Transistor Life Span.

Discussion in 'General Electronics Chat' started by lubnaan90, Mar 22, 2010.

  1. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    Hello to all,
    I have an old HI-FI power amplifier, which was manufactured in year 1985.
    I would like to replace the power transistors
    it uses for the output, since i believe that the transistors have served for long duration & its time for them to retire.

    All i would like to know , will replacing the transistors with the new ones give better output or any change in the output ?
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You may have problems locating transistors that are a good match for the originals.

    If you really want to protect your amplifier, you will test/replace all of the electrolytic capacitors. If they short out, they can do damage.

    You should also inspect all of the resistors. Those that are the old-style brown-body cylindrical resistors may have changed in resistance value considerably since they were made. This can cause the bias on your transistors to be incorrect, and lead to failure due to overheating.

    More modern resistors have light tan dumbbell-shaped bodies; they are carbon film.
    Modern metal film resistors have light blue dumbbell-shaped bodies.

    The best place to start is to find a schematic for your power amplifier, and document as best you can what the voltages are - while it is still working.
     
    PackratKing likes this.
  3. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    Hello,
    Thank you for your reply.

    The amplifier uses two transistors per channel , KT818B & KT819B (russian letters) , both of them are soviet made transistors , to date they are still in production.
    I can still buy them , so there is'nt a problem locating the same transistors.

    As for the capacitors , I already replaced all the soviet made electrolytic capacitors in the amp with the modern ones, Then too , thank you for your advise.

    I inspected the resistors , they all seem to be working fine .

    Overall, the current observation on the amplifier is, that its working fine. All i need to know is, if i change the power transistors , with there be a change in the output quality ?
     
    Last edited: Mar 22, 2010
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    I would not change them unless it was absolutely necessary.

    Transistors are sort of like people; they are all similar, but yet all different.

    You could take 100 transistors that have the same part number, and they would have slightly different characteristics. A well-designed amplifier would take these variations into consideration, but I really do not know anything about the design or quality of Russian entertainment systems from the 1980s - or even what they look like.

    Your amplifier may need adjustment after so many years. As I have already suggested, passive components can change values over time.

    If you change the transistors, you will possibly need to adjust the biasing for the new transistors. Since I don't know if you have a schematic and I DO know that I cannot read Russian Cyrillic, I would not be much help.

    Perhaps the best thing you could do is to monitor the temperature of the transistors. If they begin to heat excessively, it is time to turn it off and determine why they are heating up.
     
  5. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    your Points Noted.

    Here Plz check this , http://forum.allaboutcircuits.com/showthread.php?p=225313#post225313
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Are you still having the problem with the one channel having very low sound?

    If so, I would start comparing voltage readings between the good board and the bad board in order to figure out what component or components are causing the problem.
     
  7. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    I was facing the problem yesterday when i bought the amplifier,
    After replacing all the old electrolytic capacitors to modern ones , the amplifier is working fine with both the channels offering equal output.
     
  8. Darren Holdstock

    Active Member

    Feb 10, 2009
    262
    11
    Sound advice from SgtWookie, as always. I've done a fair bit of work in this area and would agree that the electrolytics would be first on the list for replacement, they really don't last long at all. Some nice low-ESR units for the output transistor reservoir caps would make a noticable difference to the bass response; hopefully lubnaan90 is reaping the benefits here already.

    Transistors do slowly age as dopants slowly migrate hither and thither, but modern silicon devices are pretty good in this respect, even after 25 years. The previous generation of germanium devices grew old less gracefully, which can be an issue when renovating some very early transistorised circuits - modern Ge devices aren't quite the same, and any new-old-stock (NOS) devices will have drifted.

    There is a chance that arbitrary replacement of amplifier output transistors will degrade the sound quality slightly if the originals have been specially gain-matched. This is a technique used only in the most expensive amps as you either have to buy a big batch of transistors and sort them into closely matched groups (the number of devices in a matched group equalling the number of devices in the amp output stage) - a time consuming and wasteful process - or you can find a supplier prepared to offer this service for you, at a hefty premium.

    Metal film resistors are generally always preferable to carbon film in audio circuits, except where they have to deal with power transients (e.g. zobel network, PSU current sense resistors, emitter resistors), in which case use CF or higher power MFs. The Sarge has a very good point about old carbon resistors, particularly the tubular carbon rod types, those things will change value if you put a voltage across them. They still have a place in the anode circuits of valve amps where their intrinsic non-linearity adds a bit of (wanted) distortion, but you don't want them in a hi-fi amp unless you like the sound of frying bacon. Modern carbon film resistors are much better in most respects, and these are what an '85 amp should be fitted with.

    If you really want to make a big improvement to the sound, replace the cap in the feedback network. Tapped off the speaker output you will find 2 resistors and a capacitor, all in series, and the junction between the two resistors will lead off to one of the small transistors in the long-tailed pair of the error amplifier at the input stage. Often this cap is an electrolytic (a bipolar one if you're lucky), and within a feedback loop is absolutely the worst case for such a non-ideal component. Replace this with a nice polypropylene cap and be amazed. But you'll appreciate why electrolytics are used in commercial designs when you see the price and physical size of polyprops.

    If you like that tweak then replacing any 'lytic coupling caps in the signal path with nicer dielectric types will also make a noticable difference, but not quite to the extent of upgrading the cap in the feedback loop. Polyprops are overkill here, anything nicer than an electrolytic/tantalum/ceramic etc. will be an improvement.
     
    PackratKing and SgtWookie like this.
  9. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    Thank you for your advise & recommendations,
    Will go ahead as per your instructions & will update.
     
  10. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Thanks for clarifying the point, Darren.

    Please be aware that although our original poster, lubnaan90, seems to have a remarkable comprehension of the English language - abbreviations make understanding difficult for others who are not as well versed.

    I can tell you that old carbon resistors will change value even if they were originally 1% military specification resistors, unused and sealed in an aluminum-foil-lined airtight package since the late 1960's. Some of them I measured right out of the packet as being as low as 70% of their marked value, some as high as 140%, with one more than 200% over specification.

    I have quite a few other brown-body tubular/cylindrical carbon film resistors in my collection from the early 1980's. These were 5% resistors with a 25ppm (parts per million) temperature specification. Many of them measure very far away from their original specifications.

    The bottom line is:
    Brown body resistors make great forms for winding small values of inductance on. However, they make for very poor resistors.
     
  11. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    The amplifier was working fine with the capacitors updated.
    I did some further updating in the amplifier that includes, applying new thermal paste between power amplifier transistors & the heat sink (since they were dry), updated some resistors aswell.
    Now i wish that I shouldn't have done the above mentioned, Since i was kind of careless while removing the power transistors from the giant heat sink.
    While removing the transistors , i noticed a thin sheet of plastic between the transistor & the heat sink (which i thought was totally unnecessary & lost them).
    Well, I installed the new transistors (without the plastic sheet) with new thermal paste, Finished installing, Switched the amp on & boom , the transistors blow up & the transformer begins to vibrate (with vibrating sound), obviously i instantly disconnected the power supply.
    Note: the heat sink was connected to the chassis when the above incident happened, i tried the same test with the heat sink away from the chassis (with new transistors installed) the amp worked fine .

    Now i think that the thin sheet of plastic played important role, since i don't have that plastic sheet which was originally placed there , is it ok to use ordinary plastic sheet between the heat sink & the transistor or is there a specialized plastic for such purpose ?
     
  12. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    That is unfortunate.

    You will need to obtain replacement insulators for the transistors. It is not a typical plastic; it is engineered to be both a good electrical insulator, and a good conductor of heat. They are usually made from mica.

    Note that if the transistors were held against the heat sink with metal screws, there would also have been insulation washers between the screw heads and the transistor mounting tab(s). These washers usually have a step in them to keep the washer centered over the hole. You need to use that same type washer.

    Since you shorted the supply out, you may also have damaged the rectifier bridge in the supply. Check things over carefully before you attempt to apply power again.
     
  13. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    Hello,

    Thank you for your reply.

    Luckily the power supply is Fine , since it uses 5A protection fuses which blew up during the Short out process (the transformer vibrated for about 2 seconds) .
    I have replaced The Fuses & power transistors , The amp seems to be working fine. Presently the amp circuit board (along with the heatsink) is out side the chassis so the Heat sink doesn't contact the chassis.

    Regarding the insulation washers between the screw heads and the transistor mounting tab, you are absolutely correct , The washer is present on all screws for all the four power transistors.

    I have ordered some Insulators for TO220 Package (which i will collect tomorrow).
     
  14. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    Hello to all ,

    Completed renewing the amplifier , Changed the power transistors , changed some Capacitors & some resistors , Also installed some New A/V type connectors replacing the old Soviet Connectors which were so irritating & impossible to use.

    A new problem arose though, The amplifier Board has two detachable boards in it (plz check the attached pic) Each one for individual channel.
    The problem is with the left channel part, a 100 Oms Resistor gets extremely heated while the music is being played , its takes around 3 seconds for it to heat up & create irritating distortion (bass) in the left channel , I tried Switching the boards , then too the same problem only in the left channel.
    Note : Detachable board from the left channel when placed in the right channel slot , it works absolutely fine. Also to be noted that the resistor only heats when there is Load on left channel (when the Loud speaker is connected)
     
    • sss.jpg
      sss.jpg
      File size:
      318.1 KB
      Views:
      32
    Last edited: Mar 25, 2010
  15. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    Thank you for your quote.
    I did buy some silicon insulation pads for my TO-220 Package power transistors. They are well connected & working , just a little problem in the left channel (please read quote no.15).
     
  16. Bychon

    Member

    Mar 12, 2010
    469
    41
    This is all very good advice. To summarize, electrolytic caps die as the years go by, no matter what circuit they are used in. They are always the first thing to attack in an old amp. Second is the old carbon resistors. They tend to change value over the years, but they are very easy to measure and see if they are off-spec. Go looking for bad resistors second. Third, poly-anything caps have very low ability to absorb energy from the sound signal. That's why they are famous for use in antique tube type guitar amps. They actually make a difference in how it sounds.

    Fourth is the transistors. In a low power circuit, there is no surprise in finding 30 year old transistors with absolutely nothing wrong. In high power circuits, the atoms in the transistors actually migrate over time, leading to lower gain, bias changes, and eventual death. Sometimes the death is is an open circuit, sometimes death is caused by voltage punch through. It's all about heat. The slow changes in transistors are more difficult to measure than passive components like resistors. It is O.K. to replace all the transistors that get hot IF you know how to make sure the new ones can be biased correctly and IF you know if they are matched pairs. The difficulty in measuring, the requirement to understand how to adapt the new transistors to the circuit, and the fact that transistors are less likely than passives to go bad are the reasons why replacing transistors that aren't bad is about the fourth thing to consider.
     
    lubnaan90 likes this.
  17. lubnaan90

    Thread Starter Active Member

    Jan 13, 2010
    196
    2
    Thank you for your advises, Will definitely note your points.
     
  18. Darren Holdstock

    Active Member

    Feb 10, 2009
    262
    11
    Ouch, sorry to hear about the unfortunate blowup lubnaan90. An ideal opportunity to get to know your amp better though ;).

    Re: Post 15 - am I right in thinking you have swapped over the left and right circuit boards and the problem remains in the left channel only? That would narrow it down a bit. It could be that the new output transistors are now running with too much bias, see SgtWookie's notes earlier. The usual biasing arrangement is to have a small transistor acting as an amplified diode, connected between the bases of the top and bottom output transistors. A variable resistor sets the voltage across the amplified diode transistor c-e junction, and the transistor itself is connected to the heatsink so the bias (approximately) tracks the temperature of the output transistors.

    I don't see any transistors other than the output pairs connected to the heatsink, so I'm assuming the bias is either set by a diode/resistor/capacitor circuit, or it's an amplified diode without the thermal coupling. Either way, a bias circuit without good thermal coupling would be prone to "thermal runaway" so would need to be set a little lower than otherwise.

    Normally the output bias would be conveniently measured across the emitter resistors, but I can't see any in this circuit, so you'll have to find a clever way of measuring the currents through the c-e path of the output transistors. You might have to break a wire or cut a track, but don't insert emitter resistors unless you plan to keep them permanently in place as their presence will affect the bias.

    I think that one variable resistor (on each detachable board) sets the bias, but don't rely on my guesses, it might be time to reverse-engineer the output stage and work out the circuit for this section. There's a good chance that the overheating 100 ohm resistor feeds the bias circuit, and the bases of the output transistors too. You don't have to work out the entire amplifier circuit, just the output transistors and everything directly connected to them - post that it will be much easier for people to diagnose problems.

    My apologies, lubnaan90, for any jargon and abbreviations in previous posts, the Sarge is quite right, these don't always translate well.
     
Loading...