Review and Measurements of the PS Audio Stellar Gain Cell DAC

[digicidal]

I could go on, but the bottom line is that lab quality results require special measures to be taken to reduce variables such as power irregularities. Why? Because when you're measuring down to as low as -122 dB (which IIRC is one of the better readings that Amir captured in his testing), it doesn't take much to throw readings off.

I agree with that in principle, however with the noise floor of the APx555 still being a good 20-30dB below that level I think there's little problem. The thing to keep an eye out for is when the low-side results become buried... as that would indicate that either the measuring floor has risen or that the testing level has dropped to a point where more accurate devices would be required to provide accurate measurements.

It would be interesting (to me at least) to see if a pure-sine conditioner/generator would make any difference - however, I believe that @amirm stated in the "introduction to my new gear" post that the baseline values were nearly identical to AP's specifications so I very strongly doubt it would matter. That might be an overly simplistic view, but I'm not an EE so this definitely isn't an authoritative response on my part - but it makes sense to me at least. I know I didn't see even the tiniest ripple around 60Hz in those tests (looped AP-AP graphs)... so I think power issues are a null factor.

Edit = I'd say until we have DAC's hitting down in the area between the yellow lines... there's no reason to go "snipe hunting" for that last remaining .0001% of capability.

 

[GrimSurfer]

I agree with that in principle, however with the noise floor of the AP555x still being a good 20-30dB below that level I think there's little problem. The thing to keep an eye out for is when the low-side results become buried... as that would indicate that either the measuring floor has risen or that the testing level has dropped to a point where more accurate devices would be required to provide accurate measurements.

I would be interesting (to me at least) to see if a pure-sine conditioner/generator would make any difference - however, I believe that @amirm stated in the "introduction to my new gear" post that the baseline values were nearly identical to AP's specifications within the nominal margins of any device that sensitive that is. That might be an overly simplistic view, but I'm not an EE so this definitely isn't an authoritative response on my part - but it makes sense to me at least.

Oh I'm not talking about the APx555. I'm sure that has a superb fully regulated power supply to keep noise out of its circuitry.

I'm talking about consumer audio devices that don't cost $28k and have relatively simple power supplies whose performance varies with voltage, current, and line noise. Indeed, one could even go so far as to say that, when powered by a line of unknown quality, devices using an EI transformer (properly oriented and internally shielded of course) may have an unfair advantage over toroid powered devices using the same line power.

Since it is the consumer audio devices being tested (on different days and times of the day), it would be best to manage or eliminate the variables of the power feeding them. Several ways this could be done, but I'm getting ahead of myself.

My point is that without controlled power, it might not necessarily be a fair (or accurate) comparison. Most likely the differences won't be huge, but they could be measurable and, therefore, affect which bin a component goes into for devices on or near the line.

 

[digicidal]

Oh I'm not talking about the APx555. I'm sure that has a superb fully regulated power supply to keep noise out of its circuitry.

I'm talking about consumer audio devices that don't cost $28k and have relatively simple power supplies whose performance varies with voltage, current, and line noise. Indeed, one could even go so far as to say that, when powered by a line of unknown quality, devices using an EI transformer (properly oriented and internally shielded of course) may have an unfair advantage over toroid powered devices using the same line power.

Since it is the consumer audio devices being tested (on different days and times of the day), it would be best to manage or eliminate the variables of the power feeding them. Several ways this could be done, but I'm getting ahead of myself.

My point is that without controlled power, it might not necessarily be a fair (or accurate) comparison. Most likely the differ noes won't be huge, but they could be measurable and dictate which bin a component goes into.

Perhaps, but I have a couple issues with that from an application standpoint:
1) Assuming it did make a meaningful difference - if the device isn't tested exactly as the consumer would use it... isn't that by definition a "broken design"?
2) Assuming it didn't - aren't we giving tacit approval to all manner of snake-oil simply by implication that it might?

Regardless, I think that avenue would have to be proven broadly effective in more than just an edge case or two before I would want to see it implemented across all reviews and devices. It seems to actually add more potential for disagreement than eliminate - to me at least.

"As long as you also have a professional grade power regulation device... this $400 DAC is great... if not it's garbage" - hyperbole of course, but how would you use that kind of result to comment on the actual engineering of the device? They did 80% of an amazing job, and 20% of a shit one?

 

[GrimSurfer]

I never threatened to report you and I did not make a snide comment.

Anything to add on the tech discussion, now that the SNR has gone up a few dB?

 

[GrimSurfer]

Perhaps, but I have a couple issues with that from an application standpoint:
1) Assuming it did make a meaningful difference - if the device isn't tested exactly as the consumer would use it... isn't that by definition a "broken design"?
2) Assuming it didn't - aren't we giving tacit approval to all manner of snake-oil simply by implication that it might?

It's impossible to test it as the consumer would operate it. What about the guy in Louisiana who has his system in a crawdad shack, or the guy who lives way out in the country (with either really clean power because he's miles from anyone or it's 7 am and he's listening to tunes after turning on the milking machine)?

To Amir's credit, he tests things against a manufacturer's specification. You can bet your bottom dollar that the manufacturer is using reasonably clean power.

This isn't about snake oil. It's how the national power grid, local utilities, and people's homes work. There are variations that need to be brought as close to zero in order for Amir to say, with any authority, that device "x" has a SINAD of "y".

Amir knows what he's going. His APx555 is unquestionably good. But he doesn't control the electricity ATM, and this is a real variable managed by real people in real testing labs... it is not a snake oil issue, especially when you're testing down to the levels we read about on ASR.

 

[GrimSurfer]

Nope. Now I will read and learn.

I would suggest that you should count your blessings as they come instead of badgering ppl to give you free lessons.

By all means, ask a question but don't talk about SNR unless you have a specific issue that needs to be addressed.

 

[digicidal]

I'm as guilty as anyone - I actually forgot this thread was originally about the PSA GCD, and not one of the threads about hi-res or psychoacoustics. DOH! I think the last several pages should probably be moved to one of those areas instead. I'll abstain further, despite having interest in the discussion (at least most of it).

Amir knows what he's going. His APx555 is unquestionably good. But he doesn't control the electricity ATM, and this is a real variable managed by real people in real testing labs... it is not a snake oil issue, especially when you're testing down to the levels we read about on ASR.

OK, I lied a little but I'll just edit this one. I partially agree, but in any electronic device where signal precision is a metric of concern... power regulation capable of handling all but the most egregious issues should be the first area of engineering focus... otherwise the rest is just meaningless waste of resources.

 

[GrimSurfer]

I didn’t mention SNR once. Once you asked a substantive question I answered it.

You asked for the SNR to be lifted... then backed yourself away from the discussion. So one must ask whether your intervention was that of a self appointed cop or a passive aggressive troll. It certainly wasn't as an interested party wishing to discuss something of interest/importance.

And, yes, you did talk about noise and SNR (see your below, emphasis added)... so I'm now thinking you might be doing a little trolling yourself...

Yes, he was helping me understand something, in a very honest and polite and expert manner, if you bother to read about the next seven posts up from the link. That’s what intelligent and humble discourse looks like. You might take notes.

1). Because balanced inputs reject common mode NOISE, break ground loops, and increase S/N RATIO by 3 dB, if properly implemented, it is better to test a DAC by the XLR outs, if it has them and if they are properly implemented.

2). No, power conditioning would not be helpful to @amirm . The device should be judged on how it handles the power that it receives, and power conditioning is unlikely to make any difference at all.

 

[amirm]

Let's dial down the back and forth please.

 

[SIY]

Really, most modern active devices have good PSR and bypassing takes care of the rest. I just have never found an actual power line issue in the last 5 places we've lived. I'm sure someone, somewhere may have some issues, but that's the exception and would manifest itself in other things like computers and TVs, so it's not like it would be an audio surprise.

 

[GrimSurfer]

Really, most modern active devices have good PSR and bypassing takes care of the rest. I just have never found an actual power line issue in the last 5 places we've lived. I'm sure someone, somewhere may have some issues, but that's the exception and would manifest itself in other things like computers and TVs, so it's not like it would be an audio surprise.

Yeah, you're probably right @SIY. But Amir's testing is attracting the attention of the industry. At the moment, nobody's interests have been seriously threatened, so industry criticism has been restricted to a few threads by PS Audio.

As this project grows, peoples' noses will get out of joint. When that happens, we'll see greater scrutiny of his test methodology before the "cease and desist" letters start to flow or summons are served.

Amir is managing things very well so far. He's using industry accepted test procedures, has outstanding gear, and demonstrates a measured and balanced approach to his work. But one day his measurements will place the right piece of gear by the wrong company on the another side of the Pantone scale. So best prepare for that day by being super squeaky clean instead of losing the end game because his power supplier (I'm assuming PG&E in the PNW) goofed or his test results couldn't be repeated within a single dB because they happened to occur on the only noisy day of the year.

None of this is binding. Just my 2 cents to raise the SNR of the discussion by weighing in on an issue that had been raised herein.

Now wrt power supplies, I have read a few reviews where it is clear that the devices that Amir is testing are somewhat inadequate. So I think that some of the noise he's seeing (beyond the normal 50/60 Ha and higher harmonics) could be line related. It's a very real possibility.

 

[amirm]

So best prepare for that day by being super squeaky clean instead of losing the end game because his power supplier (I'm assuming PG&E in the PNW) goofed or his test results couldn't be repeated within a single dB.

Any company whose product don't work well with our AC feed and needs a special one to work well, is a bad design. So either way, the outcome is the same.

I do plan to do some testing of this in the future. For now, I subjected one piece of well designed headphone amp to really bad AC and it made zero difference to its measured performance. As to be expected.

 

[GrimSurfer]

Any company whose product don't work well with our AC feed and needs a special one to work well, is a bad design. So either way, the outcome is the same.

I do plan to do some testing of this in the future. For now, I subjected one piece of well designed headphone amp to really bad AC and it made zero difference to its measured performance. As to be expected.

I agree that gear should operate well when supplied by normal house power. But "working well" is different than meeting spec. Lawyers send their kids to Ivy League colleges on the difference, if you take my meaning.

Glad to hear that you're starting to look at this. My intent is not to add to your already busy day. It's to highlight something that might be called into question in the future.

Heck, we've already seen the kooks come out in force after you started reviewing AVRs. They're not the big threat, but there are other, more serious players, who are.

Edit: Just so that ppl tuning in don't think I'm talking out of my ass here, I will provide the basic QoS info that my provider publishes for a city in the NE of 1 million people within a service area containing roughly 3 million people.

Normal Service:
120V / 240V
+4.17% ; -8.33%
Min: 110V / 220V
Max: 125V / 250V

Abnormal Service:
120V / 240V
+5.83% ; -11.67%
Min: 106V / 212V
Max: 127V / 254V

It doesn't take a math wiz to see that the QoS standard is biased towards providing slightly lower, rather than higher, nominal voltage to customers.

No mention of DC or RF interference limits anywhere in QoS standards. They do, however, advise customers of their right to inspect for such if other users advise there is a problem.

The QoS agreement is couched in more legalese than the Magna Carta. Essentially, it says that they absolve themselves of any responsibility and hold customers to account for disturbing the grid.

Anyone familiar with government or utilities will understand that overpromising and under delivering is part of the unwritten business plan (if you disagree with this, Google and see if all of PG&E's customers in NoCal have power ATM). So what you think is clean power may not always be so. Regardless, it's safe to say that it's clean enough for daily household use. But beyond that depends on a whole bunch of issues.

Now power supply caps can, and do, smooth things out quite a bit. So it's not all doom and gloom but something to be mindful of (and, perhaps, research with your service provider) if your work is dependent on clean and consistent power.

 

[RayDunzl]

Any company whose product don't work well with our AC feed and needs a special one to work well, is a bad design.

Maybe take some measurements of your utility power at random (or testing) times, for reference, to demonstrate that it "looks good" or "is challenging".

 

[GrimSurfer]

Maybe take some measurements of your utility power at random (or testing) times, for reference, to demonstrate that it "looks good" or "is challenging".

Handy little tool with data logging capability over periods up to 24 hrs. I'm sure there are a variety of tools with this capability from which to choose.

https://www.homedepot.com/p/Extech-...ta-Logger-DL150/205556939#product_description

 

[617]

These PS Audio reviews are toxic and I think we should examine the way we approach reporting and commenting on measurements of product which are sold to 'enthusiasts'. ASR is not as cultish as other audio corners on the internet as it tends to attract people who read research for fun and sometimes for a living, but the discussion here is nevertheless tainted.

Where is the Glee and i-told-you-so coming from? If, as a community, we had an interest in the objective assessment of audio products, learning that an expensive DAC measures poorly would be the shortest and least interesting thread on the forum. We all know people buy things for reasons other than performance. PS Audio deceiving people about DAC distortion may be obnoxious and even unethical, but it's not worth this level of bickering seen here. Let's all get a grip here.

I like this forum because it is informative and enables me to buy really good audio equipment for very little money, allows me to chat with people who know more than me, allows me to share knowledge I have with people interested in learning. If this forum becomes a crusade against PS Audio or any number of other companies selling crap to insufferable people, count me out.

I was at my brother's place the other day and he made me listen to his little dot tube amp through some pretty unappealing grado headphones. It wasn't hard for me to understand why he liked it, I didn't particularly, but if I acted the way some people do here, I'd rightly be labeled a dick.

I don't know. I feel like some people feel slighted by the audiophile industry, like they've been excluded from it or swindled. If you bought a ps Audio gain cell and are angry about it, sell it to the next mark on audiogon and move on. Yawn at the bad measurements and cheer for the good ones. It's far more difficult to create good audio stuff than bad.

 

[GrimSurfer]

Pretty much agree with everything you're saying here @617 , with the following exceptions...

If your BIL was badgering you to follow his lead based on mysticism or some other form of audio tomfoolery, I would offer that giving him a blunt piece of your mind might be in order.

But where you really lost me was here:

If you bought a ps Audio gain cell and are angry about it, sell it to the next mark on audiogon and move on.

Not sure if I could do that. As much of a d!ck as I can be (and I can be a big one when the situation warrants), I draw the line at turning my bad purchases into other people's' problems.

 

[617]

Pretty much agree with everything you're saying here @617 , with the following exceptions...

If your BIL was badgering you to follow his lead based on mysticism or some other form of audio tomfoolery, I would offer that giving him a blunt piece of your mind might be in order.

But where you really lost me was here:

Not sure if I could do that. As much of a d!ck as I can be (and I can be a big one when the situation warrants), I draw the line at turning my bad purchases into other people's' problems.

For all you know the next buyer was buying it to measure it and slag it on his website. I see what you're saying though. Even if you don't sell it, audio is a hobby where you waste money. It's not like EBT covers phono cartridges.

 

[amirm]

Maybe take some measurements of your utility power at random (or testing) times, for reference, to demonstrate that it "looks good" or "is challenging".

I have done that twice. Once it had a clear glitch in the sine wave. Another time it was very clean.

 

[GrimSurfer]

I have done that twice. Once it had a clear glitch in the sine wave. Another time it was very clean.

Interesting. What was the glitch, if you don't mind my asking? Time of day?