March Audio P502 Stereo Amplifier Measurements

[PierreV]

The issue is not actually an issue .
The measurement results will depend on how the amp is configured. For example if I had chosen an earthed configuration as opposed to class II this would almost certainly would not be seen.

So, in layman's terms, this wouldn't be an issue with a speaker because it has no power supply and no ground by itself, therefore none of this should flow to it anyway. And, in an earthed configuration, all the nasties would go there. Is that roughly the idea?

 

[pos]

The measurement results will depend on how the amp is configured. For example if I had chosen an earthed configuration as opposed to class II this would almost certainly would not be seen.

Is there an advantage in going class 2 compared to earthed here? Avoid issues with unbalanced connections?

 

[March Audio]

So, in layman's terms, this wouldn't be an issue with a speaker because it has no power supply and no ground by itself, therefore none of this should flow to it anyway. And, in an earthed configuration, all the nasties would go there. Is that roughly the idea?

Not quite that simple but not a million miles off. Re earth it's a common misconception that noise flows into earth and it "disappears". Earth is just a conductor. We just end up completing a circuit through it instead of the signal wire. Can I take this one in detail tomorrow, getting late here

 

[March Audio]

Is there an advantage in going class 2 compared to earthed here? Avoid issues with unbalanced connections?

That's the idea, minimise the potential for those issues.

 

[DonH56]

Trying to minimize how deep into the mud I step, I'll note that ground (return) reference issues are a huge problem in every facet of electronics testing, especially with balanced/differential circuits. There are many "sneak" return paths that can corrupt the signal when just connecting test equipment "normally". What is normal for test equipment is not always (hardly ever) normal for actual use. This is a problem for wide dynamic range (audio) and wideband (RF/mW/mmW) systems and something most of us in the design and test world fight routinely. Mostly cause and effect are well known and solutions implemented without thought. It is easy to run into problems with a new device when connecting just like everything else before figuring out it needs something different. When looking 100+ dB down an unanticipated instrument return ("ground") path can easily corrupt the readings. Ditto when measuring GHz signals (my world) without an adequate RF shield. Both the device and instrument can be working perfectly but do not play well together until certain connections are made in a certain way. This is a test setup problem.

Gross examples many of us have done include (without admitting guilt, of course) grounding a "hot" chassis through an oscilloscope probe, at best blowing a fuse and at worst (normal) destroying (literally) the probe; running the output of a bridged amp into a 'scope or distortion analyzer that grounds one side, thus shorting the amplifier's output, at best shutting it down as it protects itself and at worst blowing the amp into a fireball (don't ask how I know this); carefully using an isolation transformer to "float" a piece of gear for testing so all the equipment including the device under test is happy, but in the process creating a big antenna of the "ground" plane so all sorts of noise is coupled into the measurements; etc. Etc. Etc!

FWIWFM - Don

 

[miero]

Are those Hypex OEM modules used in this amp different from NC400 in this regard? Or they are wired in a different way? Because amir's measurements of NC400 were almost perfect.

 

[March Audio]

Are those Hypex OEM modules used in this amp different from NC400 in this regard? Or they are wired in a different way? Because amir's measurements of NC400 were almost perfect.

Unknown how the NC400 was wired, but if it used a safety ground connection, which IIRC it did, you would almost certainly not see the noise.

I think I will need to draw diagrams tomorrow of what's happening, might make it easier to understand.

 

[AustinsGuitar]

All in all this thread has been kinda sad. People bashing march employee for looking into issues and trying to cover it up. He seems like he is genuine to me. Grounding issues are super minor on this thing, it's not a big red flag kinda like "well thats not super good." Doubt anyone could spot the distortion in listening sessions.

Also amir's review wasn't super glowing, however he didn't say the thing was garbage. It has serious power for a small device and good SNR numbers.

Only thing I am worried about is heat. Can these things really run at 500 watts for long periods of time without an internal fan(or does it have a fan?)? would be interested to know.

 

[pma]

On the contrary, proper wiring, shielding, grounding ... is a key! So, the audio chain has always to be measured as a complete chain, from sound source to amplifier output and even then with speakers connected. No special clinic conditions to be made for a single component. Even then, excellent results without mains and grounding spuriae may be achieved. These wiring and grounding issues often overdub single component's parameters and behavior.

 

[CtheArgie]

@March Audio , I wonder if you had MONO amplifiers, you would (should?) avoid this issue completely. Or am I wrong....?

 

[Samoyed]

I find March to be a straightforward guy. He has endured a number of unjustified attacks.

 

[March Audio]

On the contrary, proper wiring, shielding, grounding ... is a key! So, the audio chain has always to be measured as a complete chain, from sound source to amplifier output and even then with speakers connected. No special clinic conditions to be made for a single component. Even then, excellent results without mains and grounding spuriae may be achieved. These wiring and grounding issues often overdub single component's parameters and behavior.

Yes they are which is why you need to get them correct when measuring.

 

[March Audio]

@March Audio , I wonder if you had MONO amplifiers, you would (should?) avoid this issue completely. Or am I wrong....?

Mono makes no difference

But I need to point out again its not an issue, this is a test configuration problem.

 

[Hemi-Demon]

Mono makes no difference

But I need to point out again its not an issue, this is a test configuration problem.

Interesting slant. So amirm was incorrect in the testing methodology for this product. Did the methodology change versus the other amps tested here, that don't exhibit the same level of noise or channel imbalance?

Or are your comments inferring that there is a ground to earth, standardization issue in the DIY versions of the amp technology?

What's up with the channel imbalance though, is that a testing methodology issue also? I want to buy one, but I would like to clear up the inconsistency, since I don't want to risk shipping back to Aus from the US.

 

[March Audio]

All in all this thread has been kinda sad. People bashing march employee for looking into issues and trying to cover it up. He seems like he is genuine to me. Grounding issues are super minor on this thing, it's not a big red flag kinda like "well thats not super good." Doubt anyone could spot the distortion in listening sessions.

Also amir's review wasn't super glowing, however he didn't say the thing was garbage. It has serious power for a small device and good SNR numbers.

Only thing I am worried about is heat. Can these things really run at 500 watts for long periods of time without an internal fan(or does it have a fan?)? would be interested to know.

The units typically run at 38 deg C after a long soak. The P252 runs at 32 deg and the P122 28 deg. We check this during the 24 hour soak test. So the P502 is warm to the touch but not hot. It's body temperature. The whole case acts as the heat sink so there is more surface area than you might first think to dissapated heat.

The thing to remember with class d is unlike class A/B or worse class A, it is comparatively highly efficient. Class A is about 25%, class A/B maybe 60% whilst class d can be over 90%. This efficiency directly translates into how much power is wasted and turned into heat.

The other thing to consider is that music, as opposed to sine wave test signals, actually has a low RMS value. This can typically be around 1/5th of the peak value. So that means a 500 watt amp that has music peaking at that limit may only have an RMS power output of around 100 watts. So in reality there is again much less power and heat to dissipate.

In a simplistic calculation (there is more to it) 90% efficiency means in this example only 10 watts to dissipate as heat. The class B would be 40 watts and class A 75 watts!

 

[Samoyed]

Interesting slant. So amirm was incorrect in the testing methodology for this product. Did the methodology change versus the other amps tested here, that don't exhibit the same level of noise or channel imbalance?

Or are your comments inferring that there is a ground to earth, standardization issue in the DIY versions of the amp technology?

What's up with the channel imbalance though, is that a testing methodology issue also? I want to buy one, but I would like to clear up the inconsistency, since I don't want to risk shipping back to Aus from the US.

I think you mean “implying” rather than inferring. Anyway, I hope for an explanation in due course.

 

[March Audio]

Interesting slant. So amirm was incorrect in the testing methodology for this product. Did the methodology change versus the other amps tested here, that don't exhibit the same level of noise or channel imbalance?

Or are your comments inferring that there is a ground to earth, standardization issue in the DIY versions of the amp technology?

What's up with the channel imbalance though, is that a testing methodology issue also? I want to buy one, but I would like to clear up the inconsistency, since I don't want to risk shipping back to Aus from the US.

Incorrect is a strong term, its more that an amp with a ground referenced output can run into this problem with the measurement. Amir didn't know it was ground referenced and he is right to be cautious about connecting the screen. Some non ground referenced amps don't like this and can even go bang as a result.

I don't believe the testing methodology has changed, however if an amp uses a safety earth connection those currents that caused mains pick up take a different path. This will reduce and can eliminate the issue seen. In this test configuration they flow in the measurement system signal low wire and through its input amplifier. This can't happen in a speaker. As I showed in previous posts connecting the measurement screen makes those loop currents take a different path and eliminate the issue.

It's worth taking a look at a bunch of the amp measurements and many display a surprising amount of 60Hz pick up.

I think the difference between the channels would certainly reduce and maybe even be eliminated. I will show you Amir's plot with the mains spuria removed later.

There is no ground earth standardisation issue. The Hypex modules are class Ii compliant which means they don't need a safety earth. However the rest of the amp build, wiring etc must confirm to the relevant rules if no safety earth is used. As such most people don't bother and connect a safety earth. That's fine, but there is an issue here also. When using single ended RCA sources where the screen connection is also the low signal conductor, you can get exactly the same issue of ground loops putting mains noise on your signal.

This ground loop problem has been covered several times on ASR and in fact again by PMA today.
https://www.audiosciencereview.com/forum/index.php?threads/ground-loops-101.7162/

https://www.audiosciencereview.com/...measurements-and-solutions.10924/#post-306450

 

[amirm]

@March Audio I don’t understand well enough to contribute much to this thread but is your explanation something that Amir could adjust for and then re-test?

I have to rip apart my XLR leads to do as Alan has done. I wasn't in a mood to do that. I did however try grounding the negative output of the amp and it did have a significant impact on power supply components. So the indications are that this is the issue at hand.

 

[March Audio]

Can I just point this previous test out.

Its of the lower power 252 module. Similar scatter of mains spuria.

and this

or this A/B amp

There are many more.

Its not just my amp seeing the problem, so I think its fairly conclusive its for the reasons explained.

 

[restorer-john]

running the output of a bridged amp into a 'scope or distortion analyzer that grounds one side, thus shorting the amplifier's output, at best shutting it down as it protects itself and at worst blowing the amp into a fireball (don't ask how I know this)

Ah, I now know you've seen the light! That massive spark as you reach across to connect a BNC at the same time you realize it was a bad idea to "test" that BTL amplifier you are now fixing...

My monitor above the test bench had a sign I made until recently "Always remember your common ground"