Class A reliability

[greg788]

Due to the heat output, are class A amplifiers inherently less reliable in the long run than Class A/B or B? I'm assuming the amplifier is kept turned off between listening sessions and proper heatsinking and design, of course.

If they are less reliable, what are the typical components to fail -- caps, resistors, etc. ?

[teemuk]

I have no first hand experience with class-A amplifiers but theoretically you are correct: Excessive heat would reduce component life. If heatsink is too hot it is definitely a sign that the output devices are more stressed. Elevated temperature will also decrease the age of capacitors ...and so does ripple which is generally higher in class-AB amplifiers.

However, in my opinion the question how much an amplifier is allowed to heat up is just a question of design not topology. Aren't those Class-A amplifiers usually pretty well heatsinked anyway? Also, the output devices are just one example of heat sources: For example, I've seen some class-AB amplifiers that had burning hot series resistors in the regulator circuit - they were placed right next to main filter caps.

Anyway, most of the reliability issues that I have experienced are not related to heating but mechanical problems - i.e. worn or dirty potentiometers and switches - and in some rarer occasions poor solder joints, which are usually broken by the tension directed to PC board-mounted components like jacks and switches.

[jnb]

I agree with teemuk, It's all about how something is designed. There is no reason a class A amp can't run cool.

Since class A runs at a closer to steady current/temp than class AB, it might even be argued that it can be more reliable

[Christer]

One thing that might perhaps matter is that for class A, the output devices has a constant average power dissipation, while for class AB/B the average power dissipation varies with the output power. That might cause more thermal stress in the latter case. However, average power dissipation might not be sufficient to look at if the junction temperature varies with instantaneous power dissipation to considerable degree, which depends on the thermal impedance of the package and the heatsinking.

[sandyK]

I have been running a modded version of the Silicon Chip 15W Class A since 1999, with no problems. The heatsinks do get quite warm , so it is a good idea to provide some ventilation.The metalwork I am using has no ventilation in the lid, so I have placed a couple of washers under each of the lid mounting screws
so that the internal temperature doesn't get too high. The big 300mm heatsinks either side keep the O/P devices safe.
SandyK

[john_ellis]

Hi all

there are two guidelines to remember for transistor reliabiity.

1. Don't let the transistors get too hot
2. Don't let the thermal cycles become too many

When RCA were in business they published thermal cycle rating charts showing that TO-3's were better than nearly any other transistor and could cycle 10,000 to 100,000 times without the chip cracking or falling off the header.

I've never seen anyone else publish any similar reliability data. I suspect that plastic packages are worse than TO-3 still, but it would be good to see some data.

MOdern transistors tend to use the thinnest silicon that is possible. This reduces the thermal resistance between the junction and the case, but also reduces the temperature stress across the package. This may help with thermal cycles in modern plastic packages.

Mean-time-to-failure figures are also often absent from most data sheets. While it is true that a constant dissipation will create less stress than a cycled transistor, it is still important to keep junction temperatures cool.

Class A circuits with high dissipation may well wear out transistors sooner than AB amps with modest dissipation.

But on the other hand if you switch your hifi on once per day, 10,000 cycles is 30 years... so certainly old RCA TO-3's should still be going strong?

cheers
John

[jacco vermeulen]

Originally posted by john_ellis
When RCA were in business they published thermal cycle rating charts showing that TO-3's were better than nearly any other transistor and could cycle 10,000 to 100,000 times without the chip cracking or falling off the header.

The RCA charts i saw did not include all other packages available at the time.

A low bias stand-by switch for Class A amps not only saves on the bill, dramatically reduces the time to reach thermal stasis, it also increases the thermal cycle rating.
Might even beat a class AB that goes fully on and off for the durability of the semis.

[Jeb-D.]

It's no less reliable if done right. One good thing about class A is that it becomes fairly efficient at full volume

[sandyK]

Even a 15W/Ch. Class A gets damn hot on a hot summer's
day , though. Near 100W heat to get rid of in a cabinet.

SandyK