I was able to acquire a dead Krell KMA-160 for $0. The driver board of the amplifier seems to have set itself on fire. All the components in the vicinity of the last driver stage have been destroyed, including the transistors, several caps, and a few resistors, one of which is merely scorched, the other if which actually exploded. In addition, a diode in a power supply snubber -- the driver board has a power supply separate from the main power supply -- has exploded. Scary stuff indeed.
What's particularly interesting is there's nothing special about this amp. The topology is unremarkable and the quality of construction screams DIY. In fact the build quality of my Aleph-X monoblocks is, IMHO, quite a lot better than this thing.
Of course, the Krell precedes my efforts by 15 years
I hope to get photos and schematics up soon. So far I don't see any exotic parts. There's plenty of MJE 15030/15031 driver transistors, as well as A968/C2238 (these are not marked 2SA/2SC but they do have a great big capital "T" printed on the case. Are they Toshiba?) The small-signal BJTs are A970 & C2240 like you might expect. The power supply seems really basic, but I'll report more after I've diagrammed the whole thing.
Oh, and on the back of the board is the gross rework hack. Eww.
Anyone else have experience repairing, or improving, this unit?
What's particularly interesting is there's nothing special about this amp. The topology is unremarkable and the quality of construction screams DIY. In fact the build quality of my Aleph-X monoblocks is, IMHO, quite a lot better than this thing.
Of course, the Krell precedes my efforts by 15 years
I hope to get photos and schematics up soon. So far I don't see any exotic parts. There's plenty of MJE 15030/15031 driver transistors, as well as A968/C2238 (these are not marked 2SA/2SC but they do have a great big capital "T" printed on the case. Are they Toshiba?) The small-signal BJTs are A970 & C2240 like you might expect. The power supply seems really basic, but I'll report more after I've diagrammed the whole thing.
Oh, and on the back of the board is the gross rework hack. Eww.
Anyone else have experience repairing, or improving, this unit?
I am VERY interested in seeing the final schematic and any photos you take.
Yes these early Krells like the 80's the 100's and 160's were all hand built if i am not mistaken. Krell was very new at that time. 1984 i think. but then again, look at the very early Levinson stuff as well, they had MFG problems as well.
BUT, Dan D Agostino is still the king of SS amps in my book! I dream of owning a KSA-250 some daY! Droool!!!
Zero
Yes these early Krells like the 80's the 100's and 160's were all hand built if i am not mistaken. Krell was very new at that time. 1984 i think. but then again, look at the very early Levinson stuff as well, they had MFG problems as well.
BUT, Dan D Agostino is still the king of SS amps in my book! I dream of owning a KSA-250 some daY! Droool!!!
Zero
Hi jwb,
There is nothing wrong with the type of outputs used. Just make sure they are matched.
X.G. is correct in stating that Japanese transistors only have one manufacturer per type number. Use real parts, do not go anywhere near ECG's & NTE's. At least you will know what is in your hand for sure. Also, don't use any sub manuals for cross referencing to a different part #. The ECG/NTE manual is known as "the book of lies" up here. For good reason.
The fire is normal when the unit has been struck by technician. Sometimes, amplifiers just die that way. I had a Bryston 4B that went that way, burned a hole in the PCB. The guys at Bryston said this was impossible, their amps never fail. Right.
So, take your time and be careful. Use a variac so you don't lose anything else.
-Chris
There is nothing wrong with the type of outputs used. Just make sure they are matched.
X.G. is correct in stating that Japanese transistors only have one manufacturer per type number. Use real parts, do not go anywhere near ECG's & NTE's. At least you will know what is in your hand for sure. Also, don't use any sub manuals for cross referencing to a different part #. The ECG/NTE manual is known as "the book of lies" up here. For good reason.
The fire is normal when the unit has been struck by technician. Sometimes, amplifiers just die that way. I had a Bryston 4B that went that way, burned a hole in the PCB. The guys at Bryston said this was impossible, their amps never fail. Right.
So, take your time and be careful. Use a variac so you don't lose anything else.
-Chris
I'm almost done with the schematics. Here's a picture I took of the damaged board. Everything to the right of the molex connectors is the regulated power supply. There's two molex connectors because this same board is flipped and installed on the other side of the amp in the stereo model. For the same reason there are empty pads for a trimmer.
Note the scorched transistor and the molten capacitor. I think the upper transistor is a 2SA968, but I can't tell. I'm guessing because it's complementary to the 2SC2238 below it.
The input comes in on the left from the XLR on the rear panel. Note how the amp is DC-coupled and there is no input protection whatsoever. The gates of those poor FETs (2sk163, anybody have a datasheet?) are just asking to get blown up.
Note the scorched transistor and the molten capacitor. I think the upper transistor is a 2SA968, but I can't tell. I'm guessing because it's complementary to the 2SC2238 below it.
The input comes in on the left from the XLR on the rear panel. Note how the amp is DC-coupled and there is no input protection whatsoever. The gates of those poor FETs (2sk163, anybody have a datasheet?) are just asking to get blown up.
An externally hosted image should be here but it was not working when we last tested it.
I sure wish i could "Aquire" some dead Krell amps!!!! what a find! I sure hope you bought the person that you got those amps from at least a case or 2 of guiness!!!!
Anyone else want to off load any old Krell, Levinson, Threshold, Classe, Bedini, Sumo, etc etc. power amps for "$0.00" Dollars or very little dollars PLEASE let me know!!!
I will make sure your favorite beverage truck makes a stop by your house!!!!!
And Mondo thanks to JWB for enlightening us with this privvied info! Now if i could just get the schematics for a KSA-250, i would be in heaven!!!!
JWB, please keep going!!! im dying to see those schematics!!! Thanks for the photo. What a great piece of art, you should call it "Amp PCB scorched by fire"
And Please tell me that you plan on rebuilding these babies...right?
Zero
Anyone else want to off load any old Krell, Levinson, Threshold, Classe, Bedini, Sumo, etc etc. power amps for "$0.00" Dollars or very little dollars PLEASE let me know!!!
I will make sure your favorite beverage truck makes a stop by your house!!!!!
And Mondo thanks to JWB for enlightening us with this privvied info! Now if i could just get the schematics for a KSA-250, i would be in heaven!!!!
JWB, please keep going!!! im dying to see those schematics!!! Thanks for the photo. What a great piece of art, you should call it "Amp PCB scorched by fire"
And Please tell me that you plan on rebuilding these babies...right?
Zero
2sk163, anybody have a datasheet?
2SK163 N-FET 50V 0.03A 0.4W
J113 is full analog
can find it on
www.semiconductors.philips.com
www.fairchildsemi.com
www.onsemi.com
2SK163 N-FET 50V 0.03A 0.4W
J113 is full analog
can find it on
www.semiconductors.philips.com
www.fairchildsemi.com
www.onsemi.com
I completed the reverse engineering and now I'm just making nicer drawings. Here's the power supply. No surprises, it's a basic zener-regulated supply. Outputs are ±85V and ±38V. This supply runs the frontend and the output drivers, but the output itself runs from a separate, unregulated, bulk supply.
An externally hosted image should be here but it was not working when we last tested it.
Here's the output stage. Points A, B, C, and D are driven from the input, a pair of FET followers and dual differential voltage gain stages. There's also a DC servo. Both of those drawings to follow.
An externally hosted image should be here but it was not working when we last tested it.
I decided to draw the input stage with the CAD program. Tired hands...
An externally hosted image should be here but it was not working when we last tested it.
Here's the servo. That's all the info you need to build your own KMA-160. Just add a suitable power supply for the output stage.
An externally hosted image should be here but it was not working when we last tested it.
Some notes:
C1 and C2 in the output appear to have been added as part of a rework. They are on the wrong side of the board and stuck in the holes for some other components.
R3 and R5 in the input were certainly added in the rework. Initially the bases of Q8 and Q18 were driven directly from the divider R1/R2. There is a trace under R3, which has been manually cut to accomodate the change.
R56 is used to adjust the output DC. The 5K pot in the output is used to trim the bias.
The value of the zener on the servo is unknown. It's less than 25V.
There are many places where a 4K7/4K7 divider is used to bias the base of a transistor. For example, R44/R45 in the servo. This divider is shown many times but in the real amp there's only one such pair of resistors for each rail.
The amp I am working with was destroyed when 87V zener in the power supply exploded. Apparently, the amp didn't shut down, and the DC feedback circuit quickly destroyed itself. It might be advisable to use two zeners in the high-voltage supply, for redundancy. Also there should be a shutdown circuit in case of power supply failure. The Krell has neither.
The input of this amp is DC-coupled without protection. You could easily destroy the input FETs with a hot-plugged cable. I would advise adding zener protection on the input, and possibly also coupling capacitors.
C1 and C2 in the output appear to have been added as part of a rework. They are on the wrong side of the board and stuck in the holes for some other components.
R3 and R5 in the input were certainly added in the rework. Initially the bases of Q8 and Q18 were driven directly from the divider R1/R2. There is a trace under R3, which has been manually cut to accomodate the change.
R56 is used to adjust the output DC. The 5K pot in the output is used to trim the bias.
The value of the zener on the servo is unknown. It's less than 25V.
There are many places where a 4K7/4K7 divider is used to bias the base of a transistor. For example, R44/R45 in the servo. This divider is shown many times but in the real amp there's only one such pair of resistors for each rail.
The amp I am working with was destroyed when 87V zener in the power supply exploded. Apparently, the amp didn't shut down, and the DC feedback circuit quickly destroyed itself. It might be advisable to use two zeners in the high-voltage supply, for redundancy. Also there should be a shutdown circuit in case of power supply failure. The Krell has neither.
The input of this amp is DC-coupled without protection. You could easily destroy the input FETs with a hot-plugged cable. I would advise adding zener protection on the input, and possibly also coupling capacitors.
X.G. said:
OK, here are some pictures. The first is the main circuit board, containing most of the circuits diagrammed here.
An externally hosted image should be here but it was not working when we last tested it.
This is what a KMA-160 looks like when you take it to pieces.
An externally hosted image should be here but it was not working when we last tested it.
The main power supply. Note the quad supply, for stereo operation.
An externally hosted image should be here but it was not working when we last tested it.
Here's one of the four output boards.
An externally hosted image should be here but it was not working when we last tested it.
And this is another output board mounted on a heatsink.
An externally hosted image should be here but it was not working when we last tested it.
Enjoy.
wish to finish it in a week.
nice job JWB
What will you use for the power Fets, Q1,2,3,4, in the output stage? Are the VN and VP0210N5 available anywhere?
YOu didn't specify the voltage of the Bulk supply for the output stage. Do you happen to know what it is.
I'd sure like to built this amp but I think I'd better try something a little simpler first. I'll be keep this info you posted for future reference.
Regards,
What will you use for the power Fets, Q1,2,3,4, in the output stage? Are the VN and VP0210N5 available anywhere?
YOu didn't specify the voltage of the Bulk supply for the output stage. Do you happen to know what it is.
I'd sure like to built this amp but I think I'd better try something a little simpler first. I'll be keep this info you posted for future reference.
Regards,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.