Several years back I recapped the power supply per THIS thread (which includes a partial schematic - incomplete but the best I've seen). In the past year or so I noticed a channel imbalance. On the bench I measured it at 6dB (not 100% sure cause I'm kind of a nOOb at all this). The KBL sounds fine. The imbalance is not very much so I've been compensating with the "symmetry" control (aka balance). But now I'm sick of it, so several questions.....
The first question is what to do next? My original plan was to try to chase down the signal in an attempt to find the failed components(s). But I'm not enough of an engineer to really know what I'm doing and it occured to me that the board has about 26 electrolytic caps that are all original. My inclination is to just recap the whole thing and see what happens. Would that be the best starting point? And is it possible that this just might fix things?
Second - best caps to use? The originals are all Nichicons: 470uf/100v/VX(M), 1000uf/50v/VR(M) and 100uf/10v/VX(M). Better choices?
Third - the board has 10 tiny trim pots. What to do about those after recapping? Reset somehow? But on what basis? I could try to measure some voltages on the working channel before recapping and then reset them to that after recapping? Or just leave alone if it sounds OK? Or what?
Any comments or suggestions would be greatly appreciated. I'll post results when it's all over. Thanks and cheers,
The first question is what to do next? My original plan was to try to chase down the signal in an attempt to find the failed components(s). But I'm not enough of an engineer to really know what I'm doing and it occured to me that the board has about 26 electrolytic caps that are all original. My inclination is to just recap the whole thing and see what happens. Would that be the best starting point? And is it possible that this just might fix things?
Second - best caps to use? The originals are all Nichicons: 470uf/100v/VX(M), 1000uf/50v/VR(M) and 100uf/10v/VX(M). Better choices?
Third - the board has 10 tiny trim pots. What to do about those after recapping? Reset somehow? But on what basis? I could try to measure some voltages on the working channel before recapping and then reset them to that after recapping? Or just leave alone if it sounds OK? Or what?
Any comments or suggestions would be greatly appreciated. I'll post results when it's all over. Thanks and cheers,
I think replacing all electrolytic capacitors is a desperate and very bad move.
Blindly replacing any component hoping to solve a problem is a no go.
You need to localize the problem first.
Could be you need a signal source (signal generator or a CD player) and an oscilloscope.
Some basic knowledge is required!
Blindly replacing any component hoping to solve a problem is a no go.
You need to localize the problem first.
Could be you need a signal source (signal generator or a CD player) and an oscilloscope.
Some basic knowledge is required!
Maybe I'm underestimating myself - I've got a scope, audio AC voltmeter and signal generator and I know how to use them. Thus far I've determined that the imbalance is someplace beyond the tape monitors and final output (though schematics are incomplete so there may be no circuity before the tope mon anyway, just switching). But without any specs, clear test points etc I feel like I'm flying blind.
But in any case It's not clear to me why recapping is desperate or bad. About 1/3 of the caps are in line with the power supply, so how to object to changing those? And the rest have all been operating in a rather hot, always-on environment for the last 31 years. Seems like a replace-on-principal thing to me. What am I missing?
But in any case It's not clear to me why recapping is desperate or bad. About 1/3 of the caps are in line with the power supply, so how to object to changing those? And the rest have all been operating in a rather hot, always-on environment for the last 31 years. Seems like a replace-on-principal thing to me. What am I missing?
If the capacitors don't bulge and don't leak, it's best to leave them as-is.
They won't be the source of your problem.
There is a gain set relay in each splitter circuit that could be the source of your problem.
I would check those relays carefully, and probably just replace them both at this age.
The 6dB gain difference that you find seems to confirm this.
Schematics:
https://www.hifiengine.com/manual_library/krell/kbl.shtml
They won't be the source of your problem.
There is a gain set relay in each splitter circuit that could be the source of your problem.
I would check those relays carefully, and probably just replace them both at this age.
The 6dB gain difference that you find seems to confirm this.
Schematics:
https://www.hifiengine.com/manual_library/krell/kbl.shtml
Last edited:
If the problem is a level difference between both channels, and frequency response is similar, I don't expect capacitors are the cause.
Before replacing anything, you need to find where the problem is.
Could be switch contacts, relay contacts, solder joints, resistors, etc...
You could consider recapping after the problem is solved, but even then I would not recommend it.
Ok, thanks all. Here's a related question.
It's well established that vintage auto electronics frequently fail from bad solder joints. The cruise control module in 70s-80s Mercedes is notorious for this. Resolder the joints and odds are probably 50:50 that the unit will be resurrected. To be sure, car electronics are subject to vastly greater environmental stress than home audio.
So - how often does this occur with home electronic stuff? Now I'm wondering what harm could come from reflowing the joints (which seem to hand soldered by the factory, not wave solder or whatever). Cheers,
It's well established that vintage auto electronics frequently fail from bad solder joints. The cruise control module in 70s-80s Mercedes is notorious for this. Resolder the joints and odds are probably 50:50 that the unit will be resurrected. To be sure, car electronics are subject to vastly greater environmental stress than home audio.
So - how often does this occur with home electronic stuff? Now I'm wondering what harm could come from reflowing the joints (which seem to hand soldered by the factory, not wave solder or whatever). Cheers,
Auto use has high vibration and wide temperature variation.
Home problems would have more moderate thermal cycling.
Home problems would have more moderate thermal cycling.
I think that resoldering suspect joints and pads on a pcb causes no harm at all.
(if done correctly of course).
Almost all consumer electronics develop bad solder joint after some years.
Especially around components that get hot.
It is one of the major error causes with older equipment.
Just the kind of encouragement I was hoping for, thanks. Makes perfect sense to me. The KBL does indeed run hot, VERY hot compared to most line level devices. Cheers,
Some progress to report. Per your encouragement I tried to actually diagnose the problem. I don't entirely know what I'm doing so any comments on the approach & thought process would be welcome - I'm trying to get better with this stuff. Here goes.
The schematic is the right channel phase splitter. The outputs go on to feed other stuff (not shown). First, it was suggested that the imbalance problem might arise from old, failed relay(s) so I replaced them (K3 below). Didn't help so I tried to figure it out better. A 1khz sine wave was fed to the preamp from a signal generator. I worked my way down the signal path from the preamp outputs with a scope to compare the signals on the respective channels. In hindsight I should have done it from the input side. It turned out that the imbalance went all the way back to Q90 which turned out to be shorted (Gate <-> Drain). How/why would this happen?? No idea, but it was.
I got replacements on order. Now more questions arise. I also ordered replacements for Q88 on the theory that replacing both phases would be a good idea cause of aging and so maybe better matching. And then, if they're replaced on one channel, why not do the other channel as well to try to keep things more matched on both side??
So, does this make sense? Anything else I ought to be thinking about? All the comments have been very helpful so bring it on! Thanks and cheers,
The schematic is the right channel phase splitter. The outputs go on to feed other stuff (not shown). First, it was suggested that the imbalance problem might arise from old, failed relay(s) so I replaced them (K3 below). Didn't help so I tried to figure it out better. A 1khz sine wave was fed to the preamp from a signal generator. I worked my way down the signal path from the preamp outputs with a scope to compare the signals on the respective channels. In hindsight I should have done it from the input side. It turned out that the imbalance went all the way back to Q90 which turned out to be shorted (Gate <-> Drain). How/why would this happen?? No idea, but it was.
I got replacements on order. Now more questions arise. I also ordered replacements for Q88 on the theory that replacing both phases would be a good idea cause of aging and so maybe better matching. And then, if they're replaced on one channel, why not do the other channel as well to try to keep things more matched on both side??
So, does this make sense? Anything else I ought to be thinking about? All the comments have been very helpful so bring it on! Thanks and cheers,
In my KSL, I had the balance pot shorted out when it was in for a recap. Never missed it and its problems..
Are you sure Q90 is defective?
FET's are tricky to test with the diode test of a digital multimeter. Google the right procedure.
If Q90 is indeed defective, you will have to replace it. Success!
Well, as sure I can be. I did google the right procedure. I also pulled the corresponding JFET from the good channel to compare. I know, one is P-channel and the other is N-. I took that into account and Q90 really does seem to be shorted. But the proof will be in the pudding and revealed when replacements arrive. Is there much likelihood that Q90 shorting would take something else out with it?
Did you solve the problem with the KBL? If not you may want to check the volume control. The Spectrol wire wound VC does wear out and mistrack. I replaced mine with a P&G model that was removed from a KRC that had been upgraded to HR status.
I actually did repair the KBL, just forgot to post, too much going on. Anyway, it really was Q90. I replaced that, it's companion on the other side of the circuit as well as their fellows on the other channel just for good measure. Also recapped the whole thing. The problem was fixed and it sounds as good as it ever has. As for the volume control, it seems fine. No noise thus far, thank good! Thanks to all and cheers,
Good to know. Mine sounds great after recapping and a dead power supply which needed recapping. And the volume control was worn out which I replaced. The only problem it seems is the gain switch doesn't seem to work and the power LED is too bright even after adding many resisters.
I went through the PS thing about 5 years ago. A diode, resistor and a couple regulators died. So now it's all pretty much resurected. As I recall there relays in the gain path. I wonder if that could be it. Shouldn't be hard to test without removing them. They're available on eBay and not to expensive.
I'll delve further into it although it's no big deal. It's too loud without anymore gain. The VC can't be much higher than 1/8 the way up.
Hmmmm. Even if it's usable it would bug me knowing that something wasn't right. The VC thing is most definitely whacky. I'd be afraid of blowing ujp my amp by making a mistake. Can replacement VC be wrong? Or you need a series resistor? I noticed that on my KBL the VC had a couple resistors on the VC. I didn't pay attention cause that wasn't a problem, but they were there for whatever reason.
It was pretty simple changing the VC. I changed one wire at a time and I believe the resisters were there to change the impedance of the Spectrol to match 10k. They weren't needed on the new one. The Spectrol has extra posts that were not on the P&G. I don't believe you can blow anything up if you get it wrong.
Here's what Krell told me:
White wires = ground FYI: One set of one of each color required for one channel of volume control.
Red wires = Hot or input
Blue wires = wiper or output
You can choose which ever set of white, red and blue wires you want to assign for either the left or right channel control.
All you have really have to do is find out how the current pot is wired in and you should be able to easily substitute the P&G in it's place.
Of course you can use a different model VC. But Krell happened to have an old P&G they took out of a KRC and they gave it to me years ago. I doubt they have any left. But you might check.
Here's what Krell told me:
White wires = ground FYI: One set of one of each color required for one channel of volume control.
Red wires = Hot or input
Blue wires = wiper or output
You can choose which ever set of white, red and blue wires you want to assign for either the left or right channel control.
All you have really have to do is find out how the current pot is wired in and you should be able to easily substitute the P&G in it's place.
Of course you can use a different model VC. But Krell happened to have an old P&G they took out of a KRC and they gave it to me years ago. I doubt they have any left. But you might check.