Review and Measurements of Meier Corda Jazz Amp

[trl]

I wonder what happens if injecting the audio signal (one channel only) between VolCtrl_B and GND and placing a scope between TOUT and GND. This should basically invalidate the active ground, right?

 

[dudewithcans]

The presence of channel mismatch shows that it's got analogue volume control, which can be quite bad at a low volume setting.

thats true, but i guess thats why more expensive devices with analog volume control have parts like RK27 to avoid that, just like the corda rock does? but completely agreed, i would say im a fan of that corda jazz and classic poti + AD converter design cause they have that nice analog feeling with the pots, but actually its digital volume control (without any channel deviation, just like amir measured). i think thats pretty nice.

I don't think we're in major disagreement on anything in this discussion. Let's leave it

agreed, friend!
and to point this out, by no means have i any intend to defend the corda jazz, rock, or anything produced by meier audio (heck, im not even owning that device, my ears trust in lake people). im just sayin what i think about it.

 

[solderdude]

I wonder what happens if injecting the audio signal (one channel only) between VolCtrl_B and GND and placing a scope between TOUT and GND. This should basically invalidate the active ground, right?

I don't think so.

Assume Vctrl LB = 1Vtop.
Then at R3-R24 there will be 0.78V.
At Output U3 there would be -0.28V.
These signals are 'mixed' by R24 and R26 acc. to their values.
R26 current would be -7.823uA
R24 current would be 29.1uA
So on high gain there would be 21.28uA across R32 = Tout = 1.192 V
On Sout there would be 0.435V in counter phase so 1.627V across the load.

Quite possible the above isn't correct though.
Don't feel the need to simulate.

 

[MRC01]

I drawed a (simplified) schematic of Jazz (without the feedback network). The volume control is achieved by changing the parallel feedback resistor between VolCtrl_A and VolCtrl_B. I don't think it is a clever design as it will change the gain of the first 2 OPAMPs during volume adjustment.

Wow, excellent discussion and insights here!
Does that mean the distortion at high volume settings (past 2:00 on the knob) is more likely to happen in low gain, than in high gain?
I can definitely hear it at low gain, but at high gain it's too loud for me to listen that high.
It looks like this amp has some creative design ideas, but the implementation was not fully thought out or tested.

 

[pika02]

Wow, excellent discussion and insights here!
Does that mean the distortion at high volume settings (past 2:00 on the knob) is more likely to happen in low gain, than in high gain?
I can definitely hear it at low gain, but at high gain it's too loud for me to listen that high.
It looks like this amp has some creative design ideas, but the implementation was not fully thought out or tested.

I simulated at low gain mode. The circuit works fine while input voltage is 6Vpp and the volume is max, but the total gain of jazz is lower than 1.

 

[Deleted member 2944]

I drawed a (simplified) schematic of Jazz (without the feedback network). The volume control is achieved by changing the parallel feedback resistor between VolCtrl_A and VolCtrl_B. I don't think it is a clever design as it will change the gain of the first 2 OPAMPs during volume adjustment.
View attachment 24442

Where is the crossfeed disabling scheme in that??

 

[pika02]

Where is the crossfeed disabling scheme in that??

I just not wanted to draw that things.

 

[sssn]

Probably designed just to be different than the rest ?

Funnily enough, that's what NwAvGuy said about Jan Meier's designs as well:

[...]Meier has done some interesting things. My personal opinion is he keeps trying to re-invent the wheel when there's no need to. His "virtual ground" arguments, for example, make no sense to me. [...]

https://nwavguy.blogspot.com/2011/08/o2-summary.html (Scroll down, it's a comment)

I still respect Jan though, he is knowledgable and innovative. He's also not afraid to politely engage critics, very unlike a certain US company.

 

[MRC01]

I simulated at low gain mode. The circuit works fine while input voltage is 6Vpp and the volume is max, but the total gain of jazz is lower than 1.

Ah, so in low gain mode, unity gain must be around the 2:00 position of the knob.

 

[amirm]

Dear friends,

the Jazz is one of my best selling amplifiers. It is loved by many and noone ever complained about its performance. Search the internet and you will find quite a few opinions/reviews.

Welcome to the forum Jan. And thank you for detailed explanation. Our spam filter quarantine your post from public view while I was sleeping.

I like to express my appreciation for joining the forum and expressing your point of view. It raises your status in my eyes by a good mile.

 

[MRC01]

Amir, if you had the right connector, it would be nice to see how the Jazz measures with the ground channel properly wired. Would output level, S/N and distortion improve by 6 dB as we suspect? Perhaps something like this, then measure it as you would a standard 4-pin balanced headphone output.
https://www.amazon.com/gp/product/B00KQRN306

 

[trl]

https://6moons.com/audioreviews/meier3/graph_big.gif - there's no mains hum inside Concerto, so like Jan said, a bigger case helps, given that the distance between transformer and audio components and PCB traces increases (also, a smaller metallic case will bounce the EMI field inside more than a bigger case). Too bad there are no more AP measurements in there.

Also, mains hum is pretty low and definitely inaudible in Arietta too: http://borisgermanov.blogspot.com/2014/03/meier-audio-corda-arietta-headphone.html.

 

[Richter Di]

Unfortunately so. What a shame! I now have doubts his new expensive flagship will be all that better.

The Soul sounds amazing. I own one of the prototypes which by the way was elected "best amp of the show" by Headfonia on the CanJam Europe event in Berlin. After additional modifications the SOUL was elected "best amplifier of the show" by the visitors of the Audiovista-show in Krefeld last year.

 

[Richter Di]

I think it's one of each. Amir showed one that had the FF, and the other is mine which does not have it.
In case anyone's wondering, FF is essentially frequency emphasis in the internal gain-feedback loop. The amp boosts frequencies from around 100 Hz to 2 kHz (or attenuates freqs outside this range), so this freq range is exaggerated in the internal gain-feedback loop. Then it does the reverse before the output signal leaves the amp, so frequency response remains flat. The idea is to frequency-shape distortion, minimizing it in the critical hearing frequencies. Whether it works in practice, measurements should be able to indicate.

To my knowledge ff is hard to measure. I did some blind testing with ff versus conventional and I was able to easily pick the ff. It has much more transparency. In my opinion ff is a big step forward.

 

[amirm]

To my knowledge ff is hard to measure.

Why is it hard to measure? And if so, how was the design verified?

 

[MRC01]

If I understand Jan's description of FF (now that Jan is here he can correct me if I'm wrong), the concept is similar to "pre-emphasis" in the CD Redbook standard, but applied in the amp's internal gain-feedback loop to the frequency range most sensitive to human hearing (around 100 Hz to 2 kHz). If so, the difference it makes should be measurable as lower distortion in that frequency range.

PS: you might also measure higher distortion outside that frequency range! The motivation is, our sensitivity to distortion is frequency-dependent. We can't hear 1% distortion at 30 Hz, but we can hear it at 1,000 Hz. Same reason why frequency shaped dither sounds quieter, even when it actually adds more noise. That's another measurement vs. perception subtlely. More noise and distortion doesn't necessarily mean perceptually worse; it could be worse or better, depending on its frequency distribution.

 

[Richter Di]

Why is it hard to measure?

You would need to ask Jan about why it is hard to be measured. I remember some discussions with him at the time he developed the idea and later when he had the A/B comparisons. I guess sometimes it is hard to measure subtle differences. But ask Jan. He is extremely open - yes he is Dutch ;-)

And if so, how was the design verified?

Good question. I guess by though process and later tries. Jan has an extreme sensitive hearing. He can pick up things much faster and more reliable than other people I know. Sometimes its almost frightening. You know that he has an expensive grand piano at home and plays extremly well? He is not just a technical guy but a lover of music and a musician. What a combination.
I can also asure you that the test I did with him regarding ff was double blind. He just gave me a Corda Classic amplifier and a Corda Daccord DAC both with a switch. We implemented it in my gear (Audionet ART G2) and I listend with my headphones at that time (HD 800 and HE 500) to the two and switched between the two switches. He told me nothing and he was not in the room but was chatting with family. After half an hour I returned to them blown away. I could pinpoint on both levels which of the two switch positions sounded better (I heard stuff on my CDs I had not heard before) with each of the stages. And yes, it was ff which I heard for the first time.

 

[MRC01]

If people reliably detect a difference in DBT, you know the difference is real, so it is in principle measurable. But that doesn't necessarily make it easy to measure.

For example, hypothetical scenario: Suppose Mary says, "amp A sounds cleaner than amp B". Joe measures the noise vs. power curves of A & B with the volume knob at a fixed setting, varying the input signal amplitude. The curves look the same, or maybe amp A actually has more noise, so Joe tells Mary she's a victim of expectation bias. But it might be that Mary listens at lower volume levels, and amp A actually has lower noise than B at those lower volume settings (maybe its internal gain-feedback-volume circuit works differently than amp B). Mary heard something real, that was in principle measurable, and what Joe measured was also correct, but Joe gave it an incorrect interpretation due to faulty assumptions.

In this hypothetical scenario what Mary heard is in principle measurable, but to measure it Joe would have to give the amps a fixed input signal and vary the power output with the volume knob. Then he would measure the difference Mary heard. But he didn't know that measuring it differently like this would make a difference.

 

[amirm]

I can also asure you that the test I did with him regarding ff was double blind. He just gave me a Corda Classic amplifier and a Corda Daccord DAC both with a switch. We implemented it in my gear (Audionet ART G2) and I listend with my headphones at that time (HD 800 and HE 500) to the two and switched between the two switches. He told me nothing and he was not in the room but was chatting with family. After half an hour I returned to them blown away. I could pinpoint on both levels which of the two switch positions sounded better (I heard stuff on my CDs I had not heard before) with each of the stages. And yes, it was ff which I heard for the first time.

So I understand, you tested the switch in one position and then another and determined one of the positions sounded better? And what made it blind was that you didn't know what the switch did?

 

[jasonhanjk]

Crosstalk should be worse when using 32 ohm load, since this is a 3 channel configuration.