KSA50 Mk2
Hi all,
Courtesy of Jozua, I removed the board of his original KSA50 Mk2 (fan is bottom-mounted) for some much-needed restoration and tweaking. While busy I thought I'd reverse-engineer the circuit diagram since there are so many KSA50 diagrams and variations floating around. Attached is a picture of the removed board.
What I've found is that the diagram is identical to a hand-drawn one posted somewhere in the thread in 2004 or 2005, which I reattach for convenience. Besides the different transistors, the differences to the clone are minor. The only part differences are the feedback resistors, which are 12k1 and 475R instead of 22k and 1k5, leading to a gain of 28.5dB compared to 24dB. The rest of the diagram and resistor & capacitor values are the same.
Another difference is much higher power rails, at 50V loaded, 52V unloaded. This choice of higher rails could perhaps be due to the higher gain.
As you can see the board's negative section has seen some extensive trauma & prior repair, and the white 2.2uF capacitor at the bottom along with the two tack-soldered resistors are a DC-blocking RC filter at the input. Whether this was a private mod or an afterthought from Krell I'm not sure - on the KSA100Mk2 there also is a tack-soldered capacitor at the bottom of the board so it's not unprecedented.
Adam, I built Jozua's clone with a combination of Dale, Holco and Riken resistors, with using the Rikens and Holcos for the critical parts and Dale for the rest. Boards were the original Delta Audio ones; I believe the Pinkmouse version (which I also built but never got round to test) should perform better. The feedback cap was a single 1000uF 16V Black Gate NX. Wire was van den Hul for speaker and interconnect, Kimber for sigal hookup and generic stuff for power. Caps were 68,000uF Alcan types (cheap). The intention was for it to be upgradeable to a KSA100, with 4 pairs output devices on 2 separate heat tunnels, making it run very cool since the bias is half what the heatsinks can handle. I attach two low-res pictures I took during assembly in 2005.
Regards,
Pierre
Hi all,
Courtesy of Jozua, I removed the board of his original KSA50 Mk2 (fan is bottom-mounted) for some much-needed restoration and tweaking. While busy I thought I'd reverse-engineer the circuit diagram since there are so many KSA50 diagrams and variations floating around. Attached is a picture of the removed board.
What I've found is that the diagram is identical to a hand-drawn one posted somewhere in the thread in 2004 or 2005, which I reattach for convenience. Besides the different transistors, the differences to the clone are minor. The only part differences are the feedback resistors, which are 12k1 and 475R instead of 22k and 1k5, leading to a gain of 28.5dB compared to 24dB. The rest of the diagram and resistor & capacitor values are the same.
Another difference is much higher power rails, at 50V loaded, 52V unloaded. This choice of higher rails could perhaps be due to the higher gain.
As you can see the board's negative section has seen some extensive trauma & prior repair, and the white 2.2uF capacitor at the bottom along with the two tack-soldered resistors are a DC-blocking RC filter at the input. Whether this was a private mod or an afterthought from Krell I'm not sure - on the KSA100Mk2 there also is a tack-soldered capacitor at the bottom of the board so it's not unprecedented.
Adam, I built Jozua's clone with a combination of Dale, Holco and Riken resistors, with using the Rikens and Holcos for the critical parts and Dale for the rest. Boards were the original Delta Audio ones; I believe the Pinkmouse version (which I also built but never got round to test) should perform better. The feedback cap was a single 1000uF 16V Black Gate NX. Wire was van den Hul for speaker and interconnect, Kimber for sigal hookup and generic stuff for power. Caps were 68,000uF Alcan types (cheap). The intention was for it to be upgradeable to a KSA100, with 4 pairs output devices on 2 separate heat tunnels, making it run very cool since the bias is half what the heatsinks can handle. I attach two low-res pictures I took during assembly in 2005.
Regards,
Pierre
Attachments
Last edited:
Gents
There are some PSU upgrades in the pypeline for this clone and I am hoping that it will match or exceed the commercial KSA50 sound quality.
Extended listening sessions with the clone indicates that it has exceptional bass and tops and overall detail. I am hoping that the stiffer PSU will provide a more richer midrange. If not the implication might be that greater emphasis needs to be placed on using the same components as the original.
Regards
Jozua
There are some PSU upgrades in the pypeline for this clone and I am hoping that it will match or exceed the commercial KSA50 sound quality.
Extended listening sessions with the clone indicates that it has exceptional bass and tops and overall detail. I am hoping that the stiffer PSU will provide a more richer midrange. If not the implication might be that greater emphasis needs to be placed on using the same components as the original.
Regards
Jozua
It's not my schematic, it's one that got posted here a few years ago. I'll redraw it when I have time and post it.
I smile every time I see this thread come up... 813 pages and still going. I think that alone defines this amp as one of the greats...
Yes, I built both and I think quite a few other people upgraded their amps from Delta Audio to PinkMouse.
Hi Pierre
if we replace mpsa42 and 92 with 2sa970 & 2sc2240
(bent the wire to meet the different e - c - b)
how big different result of this two amp ?
- in frequency measurement ?
- in your ears 😀
are we really need pre-driver there ?
Oh yes people always said how good ksa50
but which one better KSA 50 or KSA 100 ?
if i look there, ksa 100 looks more promising
how hot heatsink in ksa50? and in ksa 100?
how big transformer for KSA 100 ( A & Volt ) ?
Thank you pierre, best regards
if we replace mpsa42 and 92 with 2sa970 & 2sc2240
(bent the wire to meet the different e - c - b)
how big different result of this two amp ?
- in frequency measurement ?
- in your ears 😀
are we really need pre-driver there ?
Oh yes people always said how good ksa50
but which one better KSA 50 or KSA 100 ?
if i look there, ksa 100 looks more promising
how hot heatsink in ksa50? and in ksa 100?
how big transformer for KSA 100 ( A & Volt ) ?
Thank you pierre, best regards
Last edited:
The PM boards never went into a chassis, they were tested to work but nothing more so I can't comment. Technically the 2S parts should be better but whether the difference would be audible I can't say.
The KSA100 is much more expensive to build as you need larger power supplies and twice the cooling and output devices. Some of the parts are also hard/expensive to find. Which of the two sounds better depends on personal taste and speaker load.
KSA100 uses +-55VDC rails, and 800VA minimum transformers. That being said, I'd love it if at least one more KSA100 gets built, of all the boards sold I think only one set actually made it into a completed amplifier.
The KSA100 is much more expensive to build as you need larger power supplies and twice the cooling and output devices. Some of the parts are also hard/expensive to find. Which of the two sounds better depends on personal taste and speaker load.
KSA100 uses +-55VDC rails, and 800VA minimum transformers. That being said, I'd love it if at least one more KSA100 gets built, of all the boards sold I think only one set actually made it into a completed amplifier.
pocoyo
My sonic impression of both the 50 and 100 is the 50 excells on the bottom thru the mids but seems rolled off at the very top compared to my 100. A 10k squarewave shows a slightly smaller bandwidth than the 100. I used the mpsa42/92 matched out off 200 pcs which is nearly impossible to get if you,ve tried. I use Quad esl/Vandersteen 2wQ, ML sl3.s and older vandersteens 2a's. But for critical listening the 100 is clearer and slightly more detailed on the Quads, and the 50 doesn,t quite fleshout on the ML's due to its current/low imped. hungry nature. A excellent match for the Vandy's.
The 50 watter is cut from the same cloth more so than the 100, but the 100 is more detailed overall though not in the bass, even though no slug there.
My (50) pink mouse mono block versions uses all Dales, Pana FM,FC, Wima MKP 10, Avel 800va 134k caps dual bridge 1302/32xx outputs 41v loaded rail. Been going strong now 4 years in my video system.
I have 4 more boards and will make this time with all PRP Mkp 1840/41 and I think most importantly use the 2sc2240 combo as in the 100 uses. this might bring out a balance I want out off the 50.
these are both great sounding amps either way
Regards
David
My sonic impression of both the 50 and 100 is the 50 excells on the bottom thru the mids but seems rolled off at the very top compared to my 100. A 10k squarewave shows a slightly smaller bandwidth than the 100. I used the mpsa42/92 matched out off 200 pcs which is nearly impossible to get if you,ve tried. I use Quad esl/Vandersteen 2wQ, ML sl3.s and older vandersteens 2a's. But for critical listening the 100 is clearer and slightly more detailed on the Quads, and the 50 doesn,t quite fleshout on the ML's due to its current/low imped. hungry nature. A excellent match for the Vandy's.
The 50 watter is cut from the same cloth more so than the 100, but the 100 is more detailed overall though not in the bass, even though no slug there.
My (50) pink mouse mono block versions uses all Dales, Pana FM,FC, Wima MKP 10, Avel 800va 134k caps dual bridge 1302/32xx outputs 41v loaded rail. Been going strong now 4 years in my video system.
I have 4 more boards and will make this time with all PRP Mkp 1840/41 and I think most importantly use the 2sc2240 combo as in the 100 uses. this might bring out a balance I want out off the 50.
these are both great sounding amps either way
Regards
David
find the average dissipation of each driver transistor.
Find the Rth s-a of the heatsink.
Calculate the delta Ts of the sink.
Calculate the delta Tc of the transistor.
Add on Ta to find Tc. Remember Ta for the driver heatsink is the internal chassis temperature, not the room ambient temperature. This internal temperature could be 10C to 30C hotter than the room temperature.
Check rating of transistor at this raised Tc.
Find the Rth s-a of the heatsink.
Calculate the delta Ts of the sink.
Calculate the delta Tc of the transistor.
Add on Ta to find Tc. Remember Ta for the driver heatsink is the internal chassis temperature, not the room ambient temperature. This internal temperature could be 10C to 30C hotter than the room temperature.
Check rating of transistor at this raised Tc.
In short the answer is HECK NO!
75 x 30 x 74 is not even remotely close to big enough for one channel.
75 x 30 x 74 is not even remotely close to big enough for one channel.
For the passive cooling with conrad heatsink 30cm x 8 cm x 10 cm
get hot in 1.5A BIAS, ksa 50 NEED 1.9a BIAS
Tangmonster -
Adi's proposing that for the drivers only, not the whole amp.
My gut says half of that profile should be suitable for driver use, but needs to be verified following AndrewT's analysis.
Adi's proposing that for the drivers only, not the whole amp.
My gut says half of that profile should be suitable for driver use, but needs to be verified following AndrewT's analysis.
😀😀😀
maybe some builder like me was too emotional in here
until i didnt read that for DRIVER heatsink 🙂
Dear ADI 2009
You can use bottom chassis as driver heatsink
like some builder did for their KSA 50
Best regards, jeffry
Any driver PCBs available ?
Hi everyone,
I've been collecting parts for my KSA-50 for a while now. I started with the essentials like the heatsinks. A lucky bit of ebaying has landed me some nice bits. I have an original KSA50 MK2 chassis with heatsink and fan, 2 x avel lindberg trafos.
Some of the output devices went into melt down mode and took out the drivers and pcb, apparently it was scorched so badly it was not possible to reuse.
This isn't the first Krell amp I've heard thats gone into self destruct mode. Is there something we can do to prevent this from happenning ?
The capacitors and bridges were sold before I could bid on them, although they were probably worn out from 20+ years of use, the mounting hardware would have been useful.
Lets hope I can bring her back to her former glory. I still need the following parts : psu capacitors, rectifiers, soft start, speaker delay, DC protection, decent speaker binding posts, driver board and components and various mounting hardware.
Now I'm searching for a set of replacement pcbs. I have been looking at Abin's all-in-one version which looks great, but I think it'll need too much cutting to fit into an original KSA50 chassis and heatsink.
Any one have any spare pcbs that I can use to restore my KSA50 ?
Regards
fs
ImageShack Album - 9 images
Hi everyone,
I've been collecting parts for my KSA-50 for a while now. I started with the essentials like the heatsinks. A lucky bit of ebaying has landed me some nice bits. I have an original KSA50 MK2 chassis with heatsink and fan, 2 x avel lindberg trafos.
Some of the output devices went into melt down mode and took out the drivers and pcb, apparently it was scorched so badly it was not possible to reuse.
This isn't the first Krell amp I've heard thats gone into self destruct mode. Is there something we can do to prevent this from happenning ?
The capacitors and bridges were sold before I could bid on them, although they were probably worn out from 20+ years of use, the mounting hardware would have been useful.
Lets hope I can bring her back to her former glory. I still need the following parts : psu capacitors, rectifiers, soft start, speaker delay, DC protection, decent speaker binding posts, driver board and components and various mounting hardware.
Now I'm searching for a set of replacement pcbs. I have been looking at Abin's all-in-one version which looks great, but I think it'll need too much cutting to fit into an original KSA50 chassis and heatsink.
Any one have any spare pcbs that I can use to restore my KSA50 ?
Regards
fs
ImageShack Album - 9 images