Tubes, glow and tube amps

Nice close ups! Makes my mouth go Woo Woo Woo! :sunglasses:

1 Like

Anyone here with a Elrog 5U4G notice any changes with burn in?

I didn’t notice any changes. I saw your post on HF. The changes I’ve found revolve around tube combinations rather than burn in of any one particular tube. For me, tube changes have a rather small impact to the sound of the Stellaris. If I was to subjectively apply a percentage - 5-10%.


No changes with burn in, actually with all the rectifiers I have tried never noticed any change with more usage compared to the power tubes or I put tubes which ar @genefurit mentioned in the 5-10% range. For example on a NOS 2A3 set, burn in helped quiet a slightly noisy tube when first used.
New production became crisper/tighter, in a break in sense.


Mine did a bit of a roller coaster over the first ~150 hours. Certainly never a complete character change, but it was kind of coming into and out of “ultra clarity”. Hard to describe, as it was over a long period, so couldn’t really A/B, but I distinctly remember the amp having periods of sounding so good, I thought I had stuck my head through a curtain and was listening to live music. And then it would pull back. It has stabilized now, and never returned to the best of the “ultra clarity” swings, but it’s better than the times when it swung slightly more bland.


In my experience that’s habituation (aka “brain burn in”) rather than a function of a tube or new product. With virtually every major upgrade (headphones, amp, DAC, tube, pads) I go through a period exactly as you describe. Sometimes it sounds great and sometimes it sounds awful. Sometimes it’ll ruin my sleep, make my ears rings, or give me a headache.

This typically lasts for 20 to 100 hours of listening time. It has no apparent relationship to physical/device burn in except for new pads, which squish flatter. It doesn’t matter if I leave the device running for a couple days (with music playing) or not. My awareness of the effect usually steps down with each listening session (e.g., 3 to 5 times over a week or two, typically). IMO it follows from the new sound profile / audible frequencies / harmonics stimulating a slightly different set of receptors and neurons.

My fatigue-oriented test and evaluation playlist was created in response to this observation.

1 Like

That’s interesting. I certainly believe in burn in. My Utopia 22 wasn’t great out of the box. The bass was meh, and it wasn’t cohesive at all. I let them run for about 24hrs, came back and it sounded much different.

The tubes that came with my WA33 needed a decent amount of burn in too. I definitely noticed a change.

Guess we all hear things differently though.


The problem with human learning / training / experience / exposure is that it’s a one-way street. You only have one chance to make a first impression, and only one chance for a first listen. As such, it’s methodologically difficult pin down the exact processes involved. After that, you still may or may not “like” the sound. We do know that humans have many learning biases and perceptual quirks, so I always suspect the human before a simple electrical device (with typically quite stable measurements).

It’s likely that the exact cause of changes (human vs. device) could be determined through a combination of blind testing, before-vs.-after equipment measurements, and human measurements over time (e.g., EEGs, fMRIs, etc.). But these have high costs, are very complex, and who cares at the end of the day?


I had a similar experience to you @Rhodey. I’d bought some vintage tubes that had never been used before and they sounded awful, to the point that I emailed the vendor asking if there might be something wrong with them since these particular tubes were recommended a lot on various forums.

I played music overnight without listening and they sounded great the next day, so I doubt it was brain burn in.

I’ve never experienced that before with tubes (either I like them forever or I hate them forever) but this was the first time I’ve tried unused tubes.


That’s interesting. I still think there’s something there with equipment being new and needing to settle in. They’re moving parts. It’s like a vehicle warming up or that initial break in period over the first 3-5k miles.

Focal even mentions run in time for about 20 hrs. Can’t remember the exact amount of time. I looked it up after someone mentioned it to me, because I was saying the Utopia changed quite a bit over 24 hrs. And I don’t feel it was my mood. It was pretty drastic, especially in the bass and openness of the signature.

But again, we all hear different.


Human variations, training, and perceptual weirdness WILL be a factor 100% of the time.

Electrical device changes MAY be a factor 0% to 100% of the time.

Pads – 100% factual physical changes over time
New tubes – 0% to audible changes
Dynamic drivers – 0% to audible changes
Other stuff???

1 Like

So you don’t think new tubes have a break in period at all?

And many companies talk about break in, even with dynamic drivers. Pretty sure ZMF even recommended some burn in.

And do cables make a difference?

Does the source? Are 1s and 0s, just that? Is a PC as good as a dedicated source?

1 Like

Tubes: As I wrote above, they involve 0% to audible changes. Some tubes hum at first, so yes, they can change.

Dynamic driver headphones and headphone pads can change. Some attribute this to the physical break down of the pad and the increased flexibility of the driver supports or surrounds. As a kid I and others put fingers through brittle old paper cone drivers, a sign of obvious decay.

Cables: The conductive material (e.g., copper vs. silver) is often obvious. However, some vendor claims are suspect. I therefore put cables in the “Other stuff???” bucket above. They are prime candidates for double blind tests as the impact ranges from nil to audible. It’s very, very hard for people to execute blind ABX tests because the ears, brain, and nerves take a while to habituate, and people experience fatigue over time. Many of us indeed fail rapid blind swaps. Again, I developed my fatigue evaluation playlist method because you can’t fake pain, tinnitus, and hissing in the ears. Pain might arise in seconds or over several hours per the setup (or never), but it’s the most reliable method I’ve tried.

Sources: Depends on the implementation. Again, check with multiple methods including electrical measurements and blind testing. Some people get goosebumps watching a vinyl record spin around…they love analog but may fail a blind test. Long USB cables do generate many errors in my experience, so no, digital is not just 1s and 0s. Other processing factors, such as Tidal’s high range filtering (not MQA), can be quite audible too.


I’m with you, @Rhodey, and respectfully disagree with @generic. I’ve found different gear to have different burn in behavior, often going contrary to expectations. Even going beyond burn-in, I can often pick out differences in equipment that is often put in the “snake oil” bucket (digital cables, power cables, tube dampers), and sometimes my findings are quite contrary to cost or expectations. The prime example I give is that my stock NUC (music server) power brick was quite obviously better than the very expensive Ferrum Hypsos. Disappointing? I guess, but I certainly didn’t just imagine the Hypsos into being better. My Chord DAVE stock USB cable was also better than other very expensive USB cables I tried (AQ Diamond, Curious Evolved). Go figure.


All I ask is for double blind testing.

1 Like

Tubes change electrically for at leat the first hundred or so hours, it’s quite measurable on a tester.
Transconductance actually rises initially before beginning it’s long slow deterioration to failure.

It’s been my experience that tubes do change, and that modern Chinese made tubes are particularly bad in this regard. A lot of NOS tubes aren’t actually N, so I rarely here a significant difference on those.


Indeed. “NOS” is often worn out trash. Tubes do change and are variable and this can sometimes be heard (or not). Still, 100% of humans are variable too.


I am a data analyst and scientist by training, so I’m generally with you on this. However, it’s generally very hard to do this in audio. The most practical matter is cost in terms of getting duplicate gear. After that, you need someone willing and able to help do blind setups. Also, any test equipment you introduce (e.g. a switch of some sort) is, by definition, another component in the chain which will have an audible effect - no matter how “transparent” it’s supposed to be. Finally, some of these nuances, I think do actually take time to settle into your head, and switching it out is only obvious after you’re used to a sound profile.

Another great example of something I caught very unexpectedly: At some point, I moved my NUC server from being in my rack, receiving the internet via wifi, over to my router, plugged in via ethernet. I didn’t think this should matter at all, as my streamer at the time (iFi Zen Stream) was grabbing the signal via WiFi and was not connected to my NUC server in either setup. Like - I literally give the change no thought. I wasn’t “testing” anything. However, when I booted everything up and started playing music, I immediately thought “wait, something isn’t right. This isn’t what it used to sound like”. Sure enough, moving my server back to my rack - again NOT connected to the rest of my chain - fixed the issue. My best guess is that the ethernet connection to a noisy router was doing something inexplicable. Maybe plugging it into my regenerator was the culprit. Maybe both. I don’t know, but the change was subtle, but obvious.

Anyway, I have given the double blind thing some thought out of pure frustration with folks not believing what seems audibly obvious to me, and many others. I have actually bought a pair of identical Mytek DAC/amps, and when I find some free time with my wife - which is in incredibly short supply these days - I will try to film some blind testing. I did a sample run with power cables and picked the better one without any issue (blind - my wife chose the cables and didn’t show me), but I’d like to have taped proof as well, with a large sample.




Yes. Per the difficulty of testing I hesitate to generalize about changes beyond fatigue.

This is a common, even dominant, viewpoint in the audio industry. That video, a tour of the system owned by Stereophile’s Michael Fremer, duplicates much of what you say.

Concur. See above regarding my fatigue tracks playlist.

Where does one draw the line on believing others on their words? Even when many are $elling products, paid to write reviews, or given equipment to provide testimonials?

Do we draw the line with crystals sitting on top of amps far away from the electrical circuits?

While I believe that many ‘objectivists’ are deeply naïve and go too far in the other direction, one must default to what is known about humans versus electric equipment. Humans are biased and suffer from illusions, and they always experience both training and fatigue effects. As such, one should expect all humans to hear “differences” from Time A to Time B or with Setup X versus Setup Y. As bio-chemical humans are known to be wildly more variable than electric devices, the most probable cause in boundary cases is that the human changed rather than the equipment.

1 Like