If you're looking for the smooth highs, I highly recommend Black Gates. They are very costly and hard to come by. Nothing comes close. Another cap worth looking at is Elna.
Although the choice of capacitor can affect the sound, various design decisions can make the change in sound more noticeable or less so.
Larger capacitance values (and larger current draw from the downstream circuitry) will increase the amplitude of the current spikes needed to charge up the capacitance, and the larger amplitudes will lead to greater high-frequency content (noise) being both conducted and emitted from the power supply.
Countermeasures include adding an R-C snubber across the transformer secondary windings, as close to the transformer windings as possible.
Twisting the wires from transformer secondary to rectifier, and keeping the wire lengths short.
Inserting a low-resistance high-current inductor (50~100uH) between rectifier and initial electrolytic, and again keeping the wire lengths short. Note that the inductor DCR may reduce or modulate the power supply voltage due to the current draw of the downstream circuitry, so tread carefully here.
Together these countermeasures will take off the edge of the capacitor charging spikes, and reduce the amount of radiated and conducted HF noise produced by the power supply.
Hopefully this will make the “capacitor sound” less noticeable.
Larger capacitance values (and larger current draw from the downstream circuitry) will increase the amplitude of the current spikes needed to charge up the capacitance, and the larger amplitudes will lead to greater high-frequency content (noise) being both conducted and emitted from the power supply.
Countermeasures include adding an R-C snubber across the transformer secondary windings, as close to the transformer windings as possible.
Twisting the wires from transformer secondary to rectifier, and keeping the wire lengths short.
Inserting a low-resistance high-current inductor (50~100uH) between rectifier and initial electrolytic, and again keeping the wire lengths short. Note that the inductor DCR may reduce or modulate the power supply voltage due to the current draw of the downstream circuitry, so tread carefully here.
Together these countermeasures will take off the edge of the capacitor charging spikes, and reduce the amount of radiated and conducted HF noise produced by the power supply.
Hopefully this will make the “capacitor sound” less noticeable.
You Sir are my hero. FK and NX do not come in 100v ratings, so this is a non-option, but i gradually warm up to the idea of a compound capacitor.
What gauge wire did you use for the chokes? Did those fit inside the amp case? Did you construct a pcb to carry the caps?
Amazing achievement.
Wow, thats crazy. The question is then, whether the amplifier circuit worth the power supply 😀
A lots of great advice, thank you Mr Carr.
As much as i like using chokes and especially choke input supplies, this solution may very hard to implement in a device designed to provide unlimited power into 1ohm loads.
As for the snubbers - great reminder but this will have to wait until all caps are in place.
There is no question of preserving collectors value, but i don't really enjoy doing dramatic irreversible changes inside commercial gear. Yes, the liberal use of thick brass beams threaded and bolted onto other brass beams instead of copper wiring drives me nuts but i am not going to completely redesign the amp.
+1 to the remarks by AndrewT, but a regulated supply for the low-current voltage gain stages will make the amplifier sound 'huge', and at less cost than beer-can sized caps.
The Electrocompaniet Ampliwire was the first amplifier I heard with a regulated front end, it sounded huge, even though it was only 25W or so.
It doesn't seem like anyone much cares about the sound of big caps, but for the sake of completion, here is the conclusion 🙂
I did go with the Epcos/Siemens from Mouser and apart from a very unpleasant surprise - 6mm threads instead of 5mm - all went well. The caps have a very neutral tone with decent bass and high frequency extension out of the box. Did not have sufficient patience to wait for a break in as i really wanted to have the Krell job finished.
And a big job it was. All electrolytics on the boards were replaced with a mixture of Mundorf, Panasonic, Rifa and Roe. Output relays changed, all TO3 transistors reseated with new Keratherm pads. Input/output termination changed to Vampire/WBT.
The KSA250 is now in a completely different league. Gone is the haze, the slow bass, the glossed over leading edges. Yet, the tonal balance is still pleasingly warm.
As for the Siemens/Epcos/TDK caps - highly recommended.
I did go with the Epcos/Siemens from Mouser and apart from a very unpleasant surprise - 6mm threads instead of 5mm - all went well. The caps have a very neutral tone with decent bass and high frequency extension out of the box. Did not have sufficient patience to wait for a break in as i really wanted to have the Krell job finished.
And a big job it was. All electrolytics on the boards were replaced with a mixture of Mundorf, Panasonic, Rifa and Roe. Output relays changed, all TO3 transistors reseated with new Keratherm pads. Input/output termination changed to Vampire/WBT.
The KSA250 is now in a completely different league. Gone is the haze, the slow bass, the glossed over leading edges. Yet, the tonal balance is still pleasingly warm.
As for the Siemens/Epcos/TDK caps - highly recommended.
- Status
- Not open for further replies.