I do hope that Nelson Pass reads your opinion. 😀
Regarding the original question I would suggest a simple test. Take two new capacitors and clean the ends of the leads. crimp one - ensuring a clean and tight connection - in circuit, preferably to unsoldered components/wiring. Listen for 10-20 minutes. Then take that cap out of circuit and solder the second one in the place of the first and listen again! Then report back here please. I expect to read a report which tells of some noticeable difference in sound as the soldered joints have been subjected to considerable heat whereas the dry/crimped joints have not. Could be an interesting experiment. [ May also add a dimension to the long running cables thread - especially in the case of heavy cables which have had solder pot treatment for terminal attachment.]
Last edited:
I think that if your transistors are operating any different after 'n' hours of use, you need to recheck your design. Modern semiconductors are petty reliable for a long time if the circuit design is sound. As for cables, how long does it take for corrosion to set in after removal from the package and what effect does that have on sound? Obviously gold plating would be beneficial here.🙄 I’m skeptical cables could change other than corrosion or by physical damage. I can see mechanical devices requiring some burn in time. Caps might change a bit over time, but I would think more than a few hours. I disagree with some of Carlos's pdf. Tubes are better amplifiers than transistors but work differently. I didn’t read any further, it seemed to mislead a bit into nonsense. Any knowledgeable person can pull and tweak the leg of those who are in ignorance of the subject.😉
It is not the devices themselves, but the connections to the devices that settle down during burn-in.
Ok, that makes sense. Is there any difference or reason related to this in using SMD components, assuming they are assembled properly with an oven?
Without intending to speak for JC it is my experience that the connections he is referring to are both the external leads but particularly the internal connections that run from the dye to the outside world. Look up reliability engineering, failure analysis, chaos theory pertaining to probability and failure prediction.
Burn in is one of the things manufactures of electronic components do in order to sort things out into different "batches" that meet certain specifications. Take microprocessors as an example. What happens is you make an *ssload of chips and burn them in to a given set of parameters. Some of the devices will fail early on - some will operate at a given clock rate and temp and others will still operate but at a lower clock rate or temperature. The really good ones get the [fill in the blank certification - say class 1] the others will be certified at a lower level of operation. Most of the stuff you purchase has ALREADY been burned in!!!! THAT's why most folk will never see it or need to deal with it.
When you are in the business of building bunches of stuff tho you will see it and need to adjust your manufacturing processes to account for it.
If it was not so then ponder a situation I'm sure that many here are familiar with - matching transistors - FET's - speakers. You need to match them because of variations present in the manufacturing process (and there can be all sorts of variations involved). Once the devices have been checked for operation points (matched pairs or groups) burn them in for a specified amount of time - say a 24 hour hot and cold cycle and go back and check them for still being matched. When you do something like this in batches of hundreds or thousands you will see where it does make a difference. BUT - most of the stuff we play around with has (or should have) already gone through this process.
That being said I should also present the flip side to "burn-in" and that is the concept of "burned-out"/"worn-out". Burn-in improperly performed can be detrimental if improperly performed and wind up detracting from the performance of what ever it is your playing around with at the time. BTW - QC geeks lovingly refer to this process as "shake-n-bake". Kinda fun (but scary) to see something being shook apart on a vibration table to the point parts start flying off - usually at very high speeds.
Don't get me going on applying burn-in to systems integration - that's where things start to get really interesting. 😱
Last edited:
We did the tests alluded to in my last post (*41 above). You are totally right JC. We left the soldered caps in circuit (in the eq of a phono preamp) and they did alter with soak testing as the circuit stabilized. We did not then reverse the test as we had already answered the question which we had pondered. The same is particularly relevant to heavyweight speaker cable terminations and interconnects - we also did similar tests.
On a more general note, this and other threads have dozens of demands/requests that those who have noticed such changes - and especially for JC - produce "evidence". I don't have any sensitive measuring equipment [nor even the knowledge to use such stuff]. I have had to rely on my ears over the many years I have played with this audio interest. As my ears are not now as good as they were (I am older than even JC - I mean the John Curl version of those illustrious initials😀. I also used to for many years shoot almost daily with 12 g shotguns - before ear defenders were common). The result is that I am quite prepared to accept that which JC and a few others such as Joachim Gerard have to say as they have the experience of listening to their designs and using that evidence as the final arbiter - once the required technical criteria has been met. I suspect that many of the nay-sayers actually do not have the knowledge to assimilate and understand the "evidence" even if it was to be offered.
As some of those here earn their daily bread by designing audio gear they will obviously be subject to various forms of confidentiality agreements which would be breached if they were to publish evidence. Also do not forget that there are a lot of very unscrupulous people who comb sites such as this and who would rob anyone blind given half an opportunity.
any why should they? we are folks who likes listening to our music, and nothing else.....😀
Tony, if you are where you think you are, I don't think you know what hi end hi fi is in the first place. To comment on things outside your experience is inappropriate.
John,
I'm not so sure about it.
I've done various cap tests to optimize component selection for my MyRef using sockets and burned in Silmics socketed: they changed a lot as usual in the first 40-50 hours.
I've also socketed a brand new Silmic and the difference was obvious (no need to blind test here...😀)
But I've also noticed that joints needs some burn-in too to settle.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I think John was referring to the connections inside the package from the pin to the die. This is why I was wondering if baked SMD's resulted in less effect because the entire package would be heated and cooled evenly. It is not entirely reliable to solder tiny SMD devices with a soldering iron because of extreme temperature differences causing metal fatigue in the bonding wires.
You mean Nelson agrees 😉 that's nice 🙂
As to soldering of components... that's open to much interpretation as to how it's carried out.
There's a world of difference between a board full of components with 0.8mm leaded parts soldered at 90 degrees into 1 or 1.5 mm holes, where the solder is the only connection, and soldering so the lead actually runs for a mm or two against the print.
I wish JC would have agreed to participate in the cable thread test. I can understand his reasons for not doing, but on the other hand being so sure the differences exist and are easily discernable, then why not 🙂
I still stand by my views on the original question.
In the tests to which I referred the circuits involved were hard wired; point to point. [I agree with your inferred view of PC assembly.]
Nelson Pass - as you probably know full well - advises that at least some of his amps need 30 mins or so stabilize. This too has been my experience of amps in general and the same applies to all digital sources which I have had at home. As a car engine and transmission need to get to full working temperature before stabalizing there is no reason why an amp etc...considering that there is movement of electricity occurring...should not require the same: there is absolutely no reason for expecting a nil warm-up period.
Regarding items with physical motion such as speaker drivers and cartridges obviously these do need to be run in - however there are many old hands who believe that a speaker drive unit or a cartridge are at their absolute best very shortly before the suspensions fail. If indeed these people are correct that would imply that improvement with use is more of a continual improvement than a short term run-in. Perhaps we should treat speakers with more care and only gradually work them up to high volume over a period of a couple of hundred hours or so? I know that this is true of cars. I use a petrol Audi A6 estate with 2.7T turbo engine - it was very carefully run in and did not sit at fixed revs over many miles. It has still covered less than 70,000 miles and I am totally convinced that the engine is still improving.
Incidently where are you based in the UK?
Last edited:
Hmmm... solid state circuits stabilise (or should do) within seconds at most, with all operating points carefully controlled. The quiescent current in an output stage or the tail current in a long tailed pair for example should be the same cold as hot. I know you can argue the beta and ft of transistors etc varies with temperature but that shouldn't really figure in any competent design.
I agree speakers etc being mechanical will have their characteristics vary with use, and humidity even may affect some designs of cone.
In a rather sunny Lancashire 🙂
I agree speakers etc being mechanical will have their characteristics vary with use, and humidity even may affect some designs of cone.
In a rather sunny Lancashire 🙂
And they do stabilise but as yourself pointed transistors and caps too varies with temperature and only when the normal temperature of operation is reached the circuit operate in a consistent manner.
Until that temperature is reached the circuit works, by design, but the full potential is expressed when it's warm.
🙂 Again I think so much of this comes down to "belief" and hearing what you want to hear.
I always ask why folk believe that everything improves with being on, being hot, being "burned in" etc. Mechanical devices excepted.
Why do you never hear of anything that is recommended to be kept "cool" or that it only sounds at it's best when "cold"... it's just as valid an argument. Why does the change from cold to hot not make some designs sound worse ?
Well actually you do hear of cooling... in instrumentation etc where cooling to low temperature lowers Johnson noise etc in critical components... but that is done for clearly identifiable aims.
But back to audio... Now if an amplifier for example, really was felt to perform better hot than cold etc, then surely the designer of that product could look to using different devices etc, or even use the simple technique of freezing certain devices etc to try and find out why that happens, and to design that characteristic out of the product by using more suitable devices.
Higher temperatures cause (amongst other things) higher leakage currents in caps, and semiconductor junctions.
It's just to easy to always say that some "magical" property is at work, and we hear it but don't understand it, and that it's always hotter/powered up equals better. On that basis there must also be examples where the reverse is true.
- Status
- Not open for further replies.