Alternative SINAD Chart for AVRs and Processors Showing Best Performance Voltage

[amirm]

When you do the next AVR, would you try measuring SINAD with the preouts driving an external power amp? I seem to recall you tried that once with your Purifi and for some reason it didn't work out. May be you will have better luck using an external amp that has unbalanced inputs..

I don't have any other convenient amps to test with. My own amplifiers weigh something like 100 pounds so not moving. The purifi is the only one. I will need to build a better interface for it to test.

 

[peng]

I don't have any other convenient amps to test with. My own amplifiers weigh something like 100 pounds so not moving. The purifi is the only one. I will need to build a better interface for it to test.

So you suspect the issue is related to using the Denon's unbalanced output with the Purifi's balanced input using some sort of interconnect like this:

 

[peng]

I don't have any other convenient amps to test with. My own amplifiers weigh something like 100 pounds so not moving. The purifi is the only one. I will need to build a better interface for it to test.

If our very generous @Dj7675 wouldn't mind sending you his ATI NCore amp that has RCA unbalanced inputs, I will be willing to share the one way shipping cost with him.

I just thought since so many people use their AVRs, especially Yamaha, Denon and Marantz with external power amps it is an important test so we can compare performance of the AVR with internal amps only and with external amps, in objective ways.

 

[Dj7675]

If our very generous @Dj7675 wouldn't mind sending you his ATI NCore amp that has RCA unbalanced inputs, I will be willing to share the one way shipping cost with him.

I just thought since so many people use their AVRs, especially Yamaha, Denon and Marantz with external power amps it is an important test so we can compare performance of the AVR with internal amps only and with external amps, in objective ways.

Not me that has the ATI NCore... but I did just pick up a used March Audio P252 (NC252 based) amp to try out with my X8500. If @amirm would find it useful I would be happy to send it to him before I hook it up. I think an NC252 would be a good one to test with an AVR.

 

[peng]

Not me that has the ATI NCore... but I did just pick up a used March Audio P252 (NC252 based) amp to try out with my X8500. If @amirm would find it useful I would be happy to send it to him before I hook it up. I think an NC252 would be a good one to test with an AVR.

I think that one requires an XLR to RCA adapter again, though the specs say "compatible". If he already had trouble with the Purifi, it may be better to use one that actually has RCA input connectors.

 

[Dj7675]

I think that one requires an XLR to RCA adapter again, though the specs say "compatible". If he already had trouble with the Purifi, it may be better to use one that actually has RCA input connectors.

I believe the connection on the the amps from @March Audio accept both XLR and RCA and don't require an adapter. @March Audio , is this correct?

 

[amirm]

So you suspect the issue is related to using the Denon's unbalanced output with the Purifi's balanced input using some sort of interconnect like this:

I used an all-in-one adapter but yes, that is the problem. I just purchased parts to build my own RCA to XLR converter per Purifi/Hypex recommendations. If all goes well, then I can post those results with the Denon AVR.

 

[superczar]

I have to say that the need for level matching combined with the pre-out voltage variances on consumer electronics has got me flummoxed !

From what I can gather, devices with balanced connectors use a 4 V (3.8V?) rms standard.

AVRs are more like 1.4V and I presume most amps that use a RCA connector have their gain setup to match with that.

But then I guess there are probably no real standards followed around this in the consumer audio segment.
Measured db levels from my speakers were typically 6-7dB lower with a Marantz AVR paired with a Marantz Amp vs the direct output from the AVR

To complicate things further, An amp i recently got (Crown XLS) has a setting for 0.775V sensitivity - what sort of use case would that serve?

 

[Vasr]

From what I can gather, devices with balanced connectors use a 4 V (3.8V?) rms standard.

But then I guess there are probably no real standards followed around this in the consumer audio segment.

The second more than first.

Parasound A23+ amp for example specifies 1V sensitivity unbalanced and 2V (1v per leg) balanced.

From what I understand...

The old school designs (except pro audio equipment) are all in the 1-2v unbalanced region in pre-amp outs and amp sensitivities and twice that for balanced (the same value per leg). Players came out with 2V unbalanced standards so the pre-amps which they were supposed to connect with could accommodate them as inputs. Connecting them directly to amps wasn't part of that equation.

New school designs (or ones with pro audio legacy) target 2-4V sensitivity to achieve lower noise profiles in amplification.

Commendable designs give adjustable input voltage settings on pre-amps and gain adjustments on amps to prevent this chaos.

 

[peng]

The second more than first.

Parasound A23+ amp for example specifies 1V sensitivity unbalanced and 2V (1v per leg) balanced.

From what I understand...

The old school designs (except pro audio equipment) are all in the 1-2v unbalanced region in pre-amp outs and amp sensitivities and twice that for balanced (the same value per leg). Players came out with 2V unbalanced standards so the pre-amps which they were supposed to connect with could accommodate them as inputs. Connecting them directly to amps wasn't part of that equation.

News school designs (or ones with pro audio legacy) target 2-4V sensitivity to achieve lower noise profiles in amplification.

Commendable designs give adjustable input voltage settings on pre-amps and gain adjustments on amps to prevent this chaos.

I know you know it, but for those not familiar with the sensitivity spec, I would like to add a cautionary note that the Parasound A23+'s 1 V input sensitivity spec is the input voltage required for 28.28 V output, not for rated output. That means about 100 W into an 8 ohm load (28.28^2)/8 = 99.97 W.

The complete spec is clear about that, but one overlooked the first line by mistake, one might think incorrectly that 1 V unbalanced would drive the amp to its rated output when it won't.

https://parasound.com/a23+.php

"Input Sensitivity for 28.28 V Out (8 Ω Load)
Unbalanced: 1 V
Balanced: 1 V per leg"

"Power Output – Stereo Mode (0.06% THD)
160 watts x 2, 8 Ω both channels driven
240 watts x 2, 4 Ω both channels driven"

Again, 1 V will drive it to about 100 W. It will take about 1.27 V to drive it the rated 160 W output into an 8 ohm load. Other manufacturers, such as Monolith, Anthem for examples, would specified input sensitivity as the voltage required to drive their amp to their rated output.

 

[Vasr]

I know you know it, but for those not familiar with the sensitivity spec, I would like to add a cautionary note that the Parasound A23+'s 1 V input sensitivity spec is the input voltage required for 28.28 V output, not for rated output. That means about 100 W into an 8 ohm load (28.28^2)/8 = 99.97 W.

The complete spec is clear about that, but one overlooked the first line by mistake, one might think incorrectly that 1 V unbalanced would drive the amp to its rated output when it won't.

https://parasound.com/a23+.php

"Input Sensitivity for 28.28 V Out (8 Ω Load)
Unbalanced: 1 V
Balanced: 1 V per leg"

"Power Output – Stereo Mode (0.06% THD)
160 watts x 2, 8 Ω both channels driven
240 watts x 2, 4 Ω both channels driven"

Again, 1 V will drive it to about 100 W. It will take about 1.27 V to drive it the rated 160 W output into an 8 ohm load. Other manufacturers, such as Monolith, Anthem for examples, would specified input sensitivity as the voltage required to drive their amp to their rated output.

Good catch. I just picked it as an example of old school designs requiring in the range of 1-2v unbalanced as the norm, which it is even for full power.

 

[amirm]

Note that you need higher input voltage to get the peak power the amp is capable of. Fir my testing I routinely go up to 2 volts input to get peak/burst power rating of amps.

 

[pozz]

@Dj7675 @peng @Vasr

From memory you were the most vocal guys on this issue. Adding output voltage across all recorded SINAD figures in the database is pretty much done. So that can be counted as solved.

But... in certain AVRs, if you happen to select the wrong combination of settings, SNR/SINAD inexplicably shrinks.

The Marantz SR6104 showed this issue if you turned off surround channels. The Arcam AV40 showed better performance for surround channels than the front R/L, something minor, while the Arcam AVR850, in a total non-sequitor, progressively delivered better results only if the speaker size was set to "Large" and the sub and all other outs were enabled (yet, weirdly, it delivers the highest pre-out voltage only if speakers are set to "Small"). The performance of Denon AVR-X4700 greatly improved if you turned off the amplifiers. The JBL SDP-55 didn't react well to having an HDMI cable plugged in, regardless of whether or not that input was in use.

So on the one hand fiddling which logically should increase performance reduced it, and, on the other, fiddling helped.

In my mind I can either go with the best combination Amir landed on during testing, or with the most un-fiddled-with and probably the most likely use case. I prefer to go with the latter, regardless of the effects on performance (better in the case of the Marantz, worse in the case of the Denon, ??? in the case of the Arcam). This is of course a judgment call on my part and we can talk over the specifics if anyone notices a number that looks off.

Does that make sense? It would save the trouble of figuring out how to display all of the random special circumstances in circuit design, both intentional (special "modes", settings) and unintentional (incorrect grounding), wherever this comes up, and it's generally easier to read the reviews than to capture all of their findings in the database.

 

[Vasr]

@pozz, I am not sure I understand all of the above.

From a consumer understanding/needs perspective, the simplest, in my mind, is to have a standardized voltage column (whether the device is optimal there or not) and then another column of optimal listing the voltage/SINAD when (and only if) it is measured to be better. Otherwise second column left blank.

The database can have all the data, perhaps the above is the display capability/filter? Possible?

 

[pozz]

@pozz, I am not sure I understand all of the above.

From a consumer understanding/needs perspective, the simplest, in my mind, is to have a standardized voltage column (whether the device is optimal there or not) and then another column of optimal listing the voltage/SINAD when (and only if) it is measured to be better. Otherwise second column left blank.

The database can have all the data, perhaps the above is the display capability/filter? Possible?

Things are both more and less complicated than that. There is no other data available outside of the screenshots you see in the reviews. What I can record is input, output, voltage and SINAD, as shown here:

The design of the interactive tables (not yet updated to include voltage figures) reflects the above. Let me give you an example:

Denon AVR-X4700

• Direct mode, HDMI In/RCA Pre-out: 2.0 Vrms out, 76dB SINAD.

Everything the same but the internal settings changed to turn off the power amplifier subsystem: 97dB SINAD.

JBL SDP-55

• SPDIF In/XLR Out: 3.8 Vrms out, 98dB SINAD.

Everything the same but an HDMI cable is plugged in: 94dB SINAD.

There are a bunch of instances across the tested devices. It is not dependent on voltage but the logic (or absence thereof) of device settings or circuit design.

Typically this isn't a problem. It's easy to record the results of each input/output combo and let them stand for themselves, which is how most audio devices function. The issue is the additional variable of special cases, like device settings, modes, quirks and problems, of which AVRs have more than their fair share.

For me the simplest way out is to ignore the special cases. There are just a handful across ~700 tested devices. If so, the only question that remains is my judgment of the most representative figure for a particular input/output combo. My inclination is to go with whatever's presented as the initial setup and leave out the special case, the assumption being that if a member is interested, they will read the review. So for the Denon: 76dB. For the JBL: 98dB.

 

[Vasr]

Can there be just a flag/asterisk in the data/presentation for those items that have additional tests/numbers/conditions with a general footnote to see review for the marked items?

 

[pozz]

Can there be just a flag/asterisk in the data/presentation for those items that have additional tests/numbers/conditions with a general footnote to see review for the marked items?

Not a bad idea.

 

[peng]

For me the simplest way out is to ignore the special cases. There are just a handful across ~700 tested devices. If so, the only question that remains is my judgment of the most representative figure for a particular input/output combo. My inclination is to go with whatever's presented as the initial setup and leave out the special case, the assumption being that if a member is interested, they will read the review. So for the Denon: 76dB. For the JBL: 98dB.

Obviously it would be very cluttered if all special cases are included, but I would like to see the case of internal amps disconnected included as an extra column, selectable of not.

There is no need to specify if it is done via the amp assign feature, that only disconnects the front left/right channels because the users will, or should know if the unit has full preamp mode or not.

So for the Denon it would be 97 dB amps disconnected 76 dB normal, or just show 76dB/97dB*, * indicate amp disconnected. Special cases such as at reduced voltage and higher than the standard 2V/4V should be ignored in the master data-base, as you said, "they will read the review".

This is a huge time consuming project for you, and will still be timing consuming if you intend to keep it up to date. Thank you very much for doing it.

 

[pozz]

Obviously it would be very cluttered if all special cases are included, but I would like to see the case of internal amps disconnected included as an extra column, selectable of not.

There is no need to specify if it is done via the amp assign feature, that only disconnects the front left/right channels because the users will, or should know if the unit has full preamp mode or not.

So for the Denon it would be 97 dB amps disconnected 76 dB normal, or just show 76dB/97dB*, * indicate amp disconnected. Special cases such as at reduced voltage and higher than the standard 2V/4V should be ignored in the master data-base, as you said, "they will read the review".

This is a huge time consuming project for you, and will still be timing consuming if you intend to keep it up to date. Thank you very much for doing it.

I think it will be easier to finish the work at hand and to consider the special cases after. AVRs are just one of the tested devices after all.

I cannot exclude other voltages (that would exclude too much of the testing that's already been done), although perhaps a voltage filter could be added.

 

[peng]

I cannot exclude other voltages (that would exclude too much of the testing that's already been done), although perhaps a voltage filter could be added.

If you don't mind adding a voltage filter then please do. I assume to keep it simpler, the choices would be something like "standard 2/4V" (understood to be unbalanced/balanced), and "reduced V" without specifying the actual voltage, otherwise it could be 1.2, 1, 1.4 V just too much to be useful. So I filter for reduced voltage and see a much better number, I would just read the review to see it is 1 V or 1.4 V.

Thanks again for even considering..