DIY Project Amp Burn-In

[DreadPirate]

What are the various opinions (there always seem to be various conflicting opinions around here) regarding the early burn-in, if at all, of a new build or significant re-build (outputs) of an amp? Should I leave on for a day or two? Cycle several times? How long before I pull the test speakers off and connect some "real" ones?

[h_a]

In commercial products burning in is done to check for early failures (see for example, Bryston).

At which point you disconnect the speakers is up to you; personally I use a DC-protection, so I don't have to worry too much about this. Certainly a little sinewave testing helps otherwise also.

Have fun, Hannes

[RocketScientist]

If you're worried about an early failure, I'd thermally cycle the output transistors several times (work the amp hard and then shut it off until it's cold and repeat). And I second DC protection being a good idea regardless.

If you're talking about the other kind of "burn in" favored by audiophiles for improving the sound quality, I wouldn't worry about it too much. While electrolytic capacitors have been shown to slightly improve their characteristics after some use, there's no real evidence transistors improve with more hours of use.

[DreadPirate]

I'm mostly concerned about the possibility of early failure due to component failure or quality of assembly. For a receiver (ie Marantz 2285B) with some type of internal speaker relay, should I still have something between the outputs and the speakers? Will this really catch output failure in time to prevent speaker damage? Why don't high end speakers already have something like this, seems a pair of $3000 speakers would come with some type of self-protection.

[nigelwright7557]

I have seen failure curves for electronic equipment and usually they fail mostly when very new or much later in their life.

[h_a]

Will this really catch output failure in time to prevent speaker damage?

Depends on the protection

Of course it will if the protection deserves its name.

Why don't high end speakers already have something like this, seems a pair of $3000 speakers would come with some type of self-protection.

Pro gear has usually a large elco in series to the amp; belt and suspenders way of protection in the pro-world (means both the amp and the speaker has built-in DC-protection).

I guess a large elco does not sell in the hi-fi world.

Have fun, Hannes

[RocketScientist]

Unfortunately, most speaker protection circuits either need power to operate (something speakers don't usually have) or they degrade the sound. Even fuses degrade the sound and don't offer much protection unless they're sized so small they tend to blow during regular use.

So the best place to protect speakers is in the amp/receiver. And most commercial products, with the notable exception of lots of high-end products and a few others, have at least DC protection with a speaker relay. If your Marantz has a delayed turn on relay for the speakers it very likely already has DC protection. So no worries.

I read a story about an engineer (working for a manufacture) trying to determine just how fast the protection needed to be to save a speaker. So they simulated a failure of a 100 watt amplifier and the woofer literally caught fire in under a second with 40+ volts of DC applied. Their conclusion was "the faster the better". Which is to say fuses are often too slow when sized to meet other demands.

[nigelwright7557]

Originally posted by RocketScientist
I read a story about an engineer (working for a manufacture) trying to determine just how fast the protection needed to be to save a speaker. So they simulated a failure of a 100 watt amplifier and the woofer literally caught fire in under a second with 40+ volts of DC applied. Their conclusion was "the faster the better". Which is to say fuses are often too slow when sized to meet other demands. [/B]

I use a PIC microcontroller in my amps to disconnect the relay if it sees DC for more than 500mS.
The micro also holds off the relay for 3 seconds on power up.

[h_a]

Douglas Self recommends much less than a second and his protections (at least the ones in his book) act within about 100ms.

Have fun, Hannes

[RocketScientist]

Yeah, based on what I've seen 500mS is risking damage and even flames Microcontrollers let you do some fun things, but they're not as failsafe as a simple analog circuit and they also create lots of RFI hash that can find it's way into the amplifier where it either directly causes noise or gets demodulated and causes problems.