Personally, I simply "use it gently, as expected" : this is the burn-in... 😉 I'll mention my U-OTL project as an example :
At first, there was a prototype in early 2020 on which tests and developments have been carried out satisfactorily :
The final project was operational in late 2020 :
In 2024, I performed a check-up confirmed that the numbers did not vary one iota, compared to those measured during commissioning in 2020 :
So I consider that the Burn-In has been done since - for how long ? I don't know, in fact ! 🤔
it was done "in masked time", so to speak... 🙂
T
Last edited:
Well yes, that's what I would call a real Burn-In, by analogy with mechanic devices...
Does it changes the sound ?
Does it changes the performance ?
I do not any definitive answer, even if I build Hi-Fi and Musical Instrument as an amateur since 1980 : I's say Why not ? In the sense that sometimes you can notice improvements and sometimes not - by ear and/or measurements.
One thing is a certainty, though : used or not used, any equipment ages ! This often make a big difference... I saw this many times, and - frankly - this is quite logical, unsurprising, right ? 😉
I remember having on my bench a dead-mint-preserved-unplayed Black Face Fender Deluxe Reverb from 1964, which sounded simply awful, compared to my 70s well-played-recovered Silverface Deluxe Reverb...
T
For fear of putting oil to the fire, and admitting I am not an expert in this burn-in business, I can offer my own experience: Built around ~80 phono stages for a friend who sells them (yes it is a commercial activity, and I keep it completely separate from diyaudio), and I always test them when finished against my prototype which - by now - has seen multiple years of real usage. And sometimes I get one back for service, and the listening tests confirm the same thing.
The results I hear are very clear, and so they are to my 17year kid and some others - brand-new and still warm from soldering they sound very different from run-in. If I can trust my ears, there is one short-term thing happening (it takes about 30-60min after turn-on to sound well), and a long-term thing that takes about 100hrs of usage (not just idling - tried that), to sound well.
Don't know why that happens. But as far as I am concerned, it happens.
How big is the difference? It is remarkable even to some untrained ears.
Did I do blind testing? Not really. I did inform listeners what the test was about, and wasnt exactly hiding what I was doing, they all could identify new from run-in.
And yes, they measure exactly identical, including tonality and distortion 😀
I've just had a quick listen to some TO-247 MosFets (3 x different IRFP140, 2 x different IRFP240, 1 x IRFPC50, 1 x IRFP43107). What I could grab here. As a small preliminary exercise for a large-scale test. And also 3 TO-247 BjTs.
All MosFets already sounded very bright and lean, very nervous in the treble, generally very stressful. I didn't want to listen to music with them. The BjTs sounded more relaxed, finer, much more balanced, more deepths, albeit a little "slower". The most balanced MosFet was the IRFPC50.
I would therefore rework the device (#1) not only in terms of unnecessary transitions and materials, but also in terms of MosFets. Have a long MosFet test session;-)
All MosFets already sounded very bright and lean, very nervous in the treble, generally very stressful. I didn't want to listen to music with them. The BjTs sounded more relaxed, finer, much more balanced, more deepths, albeit a little "slower". The most balanced MosFet was the IRFPC50.
I would therefore rework the device (#1) not only in terms of unnecessary transitions and materials, but also in terms of MosFets. Have a long MosFet test session;-)
This thread has certainly moved far away from your initial question. I hope, though, that you will come back after a couple more weeks to report your own subjective experience with your new amp.
When I built my first ACA I was a bit underwhelmed when I installed it in my system and fired it up. I had, of course, heard of burn-in and wondered if there would be improvement after a period so I figured this would be a great opportunity to do a little test. My ability to do objective testing is, however, limited to taking measurements with my SPL meter and my response to your question is limited to that one factor, only.
For my test, I chose a volume level on my preamp, and a particular piece of music and recorded the SPL reading at my listening position. It was slightly less than 60 dB on average. Over the next several days, I repeated the test noticing small increases over time. After a couple of weeks (sorry, but I don't know how many actual hours of listening), the SPL was closer to 70 dB and hadn't changed over the course of several listening sessions. That was an objectively measured change that occured over time, and I would think of that as burn-in. But, that is also where I stop with any claims of "improvement". I will say, my appreciation of the ACA grew over the same time period and even beyond, but that may be due to any number of factors including my brain adapting to differences between the ACA and my previous amp.
On a side note, I've recently added a second ACA and am running the two amps as monoblocks. Very nice (but, unmeasurable by me) improvements. I'm not an engineer, so I don't understand why having the potential of 15 wpc versus 7 wpc causes the differences I hear when I'm still only using less than 1 wpc to drive my highly efficient speakers to my preferred listening level, but I'll take it.
So, that's my two cents.
Then you have probably introduced expectation bias, and most listeners would not want to sound uneducated or silly by saying the burned-in one sounded worse tp them.
If you wish to gather meaningful statistical data from listeners, I'm fairly sure that you know exactly how to test for it, if you dared go down the scientific route!
Such a large difference is pathological - there would be some other factor.
That said, at PL the product is measured and adjusted before and after a 3 day burn in
and we find this adequate. This still means that you can expect subtle differences over an hour
or so from initial turn on as they warm up...
That said, at PL the product is measured and adjusted before and after a 3 day burn in
and we find this adequate. This still means that you can expect subtle differences over an hour
or so from initial turn on as they warm up...
There are some here, actually many, who would not dare go down the scientific route as you suggest. Probably because it might shatter their long-held beliefs, which are very hard to give up.
So instead, we get responses such as:
"Trust Your Ears"
"You Can't Measure Everything"
"I Know What I Hear"
Unfortunately, people are willing clamber onto these rather easy explanations rather than face the reality that their equipment doesn't really change after the initial burn-in.
Maybe I wasn't terribly clear, but "I did dare go down the scientific route" - as I am an engineer by training and trade, that was rather natural for me.
And I repeatedly experienced what I described above, and so did my co-listeners. In multiple test sessions, I did not tell them which was which, still audible. And, the rest of the system needs to be up to the task, as the differences are subtle, not big.
But it doesnt matter - seems we are now once again down that beaten track, talking about expectaction bias, statistically relevant double-blind testing, can't measure it, etc etc etc........ So many threads have veered into that topic, where unbiased discussion is no longer possible it seems.
Time to let go that preconceived-notion bias.
And I repeatedly experienced what I described above, and so did my co-listeners. In multiple test sessions, I did not tell them which was which, still audible. And, the rest of the system needs to be up to the task, as the differences are subtle, not big.
But it doesnt matter - seems we are now once again down that beaten track, talking about expectaction bias, statistically relevant double-blind testing, can't measure it, etc etc etc........ So many threads have veered into that topic, where unbiased discussion is no longer possible it seems.
Time to let go that preconceived-notion bias.
"Scientific" is not when is not done, what is claimed to be done:
In case of audio: audio:
listening, hearing;-)
In case of audio: audio:
listening, hearing;-)
The reason for being on that beaten track is that despite being an engineer you did not give any possible explanation for this phenomen. You only repeated classic audiophile clichés (sans "wife in the kitchen slicing onions"). If you have designed your device you surely should have some idea what is causing the extraordinary behaviour. Extraordinary in the sense that you claimed the sound only improves, not gets worse, which should be as likely outcome. If you truly want to discuss this issue why not give us an educated guess of the cause maybe accompanied with a schematic?
Actually, you didn't go down the scientific route at all, by your own description of it:
"Did I do blind testing? Not really. I did inform listeners what the test was about, and wasnt exactly hiding what I was doing"...
I don't see any way that you can possibly call that a scientific route.
What was your actual protocol, including the number of tests and results, as these two comments are conflicting.
I am completely unbiased in the matter of burn-in (as I am with all things unproven), however science didn't progress through personal assertions. As an engineer you surely know that experience/perception with no data or attempt at explanation is largely worthless. If you can provide a link to statistically viable burn-in test data I would be very interested to read it. If burn-in is indeed real, designers would be falling over themselves to design equipment that either a) Does not require burn-in and sounds at its best straight out of the box, or b) Designing equipment especially to fully maximise the effects of burn-in. Both would be great selling-points!
As with religions, very many hifi myths and 'improvements' remain unfalsifiable as they are based on a kind of personal revelation, so I expect few of them to ever be proven either way.
The very few double-blind tests published indicate that the ears are extremely poor at analysing sound when the equipment under test is unknown and not visible. Gold Ears become Cloth Ears, as I have stated before.
There are days when I listen to a piece of music and simply enjoy it; another day listening on the same track on the same system it will move me to tears. Supposing I were to suggest that it was the phase of the moon which made the equipment sound so much better and which elicited my emotional response? I can hear 'engineers' yelling 'nonsense' or prove it' ringing in my ears already...
No. I reported my experiences, nothing more, nothing less. You don't believe my and my co-listeners impressions, thats fine.
The schematic is rather straightforward, opamp with RIAA in the feedback loop. There are two capacitors which have evolved over time, plastic foil first and later versions use Silver Mica.
My suspicion is that these capacitors, and in particular, their dielectrica, change their behaviour over time. This is a phenomenon known from isolators in high-voltage applications, where mobile charges start to move in the presence of strong electric fields. There, the action is not so subtle and can lead to a mechanism called "partial discharge", which then even can impact reliability.
What is still puzzling is the fact that this has an audible impact on sound, even though the measurement results do not change. Menno van der Veen published a very interesting paper on transformers back in 2007 (https://www.mennovanderveen.nl/images/onderzoek-ontwikkeling/publicaties/download_1.pdf) , talking about a similar issue, only with magnetics instead of capacitors. He also published a good paper on capacitor distortion (https://www.mennovanderveen.nl/images/onderzoek-ontwikkeling/publicaties/download_4.pdf) , pointing towards even other sources of nonlinear behaviour.
What I gather from these, is that component changes can influence signals at low levels, and even more so for small signals, making this very hard to pinpoint with measurements. But it seems that our ears can detect some of this. When we are talking about a signal that is down -50dB and the impact happens at -60dB, we are close to the noise floor of many test systems.
Has anybody here made similar observations?
Why would this change of capacitor dielectric result only in sound improvement, not worse sound?
Mobile charges would cause fluctuations in the electrical signal passing through the capacitor while they are moving. Once they have settled down and are trapped in a place where the electric field required to move them again is stronger than the signal level, it does not happen any more.
Do you have a better idea?
Do you have a better idea?
That fails to explain the sound improvement which you stated happens always.
Unimpregnated mica sheets can be impacted by moisture resulting in changes in dielectric properties. However silver micas are epoxy impregnated and very stable. Change of dielectric properties due to moisture in silver micas could potentially happen if the impregnation is not stable or has been damaged. Components exposed to moisture should be baked. This however is not burn-in but normal manufacturing practice.