• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

March Audio P502 Stereo Amplifier Measurements

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
My lord, is this enough to effectively power Magnepans to 90db 1ft distance at least??
No problem at all powering Maggies. Its a kWatt in a shoe box. Its interesting to note that at the Melbourne HiFi show recently the Oz distributor of Maggies was in the room opposite ours and using class d amps (Nuprime ...spit .. only joking :) ) . Contrary to some beliefs class D has no issues with Maggies.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,498
Location
Monument, CO

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Had a few conversations with Amir and I think I have got to the bottom of the 60Hz issue. Its an interesting one :) !

I have performed a few more tests this morning and I can replicate the issue Amir sees. It does appear to be related to the measurement system screen connection.

So, whilst the amp input is balanced, the output is ground referenced. So the modules speaker out negative is connected to the whole system screen, xlr pin 1 and the chassis. Being powered by an SMPS I have no doubt there are mains leakage currents through Y caps into the ground, chassis and therefore the speaker negative.

In Amirs test set-up my understanding is that the AP input balanced screen was not connected. If I do the same I get a very similar problem.

Excuse my lash up test system.

measurement screen disconnected - note its 50Hz mains here in Oz as opposed to 60Hz in the US.
am4 off.jpg

1579146179845.png



screen connected to speaker negative (system ground reference) - problem gone!
am5 on.jpg

1579146253634.png



Note that there is no mains safety earth connection, the Hypex modules are class II double insulated and amp construction follows relevant guidelines. I went this way to minimise the potential for ground loops from single ended source connections. Note that a mains safety earth connection may reduce the issue but is not a correct and potentially problematic solution. See below.

If I release the measurement screen but connect my scope ground clip (which as you know will be connected to mains safety earth) to the amp ground (in this case the screened cable chassis connection) the problem also disappears. Interesting to see some 8kHz usb packet noise spring up :) This will be due to the screen connection going a torturous route through the mains wiring, back to the PC and probably through the USB screen to the measurement system.

am6.jpg

1579147884661.png


So on the face of it this seems a similar to the problem you can get connecting single ended audio sources into balanced inputs. Its really a measurement anomaly and not an amp fault or design issue. It's really the age old "ground loop" problem. It won't manifest in the speaker.

Hypex talk about this where the screen (in this amplifier audio input instance) should be taken to the far source end and terminated with the balanced low so the ground noise currents dont flow in the signal low. In our example the measurement system is the "amp" and the P502 is the single ended audio source

1579146792694.png


BTW it should be noted that unless you know the amp design you need to be cautious connecting screen to an output. I gather that some class D designs are floating and can go bang! So sensible starting point for Amirs testing.
 
Last edited:

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,498
Location
Monument, CO
NIce job of troubleshooting, Alan!
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,592
Likes
10,728
Location
Prague
@March Audio , your plots are in dBFS, which is a relative number. 0dBFS equals to what rms voltage level? I could find it with my tablet screen.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
@March Audio , your plots are in dBFS, which is a relative number. 0dBFS equals to what rms voltage level? I could find it with my tablet screen.
that would be around 10 volts FS. I wasnt trying to get any meaningful numbers out of it I just wanted to look for the 50 Hz and spuria.

So ref Amirs test at 4.5 volts out we are talking about a FFT noise floor around -128dB
 
Last edited:

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,592
Likes
10,728
Location
Prague
that would be around 10 volts FS. I wasnt trying to get any meaningful numbers out of it I just wanted to look for the 50 Hz and spuria.

OK, understood.
2nd, if the problem was in the meas sys screen connection, why one of the channels in @amirm measurements was significantly better than the other one?
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
OK, understood.
2nd, if the problem was in the meas sys screen connection, why one of the channels in @amirm measurements was significantly better than the other one?
Good question. If you look closely the problem is still there on the other channel. There are many variables that could explain the absolute difference, cable/connection/ pcb trace impedances, board layout, component cable positioning etc.

I have asked Amir to contact the owner and ask him to contact me. We are certainly happy to look at the amp to investigate further if they want. Obviously it will need to come back to base (our expense of course).
 
Last edited:

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,592
Likes
10,728
Location
Prague
2 years ago I measured noise of HYPEX NC400 amplifier kit. Measurements are calibrated, in dBV. Noise RMS output voltage is expressed in dBV(A). For open input, it was 151uV. For shorted input, it was 17.6uV. Both A weighted.

Hypex noise open.png

Input XLR connector open


Hypex noise.png

Input XLR connector 1-2-3 pins shorted


Edit: it seems that my measurement of noise exactly fits to datasheet parameter. They say 23uV unweighted, measurement BW 20kHz. I have 17.6uV A-weighted.

1579160975821.png
 
Last edited:

Matt Smith

Member
Joined
Jan 16, 2020
Messages
5
Likes
2
I am on the fence about buying one of these. But first I would like to find out why there is these problems. @March Audio why don't you share the measurement logs from your database on this particular serial number? If the results match, we will know if it left your factory this way, or it's something that happened later.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
I am on the fence about buying one of these. But first I would like to find out why there is these problems. @March Audio why don't you share the measurement logs from your database on this particular serial number? If the results match, we will know if it left your factory this way, or it's something that happened later.

Hello new member :rolleyes:

Previously I have not kept the detailed measurement data such as FFT plots from the production tests so I cannot refer back to it. However it would have simply shown it to be within normal parameters so it would have left here working normally. As such there is not much real benefit to keeping the data. We know exactly the same either way.

The main issue we have here is a measurement one as explained above.
 
Last edited:

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,592
Likes
10,728
Location
Prague
2 years ago I measured noise of HYPEX NC400 amplifier kit. Measurements are calibrated, in dBV. Noise RMS output voltage is expressed in dBV(A). For open input, it was 151uV. For shorted input, it was 17.6uV. Both A weighted.

Edit: it seems that my measurement of noise exactly fits to datasheet parameter. They say 23uV unweighted, measurement BW 20kHz. I have 17.6uV A-weighted.

According to NC502 datasheet the output noise should be 40uV / 20kHz BW. I believe in Bruno's specs (as to one of very few) and I think both @amirm and @March Audio should be able to provide us with noise measurement to uncover that mains spuriae issue a bit.

1579162786590.png
 

Matt Smith

Member
Joined
Jan 16, 2020
Messages
5
Likes
2
Hello new member :rolleyes:

Previously I have not kept the detailed measurement data from the production tests so I cannot refer back to it. It would have simply shown it to be within normal parameters so it would have left here working normally. As such there is not much real benefit to keeping the data. We know exactly the same either way.

The main issue we have here is a measurement one as explained above.

So you send out all of your amps without even testing them first? Is this industry standard practice? In my 35 years in this hobby this is a first for me. @restorer-john is this typical? Measuring and keeping precise logs before delivering products should be table stakes. But maybe I'm old school.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,592
Likes
10,728
Location
Prague
Not with you. You can see the FFTs above.

You have not provided a standard calibrated noise measurement yet and your explanation of the mains issue in one channel is not trustworthy, to me. To me, it is rather an attempt to find some kind of similar-to-eye plot.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
So you send out all of your amps without even testing them first? Is this industry standard practice? In my 35 years in this hobby this is a first for me. @restorer-john is this typical? Measuring and keeping precise logs before delivering products should be table stakes. But maybe I'm old school.

OK troll you are back again. Yawn. My goodness you need to change your MO. Sign up and then first post start stirring. It is so obvious. Is this the 3rd or 4th time?

I have reported your post. @Thomas savage can you ban this troll, same guy as last time no doubt.

For others benefit all products go through a suite of technical tests and a 24 hour soak test which they pass or fail. Of course they dont get shipped if there are problems.
 
Last edited:

Matt Smith

Member
Joined
Jan 16, 2020
Messages
5
Likes
2
OK troll you are back again. Yawn. My goodness change your MO. Sign up and first post start stirring. It is so obvious.

I have reported your post. @Thomas savage can you ban this troll, same guy as last time no doubt.

For others benefit all products go through a suite of technical tests and a 24 hour soak test which they pass or fail. Of course they dont get shipped if there are problems.

Wow you have problems man. I won't be buying from you!!
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
You have not provided a standard calibrated noise measurement yet and your explanation of the mains issue in one channel is not trustworthy, to me. To me, it is rather an attempt to find some kind of similar-to-eye plot.
Thats utter nonsense and somewhat offensive. It was observable in both channels. Look I know you want to blame everything on your pet theory of op amp "air borne EM sensitivity", and you are a bit miffed that I have challenged your tests in other threads, but this is a simple and logical explanation that makes complete sense. It is demonstrable, as I did above. Do you want a video?

Switch mode PSUs have Y caps for EMI protection that also leak mains current to ground. Blue discs below. Did you consider this when you performed your tests of the NC400? Where will that go and what do you think will be the result without a low impedance screen connection on the measurement system?

y cap.jpg
 
Last edited:

Matt Smith

Member
Joined
Jan 16, 2020
Messages
5
Likes
2
Thats utter nonsense and somewhat offensive. Look I know you want to blame everything on your pet theory of "air borne EM sensitivity" but this is just a simple explanation that makes complete sense and is demonstrable, as I did above.

Switch mode PSUs have Y caps that leak mains current to ground. Did you consider this when you performed your tests? Where will that go without a low impedance screen connection?

View attachment 46050

Wow he's only trying to help!
 
Top Bottom