Connect with us

blown output transistors

Discussion in 'Electronic Repair' started by Dave, Nov 29, 2006.

Scroll to continue with content
  1. Dave

    Dave Guest

    Last week I posted about an Emerson Research "home theatre" system that I
    was given. Two 5A fuses on the CT transformer secondaries blown. This unit
    has outputs for two rear speakers (L/R), a center channel (DIN connector)
    and two front satellites (L/R). Now the inputs are analog RCA audio so this
    box must be just simulating 5.1 using a delay or some other trickery.

    What I'm looking for here is an understanding of how the various output
    components interact... in particular how the NPN/PNP pairs function

    Anyway, there are six output transistors, 3 x TIP41C (NPN) and 3 x TIP42C
    (PNP). It _APPEARS_ by looking at the board traces that one set goes to
    the rear output RCA connectors (can't really see how it's connected as the
    connector block takes up a lot of real estate on the board), one set goes to
    the satellite RCA outputs, and one set goes to the subwoofer single output
    which is enclosed in the cabinet with the PCB board. The 41/42 pair
    adjacent to the subwoofer output are the only two (of the six) output
    transistors which are NOT blown.

    Now, another poster suggested that I check the bias and driver transistors,
    but I'm not sure which ones these are. here's what I've got, I wish I were
    skilled in ascii art but I'm not so bear with me.

    The output transistors are all in a row along a large aluminum heatsink
    which runs the width of the PCB. They alternate --- In between each
    PNP/NPN transistor is a smaller TO-92 NPN which is adjacent to the heatsink
    as well, sort of stuck there with thermal grease. Adjacent to each TO-220
    Tip transistor but on the side opposite the heatsink, each output transistor
    is paired with another TO-92 Like so:

    blown pair blown pair good pair
    || || || || || || || || || || || || || || || || || || ||
    | | | |
    | |
    c945 a733 c945 a733 2sc2245 2sa965

    I pulled ALL of these transistors and tested them. Of the TO-92's, all test
    good except one of the a733's. My question is this: do I need to replace
    BOTH a733's, i.e. to maintain matched set characteristics? I'd be looking
    at an NTE290A as a replacement. Or are the c945/a733 some sort of matched
    pair. I see that NTE offers either a) two matched 290A's or b) matched sets
    of 289A/290A (289A is NOT listed as a replacement for 2sc945).

    As I noted at the beginning, I'd like to understand which transistors drive
    what if it's at all possible to determine from the information I've
    provided. I can trace out the relationship between the transistors if
    necessary but maybe the number, placement and type of units is enough.

    Any replies greatly appreciated.

  2. Guest

    I have little idea what you know and what you dont... npn/pnp pair is
    your standard output pair, normally in more or less the same
    configuration, but for any more details you'd need to either trace the
    circuit yourself or get a cct diagram.

    Little chance of needing matched pairs for anything. Transistors
    distort so much they get lots of nfb applied to them, so it doesnt
    matter whether theyre matched.

    not really.

  3. dave

    dave Guest

    Well, typically these output transistors are usually arranged in a push pull
    type circuit. One transistor amplifies the positive side of the waveform
    and the other does the negative side.

    But I have to ask why. Emerson is junk. Not worth spending time and money
    on. Buy a new one.

  4. Bill S.

    Bill S. Guest

    With only 3 amp channels, it probably just has a low pass filter to
    drive the subwoofer.
    My guess, TIP41/42 are outputs, 2SC945/2SA733 are drivers connected
    as a darlington configuration with the outputs, 2SC945 is bias
    regulator connected between the bases of the drivers. 2SC2245/
    2SA965 perform the same driver function but the subwoofer amp
    runs at different supply voltages. No need for matching.

    What you have drawn is just the current amplifying stages, there
    are more transistors or IC's upstream that may or may not be
    damaged. And there's a good chance you have some open emitter
    resistors on the outputs.
  5. Dave

    Dave Guest

    Well I'd like to hook it up to my PC at work. I've got a pair of Polk Audio
    satellites but they've got no bass, I mean NONE. I don't expect high
    fidelity out of it, but it was entirely free and, given the fact that I'll
    be listening at low volume, it's entirely adequate for my humble needs. So
    far my investment is two 5A fuses, four power transistors and one driver
    transistor, maybe $10 all together. Given that this "system" likely cost,
    oh, maybe $80, if you'd like to chip in the extra $70 I'd be real happy to
    "buy a new one".

    That's why.

  6. Dave

    Dave Guest

    How can you output 2 x rear, 2 x front, 1 x ctr, 1 x subwoofer with ony 3
    channels? I don't see how it could be done unless the two input channels
    are amplified and sent to both front and rear outputs, PLUS are combined and
    sent to center channel, leaving one amp channel left over to drive the
    subwoofer on its' own...
    There are only 1 or 2 other transistors on the board but several IC's, a
    couple of 14- or 16-pin and one bad-lookin' dog on the underside which looks
    to be SMT, many pins. If the outputs were shorted, would this not smoke the
    outputs, then the drivers, and so on? I tested the resistors to the bases
    of the output transistors and all were ok, will look at the
    emitter/collectors next.

    Thanks a lot for your help, your reply as to the operation is exactly what I
    was hoping for.

  7. Dave

    Dave Guest

    The closest thing I can find locally as a replacement is a 2SA817.

    The original 2SA733 has I(c) of 150mA, power dissipation 0.25W and hfe of
    between 90 and 600 (200 typical).

    The 2SA817 has an I(c) of 500mA, power dissipation of 650mW and hfe of 70 to

    Is this close enough?
  8. Dave

    Dave Guest

    Sorry, the 2SA817 has an I(c) of 300mA, power dissipation of 600mW
  9. Dave

    Dave Guest

    Okay, I rec'd my TIP41/42 replacement pairs yesterday. Before I put them in
    I figured I'd power up the unit with no output devices installed. Checked
    the bias voltage on the base of each output transistor position.

    On the two pairs that are driven with C945/A733 pairs, there is 300-400mV DC
    at the base. At the 2SC2245/2SA965 driven pair the I read a whopping
    15.5VDC at the NPN and -15.5VDC at the PNP output base socket. This can't
    be good. As the pair uses a common bias transistor, I replaced the bias
    transistor, no change.

    I can see what you all mean about troubleshooting a DC coupled amp. I've
    made a schematic from the outputs back, but I don't have anything to compare
    my values to because it's a completely different amp channel than the two
    other channels which are identical.

    Going backwards from outputs, I see outputs -> drivers -> dual op-amp IC.
    Behind the op-amps are more op-amps (quad op-amp IC TL074CN), another
    half-dozen NPN's (all c945's), two diode half-wave bridges, all kinds of
    good stuff. Is there any hope at all, given that I don't have an identical
    working channel to reference, of finding the source of this dc voltage? Or
    should I just bring it to the dump?

    I don't mind troubleshooting this piece of plastic junk as a learning
    exercise if there is a potential positive outcome (i.e. light at the end of
    the tunnel). But, maybe I'm just asking for punishment...

    Any and all responses appreciated

  10. Guest

    If its all dc coupled youre most definitely asking for punishment. I'd
    dump it and get some other chuck out to work one, one youve got a
    decent chance of fixing.

  11. Unfortunately, testing with no output transistors in place will give
    erroneous readings. I even had an HK 570i receiver flame out on me when I
    tried this back in the '80s.
    The problem is that the base-emitter junction of each output transistor is
    effectively a diode to the diving circuit, clamping at around .6 volt,
    unless the outputs are Darlintons, of course...

    Mark Z.
  12. Guest

    then youve still got as set of output trannies yu can use in a project.
    Not often I'd say forget it, but with dc amps I wouldnt bother.

  13. Dave

    Dave Guest

    Yeah, that's about where I'm at with it. Given it was a cheap unit to start
    with, it's not worth it... I just have a hard time letting go once I've sunk
    my teeth into a project. However the board does have LOTS of useful
    regulators and diodes on it, plus I recover a nice 5A CT transformer.

    Thanks for the pragmatic advice.

  14. Guest

    Well, the good news is the sooner you get onto another piece of kit the
    sooner you're going to have a result.
    Plus you can use your TIPs to make a nice big power amp.
    Coincidentally, you've even got the psu, case & preamp all already

    I reckon this one has horrors instead of darlings for the output

  15. Dave

    Dave Guest

    The pre-amp/amp are *heavily* integrated, same PSU. I'd have to look pretty
    hard to find a point to break into the circuit at pre-out. I think I'll
    just build myself a 20-150Hz variable low-pass filter and 50W push/pull amp
    for the subwoofer with the TIPs. I can cut the board off (the PSU is well
    isolated off by itself) and use the relatively large empty space for my new
    amp. Conveniently, the 4-channel audio processor chip which provides input
    switching btween three sets of input jacks (even a high-gain set!) are all
    on the PSU end of the board.

    I'm disappointed I can't get it working fully as it *potentially* does
    exactly what I want.

    Oh well, success is knowing when to cut your losses.

  16. Guest

    IIUC all you've lost is your power amps, the rest works. I'm not clear
    what you mean by pre and main being integrated, I've not seen anything
    like that since the valve days. Same pcb sure, same psu of course, but
    circuit intertwined? I cant think how it would be.

    Simple no frills power amp modules are quite easy to make. Use a fast
    opamp and you can do away with output bias circuits altogether. Forget
    about all the nonessential points and you can get some very simple
    power amps done. If you cant get them into the case any way, they can
    always sit outside it on their own, using the main units psu.

    Anyway, It ounds like you'll get something out of it one way or

  17. Dave

    Dave Guest

    I'm not that clear on where the amp begins and where the pre-amp stops in
    this case. It's not like an amp with discrete components where everything
    after the tone/balance pots is amp and everything before is pre. I'd expect
    at least four of the 12 op-amps on the board (2 x dual & 2 x quad) to be a
    buffer/integrator for the subwoofer signal... I guess that's the equivilant
    of a "pre-amp" for the subwoofer. Maybe four more (or less) to combine the
    signal for the center channel output? I really know very little about
    multi-channel (I mean more than two) systems and this one is a real deviant
    because I think it's basically just some sort of delay circuit to simulate
    surround sound from using only two analog R & L inputs.

    If the amp is just the drivers and output transistors, and I've replaced the
    outputs and a good portion of the drivers, and I've still got DC bias
    problems, that would lead me to think that the source of the problem may
    well be in the preamp UNLESS there is a coupling cap between the pre-amp and
    amp stages. Hmm, if I can figure out where to make the split, maybe I can
    add such a cap which would at least tell me on which side of the fence (pre
    or main amp) the problem lies... ah, but I was going to dismantle it and use
    the components... silly me.
  18. Guest

    right, it sounds like the problem is one of circuit tracing then. When
    you use a capacitor coupled power amp(s), the dc conditions on its
    input dont matter. So you can follow the signal back from the (now
    blown) output tr pair using a probe thats a single opamp amplifier and
    a 2" speaker until you get a nice clean souding signal, and take it
    from there. Or, if it uses a pot for volume, you could tap 2 channel
    putput straight off the volume pot.

    Preamps routinely have dc bias on them, its not an issue.

    If you probe around with something very low power, you wont scare your
    ears or damage anything, and you should find where its all good. The
    signal quality may not be good within some parts of the old dead power
    amp, because the dead output trs will have made the feedback loop open
    there. Just keep going back till its good and clean.

  19. Dave

    Dave Guest

    Well, turns out it IS as simple as discrete components: the pre-amp ends
    just past the tone/balance *IC*, and the power amp starts there. I made a
    probe out of the inputs to a disposable powered computer speaker.

    Thanks for your insight.

  20. Practically all the amps I work on are dc coupled. Not a problem in most
    Virtually all your Kenwood, Pioneer,Sansui, etc going back to 1970 or so are
    DC coupled.
    Modern amps like Denon, Yamaha, Onkyo, etc are all pretty similar in
    topology. Certain components tend to go out along with the outputs
    (resistors and driver transistors for example) and symptoms such as
    premature clipping, DC offset etc have predictable causes.
    As to the OP, it does seem his piece is quite possibly a "piece" all right,
    but I'm not at all sure I'd classify it as such a difficult repair simply
    because it may be a "DC" amp. As I say, I fix such amps all the time. No

    Mark Z.
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