Do external DAC & headphone amps need "burn-in"?

[Guest]

Hi ho! I'm trying to avoid marking student assignments, and posting on LTT was the first thing that came to mind. Question set: For external, solid-state DAC and/or headphone-amplifiers... do such devices require a warm up period? Or like with a brand-new vehicle or chain-saw, a break-in period (i.e., avoid maximum RPM; avoid sudden accelerations; no towing for xxyyy km)? Other than the little relay switches, there's no moving parts nor little gaskets to seat themselves. Right?

I'm thinking of the sub-$1000 CDN variety of headphone devices. And no vacuum tubes (nor steampunk-esk nightmares).

[Max_Settings]

Some headphones will require between 48 and 150 hours of burn in to sound at their optimal. Solid state Amps don't and don't need any warm up time. Tube Amps do need to warm up every time they turn on in order to function properly and the tubes do need to be burned in when they are new. Some people say DAC chips also need a bit of burn in time.

[Max_Settings]

Also for example. Electrostatic headphones a lot of the time sound bad when they are new because they need time to charge up electrons in their drivers and cables.

[Guest]

So what I what I heard you say, is that warm-up and burn-in is more appropriate in headphones (i.e., plastics stretch, vibrating diaphragms seat themselves somewhat). WRT glass-tube amplifiers, the metals stretch/contract,...  Re: DAC chips & possible burn-in. I understand burn-in can either address initial failure rate OR optimal audio output quality (whatever the hell THAT means). I find it interesting that statistically, electronics will fail most frequently after initial assembly (circuits heat up; solder-connections loosen)... and then the failure rate drops.

Re: electrostatic headphones. I like showcasing that application with my grade 9 junior-scientists. Statics is taught poorly in my realm (e.x., rub an ebony rod... touch the 'scope... it's humid so nothing happens... yay)... these case studies spice it up.

Thanks for the post @Max_Settings , eh.

[Spuriae]

There are slight advantages to warm-up even on solid state devices, but don't expect the changes to be significant. It's more about having stable operating conditions than a real burn-in phenomenon. Off the top of my head,

• Oscillators are fairly temperature sensitive. Very expensive ones (unrealistic for normal consumers) will be temperature controlled (the IC contains a heating element and insulation), while cheaper ones will decrease in phase noise slowly as temperature stabilizes.
• Most multibit DACs have better even order linearity and lower switching error around 50C. In general R2R designs are going to be very susceptible to temperature changes since the error on the most significant bit needs to be very low. A drift of 0.001% starts eating into 16 bit resolution; 0.00008% (less than 1C at 100ppm/C) is 20 bits. High-end oscilloscopes specify warm-up time and provide temperature calibration for this reason, as they too require a very accurate MSB.
• Lots of components used to set operating points have surprisingly large drifts with temperature. The LM334, often used as a current source, is variable enough to be advertised as a temperature sensor. Zener diode temperature coefficients are all over the place. These effects can be compensated for but rarely are since they aren't that big a deal for most people (after all, it's just bias and not directly in the signal path).
• Even though they aren't as temperature dependent as tubes, transistors still drift a decent amount. Though it's a higher order effect, it's roughly a linear 0.5-1% difference in gain per degree at reasonable temperatures depending on device.

Take a look at these measurements of the Yggdrasil taken at ~2hrs on vs ~2weeks. Note especially the improved jitter and midscale error. Keep in mind also that these are small effects. For amplifiers the practical warm-up time is probably on the order of minutes.

 

 

[Guest]

On 2018-02-25 at 6:02 PM, Nimrodor said:

There are slight advantages to warm-up even on solid state devices, but don't expect the changes to be significant. It's more about having stable operating conditions than a real burn-in phenomenon. Off the top of my head,

• Oscillators are fairly temperature sensitive. Very expensive ones (unrealistic for normal consumers) will be temperature controlled (the IC contains a heating element and insulation), while cheaper ones will decrease in phase noise slowly as temperature stabilizes.
• Most multibit DACs have better even order linearity and lower switching error around 50C. In general R2R designs are going to be very susceptible to temperature changes since the error on the most significant bit needs to be very low. A drift of 0.001% starts eating into 16 bit resolution; 0.00008% (less than 1C at 100ppm/C) is 20 bits. High-end oscilloscopes specify warm-up time and provide temperature calibration for this reason, as they too require a very accurate MSB.
• Lots of components used to set operating points have surprisingly large drifts with temperature. The LM334, often used as a current source, is variable enough to be advertised as a temperature sensor. Zener diode temperature coefficients are all over the place. These effects can be compensated for but rarely are since they aren't that big a deal for most people (after all, it's just bias and not directly in the signal path).
• Even though they aren't as temperature dependent as tubes, transistors still drift a decent amount. Though it's a higher order effect, it's roughly a linear 0.5-1% difference in gain per degree at reasonable temperatures depending on device.

Take a look at these measurements of the Yggdrasil taken at ~2hrs on vs ~2weeks. Note especially the improved jitter and midscale error. Keep in mind also that these are small effects. For amplifiers the practical warm-up time is probably on the order of minutes.

 

 

...off the top of your head...

...

...

...whoah. I learned something today. Thanks @Nimrodor . I’ll check that link about the Yggdrasil.

Oh, apparently Schiit’s releasing some new products soon. I’m Curious about The announcements.

[NinjaQuick]

Burn in is more about your brain becoming used to the signature than any physical changes. A solid state amplifier will sound the same until the delivery circuit starts to fail, then it will only start to sound worse/less accurate. Heating up, as has been mentioned, does affect the conductivity and resistance of all the components in a circuit, but that is not the same thing as burn in, that's just warming the components up, once you cool it down you need to start again.

 

Headphones do not have a burk-in period. If they did, it would either be a matter of the burn in period being incredibly long, or the headphone drivers themselves would eventually rip and tear and seize and fall apart under regular use within a relatively short amount of time. It is all psycho-acoustic. Like when you own k702s and think they are amazing until you get used to them and notice the 8K peak.