Announcement

Collapse
No announcement yet.

SS amp with Burning Resistors at Power-Up

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • SS amp with Burning Resistors at Power-Up

    I've finally gotten over some inertia and I'm looking at a bunch of stuff that's been on the back burner.

    Today's amp is an old Kenwood KR-3200 receiver. Its a low powered unit that is old as dirt. It has a set of four 2SC789 resistors inside and a 25 V rail. Its probably good for 15 watts or so. I have no service data, no schematic.

    I was given the amp by a family member with a complaint that it was "broken." No other information. I noticed a missing 2A mains fuse. I replaced the fuse with a 2A SB fuse (maybe not such a good idea) and powered the amp up.

    On turning on the power, a pair of 0.47R 2W resistors in the output section started smoking. They were toasted by the time that I was able to turn the power off.

    I have no schematic, and worse yet -- no familiarity with working on SS gear. It looks like the 2SC789 are in a T0220 package, though I have no idea what the pinouts are. I haven't been able to find any data on the 2SC789, including possible substitutions. AFICT both of the smoking resistors connected to Pin 3 on the leftmost two of the four power transistors.

    I'm familiar enough with tube gear, but I'm 100% ignorant about transistor amps. I never learned about them. Am I correct in guessing that the 0.47R resistors perform biasing, and that I cooked them because the transistors are shorted and because I used a SB fuse instead of a FB fuse?

    Any insights would be appreciated, and I wouldn't be offended if anyone felt inclined to water down their comments to the level of a SS idiot.

    thanks.
    "Stand back, I'm holding a calculator." - chinrest

    "I happen to have an original 1955 Stratocaster! The neck and body have been replaced with top quality Warmoth parts, I upgraded the hardware and put in custom, hand wound pickups. It's fabulous. There's nothing like that vintage tone or owning an original." - Chuck H

  • #2
    No, the .47 ohm resistors are ballasts. They serve to limit current and in multi-xstr circuits they aid current sharing.

    The fast/slow blow of the fuse was not the real issue. You needed to either bring it up slow on a variac while monitoring the mains draw, or at least use the old light bulb trick. Wire a light bulb and socket with clip wires, and connect them in place of the mains fuse. If the amp has a short and draws a lot, the bulb will light brightly. If the amp is more or less OK - not drawing a ton of current, then the bulb glows dim, and you can then try a fuse. I prefer the variac, since I can tell with the mains at only 10 volts if the current is already ramping up then I power down and hunt.

    If you look at most any tube powr amp, they all look the same. I mean you could draw a tube power amp and phase splitter on a napkin I am sure, even if you didn't have specific values.

    SS amps are like that. They are all more or less equivalent. Oh yes, we often need schematics for part values and nuance, but still very similar.

    If this is a stereo amp, and there aer four power xstrs, then two per channel in push pull. If all four are C789, then the amp is "quasi-complementary." All four are NPN. A true complementary amp would have an NPN on the positive side and a PNP on the negative side.

    2SC789 is a TO220, 70v, 4A, 30W , NPN

    There are plenty others to sub in there. TIP41B or TIP41C would be common US types. I stock the C789, I have 2SD613 which would work. MCM sells C789 but is out of stock. Others might be 2SD525, 2SD526. I like the D613 sub in fact.

    US xstr numbers are 2Nxxxx and you have to look them up to know anything. Also some are like TIPxxx or MPSAxx, and again the number tells you little, you have to look them up.

    Asian xstr numbers all start with 2S, so 2SC789. That 2S is understood, so many parts are labeled simply C789 or whatever.

    2SAxxx are PNP and so are 2SBxxx.
    2SCxxx are NPN and so are 2SDxxx.

    The difference between A and B, or between C and D doesn't mean much at audio. Once into radio freqs then the B and D have higher freq ratings.

    TO220s are B-C-E across the three legs looking at the front - where the writing is. So the resistors connected to the emitter is proper. If teh resistors burnt ,then the transistors on that channel are likely shorted. Measure legs 2 to 3 - C to E - better see powr supply charging, and if shorted, then the xstrs have to go.

    It may be no more complex than a couple xstrs. But connected to the base of each - leg 1 - will be another smaller xstr. make sure it is not shorted in any of the three ways. Those smaller xstrs - drivers - will likely not be the same as each other. One will be NPN and one PNP. They will be easily sourced or subbed.



    If you cannot find the drawings online, Kenwood might have them for sale, and Sams Photofact probably covers it in one of their manuals. It looks like they have it for $22. I am not suer what options they still offer. But check them out:
    http://www.samswebsite.com/photofact/pf_search.asp

    Fuses never can blow fast enough to save a failing part, they are there to prevent things from catchng fire. If those power xstrs were shorted, and i am sure they already were, then those resistors were already stressed and should have been replaced anyway, so don't sweat the slo blow.
    Education is what you're left with after you have forgotten what you have learned.

    Comment


    • #3
      Hmm. I was under the impression that you had to worry about running SS gear at low "brownout" voltages, and that using a variac was a bad idea because low voltage would hurt the semiconductors. If I had known better, I would have used my variac, or if I had my thinking cap on, I could have used a light bulb current limiter -- but to do that I really would have needed to know what kind of current a SS amp is supposed to draw at idle.

      Insofar as the amp is "quasi-complimentary," does that mean that its a cheep amp, that its only MidFi, or is that this is just something that they can get away with in low powered amps?

      I have looked for substitute transistors, and I see that Allied carries the TIP41C at a cost of $0.60 each. I couldn't find the others. I'm accumulating stuff for an order with Allied anyway, so I'll probably just piggy-back 4 transistors onto my next order, and replace all 4 power transistors in the amp.


      Originally posted by Enzo View Post
      TO220s are B-C-E across the three legs looking at the front - where the writing is. So the resistors connected to the emitter is proper. If teh resistors burnt ,then the transistors on that channel are likely shorted. Measure legs 2 to 3 - C to E - better see powr supply charging, and if shorted, then the xstrs have to go.
      Could shorted speaker terminals have caused these transistors to blow up? Or the ballast resistors to smoke? I'm trying to determine why those resistors smoked for me and not for anyone else. I'm not sure, but I might have had an alligator clip from my resistive load slip off from its mount point when I did the power on test.

      It may be no more complex than a couple xstrs. But connected to the base of each - leg 1 - will be another smaller xstr. make sure it is not shorted in any of the three ways. Those smaller xstrs - drivers - will likely not be the same as each other. One will be NPN and one PNP. They will be easily sourced or subbed.
      Are these something that I should just plan on replacing, or should I only replace them if they test bad? I'm thinking it would probably be best to just order the drivers with my other order, just so that I have them on-hand. I'm going to ask something really stupid now -- can I test the little driver transistors in circuit, or do I have to remove them? I don't have anything suitable for testing a transistor that's any more complex than the diode tester on my VOM.

      If you cannot find the drawings online, Kenwood might have them for sale, and Sams Photofact probably covers it in one of their manuals. It looks like they have it for $22. I am not suer what options they still offer. But check them out:
      http://www.samswebsite.com/photofact/pf_search.asp
      I think that this is another case where the amp, even once repaired, isn't worth the price of the Photofact. I might spend a few bucks on inexpensive output transistors just to see if I can get it going, but I'd probably throw the receiver away if I had to buy the service manual. After all, its only a 15W amp and its hardly worth using for anything more than a learning exercise.

      Thanks again.
      "Stand back, I'm holding a calculator." - chinrest

      "I happen to have an original 1955 Stratocaster! The neck and body have been replaced with top quality Warmoth parts, I upgraded the hardware and put in custom, hand wound pickups. It's fabulous. There's nothing like that vintage tone or owning an original." - Chuck H

      Comment


      • #4
        Hmm. I was under the impression that you had to worry about running SS gear at low "brownout" voltages, and that using a variac was a bad idea because low voltage would hurt the semiconductors. If I had known better, I would have used my variac, or if I had my thinking cap on, I could have used a light bulb current limiter --
        Well all we are doing is powering it up. I wouldn't want to try and operate an amp at say 40VAC mains. Actually once there is enough voltage on the amp rails for stabiluty, they aer remarkably tolerant of most any voltage. Dial the mains down to 90 volts and it still works, just doesn't have as much power.

        Some amps come up lopsided, so they might want to draw some at 10-20 volts. It is just experience to decide if excvess draw is just that (add a few volts adn it settles down) or a real problem.

        Bringing an amp up on the variac slowly is the way to do it. A dead short in something will draw heavily at even 5v mains, but that voltage won't hurt the amp. At least not in the time it takes to see what it it doing on the meter.

        but to do that I really would have needed to know what kind of current a SS amp is supposed to draw at idle.
        What, from the mains? Or through the outputs? (thinking like a power tube)

        A SS amp doesn't draw much from the mains sitting there. Losses in the PT iron, maybe a few milliamps from the low voltage regulators dropping. There are no heaters and not much going on. The amps are pretty much class B, so the outputs are idle biased right about at turnoff. Not a lot of draw there.

        So I don't expect a lot of draw - 10-20 watts or something.

        Bias doesn't work the same as in tubers. WE set things by crossover notch, or my secret mains current approach. I'll share that later - I've described it elsewhere around here. SOme makers spec idle current through the ballast resistors, since unlike tubes, the xstrs all are fairly similar in a batch. But a lot don't. Call them up and ask for a current spec and they will tell you "I don't know, just adjust for Xover."

        Insofar as the amp is "quasi-complimentary," does that mean that its a cheep amp, that its only MidFi, or is that this is just something that they can get away with in low powered amps?
        Quasi is a perfectly legit way of doing it. It is not less FI than other methods, no cheaper, and you can make a high power amp that way if you like.

        In a tube amp, there is a phase inverter, so the signal is fed to two tubes both connected the same, but at reverse signal polarity. In a SS amp, the same signal goes to both sides. But one side is wired to positive and the other to negative. As the signal goes more pos, it turns the pos xstrs on harder, so more pos current through the load. Then when the signal goes neg, it turns off the pos xstrs, and starts turning on the neg ones. The circuits are more or less mirror images pos and neg.

        That is complimentary, that mirror image, the two side compliment each other. In quasi complimentary, we don't use both NPN and PNP. We use all NPN (usually, though you can do it either way). On the pos side there is no difference (colector to + rail and emitter to output bus), on the neg, instead of a PNP with collector to - rail and emitter to output bus, the NPN is wired vollector to output bus and emitter to - rail. Basically wire it backwards of the PNP.

        This if course is no longer a mirror image, and the driver xstr is wired different too. But overall it is like a complimentary amp, but those changes make it Quasi-complimentary.

        Why? I don't know. NPNs are cheaper, there is a greater selection. But there are probably other design consideration I don't know about. Peavey makes them both ways, though no quasis lately.


        Allied doesn't sell asian xstrs. Darn few at Mouser. Digikey has mor of them, but not tons. COnsumer product repair suppliers like MCM are the source for this stuff. I always buy my asian xstrs from MCM, since I don't get counterfeits there. A growing problem in power semis.

        Allied dropped the minimum order a while back, but MCM still has a minmum.

        COuld a shorted speaker line blow it up? YOu bet.

        The only current path through those emitter ballasts is through the xstrs. I cannot imagine those xstrs are not shorted. ten seconds with an ohm meter will tell you. Imagine a cathode resistor burning up - pretty much has to be a shorted tube. taking measures while in circuit can offer funny readings, but if it shows shorted, there is nothing else in the circuit parallel to them to show a short, so they are bad if so.

        For 60 cents I would likely just change them anyway, but it is worth a quick check with a meter. Doesn't even have to be the diode test. A short is a short. The drivers are connected directly to the outputs, so ther is always the potential for damage to them. Check for shorts in circuit. Check for junction drops in circuit. But remove any shorted outputs before this, since they are wired together. In the + side, for example, the E to C of the driver is wired right to the B to C of the output. A shorted output can make a good driver appear short. Once the bad output is gone, check the driver. if it shows short, it probably really is. if the junction drops sjow OK< they likely really are.

        Of course you can remove the driver for test - three legs is all it has after all.

        Remember that the small TO92s are different. US types are EBC looking at the flat side. Asian types are ECB. You can swap them, but then the legs have to be bent funny. SO much easier to replace asian with asian.

        I don't use anything more than my meter on diode test when checking xstrs: They are shorted, they have funny low drops, they are open, or they are OK. I suppose now and then one fools you, but I don't worry about it, I can usually tell it is bad in circuit from its performance even if it checks OK.

        I wouldn't buy the photofact either, but it was an option. Someday yuo might want to fix something expensive, and now you know about photofact.

        TV guys actually subscribe to photofact. Every month stacks of new manuals. I guess it is all online these days though. They cover most anything though,.

        In general my advice to shotgun these things is check the outputs and drivers, check the power resistors (your .47), and check any resistors connected to the drivers. Look close visually for burnt small resistors. Also if there is a small xstr between the output pair but also on the heatsink or at least touching it, that is the bais xstr, and should be checked. In circuit it might read low, but not shorted.
        Education is what you're left with after you have forgotten what you have learned.

        Comment


        • #5
          Well, I read about transistors last night to try to get up to speed on them. I've always skipped those chapters in my electronics textbooks. Basically, here are the conclusions that I drew. They may not be right, but they make sense to me.

          1. I think of transistors as triodes, and Base, Emitter, Collector are analogous to Grid, Cathode, Plate
          2. By using an NPN transistor paired with a PNP transistor, you can form a push-pull circuit without the need for a PI.
          3. NPN power transistors need PNP drivers, and vise-versa.

          On the subject of doing tests, I'm thinking that its probably best to troubleshoot the amp after taking out all four power transistors, kind of like debugging a tube amp with the power tubes removed. With the power transistors removed, can I tell if the rest of the circuit's transistors are bad by signal tracing through the transistors on the power board?

          If I have to remove the driver transistors in order to test them, I guess its just as easy not to bother testing them, and to just replace them with new ones while they're removed. It may not be the smart-guy approach, but since I'm ordering parts anyway its probably the most expedient way to go.

          Counterfeit cheap transistors? I hadn't thought about that, but I guess if somebody is going to counterfeit something, they'll counterfeit cheap stuff, as its less likely to be detected -- kind of like counterfeiting a $5 bill. I'm hoping that Allied does a good job of policing their suppliers.

          Regarding testing the power transistors -- am I correct in understanding that continuity between emitter and collector while mounted in the circuit means they're toasted, and any non-zero impedance means they're OK? Same for the drivers?

          I'm still trying to get a handle on how to do the in-circuit testing. I've found what looks like a good reference for testing using a DMM, but it recommends only testing out of circuit. http://www.anatekcorp.com/qdmmvom.htm

          thanks!
          "Stand back, I'm holding a calculator." - chinrest

          "I happen to have an original 1955 Stratocaster! The neck and body have been replaced with top quality Warmoth parts, I upgraded the hardware and put in custom, hand wound pickups. It's fabulous. There's nothing like that vintage tone or owning an original." - Chuck H

          Comment


          • #6
            If the part shows short in circuit, it needs to come out, then test it loose. A parallel shorted path COULD make it seem shorted, but in my experience, if it shows short it is short.

            The meter company wants you to pull the part so their meter won't be involved in any confusion. Sorta like "Only use genuine Fender repair part." When measuring junction drops with a meter, parallel resistors and stuff can skew the reading. But chances are that of you get a reasonable reading in circuit, then the part is OK. If the part is loose, then of course no confusing elements are present.

            In a tube amp, if the third B+ node was zero volts, you might well suspect a shorted cap. If your meter said short, you'd likely believe the cap was short. After all, what else could ground that point? Pull the cap and verify, of course, but the in circuit was probably trustworthy.

            I don't think good results will happen with no outputs. The amp is massively NFB'd, and without the finals, the amp will wander all over in terms of DC levels. The drivers might not be happy with the current through them alone. Bit I don't know your particular circuit.

            Basically I replace parts and back to the variac.

            Since the power rail is at one end of the transistor and the output bus at the other, ther should be no short from E to C. Otherwise the full current of the PS flows to the speaker. BZZZZZZT. SO EC shorts are generally dead parts.

            In circuit there is oftN xstr - red probe on base, then should get half a volt drop to E and to C, with open E-C.

            PNP? REverse polarity. and the imaginary diodes have their cathodes common instead of anodes.

            So in circuit, I expect those three readings to make sense: BE, BC, EC. ANd be aware that there are often other xstrs wired in parallel. Not exactly parallel, but if there seems to be a diode EC on something, look for another xstr wired there. My half volt might be .4 on a power xstr, and as much as .6 or .7 on some signal types.

            Don't rely too much on tube analogy. JFETS really are similar to triodes, but the plain old bipolar xstrs are equivalent to triodes in terms of signal flow maybe, but they work very different.

            3. NPN power transistors need PNP drivers, and vise-versa.
            NO. You wil find that in some circuits, but that is not a rule of thumb.

            In yours, the pos output is NPN, and so will its driver likely be. Your neg out put is also NPN as a quasi, and its drover will likely be PNP. In a true complimentary, both output and driver would be PNP in the neg side.
            Education is what you're left with after you have forgotten what you have learned.

            Comment


            • #7
              Originally posted by Enzo View Post
              In circuit there is oftN xstr - red probe on base, then should get half a volt drop to E and to C, with open E-C.
              Duh? oftN?
              "Stand back, I'm holding a calculator." - chinrest

              "I happen to have an original 1955 Stratocaster! The neck and body have been replaced with top quality Warmoth parts, I upgraded the hardware and put in custom, hand wound pickups. It's fabulous. There's nothing like that vintage tone or owning an original." - Chuck H

              Comment


              • #8
                That's a typo, let me see what it might have sposed to been.

                In circuit there is oftN xstr - red probe on base, then should get half a volt drop to E and to C, with open E-C.
                OK, I got distracted and something was left out.

                In circuit there is often a low value resistor from the driver emitter to the output buss. since the emitter of the driver connects to the base of the output, and the output emitter connects to the output bus (through that .47 ohm resistor), we can see that that low value resistor would be across the output's B-E. SO that low resustance in parallel makes the xstr B-E look lower than it is in resistance or voltage drop, either one. SO when reading an output keep in mind that insread of .4v BE, it might appear to be .2v. When that happens, check the actual resustance and see how it compares to the good channel. That 22 ohms or whatever might be OK.

                To check an NPN xstr, put the red probe on teh base and expect the half a volt drop to E and to C with the black probe. And then an open E-C both directions with the meter.

                That's what I meant to say. I think I managed to hit the delete button with something laying on the keys. I don't know how much was erased.

                OH poop, the whole diode thing is gone.

                Now I recall, when I posted, it gave me an error message but didn't explain it. I get those now and then, now I know what happens.


                There are several paragraphs missing. Not sure I recall exactly what was there, but the gist of it... Whatever is left out will come up eventually I am sure.

                Try again. On diode test, your meter sends out current at enough boltage to turn on a semi junction. In diode test, if you put your red probe on the abide if a diode, and the black on teh cathode, you will get teh characteristic junction drop of the diode - about half a volt.

                A transistor checks the same way, but it is like a dual diode. Take two diodes and twist their anodes together. This will rest the same as an NPN xstr. Red probe to the two anodes and check with black to either cathode. Each way gets about half a volt. Now check cqthode to cathode - should be open. That is exactly how to test a transistor
                Education is what you're left with after you have forgotten what you have learned.

                Comment

                Working...
                X