Review and Measurements of Meier Corda Jazz Amp

[Richter Di]

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?

Yes. Blind 1 was that Jan had not told me about ff or what it could be. He just said here is something would you mind trying it out. So I knew nothing. As a matter of fact, it could have been that Jan wanted to show me how easily someone can hear difference alltough a switch does nothing. With Jan, you never know. The Blind 2 was that I sat alone in my room with no one possibly influencing me. So I could not look to Jan and say something. A kind of Blind 3 was that DAC and Amp had this switch and I tried to figure out if both really had „this whatever“ and try fo figure out in case they had it in which direction the switch had to go. Again with Jan it was super possible that one switch would work in one direction and the other in the other direction. So ff is audible if your ears are able to detect it.
This reminds me about the measurable hiss of the Audeze Mobius where around 50% can not hear it. In this case you can measure the hiss coming from the DAC.

 

[Nango]

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

If you can't measure it, most likely it is snake oil.

 

[Nango]

Yes. Blind 1 was that Jan had not told me about ff or what it could be. He just said here is something would you mind trying it out. So I knew nothing. As a matter of fact, it could have been that Jan wanted to show me how easily someone can hear difference alltough a switch does nothing. With Jan, you never know. The Blind 2 was that I sat alone in my room with no one possibly influencing me. So I could not look to Jan and say something. A kind of Blind 3 was that DAC and Amp had this switch and I tried to figure out if both really had „this whatever“ and try fo figure out in case they had it in which direction the switch had to go. Again with Jan it was super possible that one switch would work in one direction and the other in the other direction. So ff is audible if your ears are able to detect it.
This reminds me about the measurable hiss of the Audeze Mobius where around 50% can not hear it. In this case you can measure the hiss coming from the DAC.

If you know what you do, you don't need blind testing.

 

[agjell]

"They didn't even get the CE logo right."

With audio equipment the manufacturer has to guarantee that the devices he sells do fullfill the CE-requirements. And the JAZZ does, please be assured of that! The fact that I may not have printed the logo exactly according to design ....

" this amp is required by law to be CE-certified"

No, this is only required for very specific products like medical implants. With low voltage audio-products the manufacturer is responsible for fullfilling the CE-requirements. A certification is not mandatory (and never done by any of the smaller manufacturers, as otherwise you no longer could afford these products).

The JAZZ is very safe. I know well of the danger of electricity to the human body. By origin I'm a biomedical physicist and my PhD thesis was on the subject of neurostimulation. I worked 10 years as research project leader with a pacemaker company.

The CE-requirements are very strict. Do you ever wonder why certain companies from the US or Asia do not sell their headamps through a distributor in Europe?

This was pointed out to me in another thread, in which I corrected my post. I forgot to correct my post in this thread, though, sorry. Regardless, the amp needs to be CE-compliant, which you say it is. But please, use the correct logo. This IS the sort of attention to detail that matters. How else are consumers supposed to tell a compliant product from a non-compliant one? With all the non-compliant phone chargers I've had crapping out or burning up I dare no longer use non-compliant electronics.

 

[solderdude]

The used logo looks like 'China Export' logo.

 

[agjell]

The used logo looks like 'China Export' logo.

This has already been discussed. You're late to the party

 

[solderdude]

It was just added pictorial support for your case, for when Jan is watching this.

 

[pika02]

China Export LOL. A funny mistake but it's really from China.
Indeed, Shanling is a top-class audio equipment manufacturer in China. They also do OEM or even ODM for some hi-end CDPs as I know.

 

[MRC01]

If you can't measure it, most likely it is snake oil.

I would rephrase that as, if it cannot be measured, most likely it is snake oil.

The difference: just because you can't measure it, doesn't mean it can't be measured. And just because it is in principle measurable, doesn't mean you can measure it. It could be that you don't understand the difference well enough to know what to measure, how to measure it properly, or how to interpret the measurement. I gave a simple example a few posts back. This entire thread serves as another example. Of course the real world can be more subtle and complex.

 

[SIY]

If you know what you do, you don't need blind testing.

Please tell me this is parody.

 

[invaderzim]

....This entire thread serves as another example. Of course the real world can be more subtle and complex.

If Jan's post makes as much sense to those in the know as it does to me then this is a great example of measurement vs real world use.
If most of the bad stuff that showed in the tests would only be noticeable at volume levels that would destroy your hearing then are they really problems at all?

 

[MRC01]

This discussion has revealed 2 ways the measurement techniques that Amir uses, which are appropriate for most amps, are not applicable to this amp, due to its design differences:

1. Not connecting the ground signal. This made output power and distortion look 6 dB worse than it really is (half the voltage, 1/4 the power, and worse S/N).

2. Measuring distortion vs. power at max or high volume. This is where most amps measure best, but since the Jazz's volume control changes the internal gain-feedback loop instead of attenuating a fixed gain like conventional amps do, it should measure better (wider bandwidth, lower noise) at lower volume settings.

For example, if you simply add 6 dB to the distortion vs. power curves that Amir already measured, the Jazz becomes comparable to the JDS Element. It might actually be better (cleaner) with the volume at a lower setting, which is where most people actually listen. We could measure and find out!

 

[agjell]

If Jan's post makes as much sense to those in the know as it does to me then this is a great example of measurement vs real world use.
If most of the bad stuff that showed in the tests would only be noticeable at volume levels that would destroy your hearing then are they really problems at all?

A very valid point. But isn't objective measurements what separates this particular forum from many other forums? My impression is that the members here come to discuss the objective performance of a product in addition to the subjective performance. Objective performance also says something about engineering skill, and is a proof of what is achievable within certain budgets. Other aspects matters as well, of course, like availability, warranty, support and robustness. All else being equal, I will still pick the unit that measures better. Paying the same for something that measures worse just doesn't make sense to me.

 

[MRC01]

Measurements are valuable, but sometimes people take an oversimplified view. Its value depends on what you're measuring, how you measure it, and how the measurement correlates to real-world usage and performance. If one doesn't consider this, he can be easily misled by measurements. There are many examples, this thread among them, so I'll mention only a few simple ones.

Amp A could measure lower noise & distortion than B. So B is worse, right? If the frequency distribution of A and B's N&D is different, B could sound perceptually cleaner. Our ears are more sensitive to noise & distortion in the midrange, than in the extreme low or high frequencies. And, some kinds of distortion are more dissonant than others. So the N&D measurement can be misleading if you don't also look at its frequency spectrum.

Amp A could measure lower noise and equal distortion, to amp B. That's better, right? Yet B might sound better if its slightly higher noise masks distortion, and is perceptually less dissonant. All else equal, under some conditions (of course not always), reducing noise can make it sound worse. I'd rather hear a low-level fuzz than spurious intermodulation tones!

So we must make and interpret measurements in light of perceptual relevance. Just because amp A has better measurement X doesn't necessarily mean it is perceptually more transparent under actual listening conditions. That depends on what X is, how it was measured, how we perceive it, and how it interacts with other measurements Y and Z.

 

[Soniclife]

Not connecting the ground signal. This made output power and distortion look 6 dB worse than it really is (half the voltage, 1/4 the power, and worse S/N).

Does the ground not have to be connected for power to flow?

 

[KSTR]

2. Measuring distortion vs. power at max or high volume. This is where most amps measure best, but since the Jazz's volume control changes the internal gain-feedback loop instead of attenuating a fixed gain like conventional amps do, it should measure better (wider bandwidth, lower noise) at lower volume settings.

Please keep in mind that this is only relevant when the noise gain of the circuit is larger than 1. If @pika02's schematic is representative I see a inverting input amp with 30k Rin and 100k(max) Rf giving a max signal gain of 3.3x (with R_src=0), Rf probably being a 100k linear or reverse log pot. The signal gain of this is much lower than 1 at low volume settings, but the noise gain which is relevant for noise and distortion is only decreasing from 4.3 to 1. Assuming a linear pot, at 1/3rd rotation (33k) where the noise gain has reached 2 (and signal gain is unity) the improvement downwards from there tops out at 2x as noise gain is always >=1.
That is, at very low volume settings the benefit of this pot-in-feedback-scheme does not scale. Of course it remains a valid approach, within the usual set of compromises as there are drawbacks, too.

 

[solderdude]

Rf appears to be resistor 'ladder ?' type, possibly using FET-switches inside the square (programmable ?) chip to switch some resistors in parallel combinations as seen in the pic below.

The 'limitation' of this method is shown in the volpot linearity plots which shows a relatively small adjustment range of just over 40dB.
This means the smallest (total) R feeback resistance is 676 Ohm. The result is that the max. output voltage of the amplification stage is reduced depending on the volpot setting. When one increases the input voltage of the amp the max output voltage will increase so max. output power is input voltage dependent as well.
The gain of the 'power stage' is merely 1.6x = 2dB (acc. to what I calculated)

To really characterize this amp one would have to measure FR and distortion at different volpot settings and input voltages belonging to the 'volpot setting' and playing with input voltage levels to determine the max undistorted output power for that setting.

 

[Dro]

50 mV SNR actually looks pretty good with 90 dB. The Monotor drops far from max level to 50 mV, the Jazz actually gets a bit better.

Looking at the specs of the ADI-2 DAC IEM out, SNR there is 115 dB at 3 dBu output, which is a noise level of roughly -120 dBV. 50 mV is -26 dBV, so assuming noise is constant and not modulated by the signal, the ADI-2 DAC would have 94 dB SNR at 50 mV.

 

[MRC01]

Does the ground not have to be connected for power to flow?

Of course a ground must be connected for power to flow. The Jazz amp carries an inverted signal on its ground channel, like a balanced amp. So even though the Jazz has an unbalanced 1/4" headphone output, it must be measured as you would a balanced output. So the question is, which ground did the AP analyzer use when taking the measurements? I think Amir is looking into this.

 

[solderdude]

The ground of the signal generator should not be connected to the ground of the measuring input.

When these are electrically connected you short the 'active ground' and at the same time don't measure the correct output as the real output signal is between the 'active ground' and output.
The output opposite the input ground thus contains the amplified signal + inverse of the 'active ground' signal, which b.t.w. is combined with the other channel. The 'active ground' thus is a L+R combined (and inverse) signal and has a certain ratio, so not a 1:1 ratio as in a balanced connection.

The input thus is single ended and the output is also single ended but the output 'ground' is not and must not be connected to the input ground.
Indeed one way of measuring would be to use the + and - input of the AP but the ground pin should not be connected to the input ground as the output signal is not truly balanced.