More Burn-in Misinformation

More Burn-in Misinformation

I was taken by surprise reading a long rant by a particular company owner against solid-state burn-in, in which he displayed considerable ignorance, and used it to convince others he was right.

It was 2006 and he might have since repented, so forgiveness would be in order, plus I have considerable respect for what he does, so I will not name to shame.

He cited the data provided for electrolytic capacitors as reason for it being madness to leave an amplifier (or other electronic item) on.

Electrolytic capacitor lifetime is usually shown as 2000 hours which is roughly 12 weeks. Therefore he deduced an amplifier would be no good after being constantly left on for 12 weeks (he didn't exactly state it this way - I'm simplifying it for you to understand).

He also seemed to be against electrolytic capacitors in general because of the fact they are "wet" and that they dry out, and that would be on reaching 2000 hours.

Another thing he noted was their leakage, but that doesn't mean the oil leaks out  causing them to dry out.

So what does the specification mean when it says 2000 hours? There are qualifiers, and one of them is the voltage rating, another the operating temperature, and ripple also enters the equation.

But let's take temperature rating. A general purpose electrolytic is rated at 85°C, which isn't far from boiling point. If run at room temperature the lifetime is going to be far greater, and page 7 on this data sheet gives a useful chart to show how much longer: https://docs-emea.rs-online.com/webdocs/15c4/0900766b815c4ed0.pdf

At 40°C at rated voltage and ripple we find the multiplier is 50. Instead of lasting 12 weeks, it's now 600 weeks, or eleven and a half years.

But, if ripple current is less, say half, then the multiplier is 70, taking it to 16 years.

And this ripple current is only maximal when operating an amplifier at rated power, which is where the "reservoir" is being emptied at a larger ratio to what it is being "filled" (charging current), so if left idling when not playing anything, could that take it to 20 years, or 30 years?

So far this has been at rated voltage, but we always provide latitude, and will use 100V capacitors on a 70 volt supply for example, extending life further.

The bit about leakage is not the moistness leaking out (as stated above). It is electrical leakage, leakage current to be precise, and this is what happens at first use.

The larger the electrolytic (in voltage and capacitance) the larger the leakage current. But this subsides over a period of time. This being an exponential curve, it will gradually reduce to zero, or as near as damn it.

When first switched on, the circuit voltages go to their places as capacitors charge, but it takes a little longer to reach the exact voltages because of leakage current. Some can take hours, or even days, to accurately settle to the voltages required in a design.

The product works however, and tests out specification wise, but left on longer the measurements often improve, but time is money - how much do you want to pay?

Leakage current eventually dissipates. If it didn't the circuit would never charge to the right voltage, but we know it does.

Another part of the argument went that with solid-state, on switch-on the amplifier is instantly up to its working "speed" or it couldn't be handling music, so burn-in is just twaddle (my words, not his). Leakage current proves this argument wrong.

But what is solid-state? Solid-state is chemicals replacing physicals. Rather than a glass bottle containing a vacuum where electrons can play given the catalyst of heat; transistors and the like pass electrons between chemicals (doped silicon). They are temperature (warm-up) dependent - one only has to view the term VT...

VT is the thermal voltage kT/q (approximately 26 mV at 300 K ≈ room temperature)

...but that's mathematical modelling, and it can hurt, so we tend to avoid it. But we cannot go around telling folks what's what if we can't be bothered finding out if what we say is true.

Very interesting stuff...well above my pay grade. 

It is interesting that the initial author had confused leakage current with electrolyte leaking out of the can. I heard something similar about the stains around caps on an active speaker PCB and explained it was more likely to be glue used as an anti-vibration measure.