I just bought a Krell KSA-150 that was sent to Krell for repair but the owner turned down the repair estimate. I made a deal for the amp, from the owner, "as is" and just got the amp today. I tore it down for inspection this afternoon and spent a few hours going through it.
Observations:
Krell gave a parts list of what they said it would need. Frankly I don't see how they could have done that, as the dust patterns in the unit were totally undisturbed. Nobody's been in that amp since it was made!
The amp modules are connected to the soft-start PC board using Phoenix push-locking spring connectors. Which frankly I don't like at all. Those will
be going away, and soldered directly to the PC board.
I found one of the connections almost burned through. Not due to catastrophic failure, but due to age and questionable connectivity.
Every failure I've found was on the soft start board. There is a design flaw apparent, with the inrush limiting resistors, big 10 watt range 700 ohm
vitreous ceramic types, being mounted too close to the PC board in a limited space right next to the relays and the protection MOV.
The heat given off by those inrush limiting resistors has totally cooked the MOV, both relays, and the nearby 100 uF 63 volt electrolytics, one of which
shorted and actually has a hole burned right through the aluminum shell. Its formerly blue shrink wrap label is a black fragment of charred plastic. Fortunately there seems to have been no fire.
The relays took a lot of thermal damage. I'll replace them even if they test good. Their PC board connections are the most majestic case of ring cracks I have ever seen.
The MOV was also cooked through. The epoxy shell of it is cracked clean
through on both sides. It will of course be replaced.
All this could have been averted by standing the inrush limiting resistors off
into a free air space and NOT so close to the components. I'll be able to
install new ones and give them an inch or so of standoff space and that'll be much better in the long run.
Due to heat issues, every capacitor on the soft start board will be replaced,
and the whole board needs to be resoldered.
All the emitter resistors on the PA rails seem to be in perfect condition, so
I suspect that there are fortunately NO damaged transistors.
This looks like it will be a pretty quick and easy fix. I'll replace all electrolytics
but not the main reservoir caps, since they're huge computer grade caps
that will probably last another 20 years or more, but all the rest looks like
maybe a 30 dollar cap order. Add an MOV and some big resistors and I
think I'll have a quick and easy Krell repair.
Pics coming soon. I didn't make any today.
The cosmetic condition of the amp is perfect.
I'm surprised that Krell actually designed a heat management problem into
the amp. There's no reason for it. ANYONE with any experience would
know that inrush limiting resistors get HOT and need ROOM TO BREATHE.
Shoving them into a corner with limited airflow, next to plastic relays
and heat sensitive electrolytic caps, was NOT smart design.
CJ
Observations:
Krell gave a parts list of what they said it would need. Frankly I don't see how they could have done that, as the dust patterns in the unit were totally undisturbed. Nobody's been in that amp since it was made!
The amp modules are connected to the soft-start PC board using Phoenix push-locking spring connectors. Which frankly I don't like at all. Those will
be going away, and soldered directly to the PC board.
I found one of the connections almost burned through. Not due to catastrophic failure, but due to age and questionable connectivity.
Every failure I've found was on the soft start board. There is a design flaw apparent, with the inrush limiting resistors, big 10 watt range 700 ohm
vitreous ceramic types, being mounted too close to the PC board in a limited space right next to the relays and the protection MOV.
The heat given off by those inrush limiting resistors has totally cooked the MOV, both relays, and the nearby 100 uF 63 volt electrolytics, one of which
shorted and actually has a hole burned right through the aluminum shell. Its formerly blue shrink wrap label is a black fragment of charred plastic. Fortunately there seems to have been no fire.
The relays took a lot of thermal damage. I'll replace them even if they test good. Their PC board connections are the most majestic case of ring cracks I have ever seen.
The MOV was also cooked through. The epoxy shell of it is cracked clean
through on both sides. It will of course be replaced.
All this could have been averted by standing the inrush limiting resistors off
into a free air space and NOT so close to the components. I'll be able to
install new ones and give them an inch or so of standoff space and that'll be much better in the long run.
Due to heat issues, every capacitor on the soft start board will be replaced,
and the whole board needs to be resoldered.
All the emitter resistors on the PA rails seem to be in perfect condition, so
I suspect that there are fortunately NO damaged transistors.
This looks like it will be a pretty quick and easy fix. I'll replace all electrolytics
but not the main reservoir caps, since they're huge computer grade caps
that will probably last another 20 years or more, but all the rest looks like
maybe a 30 dollar cap order. Add an MOV and some big resistors and I
think I'll have a quick and easy Krell repair.
Pics coming soon. I didn't make any today.
The cosmetic condition of the amp is perfect.
I'm surprised that Krell actually designed a heat management problem into
the amp. There's no reason for it. ANYONE with any experience would
know that inrush limiting resistors get HOT and need ROOM TO BREATHE.
Shoving them into a corner with limited airflow, next to plastic relays
and heat sensitive electrolytic caps, was NOT smart design.
CJ
A big PSU is necessary on the Krells, hence the soft start. I would recap the boards, with age and the high temps they deteriorate fast and lead to most of the misstatements about the sound quality of the amplifier.
I have also seen then re-biased and the sound is better than the high class -a bias used on the KSA150. On the KSA200S Krell did reduce the amt of class-a bias and the amp was much better sounding than the KSA200 before it.
regards,
I have also seen then re-biased and the sound is better than the high class -a bias used on the KSA150. On the KSA200S Krell did reduce the amt of class-a bias and the amp was much better sounding than the KSA200 before it.
regards,
Yes, I will be recapping the whole unit, except for the main reservoir caps which are expensive and don't need it.
Information on rebiasing the amp for better performance will be greatly appreciated.
I'm going to have to replace the relays on the soft start board and am contemplating putting in sockets for them to keep the heat away from the board a little, but I'm not sure if this actually provides any benefit.
The board has taken enough heat damage in that area that it's kind of fragile. Traces will probably lift very easily and I want to avoid undue stress.
CJ
Information on rebiasing the amp for better performance will be greatly appreciated.
I'm going to have to replace the relays on the soft start board and am contemplating putting in sockets for them to keep the heat away from the board a little, but I'm not sure if this actually provides any benefit.
The board has taken enough heat damage in that area that it's kind of fragile. Traces will probably lift very easily and I want to avoid undue stress.
CJ
Well, I'm well into this amp now. I've almost finished completely rebuilding one channel.
Every electrolytic cap in both channels is being replaced, and I have found that a few had started to leak but there's no board damage so I got to them just in the nick of time.
The amount of labor involved is mind-boggling. I have six hours invested in rebuilding just one channel so far, and that's straight working time. Part of that is learning curve, but still, this is a deceptive design. It's far more labor intensive to build or work on than you'd guess at first.
I checked every output transistor (why not, they have to come out anyway before you recap the driver board) and every one in the first channel is good. Second channel, not checked yet.
So far, so good. It will be nice to complete this in the next few days and hear it running.
CJ
Every electrolytic cap in both channels is being replaced, and I have found that a few had started to leak but there's no board damage so I got to them just in the nick of time.
The amount of labor involved is mind-boggling. I have six hours invested in rebuilding just one channel so far, and that's straight working time. Part of that is learning curve, but still, this is a deceptive design. It's far more labor intensive to build or work on than you'd guess at first.
I checked every output transistor (why not, they have to come out anyway before you recap the driver board) and every one in the first channel is good. Second channel, not checked yet.
So far, so good. It will be nice to complete this in the next few days and hear it running.
CJ
I've completed all component testing and nearly all replacement of capacitors.
Sum total of component failures aside from known cap age issues:
One 100 v, 100 uF cap burned up at the bottom and almost burned the PC board,
and the resulting short took out a single .33 ohm 2 watt ceramic resistor, which
happens to be the one that feeds a collector rail on the output board. This would have
taken all of the 10 transistors on that rail out of operation but they're not damaged.
This imbalance would have shut down the amp immediately, I believe.
All transistors check good.
I won't declare victory yet because there's more to be done, but it looks like the
total damage to the amp is one resistor, two heat damaged relays, one heat damaged
varistor, and a big handful of caps that were all due for replacement with a few showing
signs of failure in progress, or fully failed.
I would say based on this experience that EVERY KSA-150, KSA-250, MDA-300, or MDA-500 amp is by now well due for a total overhaul. Those caps are dying and they
will kill the amps if not taken care of before that happens.
There's probably a good market for overhauling these amps for less than Krell charges.
However, it's a lot of work. I would say it's a solid 10 hours of labor if you know your
way around these units. Maybe more.
CJ
Sum total of component failures aside from known cap age issues:
One 100 v, 100 uF cap burned up at the bottom and almost burned the PC board,
and the resulting short took out a single .33 ohm 2 watt ceramic resistor, which
happens to be the one that feeds a collector rail on the output board. This would have
taken all of the 10 transistors on that rail out of operation but they're not damaged.
This imbalance would have shut down the amp immediately, I believe.
All transistors check good.
I won't declare victory yet because there's more to be done, but it looks like the
total damage to the amp is one resistor, two heat damaged relays, one heat damaged
varistor, and a big handful of caps that were all due for replacement with a few showing
signs of failure in progress, or fully failed.
I would say based on this experience that EVERY KSA-150, KSA-250, MDA-300, or MDA-500 amp is by now well due for a total overhaul. Those caps are dying and they
will kill the amps if not taken care of before that happens.
There's probably a good market for overhauling these amps for less than Krell charges.
However, it's a lot of work. I would say it's a solid 10 hours of labor if you know your
way around these units. Maybe more.
CJ
Quick question for anyone who might know the answer:
On both channels of this KSA-150, capacitor C20 on the output driver board is mounted
in reverse polarity as compared to the solder mask legend printed right on the PC board.
Is this a mounting error, or is the solder mask marked incorrectly?
I'm not sure I'd buy the idea that an electrolytic capacitor survived 20 years in the amp
installed backwards without exploding or showing any outward signs of failure.
CJ
On both channels of this KSA-150, capacitor C20 on the output driver board is mounted
in reverse polarity as compared to the solder mask legend printed right on the PC board.
Is this a mounting error, or is the solder mask marked incorrectly?
I'm not sure I'd buy the idea that an electrolytic capacitor survived 20 years in the amp
installed backwards without exploding or showing any outward signs of failure.
CJ
This evening was reassembly and test.
A little smoke came out.
But...it's not a major problem, or so I believe. The smoke occurred right under the same cap that burned earlier, on the right side preamp/driver board, which connects directly to the collector rail through a 0.33 ohm fusible safety resistor.
I removed the cap after the smoke check and the board was burned (again, I had cleaned it thoroughly) and the cap had some smoke on the bottom of it but it had not shorted and was not the cause of the smoke.
There's a small crater in the board just beside the cap leads, showing visible copper from an internal trace A quick scrape of the edge of the board confirmed that this in fact a four layer PC board.
I BELIEVE that the chain of events was as follows:
Old weak cap leaked, and got hot. It carbonized the board over the underlying trace,
giving it a low resistance path and causing a short and additional localized heating.
When I cleaned it, I didn't clean it well enough. I thought I had, and had inspected the board at that time and found NO craters in it, but apparently that crater had simply not
yet opened up. It was still carbonized inside the board.
I recapped the board, and when I fired up the amp this evening, the hidden board
damage erupted, exposing the carbonized short to the 19 volt rail that the cap connects to.
It's possible that now, after a very thorough cleaning to remove all short circuit issues
in the PC board, I may be able to make suitable repairs to the board and get the amp
up and running. But I really need to try to determine what's connected on board's
inner layers first.
A little smoke came out.
But...it's not a major problem, or so I believe. The smoke occurred right under the same cap that burned earlier, on the right side preamp/driver board, which connects directly to the collector rail through a 0.33 ohm fusible safety resistor.
I removed the cap after the smoke check and the board was burned (again, I had cleaned it thoroughly) and the cap had some smoke on the bottom of it but it had not shorted and was not the cause of the smoke.
There's a small crater in the board just beside the cap leads, showing visible copper from an internal trace A quick scrape of the edge of the board confirmed that this in fact a four layer PC board.
I BELIEVE that the chain of events was as follows:
Old weak cap leaked, and got hot. It carbonized the board over the underlying trace,
giving it a low resistance path and causing a short and additional localized heating.
When I cleaned it, I didn't clean it well enough. I thought I had, and had inspected the board at that time and found NO craters in it, but apparently that crater had simply not
yet opened up. It was still carbonized inside the board.
I recapped the board, and when I fired up the amp this evening, the hidden board
damage erupted, exposing the carbonized short to the 19 volt rail that the cap connects to.
It's possible that now, after a very thorough cleaning to remove all short circuit issues
in the PC board, I may be able to make suitable repairs to the board and get the amp
up and running. But I really need to try to determine what's connected on board's
inner layers first.
It is most definitely a four layer PC board. In many cases it is easy to verify that a board
is multi-layer by taking a razor knife and scraping the edge of the board. This will reveal all layers in it as separate lines of copper in the matrix, provided they go to that point on the board. I did the scrape test and all four layers are visible.
I believe that the only thing for me to fix before this amp is back in operation now is to correct this board damage. But it would be very helpful to understand the trace layout of the internal layers.
I may have to beg Krell for the board masks in order to come up with a repair plan. I have to know how that second layer is laid out and where it connects.
CJ
is multi-layer by taking a razor knife and scraping the edge of the board. This will reveal all layers in it as separate lines of copper in the matrix, provided they go to that point on the board. I did the scrape test and all four layers are visible.
I believe that the only thing for me to fix before this amp is back in operation now is to correct this board damage. But it would be very helpful to understand the trace layout of the internal layers.
I may have to beg Krell for the board masks in order to come up with a repair plan. I have to know how that second layer is laid out and where it connects.
CJ
I was able to repair that issue on the board.
This amp is making me bang my head against the wall, though. I did another test run and BAM, there goes a 1000 uF, 63 volt capacitor. Blew its guts all over the inside of the amp.
Now there's a bit of ambiguity for you: I most carefully placed that cap in the same polarity as the original one. I have the photos to prove it. But when I removed the
bad cap, the holes for the cap on the board indicate that the polarity was reversed.
This was one of two caps in parallel, on the power control board. Both in the same polarity.
So I yanked both of them and installed new ones, following the indicated polarity on the board this time. Square pad, positive leg.
I fired up the amp again...and something started to smoke in the left channel. (The board repair I did was on the right channel.) I killed power and, as part of my checks,
felt the two caps I'd just replaced. They were rather warm for having been powered up for just a few seconds.
I guess I'll have to do a live voltage test on that board to absolutely verify what polarity is right. AFTER I find and clear the new burned part.
This is starting to get a bit old.
CJ
This amp is making me bang my head against the wall, though. I did another test run and BAM, there goes a 1000 uF, 63 volt capacitor. Blew its guts all over the inside of the amp.
Now there's a bit of ambiguity for you: I most carefully placed that cap in the same polarity as the original one. I have the photos to prove it. But when I removed the
bad cap, the holes for the cap on the board indicate that the polarity was reversed.
This was one of two caps in parallel, on the power control board. Both in the same polarity.
So I yanked both of them and installed new ones, following the indicated polarity on the board this time. Square pad, positive leg.
I fired up the amp again...and something started to smoke in the left channel. (The board repair I did was on the right channel.) I killed power and, as part of my checks,
felt the two caps I'd just replaced. They were rather warm for having been powered up for just a few seconds.
I guess I'll have to do a live voltage test on that board to absolutely verify what polarity is right. AFTER I find and clear the new burned part.
This is starting to get a bit old.
CJ
MOV, are you sure?
I think the before orange colloured "MOV" is just a capacitor . Look at the pcb it say C (don't remember the number). From mine i can read 440L, this is a capacitor from vishay/sprague X1 class safety capacitor 250V .01uf . Am i right? Restoring a ksa-250 myself...
Hi, yes i finished the repair some weeks ago, and it was alot of work!
You can see my repair on this link.
YouTube - Krell KSA-250 recapping
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.