The Krell KSA-50 clone made it about 7 years before an overzealous owner decided to start taking a few measurements and through sheer carelessness grounded one of the power supply rails, releasing a nice spark and ... what is the word in military terminology "report" "retort"? The fuses blew and despite repair attempts it would never work right thereafter, so it sat in the garage disassembled gathering dust for the last 5 years. You can read about this amp here.
Due to recent events leading to temporary unretirement from DIY, the amp was examined again and turned over to the warranty department for a fix. Of course the warranty department is tired and lazy so decided to half-a** it. The reason being that since the big push for these amps in this forum, fully assembled Krell clone boards have become available for cheap on ebay. The slackers in warranty planned to remove the guts and toss in a a couple of crappy chinese boards.
Before that though, the amp was carefully....over the course of a month... put back together which was not so easy. The black bolts on the face plate had started rusting, but the insides seemed ok with no visible problems. The damn thing is complicated so it took a lot of reading and tracing of wires to figure stuff out.
Here is the left channel output stage getting put back in.
The VARIAC was dusted off and the thing carefully brought up to power....and what do you know, it took a week to confirm but the thing actually works, well was mostly working.
Back in the day, I thought I was super cool using a 3 way AC inlet (having AC in, on/off switch and fuses). Dumb idea, the lousy inlet never worked right and I already replaced it 2 times. You would flip it "on" and nothing would happen. Turned "on" about 1 in every 8 tries, but turned off no problem. Hmmm... is it because I have a safety line rated capacitor connected who knows where? I have never seen an authoritative schematic with (1) on off switch, (2) fuse (3) line rated cap, and I'm not looking now. So out went the crappy inlet and in went another identical one without any AC rated cap. What? I still have a bunch of these terrible inlets sitting in a box. Like I said, lazy.
Here is the not-recommended AC three way inlet.
Wires from two emitter resistors for bias adjustment were taken out to the back panel soldered to a non-conforming DIN type socket. The socket was removed and the hole carefully filed to fit an actual DIN socket for which mating connectors exist. Now I have a mating cable and socket for easy bias measurement.
DIN Bias socket
Connected, measuring bias.
The output stage of this amp uses 4 metal transistors per rail (16 total) which are like tiny blowtorches on a large tubular vertical heat sink. A small bottom fan pushes air up, and a larger 120mm top fan also pulls air up. There is no cover and the fans are essential, but noisy.
The geeks that do PC overclocking would cringe at this setup as the inlet and outlet fans are mismatched and you are essentially cooking the top fan by blowing hot air into its mechanism. There are now more advanced, quieter fans and "static pressure" fans designed to be mounted right next to a radiator. Not having a clue I bought two of the newer Noctua fans (regular and static pressure) to try out.
Fan Assortment- Old Aspire and new Noctua NF-P12 static pressure fan and NF-S12A FLX normal fan.
I did experiments with closing off some of the surrounding open space to try to force all airflow through the heat sink fins, but it turns out that the best cooling is with the wider opening (tape removed).
Trying out fans and airflow
In the old days the Aspire fan kept things around 60C, the newer very quiet static pressure Noctua was 10C worse, and the normal Noctua NF-S12A was not quite as quiet as the static pressure Noctua but only +5C (so 65C in operation). I am going with this one now.
Due to recent events leading to temporary unretirement from DIY, the amp was examined again and turned over to the warranty department for a fix. Of course the warranty department is tired and lazy so decided to half-a** it. The reason being that since the big push for these amps in this forum, fully assembled Krell clone boards have become available for cheap on ebay. The slackers in warranty planned to remove the guts and toss in a a couple of crappy chinese boards.
Before that though, the amp was carefully....over the course of a month... put back together which was not so easy. The black bolts on the face plate had started rusting, but the insides seemed ok with no visible problems. The damn thing is complicated so it took a lot of reading and tracing of wires to figure stuff out.
Here is the left channel output stage getting put back in.
An externally hosted image should be here but it was not working when we last tested it.
The VARIAC was dusted off and the thing carefully brought up to power....and what do you know, it took a week to confirm but the thing actually works, well was mostly working.
Back in the day, I thought I was super cool using a 3 way AC inlet (having AC in, on/off switch and fuses). Dumb idea, the lousy inlet never worked right and I already replaced it 2 times. You would flip it "on" and nothing would happen. Turned "on" about 1 in every 8 tries, but turned off no problem. Hmmm... is it because I have a safety line rated capacitor connected who knows where? I have never seen an authoritative schematic with (1) on off switch, (2) fuse (3) line rated cap, and I'm not looking now. So out went the crappy inlet and in went another identical one without any AC rated cap. What? I still have a bunch of these terrible inlets sitting in a box. Like I said, lazy.
Here is the not-recommended AC three way inlet.
An externally hosted image should be here but it was not working when we last tested it.
Wires from two emitter resistors for bias adjustment were taken out to the back panel soldered to a non-conforming DIN type socket. The socket was removed and the hole carefully filed to fit an actual DIN socket for which mating connectors exist. Now I have a mating cable and socket for easy bias measurement.
DIN Bias socket
An externally hosted image should be here but it was not working when we last tested it.
Connected, measuring bias.
An externally hosted image should be here but it was not working when we last tested it.
The output stage of this amp uses 4 metal transistors per rail (16 total) which are like tiny blowtorches on a large tubular vertical heat sink. A small bottom fan pushes air up, and a larger 120mm top fan also pulls air up. There is no cover and the fans are essential, but noisy.
The geeks that do PC overclocking would cringe at this setup as the inlet and outlet fans are mismatched and you are essentially cooking the top fan by blowing hot air into its mechanism. There are now more advanced, quieter fans and "static pressure" fans designed to be mounted right next to a radiator. Not having a clue I bought two of the newer Noctua fans (regular and static pressure) to try out.
Fan Assortment- Old Aspire and new Noctua NF-P12 static pressure fan and NF-S12A FLX normal fan.
An externally hosted image should be here but it was not working when we last tested it.
I did experiments with closing off some of the surrounding open space to try to force all airflow through the heat sink fins, but it turns out that the best cooling is with the wider opening (tape removed).
Trying out fans and airflow
An externally hosted image should be here but it was not working when we last tested it.
In the old days the Aspire fan kept things around 60C, the newer very quiet static pressure Noctua was 10C worse, and the normal Noctua NF-S12A was not quite as quiet as the static pressure Noctua but only +5C (so 65C in operation). I am going with this one now.
Last edited:
More stuff
I know what you are thinking- where the effff is the top cover? There never was one so its back to frontpanelexpress.com and measuring where to drill holes and put openings for air vents.
New Front Panel- $130.00 (top panel)
Got some "modders mesh" for the grills as I needed something that would not impede airflow. After fun with toxic fumes they are all painted.
The top cover has ledges built in so the grills can sit right on them, dabbed in put a small amount of adhesive to hopfully hold them there. There are also four #4-40 threaded holes in each set of ledges for screws to hold down the grills, unused as of now.
Front Grill
Rear Grill
Amp with a drink for size comparison.
Finally, the fan controller. This has been great to have, it has 4 heat sensors clamped to various heat sinks and a controller to set the fan speeds to individual RPM levels, for most all pcm fans. It also has an optional alarm to sound beeps at a preset temp for every zone.
Over the last 10 years sensor #1 (soft start heatsink) has failed but sensor #2 (driver heatsink) and #3/#4 (output stage heat sinks) are still going strong. Maybe some day a more 'analog' type of controller will be put in if my aging hands can still hold the iron.
I know what you are thinking- where the effff is the top cover? There never was one so its back to frontpanelexpress.com and measuring where to drill holes and put openings for air vents.
New Front Panel- $130.00 (top panel)
An externally hosted image should be here but it was not working when we last tested it.
Got some "modders mesh" for the grills as I needed something that would not impede airflow. After fun with toxic fumes they are all painted.
An externally hosted image should be here but it was not working when we last tested it.
The top cover has ledges built in so the grills can sit right on them, dabbed in put a small amount of adhesive to hopfully hold them there. There are also four #4-40 threaded holes in each set of ledges for screws to hold down the grills, unused as of now.
Front Grill
An externally hosted image should be here but it was not working when we last tested it.
Rear Grill
An externally hosted image should be here but it was not working when we last tested it.
Amp with a drink for size comparison.
An externally hosted image should be here but it was not working when we last tested it.
Finally, the fan controller. This has been great to have, it has 4 heat sensors clamped to various heat sinks and a controller to set the fan speeds to individual RPM levels, for most all pcm fans. It also has an optional alarm to sound beeps at a preset temp for every zone.
An externally hosted image should be here but it was not working when we last tested it.
Over the last 10 years sensor #1 (soft start heatsink) has failed but sensor #2 (driver heatsink) and #3/#4 (output stage heat sinks) are still going strong. Maybe some day a more 'analog' type of controller will be put in if my aging hands can still hold the iron.
Last edited:
Bring the juice
I am still not convinced that a soft start which places a resistive or NTC load on the transformer primaries for a short time and then bypasses this increased load after time expires is the way to go. Even though I have not yet blown a fuse or anything using this method. I use this delay/NTC method on quite a few amps and when the NTC/resistor is removed you still get a "boooiing" as the transformer takes its standard load, even after like a 10 second delay.
Here is a video of the amp being turned on. I have left the audio/ sound on so you can hear me grunt as well as the noise of the fan. About a second or two after the power switch on the back is flipped you hear the soft start relay click and the "booooing" of the transformer going full throttle.
KSA Video
The amp uses some serious power, biased at 55W class A, each channel draws 2 amps at 120V from the wall, slightly over 4A overall after warm up, which takes about 30 minutes. And because it is class A, it does not matter if music is playing or not- you will draw slightly over 4 amps. Due to the huge capacitors it continues playing for about 5 seconds even after switched off.
I am still not convinced that a soft start which places a resistive or NTC load on the transformer primaries for a short time and then bypasses this increased load after time expires is the way to go. Even though I have not yet blown a fuse or anything using this method. I use this delay/NTC method on quite a few amps and when the NTC/resistor is removed you still get a "boooiing" as the transformer takes its standard load, even after like a 10 second delay.
Here is a video of the amp being turned on. I have left the audio/ sound on so you can hear me grunt as well as the noise of the fan. About a second or two after the power switch on the back is flipped you hear the soft start relay click and the "booooing" of the transformer going full throttle.
KSA Video
The amp uses some serious power, biased at 55W class A, each channel draws 2 amps at 120V from the wall, slightly over 4A overall after warm up, which takes about 30 minutes. And because it is class A, it does not matter if music is playing or not- you will draw slightly over 4 amps. Due to the huge capacitors it continues playing for about 5 seconds even after switched off.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.