Connect with us

stereo amp cleaning

Discussion in 'Electronic Repair' started by Dave, May 1, 2007.

Scroll to continue with content
  1. Dave

    Dave Guest

    I have a mid-80's HK integrated amp (PM-650). It's been very reliable and
    sounds good, but there are a few things that have been bugging me:

    1. The output of L & R channels are not equal... the left channel is louder
    for any given source. It's in the pre-amp section, because I can make the
    right speaker louder by flipping to "reverse stereo" so it's not the power
    amp. I've aligned/calibrated the amp using the adjustments in the service

    2. The balance knob is scratchy.

    3. Every once in awhile one or both speakers cuts out. I have to toggle
    the subsonic filter switch a few times to get it back on.

    Anyway, last night I decided to give the amp a throrough cleaning, and I
    replaced all of the electrolytic caps upstream of the reverse stereo switch.
    I liberally sprayed all of the knobs (bass/treble/balance/tape monitor
    select/volume) with MG contact cleaner and operated the knobs vigorously for
    a minute or two. I then let the amp sit for an hour. The contact cleaner
    evaporates almost instantaneously so I figured an hour was a reasonable time
    to wait for the switches to dry out. When I powered up the amp everything
    looked normal i.e. it lit up, but when I switched the speakers on (volume
    all the way down), WHOA! CRACKLY STATIC, LOUD! None of the preamp controls
    have any effect on the volume of static except my old nemesis, the subsonic
    filter which creates a hum in the right channel when switched on. Switched
    it back off, waited for four more hours. Double-checked all the solder
    joints on the caps I had replaced for solder bridges, bad joints, correct
    orientation on polar caps, etc.

    When I fired it up this time, much much less static, but a) definitely some
    residual low-level static on the right channel and b) left channel keeps
    cutting out... need to crank up the volume and it "pops" back on. Sounds
    good when it's on, and both channels very clear at high volume. Could be
    worse I suppose.

    My question is: how long does it take for this stuff to dry out? Could I
    have just moved the dirt around inside the pots and made it worse? Should I
    clean everything again?

    I seem to have transformed a infrequent mild annoyance into a show-stopper.

    Any replies greatly appreciated.

  2. DaveM

    DaveM Guest

    Since, by your own observations, the subsonic filter switch could be the
    problem, or something in the immediate vicinity of the switch. What kind of
    switch is it (slide switch, toggle switch, rotary switch)? If you can easily
    disassemble it to get at the contacts, you might find that they are very
    tarnished or even corroded; caused by airborne contaminants.
    Also sounds like the volume pots are still dirty. Use a cleaner that lubricates
    as well as cleans. A tip... when you spray the cleaner into the pot, turn the
    set so that the residual cleaner can drain back out. That allows the cleaner to
    float out any contaminants as it drains.
    I have also found transistors to cause that exact problem. I surmise that the
    wire bond from the external leads to the silicon inside breaks and becomes
    extremely intermittent, and can be very difficult to find if you aren't wary of
    this kind of fault. Lightly tap on each semiconductor in the area and see if
    the audio is affected.
    Actually, just about all types of components can become intermittent
    (microphonic) in this manner, so using an insulated tool such as a plastic or
    nylon tuning wand, tap components and circuit board to see if that affects the

    Good luck...

    Dave M
    MasonDG44 at comcast dot net (Just substitute the appropriate characters in the

    Life is like a roll of toilet paper; the closer to the end, the faster it goes.
  3. Dave

    Dave Guest

    It's a plastic push-button PCB-mounted switch, not easy to disassemble at
    all. I'd worry about breaking the plastic parts which are, I'm sure, quite
    irreplaceable. Not particularly accessible either as there is a display
    controller board above the tone control board where the switch is located.
    I was using zero residue cleaner, I'll pick up some lubricating contact
    cleaner and give it another go. There appears to be a hole in the back of
    each pot where you can see the shaft turn, plus a small (1-1.5mm) hole on
    one side. Maybe a blast of compressed air after cleaning wouldn't hurt
    This amp is totally discrete and there are no soldered wire connections (all
    wound) , so I think (hope) the fact that BOTH channels are messed up
    indicates a simple dirty pot problem... I did check voltages at maybe a
    dozen transistors, both left and right channel from PS through the pre-amp
    and amp and they were all really close to expected at idle. I don't know if
    +/- 1V is significant on a 43VDC rail or not... anyway there are quite a
    few transistors but nothing unmanageable and as there are no proprietary,
    obscure, or out of production IC's in this dog, whatever I did is completely
    fixable. I just hope that it doesn't take me 30 or 40 hours to figure it

    I may try just bridging the "open" pair of contacts on the subsonic filter
    switch with a jumper wire and see if that helps.

    Thanks for your help

  4. Guest

    The still bad switch probably needs disassembling & reassembling, or
    else replacing the individual switch or switch block. Rows of
    pushbutton things on PCBs were a standard design around for many years
    in the 70s & 80s, so finding a replacement from some scrap item is

    The pot is either still dirty, probably not, or else is worn out or
    cracked, or possibly the rivet at the end of the track is loose.

    Replacing the pot is most likely the solution, you've already cleaned
    it. Is its an oddity, rivets can be squashed, and worn out tracks can
    be painted, tho this would unbalance the volume L-R tracking,

    Re the off balance, there are reasons amps have balance controls. If
    yours has none you can fit a potential divider in the channel with the
    higher output, and set it up and replace the covers. If youre lcky and
    the amp has external connections between pre and power sections, you
    can put what you want between those 2, such as a new volume & balance

  5. René

    René Guest

    Had the exact same problem once in a mixing panel (unbalanced audio,
    one channel fading away and coming back, incidental cracle and pop
    sounds - cleaning did not help)

    turned out to be a leaky electrolyte. Found it by measuring biasses on
    every transistor in sight. Tapping the cap did not reveal anything
  6. Dave

    Dave Guest

    I re-cleaned the volume pot with some lubricating cleaner and dried it out
    with a hair dryer (on low heat setting). The crackling in the right channel
    has now all but disappeared.

    The left channel, however, will now work for about 10-30 seconds when the
    amp is first powered up, then distorts rapidly over about 1-3 seconds and
    cuts out completely after that.

    The amp is set up as follows:

    Input jacks --> input selector PCB --> volume knob/loudness PCB --> tone
    control PCB --> power amp PCB.

    I'm going to take out the input caps off the tone control and route the
    signal directly from the volume knob to the power amp. That will at least
    tell me if the problem is with the preamp or power amp. I'm pretty sure
    it's going to be in the power amp, as the second-to-last component in the
    pre-amp (tone control board) is the stereo/reverse stereo/mono switch. If
    the problem was in the preamp prior to this point, changing this switch
    should move the problem from one speaker to the other. Which it does not.
    The only downstream component from the stereo/reverse stereo/mono switch is
    the balance pot.

    When I replaced the caps on the power amp, I cleaned the board with spray
    cleaner. It was filthy, like 30 years of dust and grime. In retrospect, I
    probably shouldn't have bothered, maybe the cold temperature of the spray
    cracked off an internal lead in one of the transisotrs or something.

    The last component in the amp is "speaker protector". This is a white
    rectangular box about 1-1/4"L x 3/16"W x 1/2"H (approx, I'm not looking at
    it right now.). I don't know how this works, but could it possibly be my
    problem? Or should I be thinking transistors? I guess if I've got the
    whole amp apart it wouldn't be such a big deal to measure bias currents on
    both channels at idle... there are 14 transistors per channel including the
    My amp is integrated, but fortunately is very simple to split as the
    connection between pre-amp and amp is a three-conductor wire. I've been
    meaning to do this for awhile, I've just purchased a sub with line inputs
    and outputs... will have to cannibalize some piece of junk for a block of
    RCA jacks.

    All of my signal path connections are wire-wound, not soldered. Is there
    some tool to do this if I wanted to stay true to design and wind the
    connections on my pre-out/main-in connections? Is there any benefit to
    doing it this way? I'm guessing there probably is as it looks like a fair
    bit more work than just plonking a blob of solder on a ribbon cable lead.

    Thanks for the help.

  7. Guest

    In no particular order:

    a) cleaning pots and switches is only the first step. You need to
    provide some (slight) amount of lubrication afterwards. 100% volatile
    cleaners can cause serious friction after the skunge is removed
    rebuilding additional skunge almost instantly. The idea is to exercise
    the pot only as long as the cleaner is wet, then stop, then re-
    saturate to rinse out the residue. *AFTER* that, a wee spritz of
    lubricating cleaner makes all go smoothly.

    b) switches are even 'more so' as wear can happen in any of several
    ways including breaking of tabs or contact springs, and additional
    friction can cause failure. So, same process. Saturate, exercise for a
    few cycles whilst saturated, rinse, lubricate.

    c) Look for broken traces on the boards, look (still) for bad caps. Do
    you test the new caps *before* you install them? I have found about
    0.5% of new electrolytic caps are bad or sufficiently out of spec to
    be problematic. "Badness" typically manifests as intermittents... and
    your delay may be related. Or, one you did not replace for whatever

    d) you may believe that the power-amp section is clean, but one thing
    I have experienced that is quite peculiar but as I have seen it
    twice, it cannot be all that rare... an input transistor on the power-
    amp side has become sufficiently different from its other-channel mate
    that similar-volume signals to it are quite different in output
    volume. RARE, and the only way you will catch this is to remove the
    questionable transistors and put them either on a scope or tester that
    can determine their actual response.

    e) lastly and I have seen this only once and it was A BEAR to find. A
    perfectly fine looking 1/2 watt resistor in the signal path had
    drifted to 3X its nominal value. How did I find it? Well, every other #
    $%^&*( part checked out, so I started measuring resitance "cold"
    across both channels and comparing the results component by component.
    When I found a difference, I started lifting legs about that point.
    And there it was.

    Good luck with it.

    Peter Wieck
    Wyncote, PA
  8. Dave

    Dave Guest

    I re-cleaned with a cleaner with lubricant last night.
    Did not re-do switches. I will though.
    I don't have an ESR meter. Is testing plain ol' capacitance any help? I'd
    like to think it's a capacitor given that that's all I changed out.
    This was an overnight change, so unless the change was mechanically induced
    (definitely not ruling that out) I wouldn't be inclined to look there

    The more I think about it the less likely a pre-amp problem seems... the
    signal goes through primary amplification (a few transisitors) then through
    tone controls, subsonic filter, high cut filter, balance control --> power
    amp. I've got a switch to bypass the tone control, and switches to bypass
    the filters, plus a switch to reverse the channels between pre-amp and amp.
    None of these affect the amp output. If I've got a bad cap in the amp
    stages and am getting a DC offset at my outputs, that MAY be engaging the
    speaker protection. Or, I suppose, worst case scenario I've baked one of my
    giant expensive output transistors. I am assuming the speaker protection
    (being two leaded) is a polyswitch. The distorted sound for a second or so
    prior to speaker cut-out might match a polyswitch tripping, I don't know as
    I've only ever had amps with relay protection (or none) in the past. Do you
    know offhand if polyswitches tend to fail? I suppose probably as much as
    any other component exposed to variable voltage and current.

    I'm going to bypass the preamp section. If I can confirm that the problem
    is in the amp, I could a) check DC voltages at all transistors until I find
    one that's off or b) start swapping electrolytic capacitors from channel to
    channel. there are only four per channel in the amp section.
    Yes I've had several of those; they can try your patience.
    Thanks, I may just need it.
  9. The posts I've seen in this thread do seem to suggest a problem with
    coupling caps. Also, I've seen bad (open circuit or intermittent)
    polystyrene caps on HK's in the past.

    If you're not already using them, a sine/square generator and an
    oscilloscope would be most useful here. In addition to rapidly narrowing
    down to the affected stage, a square wave might show that you have a
    frequency response problem in the affected stage. If so, the source of the
    problem may well be more obvious.

    Remember that with coupling caps, you can often get a clue with DC voltages.
    There may be DC, or a higher DC than the other channel, on the output side
    of a bad cap, or the following transistor stage may show a low collector

    If indeed the static was not controlled by the volume or balance, then your
    problem is narrowed down a bit, assuming there is only the one problem
    involved here

    I have the service manual pdf if you need it.

    Mark Z.
  10. Dave

    Dave Guest

    Thanks Mark, I've got the service manual. I do not currently own a
    functioning oscilloscope (two fine doorstops), yes it would certainly help
    here. Ah, but these old amps are simple enough to diagnose without one.
    Nothing but resistors, caps, and transistors. And time, lots and lots of

    The problem is definitely in the amp section. I separated the pre-amp and
    amp and tested both. The pre-amp drove an external amp just fine, the amp
    exhibited the same symptoms with another preamp.

    As there are only four electrolytic caps in each channel of the amplifier,
    it's a fairly quick fix to swap them out one at a time until (hopefully) the
    amp comes back to life.

    I tested the voltages everywhere in the amp on both channels, wherever
    voltages were shown in the service manual. The affected (left) channel had
    voltages a couple of volts higher ON AVERAGE (2-2.5VDC) than the
    corresponding right channel components which were really close to expected
    values. I replaced the input coupling cap on the bad channel which did not

    Following your advice I looked at collector voltages... there is one that
    really stands out and that is Q405. Expected is -7.1V and I read -0.3V.
    This collector is tied to the emitters of Q401 and Q403 via a resistor
    network... the emitter voltages of Q401 and Q403 are also significantly off
    (low). It's a tough call when your expected voltage is less than a volt...
    Q401 emitter should be -0.6V and I see -0.2. Q403 should be -0.6 and I see
    0. The only cap in the signal path prior to these transistors is C401, the
    input coupling cap which I swapped out as a first try. Otherwise there's an
    electrolytic from the negative rail to the base of Q405... that'll be next.

    Thanks for your help.

  11. Guest

    Do you have a good transistor checker? (one that does good/bad and is
    also capable of matching?)

    Where are you (by region)?

    Peter Wieck
    Wyncote, PA
  12. Dave

    Dave Guest

    I've got a transistor checker on my multimeter, it shows an arbitrary hfe
    which, for a given transistor, never matches what's in the data sheet. Of
    course the gain in the datasheet is always at a particular voltage and
    current and I'm not sure what voltages/currents are applied by the meter...
    or whether the voltage or current is controlled at all. I'm guessing 9V at
    whatever the multimeter battery can put out for current through an unknown
    series resistor. If it's a truly fried transistor, no value is returned for
    hfe. I've had as much luck testing transistors with an ohmmeter.

    I've never really trusted it to be honest with you.

    Are you thinking driver transistor? Maybe one early in the signal path
    which could introduce DC voltage which carries right through and would
    explain the elevated voltages I am seeing in the left channel?

    I live in Penticton, British Columbia, Canada. It's about 250 miles inland
    from Vancouver and about 40 miles from the Washington border. It's a town
    of 30,000 and, as you may surmise, very little of my equipment and parts is
    purchased locally.

  13. Guest

    Exactly on the driver transistor. And "by region" I was thinking if
    you were relatively nearby I would send you my transistor checker to
    help through this process. Canada may be squirrelly what with customs
    and such, by the time it reached you, you would have found the
    problem. Then you would have to send it back.

    I have had very good luck "pairing" transistors when doing repairs/
    upgrades for 70s vintage SS amps (Dynaco, Scott, AR, Fisher), that is
    making sure the 'right-side' transistor is reasonably closely matched
    to the same 'left-side' transistor. This little instrument really does

    Not as fancy as their IT-30, but it does the trick

    Peter Wieck
    Wyncote, PA
  14. Dave

    Dave Guest

    Thanks for the link. As luck would have it, the seller lives in my home
    town of Camden, Maine. Weird.

    This tool would be really really handy and it's a helluva' lot cheaper than
    a 'scope.

    A couple of the driver transistors are not available, 2SC2603 and 2SA949's.
    NTE crosses would have to do.


  15. Guest


    And why I try to get a bit closer than a cross-reference when I use
    even direct-replacements. Unlike tubes which can vary somewhat,
    transistors can vary pretty vastly, well beyond what common sense
    suggests. And even though transistors do not really have a "burn-in",
    when they begin to fail they can drift for a brief period before final
    failure sets in.

    Peter Wieck
    Wyncote, PA
  16. Mr. Land

    Mr. Land Guest

    There used to be a spray on the market - can't remember who made it at
    the time, might have been GC - it was called "Blue Stuff". I worked
    in a stereo service department thru college, and that spray was the
    best we could find for cleaning stubbornly noisy pots and switches.
    When you sprayed it on it formed a mildly abrasive blue paste, which
    would actually scrub the tarnish and dirt off of switch contacts. I
    never saw it fail, even with the noisiest, most tarnished switches and
    controls. I think it has diatomaceous earth in it.

    Anyway, I think TechSpray carries it - might be worth a try
  17. Maybe not an issue, but the 2603 and the 949 are not a pair. The C2603 is a
    really small signal type, even smaller than a TO-92 package, The A949 is
    similar to a TO-92 but taller. Don't know the package designation. I think
    it's complement is the 2SC2229. They are readily available. I get them from
    Onkyo and B & D Enterprises. BTW if you register on you can get
    datasheets for these really easily.

    I'll try to look over your troubleshooting results with the schem and post

    Mark Z.
  18. Dave

    Dave Guest

  19. Man.

    This might be really easy. Q 405 conrols the turn-on delay. HK's don't use
    relays, so they mute the signal til the amp stabilizes. On this model it is
    Q405 and Q406 respectively. The transistor could be bad, but I would be
    especially concerned with D401, R405, C405, C407, and D403, which is a 15
    volt zener.

    Could just be solder connections relating to the above, but in any case you
    need to see that Q405 turns off hard a few seconds after turn-on. The -12.5
    or so volts at the base is critical.

    Mark Z.
  20. Dave

    Dave Guest


    First off, thanks very much for all of your insight and time.

    Okay, as I had just "upgraded" C405 and C407, I put the originals back.
    Left channel is now functional once again, although I haven't had time to
    check my voltages yet nor have I given it a good listening test. I can't
    say why, but I've got a gut feeling my voltages are still on the high side.
    How common is it for zener voltages to wander over three decades?

    I set about setting the idle currents and, as soon as I touched V404 (right
    channel), poof, right channel gone. I think the 1K pot just disintegrated
    when I moved it after 30 years of sitting in one position. Well, I must be
    doing something right, because I stumbled across a 1K 70's style (BIG) pot
    in one of my many doorstops which I am cannibalizing to make myself an ESR
    meter. I'll put it in tonight, hopefully it's in better condition than the
    one I'm taking out. One wonders at moments like this whether one is moving
    forward or backward...

    I can't believe there are so few companies that make ESR meters... the more
    I get into electronics repair, the more critical such a piece of equipment
    becomes. Ditto with a transistor tester although there seem to be tons of
    those on the market.

    I have a hard time with transistor theory... I understand what they do, but
    get quite confused when I see a whole bunch of them tied together,
    controlling each other with various feedback schemes. That's when I start
    testing components as I lack the deductive skills required to narrow things
    down, as you did, to a single transistor and associated passive components.
    Hopefully I'll get better at it... I've read some texts but don't find the
    ones I've read particularly helpful. They all seem to assume that one is in
    a laboratory environemnt at school with access to and training on the SPICE
    program, in particular. Which I do not have.

    An oscilloscope and signal generator are handy, but again is somewhat hit or
    miss (for me) when I get into a transistor network as it's not immediately
    obvious where the signal path is. It's almost easier in a newer amp with
    IC's... what you end up doing is replacing an IC with ten transistors in it
    because you know the problems's gotta' be in there someplace if you've got
    signal going in and none coming out. If the IC is not obsolete and/or
    proprietary and hence unavailable. In the past I've used the scope to
    localize distortion which saves A LOT of time and lifting of legs to test

    I am quickly learning that with vintage audio gear, if it ain't broke, don't
    even breathe on it or else...

    Thanks again for your help.

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day