comparing 2 same amps (old vs new)

[khaliss]

Hi all,

Just wanted to share an observation & perhaps get some feedback from the experienced & knowledgeable ASR members. So I purchased a second Musmys E-406 amp (original post here) with the intention of eventually having a true vertical bi-amp setup once I upgrade my main speakers (Martin Logan ESL) to a higher-end pair like the ML Classic ESL-9.

The first one I had for about 7 months now, and has probably between 600-1,000 hours of usage. It is still working reliably & remarkably well... sounds as good as the day I got it. So, before I disconnected the cables on the back, I did one more REW in-room measurement test. Then I plugged everything on the brand-new amp, without any changes to how the cabling (DAC > preamp > amp). Then after a bit of warmup, I measured with REW. Here are the graphs:

LEFT channel = red (new) | blue (old)

LEFT channel = green (new) | yellow (old)

Definitely consistent, however the difference is I actually had to lower the preamp volume by around 3db on the new amp before taking a snapshot of the graphs. It seems the newer amp is about 3db louder than the old. So I asked AI (Gemini) if this is normal (comparing old vs new), here's Gemini's answer:

Let me know what you guys think... is what Gemini AI claiming realistic? =)

 

[SSS]

Without real measurements of the Amps themselves hard to tell. Sinne they are clones the answer from Gemini seems reasonable. Although not all of it might be a reason.

 

[solderdude]

The difference (using an unreliable method) shows 0.5-1dB difference, not 2-3dB.
And that is even within measurement error seeing that sometime 'old' is above 'new' and other positions 'new' is above old.

Microphone measurements in a room are unreliable.

You can change the scale in REW so that's one thing that takes away the guess work on the level differences.

Totally acceptable and inaudible numbers b.t.w. for such knock-offs (a difference of 0.5dB or so).

It fits in the 'resistor tolerance range'.
The rest of the 'theories' by the AI are nonsensical and can safely be discarded.

 

[Petrushka]

I’m curious how you determined the setting. By the meters on the amp?

 

[Sokel]

Despite what the knock-offs do, you have to do something about this 2.5kHz-6kHz bump there, it's painful only to look at, 5db at 3.5kHz is the worst case scenario.

 

[DVDdoug]

The difference (using an unreliable method) shows 0.5-1dB difference, not 2-3dB.

But he said:

I actually had to lower the preamp volume by around 3db on the new amp before taking a snapshot of the graphs.

I'd say it's a manufacturing change/difference, not "age" or "burn-in". Normal resistor tolerances wouldn't create a 2-3dB difference but a different resistor or the wrong resistor certainly could!

And really microphone measurements create lots of additional variables and it's difficult to get the exact same measurement twice. Your curves are remarkably close.

I asked AI (Gemini)

 

[khaliss]

correct, before I did the measurement test, there's an SPL meter in REW... and based on that SPL pre-test, the new amp was measuring +3db louder (without adjusting anything else)... so I lowered the preamp by about -3db for the new unit, to see how closely the SPL graphs would be on old vs new. Everything else is the same (I didn't reposition the microphone, same preamp, same computer running the test, etc)

 

[khaliss]

I don't know why some of you have anything against AI, but I use it for work & day-to-day stuff, it has helped me tremendously on all aspect in life LOL. Why not use it for the audio hobby as well? =D

 

[khaliss]

Despite what the knock-offs do, you have to do something about this 2.5kHz-6kHz bump there, it's painful only to look at, 5db at 3.5kHz is the worst case scenario.

I typically enable the Equalizer APO software (parametric EQ correction with REW) to try to have a more desirable neutral response graph with my HTPC (which is my streamer as well), the 2.5-6khz bump is "neutralized", but I disabled EQ-APO to take raw measurements/graphs for this comparo.

Click the "REW graphs LINK" on my sig so you can see the difference after EQ-APO is enabled =)

 

[solderdude]

correct, before I did the measurement test, there's an SPL meter in REW... and based on that SPL pre-test, the new amp was measuring +3db louder (without adjusting anything else)... so I lowered the preamp by about -3db for the new unit, to see how closely the SPL graphs would be on old vs new. Everything else is the same (I didn't reposition the microphone, same preamp, same computer running the test, etc)

Then the most likely answer is the manufacturer simply changed the gain (just a matter of changing a single resistor value).
The AI (in this case) is telling porkies and simply searched for the words you typed in the question and then it searches for the words on all kinds of websites.
It does not know if the info is reliable.
Then it creates a nice 'summary' that looks nice but contains nonsense.
AI is fine for certain things but not all.

 

[khaliss]

so you are saying the other factors (that could affect the gain) suggested by Gemini AI are not actually realistic(?) Well, my plan is to burn-in the new amp for 6+ months and re-measure then. If it doesn't change, guess I can always adjust the balance knob on the preamp to equalize the gain output between the left channel versus the right... when I use these amps to bi-amp (vertically) my next speaker upgrade =)

 

[DVDdoug]

I don't know why some of you have anything against AI,

Everything it said was ridiculous.

The comment about "environmental" changes could ALMOST be true but things like slightly different microphone placement or your location in the room while measuring will have more effect in the higher frequencies and almost no effect at lower frequencies. It couldn't cause a "clean" 2-3dB difference overall.

Why not use it for the audio hobby as well?

Well... Most "audiophiles" are nuts and most audio related websites and publications are full of nonsense so AI is "learning" a lot of misinformation. This is one of the few rational-scientific audio-related resources.

 

[solderdude]

Then the most likely answer is the manufacturer simply changed the gain (just a matter of changing a single resistor value).
so you are saying the other factors (that could affect the gain) suggested by Gemini AI are not actually realistic(?) Well, my plan is to burn-in the new amp for 6+ months and re-measure then. If it doesn't change, guess I can always adjust the balance knob on the preamp (to equalize the gain output between the left channel versus the right... when I use these amps to bi-amp (vertical) my next speaker upgrade =)

All the other 'suggestions' are total nonsense and could never (if they actually could create small differences) not in a million years produce a 3dB difference (on both channels the same no less). Here's the thing... that nonsense almost certainly can be found on the web and that's where it got that info from.
Nicely formatted and aimed toward the question that was asked.
AI does not know anything about the subject of the question.

Burn in will NEVER change the gain of an amp.
When you need the amps to have the exact same gain you simply need to use a -3dB attenuator in front of the newer version.

 

[khaliss]

Everything it said was ridiculous.

The comment about "environmental" changes could ALMOST be true but things like slightly different microphone placement or your location in the room while measuring will have more effect in the higher frequencies and almost no effect at lower frequencies. It couldn't cause a "clean" 2-3dB difference overall.

I tested twice (at least on the new amp) to make sure the resulting graphs have consistency. The measurement microphone sat on top of my listening chair, it never moved, even for a fraction of an inch (there is no wind/breeze that could have moved it inside my listening room LOL).

 

[khaliss]

Well... Most "audiophiles" are nuts and most audio related websites and publications are full of nonsense so AI is "learning" a lot of misinformation. This is one of the few rational-scientific audio-related resources.

That's why I wasn't sure how realistic the AI's response was... I posted this thread to see it here to see what true audiophiles/nerds think hehe

 

[khaliss]

Burn in will NEVER change the gain of an amp.
When you need the amps to have the exact same gain you simply need to use a -3dB attenuator in front of the newer version.

gotcha... I do have the balance knob on the preamp, so I will use that to "equalize" the SPL output between L & R channels, once I go into vertical bi-amping =)

 

[Sir Sanders Zingmore]

Sort of on topic, I reckon it would be interesting to see comparison photos of the internals of both amps. I wonder how consistent the manufacturing is

 

[khaliss]

Sort of on topic, I reckon it would be interesting to see comparison photos of the internals of both amps. I wonder how consistent the manufacturing is

the amps are heavy (about 47 lbs each), but I will try to find time to remove them from the rack... and put them side by side, open the cover, and take a closer peek inside to compare both. They were purchased about 6 months apart (the new one just took longer cuz it shipped with the cheapest option, which took 5 weeks lol), the older one was via DHL which was just 5 days.

 

[threni]

That's why I wasn't sure how realistic the AI's response was... I posted this thread to see it here to see what true audiophiles/nerds think hehe

AIs don't tend to give sources; they're *literally* just saying "well, I read about it on the internet". You could cut out the middleman and ask your question directly on reddit or head-fi.

"I use it for work & day-to-day stuff, it has helped me tremendously on all aspect in life"
It's safe to use when you know enough about the domain to know when you're getting lied to.

 

[khaliss]

AIs don't tend to give sources; they're *literally* just saying "well, I read about it on the internet". You could cut out the middleman and ask your question directly on reddit or head-fi.

"I use it for work & day-to-day stuff, it has helped me tremendously on all aspect in life"
It's safe to use when you know enough about the domain to know when you're getting lied to.

i pretty much use AI daily, mostly for work... I would say it gets things right about 80% of the time, so I've come to rely on it. But yeah, that still leaves 20% wrong, so thus I started this thread Though on hindsight, it's so convenient+efficient & a real time-saver (rather than googling/researching things yourself)