battery power supply. So the power supply you are now using might perform similar to the charge-transfer power supply.
Other advantage of the charge-transfer power supply is that it can drive higher voltage loads like (tube) power amplifiers. In that case, the tube rectifier can be replaced by hybrid HV Schottky diodes, and the large chokes could be replaced by capacitance multipliers, these will also emulate the slow rise in anode voltage during power-up. This is possible because the mains interference path is interrupted / blocked, so the whole power supply becomes far less critical.
You are always welcome to come over and listen to the DI4T, you could bring your DACs for direct comparison. You will have to wait a bit for the modules, as the ones on the photograph are fully handcrafted prototypes.
Hi XXHE,
???
Just for clarity, I use plain electrolytic caps of around 2200uF.
The capacitors are used in a similar way as with conventional power supplies (check the schematic), they are charged at a rate of 50/60 or 100/120 Hz, just like with a conventional mains power supply. The charge current flowing into the secondary smoothing cap is limited, so the service life of this cap should even be longer.
In other words, the capacitor service life isn't shortened by the charge-transfer circuit.
???
Just for clarity, I use plain electrolytic caps of around 2200uF.
The capacitors are used in a similar way as with conventional power supplies (check the schematic), they are charged at a rate of 50/60 or 100/120 Hz, just like with a conventional mains power supply. The charge current flowing into the secondary smoothing cap is limited, so the service life of this cap should even be longer.
In other words, the capacitor service life isn't shortened by the charge-transfer circuit.
hi john,
i hope i understand correctly the principle of your new power supply.
my question: do you use only one switch or two switches (one before the first cap and one between the caps for alternately switching during the charge and discharge cycle) ?
is it not better to go above the 30khz as the switching frequency? (not audioable frequency!)
my respect for your knowledge and your efforts
regards
mamal
i hope i understand correctly the principle of your new power supply.
my question: do you use only one switch or two switches (one before the first cap and one between the caps for alternately switching during the charge and discharge cycle) ?
is it not better to go above the 30khz as the switching frequency? (not audioable frequency!)
my respect for your knowledge and your efforts
regards
mamal
-ecdesigns- said:
Is there any comparison to your power supply and this product?
Never Connected
Regards,
Dan
XXHE said:Everybodyis welcome here. Have some places to sleep as well.
And uhhm, don't expect any BD sound here from back then. Rather think of 10 times well ... different.
You were definitely on my schedule too
How far do you live from John?
XXHE said:
hehehe
I expect to find a certain DAC completed and working, though.
Telstar said:
You were definitely on my schedule too
How far do you live from John?
Outside traffic hours ? 70-90 minutes at most.
In traffic hours ? 150 minutes at least, with some rain 170 minutes, with snow many days (am I joking ? no !).
Telstar said:
By that time I can present you two if all goes as planned. I will be very much surprised though if "the second" superceedes the "NOS1". But we'll see.
But watch soundcheck. Never underestimate his Linux achievements !
But hey, I suggest to deal with this by email further and leave the efforts of John in its own leage. It really is you know ...
http://www.trichordresearch.com/PDF_Files/Power Supply Hi Fi Critic review.pdf
I don't care much, but I guess not.
The point is IMO : when you know your abouts a bit with supercaps (and I'm sure they showed up in this thread somewhere) you rather automatically start to think in these direction of solutions.
Quite another matter is the recognition of the polution of which I wouldn't know how to distinguish that from, well, whatever - without measuring.
What remains (for me) is the so often discussed infinite "current" of the battery vs. the mains power, vs this solution. One thing I know : for a good DAC the power supply to the I/V is the most important of all. If the DAC is accurate to 1/2 lsb, it won't help much if the PSU degrates that to 4.
Not much related to pollution perhaps, but possibly related to the solution.
Peter
Actually, it is really needed that the switch runs on 50-120Hz?
Could it be much less?
Even if a little wasted energy on the non-connected time, for these circuits that are very low current, maybe 1Hz or such is more than enough?
My TeddyRegs are slow to discharge...
I'm only guessing you know...
Cheers,
M
PS: thanks Soundcheck for you efforts on the Wiki!
Hi mr whocares,
I use one MOSFET switch only, the circuit is very simple but highly effective. The Schottky rectifier diode already acts like a switch, and disconnects the charge current path when the sinewave voltage drops below the voltage across the first smoothing cap plus the voltage drop across the Schottky diode.
The MOSFET disconnects the load during the time that the primary smoothing cap is being charged.
The ripple voltage after the MOSFET switch is already approx. 3 times lower than the ripple voltage across the smoothing cap of a conventional power supply.
Then I use a capacitance multiplier to completely remove the ripple voltage before the voltage regulator. It's also possible to integrate the voltage regulator in the capacitance multiplier, achieving a fully discrete low noise supply.
It will inter-modulate with both masterclock and derived signals, and harmonics of the high switching frequency will also affect the audio range. Even worse, the high switching frequencies can hop along all stray capacitances and spread throughout the circuits. By keeping switching frequencies as low as possible, they can be filtered out easily using capacitance multipliers and plain voltage regulators, the higher the switching frequency, the more difficult this will get.
I use one MOSFET switch only, the circuit is very simple but highly effective. The Schottky rectifier diode already acts like a switch, and disconnects the charge current path when the sinewave voltage drops below the voltage across the first smoothing cap plus the voltage drop across the Schottky diode.
The MOSFET disconnects the load during the time that the primary smoothing cap is being charged.
The ripple voltage after the MOSFET switch is already approx. 3 times lower than the ripple voltage across the smoothing cap of a conventional power supply.
Then I use a capacitance multiplier to completely remove the ripple voltage before the voltage regulator. It's also possible to integrate the voltage regulator in the capacitance multiplier, achieving a fully discrete low noise supply.
It will inter-modulate with both masterclock and derived signals, and harmonics of the high switching frequency will also affect the audio range. Even worse, the high switching frequencies can hop along all stray capacitances and spread throughout the circuits. By keeping switching frequencies as low as possible, they can be filtered out easily using capacitance multipliers and plain voltage regulators, the higher the switching frequency, the more difficult this will get.