[JoeKickass]

I'm not sure anyone 100% understands all aspects of burn-in, but I like to think of it as breaking in a pair of jeans, something getting used to a pattern of wear.

Semiconductor & capacitor properties vary with temperature, load, rf environment, etc. and modern micro-components also have quantum-scale effects so any number of complex chemical and physical interactions can take place over time.

Just because we don't understand something doesn't mean it isn't having real effects!

---
Burn-in is a real, common effect, and while it's causes are not known, the number and diversity of factors that affect semiconductor performance mean that physical changes cannot be ruled out.
I also believe the human mind helps amplify the effect, like a placebo.

Is burn-in real? Is it placebo? I suspect the answer is a mix of both.

 

[Lost Cosmonaut]

I'm not sure anyone 100% understands all aspects of burn-in, but I like to think of it as breaking in a pair of jeans, something getting used to a pattern of wear.

Semiconductor & capacitor properties vary with temperature, load, rf environment, etc. and modern micro-components also have quantum-scale effects so any number of complex chemical and physical interactions can take place over time.

Just because we don't understand something doesn't mean it isn't having real effects!

So when Schiit first designs and creates an amp, they won't know what it actually sounds like until it's burned-in? So basically, designing audio products is liking shooting in the dark--there's no real precision or fine-tuning? It's just a bunch of guesswork.. I don't get it.

 

[JoeKickass]

So when Schiit first designs and creates an amp, they won't know what it actually sounds like until it's burned-in? So basically, designing audio products is liking shooting in the dark--there's no real precision or fine-tuning? It's just a bunch of guesswork.. I don't get it.

It's an over-simplification to call it shooting in the dark. Semiconductor design equations give you things like power, gain, and current values so you know roughly what an amp will do.

But I'm pretty sure you are right that they (or any other audio device maker) won't truly know exactly what a device sounds like until they test listen to it!

I bet that adds a lot to the fun and enjoyment of their work. It also shows the value of experience, and the benefits of knowing what different designs sound like.

 

[Jimster480]

So when Schiit first designs and creates an amp, they won't know what it actually sounds like until it's burned-in? So basically, designing audio products is liking shooting in the dark--there's no real precision or fine-tuning? It's just a bunch of guesswork.. I don't get it.

If you use logic, nothing like what guys like that say will make any sense

Capacitors and other components DO NOT CHANGE after usage, they have a set of specs and they are tested prior to shipping (at time of MFG) to make sure they meet those specs.
Burn in of computer electronics is literally nothing more than myth as it would mean that every device we have would behave differently after using it. It would make designing/selling/building all computer electronics completely impossible.

Edit:
But when building audio products, you don't really know exactly what it will sound like until you build it. You have to then test components to see how it changes the sound, or tweak the components you have (things like TI Class AB Amps have some configuration options which you can change "on chip" or even externally, such as gain for a simple one).

 

[JoeKickass]

Capacitors and other components DO NOT CHANGE after usage, they have a set of specs and they are tested prior to shipping (at time of MFG) to make sure they meet those specs.
Burn in of computer electronics is literally nothing more than myth as it would mean that every device we have would behave differently after using it. It would make designing/selling/building all computer electronics completely impossible.

I think you are confusing digital with analog. Everything ages, and everything changes, it's just a matter of scale.

For burn-in we are talking about minuscule, perhaps immeasurable changes. A headphone amp like the Magni is an analog device, and we hear every detail down to it's finest resolution.

Digital devices like computers are built with huge tolerances, they are designed so that a 1 is still a 1 and a 0 is still a 0 even if your temp is -20C or 40C, your wall voltage is 105V or 125V, or you use the crappiest USB money can buy:

 

[Jimster480]

I think you are confusing digital with analog. Everything ages, and everything changes, it's just a matter of scale.

For burn-in we are talking about minuscule, perhaps immeasurable changes. A headphone amp like the Magni is an analog device, and we hear every detail down to it's finest resolution.

Digital devices like computers are built with huge tolerances, they are designed so that a 1 is still a 1 and a 0 is still a 0 even if your temp is -20C or 40C, your wall voltage is 105V or 125V, or you use the crappiest USB money can buy:

This one graph is not indicative of all of computer design.
Processor voltages, northbridge voltages, LLC (line level currents) all are very precise or the system will crash.
Modern CPU's and other IC's fluctuate thousands to billions of times per second in terms of voltages / current requirements / frequencies.
What you are showing is that digital data has a huge tolerance, and its also how data is not messed up with cables as some people suggest.

But this doesn't mean that your capacitors or transistors will be putting out different voltages or magically having different impedances as they burn in.
The tolerances are quite low (under 0.2% typically) and they can be even 0.02 or 0.01% depending on the type of component.

 

[Lost Cosmonaut]

If you use logic, nothing like what guys like that say will make any sense

Capacitors and other components DO NOT CHANGE after usage, they have a set of specs and they are tested prior to shipping (at time of MFG) to make sure they meet those specs.
Burn in of computer electronics is literally nothing more than myth as it would mean that every device we have would behave differently after using it. It would make designing/selling/building all computer electronics completely impossible.

Edit:
But when building audio products, you don't really know exactly what it will sound like until you build it. You have to then test components to see how it changes the sound, or tweak the components you have (things like TI Class AB Amps have some configuration options which you can change "on chip" or even externally, such as gain for a simple one).

The concept of burn-in makes no sense... if it's true, then there should be great variation and inconsistencies in copies of the same audio product. Not even the manufacturers would know what their product would ultimately sound like if the components are that mutable and inconstant.

 

[DJ The Rocket]

If you use logic, nothing like what guys like that say will make any sense

Capacitors and other components DO NOT CHANGE after usage, they have a set of specs and they are tested prior to shipping (at time of MFG) to make sure they meet those specs.
Burn in of computer electronics is literally nothing more than myth as it would mean that every device we have would behave differently after using it. It would make designing/selling/building all computer electronics completely impossible.

Do you have a source you can cite that confirms/explains this? It's not that I doubt you, I actually suspect you're right, which makes me want to see this confirmed/explained even more.

Or @JoeKickass, same thing. I'd be interested if you can cite a good source from your perspective as well.

 

[Jimster480]

The concept of burn-in makes no sense... if it's true, then there should be great variation and inconsistencies in copies of the same audio product. Not even the manufacturers would know what their product ultimately sounds like if the components are that mutable and inconstant.

Basically it would be a complete crap-show in terms of selling products.
I mean even studios would produce inconsistent results because of how their audio devices performed from one day to another day.

All that stuff was solved a long time ago now that mfg tolerances are incredibly small and everything is so precision mfg'd.
its the same with all the car stuff these days too, power that used to be able to be made with simple bolt ons or "shaving down extra mfg casting material" basically doesn't exist because everything is very precise from the factory.
The availability of components at such a low price is astounding.
I just ordered all the parts for the pmillet hybrid NuTube amp last night and honestly even Nicheon and WIMA Caps with very precise tolerances are $1 or less each and IC's are $0.39-$5 even for great OpAmps and other things.

Hell even when building things like tablets you can buy entire SoC's for $5 in bulk pricing or $8-9 in "singular" pricing.

 

[Jimster480]

Do you have a source you can cite that confirms/explains this? It's not that I doubt you, I actually suspect you're right, which makes me want to see this confirmed/explained even more.

Or @JoeKickass, same thing. I'd be interested if you can cite a good source from your perspective as well.

What kind of source are you looking for? Mouser is an electronic parts whosaler so you can look at something like that. Unless you mean tolerances, in which case you would have to do some digging..

https://en.wikipedia.org/wiki/Engineering_tolerance

https://sciencing.com/tolerance-resistors-6965888.html
https://en.wikipedia.org/wiki/Failure_of_electronic_components
http://reliasoft.com/newsletter/v7i2/burn_in.htm

^The last two basically explain that "burn in" is basically running the device under stress near the max tolerances of the chips to see what "fails" or how testing devices with "burn in" helps mfg's to understand "expected life". Things like characteristics changing are barely mentioned because its so silly in the eyes of actual component design.

 

[Lost Cosmonaut]

Basically it would be a complete ****-show in terms of selling products.
I mean even studios would produce inconsistent results because of how their audio devices performed from one day to another day.

All that stuff was solved a long time ago now that mfg tolerances are incredibly small and everything is so precision mfg'd.
its the same with all the car stuff these days too, power that used to be able to be made with simple bolt ons or "shaving down extra mfg casting material" basically doesn't exist because everything is very precise from the factory.
The availability of components at such a low price is astounding.
I just ordered all the parts for the pmillet hybrid NuTube amp last night and honestly even Nicheon and WIMA Caps with very precise tolerances are $1 or less each and IC's are $0.39-$5 even for great OpAmps and other things.

Hell even when building things like tablets you can buy entire SoC's for $5 in bulk pricing or $8-9 in "singular" pricing.

In other words, burn-in = wishful thinking on the part of the consumer.

 

[Jimster480]

In other words, burn-in = wishful thinking on the part of the consumer.

Basically.
The "burn in" period of things such as tubes though can occur, so if you have a tube amp it is theoretically possible. Although its usually more of a "warm up" period because even on tubes these days they don't really change over time (until they are dying).

 

[JoeKickass]

The concept of burn-in makes no sense... if it's true, then there should be great variation and inconsistencies in copies of the same audio product. Not even the manufacturers would know what their product ultimately sounds like if the components are that mutable and inconstant.

I'm not saying burn-in is random. The various things that cause it are complex, but they are still reactions to physical laws...
If all the products are manufactured the same way, with the same components, they will burn-in with the same physical and chemical reactions and sound basically the same.
So the prototype the maker develops and spends hours listening to will sound basically the same and the one you buy and break-in yourself.

All that stuff was solved a long time ago now that mfg tolerances are incredibly small and everything is so precision mfg'd.

The tolerance of a component is just how accurate it's initial value is, it will still age and change with environmental and physical properties. Everything does.

I'm an EE and I'm not saying I can explain exactly what is happening during burn-in, but I know enough to understand that all the first person accounts are probably rooted in a physical reality.

EDIT: I'll admit that the physical reality could be a biological placebo effect, but just realize that solid state analog devices aren't immune to micro-changes!

 

[DJ The Rocket]

What kind of source are you looking for?

I don't know exactly, I was hoping you would I haven't taken an engineering class since I was 20, so a 101-level textbook from the library might be most appropriate, but since I'm not going to do that I'll start with the ones you posted!

 

[Jimster480]

I'm not saying burn-in is random. The various things that cause it are complex, but they are still reactions to physical laws...
If all the products are manufactured the same way, with the same components, they will burn-in with the same physical and chemical reactions and sound basically the same.
So the prototype the maker develops and spends hours listening to will sound basically the same and the one you buy and break-in yourself.


The tolerance of a component is just how accurate it's initial value is, it will still age and change with environmental and physical properties. Everything does.

I'm an EE and I'm not saying I can explain exactly what is happening during burn-in, but I know enough to understand that all the first person accounts are probably rooted in a physical reality.

EDIT: I'll admit that the physical reality could be a biological placebo effect, but just realize that solid state analog devices aren't immune to micro-changes!

The tolerances are typically over the lifecycle of the device, yes its possible for there to be micro-changes, but nothing big enough to actually affect audio quality. I mean you would be talking about a change of less than even 0.25%, if voltage changed under 0.25% then it wouldn't likely cause even a 0.5db change at any octave so the difference would be almost always inaudible.
Typically it would be 0.02% change... once components have passed their lifecycle though its possible for things to go out of spec.