Okay thanks. The FM Series has the 105 degrees Centigrade rating.
The Krell is 100% standard as far as I know.
Yes I know.
I've read this thread till page 89 and every night before going to bed I try to read a few pages more! Not because it's sleep-inducing, but to understand how my Krell works. I find it really great how you guys get to make a brand new KSA-100 Mk2 from almost scratch. Great to experience different main actors in the progress like you and from different countries all over the world.
Last edited:
Found just after some better searching a couple of RIFA PEH169's for substituting the 30 year old Spragues:
1. 47mF/63V
2. 33mF/100V
They cost more then 120 USD with VAT a piece...
I will mainly use this KSA-100 Mk2 for powering the woofers of my speakers.
I measured the railvoltage @ 52 Volts isn't then good practice to go for the highest capacitance value?
I did not read any definitive choice for a G.B. in this thread (reached page 89 yesterday) regarding the powersupply components yet.
I'm inclined to choose a powersupply cap that has the same dimensions as the original Spragues. It's easy to swap then with those thick attached (looks like old Monstercable) cables to the DC side of the bridge.
1. 47mF/63V
2. 33mF/100V
They cost more then 120 USD with VAT a piece...
I will mainly use this KSA-100 Mk2 for powering the woofers of my speakers.
I measured the railvoltage @ 52 Volts isn't then good practice to go for the highest capacitance value?
I did not read any definitive choice for a G.B. in this thread (reached page 89 yesterday) regarding the powersupply components yet.
I'm inclined to choose a powersupply cap that has the same dimensions as the original Spragues. It's easy to swap then with those thick attached (looks like old Monstercable) cables to the DC side of the bridge.
Rivets
I would have never thought to use rivets on output transistors. Rivets, no in your pictures? In fact. One of my collector screw/nut loosened and I burnt/shorted a transistor. Yes, I used a lock washer, but....... Hey, NASA had a loose fuel tank seal and BOOM. (not being funny, but worse things happens).
I just delved into all the old pictures of MK2's I've worked with. Your model really does seem to be very original, which is rare at this age. The only possible difference I can note is the insulation of the power lugs at the front AC loom which is black in all the ones I've seen but red in yours, but as you note this is from a time when Krell was not exactly as consistent as nowadays and it may very well be what they happened to have in stock at the time. I've seen much worse variance between amps in e.g. Sunfire.
Anyhow, experimentation has helped a great deal to learn what helps and what not, and it's of course nice if you have a whole new board to experiment with so that if something goes wrong you can still drop in the original as backup. The Clone boards also have a number of improvements.
On yours, the improvements really are only to replace the feedback caps (Panasonic FC/FM/FR/FK are all good) and the main caps. The original Spragues are hellishly expensive but still sounded better than others I've tried, but I'm sure there are better options. Munforf's M-lytics were great but crazily expensive. Replacing the speaker wiring helped a bit, and if you're a little more adventurous, modifying the wiring a little so that the speaker output and the feedback to the amplifier board are a) symmetrical and b) at the same point helped too. The fans can also benefit from replacement if they're noisy or slowed down, but that depends on the operational hours and amount of dust in the environment.
Anyhow, experimentation has helped a great deal to learn what helps and what not, and it's of course nice if you have a whole new board to experiment with so that if something goes wrong you can still drop in the original as backup. The Clone boards also have a number of improvements.
On yours, the improvements really are only to replace the feedback caps (Panasonic FC/FM/FR/FK are all good) and the main caps. The original Spragues are hellishly expensive but still sounded better than others I've tried, but I'm sure there are better options. Munforf's M-lytics were great but crazily expensive. Replacing the speaker wiring helped a bit, and if you're a little more adventurous, modifying the wiring a little so that the speaker output and the feedback to the amplifier board are a) symmetrical and b) at the same point helped too. The fans can also benefit from replacement if they're noisy or slowed down, but that depends on the operational hours and amount of dust in the environment.
There's no riveting in this amp so far as I can see.
Everything is soldered or fastened with screws I believe.
Thanks. The guy I bought it from said it was completely original and he wanted to sell it to somebody who would use it in his setup and not to a secondhand dealer that offered a better price.
I have had a Mk1 version in the eighties and a pair of KMA-200 Mk2's after that so I knew a bit how it should look in the inside.
The Papst 4800X fans are better then the ones in the Mk1 version. They resonate just a little but I'm not distracted by that.
Today I got four Nippon Chemicons U36D Series from Digikey, 100mF 63V and all four them measured 83mF. They looked like pullouts with light dents and markings on their sleeves. I will send them back. I still have the old Spragues, 61mF, 50 V from my Levinson ML-2's and their date code is from 1978. When I measure them they all give 67mF after almost forty years.... Don't know about their ESR of course.
I'm now looking at RIFA Kemet PEH200 Series and Mouser has a 100mF 63V for 80 USD. That does sound tempting. I still want to know if the voltage rating of 63V is too close to the railvoltage of 52 Volts?
Thanks for the advice.
I will opt for the PEH200 100mF, 63V.
Actually the ripple current specs for the 63V 100mF (100Hz 85Centigrade, 10kHz 50C and 10KHz 40C):
34.3, 62.2 and 50.3
are better then for the 100V 47mF:
30.8, 60.5 and 47.3.
Same is the case for the ESR values:
63V 100mF (100HZ and 100kHz):
5 mOhm and 5 mOhm
100V 47mF:
7 mOhm and 6 mOhm.
ESL values are the same: 17 mOhm.
Maybe it is different when they are in a circuit?
KSA means Krell Stereo Amplifier
KMA means Krell Mono Amplifier.
PAM means Pre AMplifier
KRS means Krell Reference Series
KRX means Krell Reference Xover
That's the kind of knowledge that's good for nothing really...
This is one of the four large Sprague 61mF, 50V out of my ML-2's.
It is still 6mF better (or more) than it's initial specified 61mF.
http://www.diyaudio.com/forums/attachment.php?attachmentid=598619&stc=1&d=1486809144
This is a "new" Nippon Chemicon U36D 100mF, 80V large can that "lost" already 17mF
http://www.diyaudio.com/forums/attachment.php?attachmentid=598620&stc=1&d=1486809144
The Spragues didn't have an easy life in the hot environment of a ML-2.
Don't know about ESR values of course.
Maybe the Nippon Chemicons were shelved for a decade or so....
It is still 6mF better (or more) than it's initial specified 61mF.
http://www.diyaudio.com/forums/attachment.php?attachmentid=598619&stc=1&d=1486809144
This is a "new" Nippon Chemicon U36D 100mF, 80V large can that "lost" already 17mF
http://www.diyaudio.com/forums/attachment.php?attachmentid=598620&stc=1&d=1486809144
The Spragues didn't have an easy life in the hot environment of a ML-2.
Don't know about ESR values of course.
Maybe the Nippon Chemicons were shelved for a decade or so....
Attachments
The Sprague are rated at 50Vdc
That is too low for the ksa100 which at nominal mains voltage and fully loaded is already at 52Vdc.
When mains is high and/or the loading is reduced, the voltage on the smoothing capacitors will rise to way above 52Vdc.
You need to use 63Vdc, or higher, smoothing capacitors.
That is too low for the ksa100 which at nominal mains voltage and fully loaded is already at 52Vdc.
When mains is high and/or the loading is reduced, the voltage on the smoothing capacitors will rise to way above 52Vdc.
You need to use 63Vdc, or higher, smoothing capacitors.
I've had Sprague's, Cornell Dubiliers and Mallory's and they were all 10 to 20% above their capratings. All four Nippon Chemicon U36D's were all 17% under their capratings so I think the whole inventory from Digikey is low on capacitance.