Busy sprucing up my old Krell as winter is coming 🙂 Pretty much decided upon all the small caps but the big four are still making me scratch my head.
From what i understand normal Krell recap service does not touch the big caps and i can see why - the old Sprague 36DE 40mF/100v have not lost any capacitance (cannot measure such low ESR though) and don't look like they are about to self destruct. 95 deg rated too.
Today i gave them a listen inside another amp. They do have plenty bass, dynamics and solidity but the upper midrange and highs are dull and crude beyond belief. Which is not surprising as the Krell has always been mellow and low res on top.
So, here is the dilemma: replace or keep stock? And what to replace them with?
I have a few Rifa 169 33000/100 but they sound so dramatically different tonally that the idea leaves me a bit hesitant. Plus they are smallish.
Mouser stocks Epcos/TDK B41456 which would fit perfectly but i have no idea what those sound like.
Looking for good advice really.
From what i understand normal Krell recap service does not touch the big caps and i can see why - the old Sprague 36DE 40mF/100v have not lost any capacitance (cannot measure such low ESR though) and don't look like they are about to self destruct. 95 deg rated too.
Today i gave them a listen inside another amp. They do have plenty bass, dynamics and solidity but the upper midrange and highs are dull and crude beyond belief. Which is not surprising as the Krell has always been mellow and low res on top.
So, here is the dilemma: replace or keep stock? And what to replace them with?
I have a few Rifa 169 33000/100 but they sound so dramatically different tonally that the idea leaves me a bit hesitant. Plus they are smallish.
Mouser stocks Epcos/TDK B41456 which would fit perfectly but i have no idea what those sound like.
Looking for good advice really.
What i find interesting is the almost complete lack of subjective opinion on large electrolytic caps and not just here.
You want to know about coupling caps? Everyone and their auntie has an opinion and usually quite strong. But this is mainly the tube and speaker crowd. Do solid state fans hear less? 😛
You want to know about coupling caps? Everyone and their auntie has an opinion and usually quite strong. But this is mainly the tube and speaker crowd. Do solid state fans hear less? 😛
Maybe they are for the first 25 years of their life, but how good are they today is hard to say. Maybe caps age like ears: the highs go first 🙂
djk has some ideas, please pm him...and you might be surprised, it is not the bigs caps like you think....
The main smoothing caps have very little to do with the high frequency abilities of a power amplifier.
They are located too far away from the output devices and the long wires/traces give too much inductance to allow a low impedance at the output devices.
Look at how Krell and your comparison amplifier dealt with local decoupling of the supply rails. These local capacitors supply all the fast changing currents that flow into and out of the output devices and their driver devices.
It has become my opinion that much of the "sound" we hear is due to how good, or mediocre, or bad, the local decoupling is.
The bass and lower mid will be affected by the voltage "stiffness" of the supply rail to low frequency current changes. This will be determined by the main smoothing capacitors.
They are located too far away from the output devices and the long wires/traces give too much inductance to allow a low impedance at the output devices.
Look at how Krell and your comparison amplifier dealt with local decoupling of the supply rails. These local capacitors supply all the fast changing currents that flow into and out of the output devices and their driver devices.
It has become my opinion that much of the "sound" we hear is due to how good, or mediocre, or bad, the local decoupling is.
The bass and lower mid will be affected by the voltage "stiffness" of the supply rail to low frequency current changes. This will be determined by the main smoothing capacitors.
This is true with regards to measured response, but the subjective impression of how capacitors influence sound is different and unrelated.
Simple analogy with coupling caps - would you say a silver Mundorf sounds the same in the highs, as a Wima MKP? Do you really know why?
As for local decoupling, the test amp has a small Mundorf electrolytic and a film cap within a centimeter of the output transistors. So, yes, local decoupling is taken care of.
Out of interest the test amp uses 47mF/63v Felsic 85 which to my ears are musical and open if perhaps bass-light.
Anyway, zero interest in theoretical discussions, Ohm's law and other trivia.
Only interested in the opinions of those who have experimented with large can caps and have been able to hear a difference.
The bigger lytic - the worse sound with it, I get this idea after assembling of more than dozen various class A se amps with output capacitors. I think bigger lytic implies more pronounced electro-mechanical vibrational effects. Even with film caps - the more tight film winding - the better sound. My personal receipt for usage of lytics - first an assembly of 16pcs of 1000uF Nichikon KZ, after that one could put 2-3pcs of 10000uF lytics. Avoid the ones of more than 10000uF.
Last edited:
I agree. Sadly, in modifying commercial equipment there are no such options, so one has to choose the lesser evil.
Then probably you should look for big lytics, designed with some tricks, like slit-foil BHC, etc. Maybe some manufacturers managed to inset several small lytic windings in on big can, and actually there are small paralleled lytics inside.
Nah. At 100v there are mostly industrial caps in production and even some of those are not easy to buy.
If Mundorf Mlytic was not in such an exotic form factor i would not be starting this thread...
Still hoping someone may share an opinion on the Epcos mentioned in the first post.
If Mundorf Mlytic was not in such an exotic form factor i would not be starting this thread...
Still hoping someone may share an opinion on the Epcos mentioned in the first post.
Audiophoolery at its best. I heard the large can caps with gold plated connections sound the best but only so if the print on the sleeve is also in gold ink and when they cost at least 10 times more than industrial large can caps.
Last edited:
And that is what I am giving to the Thread.
The main smoothing capacitors affect the bass and lower mid sound.
The local decoupling affect the upper mid and treble and the ultrasonic that we can't hear.
I recommend epcos. I gave model numbers a few years back. But they are all low profile in lower voltages. I bought 25V through to 50V
Wholeheartedly agree Andrew. This is exactly how things should be in an ideal world.
Only my experience shows that if a cap does not sound pleasant and balanced on its own, no amount of local decoupling will help, although it may disguise its ugliest traits. There are those who advocate multiple bypasses for all caps, power supply and coupling, but it's just not my taste.
And i don't really have the luxury of being able to experiment with local decouping - there are 24 caps in this position throughout the boards. Maybe some lucky combo will indeed sound great...
This is the main reason i listened to the old caps in a separate, much simpler and already voiced amp and the results were not good. In fact, the test amp sounded surprisingly close to the Krell when fitted with Sprague caps, despite a dramatically different topology and active devices.
So, saying the main filter caps have no effect on the perceived high frequency response does not really help. Please note that "perceived high frequency" and actual high frequency are completely different things.
I replaced my Krell KSA 100's 30 year old Sprague big 4 x 37,500 uf filter caps with:
484,000 uf made up from 22,000 uf Panasonic 105C then 1 mhy inductors then 100,000 of Blackgate FK and NX - cost a small fortune , but now it has phenomenal bass, mids, and highs so yes the quantity, value , and type of caps do make a difference. I was surprised at how big a difference it made.
484,000 uf made up from 22,000 uf Panasonic 105C then 1 mhy inductors then 100,000 of Blackgate FK and NX - cost a small fortune , but now it has phenomenal bass, mids, and highs so yes the quantity, value , and type of caps do make a difference. I was surprised at how big a difference it made.
Last edited:
- Status
- Not open for further replies.