Marantz SR6014 AVR Review

[peng]

I don't think Marantz is doing anything useful except marketing with their HDAMs here. I was not commenting on Marantz' use of discrete component opamps in my post- I was making a general comment to the effect that when used correctly and for the right purposes, there are reasons to use discrete component gain stages. I don't think they are helping sound quality in the Marantz unit here.

Maybe Marantz had a big pile of HDAMs in the warehouse and needed to find places to use them.....

You may be right, as they wouldn't dare use those in their integrated amps anyway.

What is puzzling too is the additional, or yet another seemingly extra buffer opamp added right after the HDAMs (for the FL and FR channels only) in some newer D+M models (two that we know of so far, one being the SR6014 discussed on this thread).

See bigguyca's post for more details:

(2) Denon AVR-X4700 AVR Review (Updated) | Page 81 | Audio Science Review (ASR) Forum

That may be remotely understandable in the case of Denon AVRs that don't have the HDAMs after the volume control IC, but in the case of Marantz, if the HDAMs are doing half of what they claimed (even just as an extra buffer), then what is the point of adding yet another extra buffer, this time an OPA IC so it is not "discrete"?

Last time I checked, even the currently most expensive AVP, the AV8805, and the ex flagship AVR-X8500H don't have that "extra" opamp IC buffer stage at the end of the preamp signal path, that's based on what I saw in the SMs.

 

[audio_tony]

If I understand correctly, HDAM is a Marantz gain module which uses all discrete components.

The advantage of using discrete parts instead of monolithic ICs is that with discrete parts you can have more accurate resistor and capacitor values, closer to the design values; it is also possible to select and closely match the semiconductor devices used. This CAN result in higher linearity prior to the application of negative feedback, and there are some advantages to this. Using large amounts of feedback to overcome some inherent shortcomings of monolithic op-amps isn't without consequence. Can you actually hear the difference? I think it depends on the engineer who is doing the design, and the engineering process and standards that are applied. I've seen some very sloppy designs using ICs in circuits that are designed and built to a price. And I've seen designs with discrete parts that seem to promote the marketing idea that a gain stage with discrete parts is ALWAYS better than one using ICs.

Actually this is incorrect.

A monolithic IC like an opamp will have perfectly matched transistors, because they are all on the same substrate. The same goes for any on chip resistors or capacitors incorporated into the opamp design - they are inherently precisely matched.

As for discrete components, 1% (and better) tolerance resistors are readily available, and then of course there are Dale and other precision resistors, as well as precision capacitors available. A preamp built around a monolithic opamp IC with precision passive components will perform better than any discrete opamp built using 'matched' transistors.

The HDAM circuit is nothing more than a discrete opamp - and likewise, those too can not perform better than a monolithic IC opamp.

 

[peng]

Actually this is incorrect.

A monolithic IC like an opamp will have perfectly matched transistors, because they are all on the same substrate. The same goes for any on chip resistors or capacitors incorporated into the opamp design - they are inherently precisely matched.

As for discrete components, 1% (and better) tolerance resistors are readily available, and then of course there are Dale and other precision resistors, as well as precision capacitors available. A preamp built around a monolithic opamp IC with precision passive components will perform better than any discrete opamp built using 'matched' transistors.

The HDAM circuit is nothing more than a discrete opamp - and likewise, those too can not perform better than a monolithic IC opamp.

How about if you compare a theoretically best designed/built discrete opamp such as using hand matched parts, and an average quality opamp IC such as the NJM8080 found in the D+M AVRs/AVPs? I asked because as someone pointed out the 8080's slew rated, while I would say more than adequate for the handling preamp output level up to only a few volts, isn't that great.

 

[audio_tony]

How about if you compare a theoretically best designed/built discrete opamp such as using hand matched parts, and an average quality opamp IC such as the NJM8080 found in the D+M AVRs/AVPs? I asked because as someone pointed out the 8080's slew rated, while I would say more than adequate for the handling preamp output level up to only a few volts, isn't that great.

The NJM8080 has a slew rate of 5v/µs. To produce a full-amplitude 20 kHz sine wave you only need 2.1V/µs

Given that it is highly unlikely to get a full amplitude signal at such a high frequency in typical music, the slew rate is not an issue.

 

[bigguyca]

I wish we were provided w/ 2 Conclusions for these tests, one for high level volumes [like we’re currently given] and one for normal listening levels like the majority of us use. For instance, this 6014 that receives a “Cant Recommend” based on reference level results has basically the same SINAD as Denon 3700 & 4700 at normal volumes. We were provided a chart to see what the 6014’s SINAD looks like at normal volumes [see below], so we can at least deduce for ourselves, but many of the other AVR tests didn’t offer such graph, so we have no idea how they actually perform at normal levels.

View attachment 137777
So as you can see above, the 6014 is being judged on a volume of 82.5 on the knob, which would destroy most of our ears for those of us who don’t have large theaters. In my living room I couldn’t get the knob above 72, and in my 12x14 home theater I’m in the mid to high 60’s at max listening volume, which is LOUD…. and mid 60’s to 73.5 is where the 6014 performs the best, which you can see on the chart [96-97 SINAD].

Compare it to the Denon 3600:
View attachment 137779
We’re getting 103 SINAD… at a volume of 85, which basically none of us will ever touch. Scale back to normal levels and we’re at the same SINAD as the 6014. Incidentally, I use an Outlaw 2200 for my center which has a sensitivity of 1.7 and my center has never come close to clipping.

With the way in which these AVRs are bashed, one might think they’re nothing but distortion machines, but I had a friend over who has a simple entry level HT and I sampled my 6014/Dynaudio setup for him… and he was thoroughly impressed by the level of clarity.

The data isn't available from the graphs to draw the conclusions you are making

At .74V on the chart the input to the DAC circuitry in the SR6014 is -9 dBFS. (73.5dB - 82.5dB) At -9dBFS the distortion of the DAC IC is lower than with an input of 0dBFS. At 82.5dB the input to the DAC IC is 0dBFS and the distortion is higher.

If the volume control were set at 73.5dB then the input to the DAC circuitry would be 0dBFS at 73.5dB. The AKM DAC IC used in these products has higher distortion at 0dBFS than at -9dBFS so the distortion would likely be higher with the volume control set at 73.5dB.

Since the volume control uses a resistor-based attenuator excess noise would be added at -9dB attenuation. Distortion from the DAC IC likely predominates however. For reference, at 82.5dB the volume control has a gain of 0dB and no excess noise is added.

The graph for the X6700H clearly shows this effect, which is explained by Amir in the note below the graph. Adjusting the volume control for 2.5dB of gain (85.0 - 82.5) moves the 0dBFS input point the 85.0dB. This lowers the distortion at 82.5dB since the DAC IC now has an input of -2.5dBFS at an indicated volume of 82.5. At -2.5dBFS the DAC IC has significantly lower distortion.

The above doesn't mean that the distortion is poor at any particular level, it just means that further measurements are required.

-------------

When used outside of direct or pure direct, then added factors come into play due to digital processing. The actual distortion at a particular volume setting becomes a more complex subject.

At least Marantz equipment, and likely Denon as well, appear to use a combination of digital and analog volume control. The reason for this mix is obvious from the above. For example,

Let's say the right channel has an offset of -4dB after calibration. Using the volume control to provide this offset means that the input to the DAC IC is 0dBFS at maximum output from the digital source and distortion will be that of the DAC IC at 0dBFS input. The volume control will be used to provide the -4dB output adjustment which adds a little noise. This is an internal setting of the volume control and not seen by the user.

Alternatively if the volume offset is provided digitally, then the input to the DAC portion of the DAC IC will be -4dBFS. The output of the DAC IC will contain a bit more noise, but significantly less distortion. The volume control will now provide no output adjustment and no excess noise will be added by the volume control.

A mixture of digital and analog volume adjustment is likely provided for different combinations of settings of the actual volume control as seen by the user and channel offsets.

 

[oupee]

The NJM8080 has a slew rate of 5v/µs. To produce a full-amplitude 20 kHz sine wave you only need 2.1V/µs

Given that it is highly unlikely to get a full amplitude signal at such a high frequency in typical music, the slew rate is not an issue.

Slew rate
https://www.analog.com/en/technical-articles/op-amps-process-large-signals-with-low-distortion.html

 

[peng]

Slew rate
https://www.analog.com/en/technical-articles/op-amps-process-large-signals-with-low-distortion.html

That article is about applications in communications systems, not audio pre amplifiers. Audio tony referred to 20 kHz, not the hundreds of MHz, or even 50 MHz.

What they said in the conclusion:

"Conclusion
The ultrafast slew rate and high bandwidth allow the LT1818 and LT1819 op amps to process large signals at high frequencies with low distortion. Combined with the low noise and moderate supply current, these amplifiers are a good choice for receivers, filters, or drivers of cables and ADCs in high-speed communication or data acquisition systems."

 

[audio_tony]

Slew rate
https://www.analog.com/en/technical-articles/op-amps-process-large-signals-with-low-distortion.html

You will note that in the article, the frequency band of interest is 100kHz to 40MHz - this is not relevant to audio.

These opamps are intended for Video and RF Amplification, Communication Receivers and Data Acquisition Systems etc. but not audio.

As I said previously, very high slew rates are simply not required for small signal audio applications.

 

[oupee]

I put it here for square wave measurements. It is a very telling method. There are practical pictures so the layman will understand. The sound is not a simple sine and the slawe rate should be oversized due to detail and transient.

 

[audio_tony]

I put it here for square wave measurements. It is a very telling method. There are practical pictures so the layman will understand. The sound is not a simple sine and the slawe rate should be oversized due to detail and transient.

The reproduction of square waves requires infinite bandwidth.

A good audio system does not need to reproduce a pure square wave to sound accurate or even perform well.

There isn't a great deal of high frequency energy even in Hires recordings - there can be high frequencies (harmonics) but the amplitude is that low that a very high slew rate is not required to reproduce these harmonics.

A typical 100w speaker will distribute around 10w to 15w to the tweeter with a 100w input - this proves that the HF energy in a typical audio signal is roughly 10% of the total amplitude.

Like I stated earlier - for an opamp (running on +/-15v supplies) to reproduce a full voltage swing (about 8.5v RMS) at 20kHz requires a slew rate of only 2.1V/uS

Very few (if any?) power amps need this level of drive - most requiring about 3v RMS so approximately 1/3rd of maximum drive from the opamp.

Divide that by roughly 10% and suddenly only 0.3v (300mV) is required for maximum drive at higher frequencies.

Under these conditions, slew rate will never be an issue.

 

[peng]

The reproduction of square waves requires infinite bandwidth.

A good audio system does not need to reproduce a pure square wave to sound accurate or even perform well.

There isn't a great deal of high frequency energy even in Hires recordings - there can be high frequencies (harmonics) but the amplitude is that low that a very high slew rate is not required to reproduce these harmonics.

A typical 100w speaker will distribute around 10w to 15w to the tweeter with a 100w input - this proves that the HF energy in a typical audio signal is roughly 10% of the total amplitude.

Like I stated earlier - for an opamp (running on +/-15v supplies) to reproduce a full voltage swing (about 8.5v RMS) at 20kHz requires a slew rate of only 2.1V/uS

Very few (if any?) power amps need this level of drive - most requiring about 3v RMS so approximately 1/3rd of maximum drive from the opamp.

Divide that by roughly 10% and suddenly only 0.3v (300mV) is required for maximum drive at higher frequencies.

Under these conditions, slew rate will never be an issue.

Thank you very much, I don't feel being the lone voice on this any more. I wish more people understand more about Fourier too when they talk about square wave..

 

[hipis_1]

So is SR61014/6015 worth to buy? Music 2.0&5.1 and movies in typical living room.

 

[SynthesisCinema]

So is SR61014/6015 worth to buy? Music 2.0&5.1 and movies in typical living room.

In that price range it should be top contender with Denon X3700H, but more competition is arriving soon! Onkyo TX-RZ50 with Dirac Live, Yamaha RX-A4A (7.2) and Pioneer VSX-LX505 with Dirac.

Manufacturer refurb unit with 3year warranty 1350$. New one 1700$ on Amazon.
https://www.accessories4less.com/ma...-ch-x-110-watts-a/v-8k-receiver-w/heos/1.html

There is some X3700H and SR6014 for 999-1100$ but only 1year warranty which would feel too risky.

 

[Todd74]

In that price range it should be top contender with Denon X3700H, but more competition is arriving soon! Onkyo TX-RZ50 with Dirac Live, Yamaha RX-A4A (7.2) and Pioneer VSX-LX505 with Dirac.

Manufacturer refurb unit with 3year warranty 1350$. New one 1700$ on Amazon.
https://www.accessories4less.com/ma...-ch-x-110-watts-a/v-8k-receiver-w/heos/1.html

There is some X3700H and SR6014 for 999-1100$ but only 1year warranty which would feel too risky.

You can buy the 6014 for $999 at A4Less and add on an extra 5 yrs for $78… or 2 extra for $40. Might even be able to get Jeff to sell it for $900 and then tack on the extra warranty.

 

[bigguyca]

The NJM8080 has a slew rate of 5v/µs. To produce a full-amplitude 20 kHz sine wave you only need 2.1V/µs

Given that it is highly unlikely to get a full amplitude signal at such a high frequency in typical music, the slew rate is not an issue.

What is the NJM8080 slew rate in actual use?

The NJM8080 datasheet lists its performance, assuming I'm reading the datasheet correctly, at a gain of 20dB. This means this means the the NJM8080 likely does not have internal compensation or at least not a high level of compensation.

An external compensation capacitor to allow use of the NJM8080 down to a gain of 0dB, which is how it is how the NJM8080 is commonly used, would seem likely to reduce the slew rate given that the capacitor required current to be charged, which takes time. Compensation capacitors are one of the major limitations to slew rate.

Does anyone have actual measurements of the NJM8080 under more typical use cases such as 0dB gain?

Certainly the 20kHz distortion rises rapidly with output as shown on the datasheet. Is this rise due to slew rate limiting even at the 20dB gain used for the datasheet measurements?

 

[audio_tony]

The NJM8080 datasheet lists its performance, assuming I'm reading the datasheet correctly, at a gain of 20dB. This means this means the the NJM8080 likely does not have internal compensation or at least not a high level of compensation.

An external compensation capacitor to allow use of the NJM8080 down to a gain of 0dB, which is how it is how the NJM8080 is commonly used, would seem likely to reduce the slew rate given that the capacitor required current to be charged, which takes time. Compensation capacitors are one of the major limitations to slew rate.

The performance graphs are quoted at AV=20dB and AV=40dB.

The package doesn't have any compensation pins, however there is no mention of unity gain in the datasheet.

Does this mean it's internally compensated? Who knows!

 

[Todd74]

Are the pre-outs on Marantz/Denon digital? Which type of interconnects should I be using, digital coax or RCA audio? My amps are Parasound 2125v2, Class AB with “unbalanced stereo RCA inputs with loop output”.

Monoprice sells 2 different versions of the same line of cable:

Monolith Digital Audio Coax 1M for $12.34
https://www.monoprice.com/product?p_id=34444

Monolith RCA Audio 1M for $16.69
https://www.monoprice.com/product?p_id=18534

 

[Mr Wolf]

Are the pre-outs on Marantz/Denon digital?

All pre-outs are analogue and those found on D&M AVRs are all unbalanced RCA connectors.

 

[Todd74]

All pre-outs are analogue and those found on D&M AVRs are all unbalanced RCA connectors.

Thanks. So then get the RCA audio cables?

What about for subs?

 

[Mr Wolf]

Thanks. So then get the RCA audio cables?

What about for subs?

Subs use a single RCA cable. To save money you can buy the speaker channel RCA cables in pairs.