miker@polari.UUCP (Mike Ranta) (05/05/91)
In article <18030315@hpfcdj.HP.COM> myers@hpfcdj.HP.COM (Bob Myers) writes: >>Can anyone tell me if the power transistors used in ampifiers degrade >>over time? What I'm getting at is: will it be necessary, after x number >>of years, to replace the transistors to get that "like new sound". I >>know that with tube amps, you must eventually retube. > >I would expect other components in the amplifier to degrade far more quickly >than the power transistors; compared to some parts, the power transistors >might as well be immortal (assuming they don't fail catastrophically). This is an interesting question. Supposedly, bipolar transistors are prone to a phenomenon known as "secondary breakdown". This occures when they're driven near the limits of their "safe operating area" as the output devices in audio power amps often are if you like it loud. Secondary breakdown is where the current flow becomes localized within the silicon "die" instead of being uniform. Bipolar devices conduct "harder" as they get hotter and hence you end up with local "hot spots" on the die. Supposedly each time this happens it alters the gain and other key properties of the device. This is one of the big "claims to fame" of MOSFETS--they're immune to secondary breakdown and hence will not degrade. If the amp manufacture did their homework, the current limiting circuitry should keep the output devices within their "safe area" and this shouldn't be much of a problem. For amps that are rarely driven hard and driving relatively "easy" loads, it shouldn't be a problem regardless of the design. Also, only the output devices would be prone to this problem. Next time I'm working on an older amplifier, I'll test some of the output transistors and compare them to a new example of the same part. This is only relative, however, as you really need to test the SAME part when it's new and after several years of hard use... Comments?
frankw@tc.fluke.COM (Frank Wang) (05/17/91)
In article <3961@polari.UUCP> miker@polari.UUCP (Mike Ranta) writes: >In article <18030315@hpfcdj.HP.COM> myers@hpfcdj.HP.COM (Bob Myers) writes: >>>Can anyone tell me if the power transistors used in ampifiers degrade >>>over time? What I'm getting at is: will it be necessary, after x number >>>of years, to replace the transistors to get that "like new sound". I >>>know that with tube amps, you must eventually retube. >> >This is an interesting question. Supposedly, bipolar transistors are prone >to a phenomenon known as "secondary breakdown". This occures when they're >driven near the limits of their "safe operating area" as the output devices >in audio power amps often are if you like it loud. Secondary breakdown is >where the current flow becomes localized within the silicon "die" instead >of being uniform. Bipolar devices conduct "harder" as they get hotter and >hence you end up with local "hot spots" on the die. Supposedly each time >this happens it alters the gain and other key properties of the device. >This is one of the big "claims to fame" of MOSFETS--they're immune to >secondary breakdown and hence will not degrade. There is a so called "metalization" in all bipolars, MOSFETs, and ICs that has recently drawn much interests for QA guys in Intel and other big volume semiconductor manufacturers. The term "metalization" came from the fact that metal diffuse into contact regions and/or SiO2 regions under high temperature or long time duration. This is not a major problem in old days. But, it is becoming a problem now because the line width can go as thing as 0.35 um for those high density DRAMs. Frank Wang @ John Fluke Mfg. Co., Inc. frankw@tc.fluke.com frankw@milton.u.washington.edu -- Frank Wang @ John Fluke Mfg. Co., Inc. frankw@tc.fluke.com frankw@milton.u.washington.edu
dickw@IASTATE.EDU (Wallingford Richard) (05/18/91)
In article <1991May16.194955.487@tc.fluke.COM>, frankw@tc.fluke.COM (Frank Wang) writes: > There is a so called "metalization" in all bipolars, MOSFETs, and ICs that has > recently drawn much interests for QA guys in Intel and other big volume > semiconductor manufacturers. The term "metalization" came from the fact that > metal diffuse into contact regions and/or SiO2 regions under high temperature > or long time duration. This is not a major problem in old days. But, it is > becoming a problem now because the line width can go as thing as 0.35 um for > those high density DRAMs. > > frankw@milton.u.washington.edu On a related topic: Will all transistors eventually become useless chunks of silicon (or GaAs, or whatever) because of the extremely slow, but nonzero diffusion of n into p and p into n at room temperature? Dick -- dickw@iastate.edu
FC138001@ysub.ysu.edu (Phil Munro) (05/19/91)
>On a related topic: Will all transistors eventually become useless >chunks of silicon (or GaAs, or whatever) because of the extremely >slow, but nonzero diffusion of n into p and p into n at room >temperature? > Will all glass windows eventually become useless blobs of glass because of the liquid nature of glass?!!! I really don't have a quantitative answer to the question about Silicon, but the diffusion of dopant atoms in Si is something which is known, and as long as the temperature of the device is kept within specs, the mean time to failure would be known. In addition to these ideas about glass and design specs, we might want to note that EVERYTHING will melt down and be destroyed in the firey judgement to come. --Phil