Raspberry Pi I2S to SPDIF Hat

I will do some testing to see what the voltage drop is tomorrow. We will figure this out.

Cheers,
Michael

 

OK, tested the 3 units I have here and I get less than 10mv of drop between the barrel jack and the Pi board. Note that this is with a Pi 2 and current draw is 1A with some USB devices plugged in. Also note that I had to set my bench top supply to read 5.2V to measure 5.0V at the input to the board. So I would not put too much faith in the accuracy of the bench supply meter. Rather use a real multimeter to measure what you input truly is.

Can you also tell me what the current draw from your bench supply is?

Cheers,
Michael

 

This is the circuit between the barrel jack (and 2-wire terminal) and the Pi. Pretty straight forward. The P-FET has less than 10 miliohm resistance at 4V.



Cheers,
Michael

 

Yes, I always use a DMM to set voltage on the PSUs. The voltage readings I took are straight from my oscilloscope (Rigol DS1054Z) and should not be that far off. At the PSU, they match my DMM (Brymen 869).

Not having the same time resolution as the oscilloscope, I have not seen the Pi+Hat+ethernet+Wifi+USB(independently powered)+BNC use more than 600 mA as I look at my HP, which uses analog meters so I can see more fluctuations than a slower updating digital meter. The current knobs at maximums (both of my bench supplies are rated for 3 A).

Did you look at the voltage drops during boot? That is where my setup had the most trouble and when the RPi3 dies.

Let me test some more here since I am fairly new to the RPi3 GPIO and the oscilloscope. Maybe I made some fundamental error in how I set up my my measurements.

 

SOLVED!

Sorry for the double-post.

Ug... a stupid novice mistake. I forgot about the cable I was using to powering the RPi3. A POS that should have been replaced. Kept forgetting to get proper solder-able banana plugs until recently. "How bad can the wire be?". Answer: REALLY BAD. ALL of the wide swings in voltage was due to the POS cable. Switched them out with 10A-capable silicone-jacketed wire and *POOF* problem is gone. It was a dick-y connection at one of the banana plugs...screwed-in type...

Now I am seeing <100 mV drop from PSU to the RPi and most of that is still due to wire/junctions. Both barrel jack and terminal blocks show about equal drops.


Thanks Michael for the time and effort you put in.

 

That is very good news. A lesson to all of us that there are many seemingly low tech pieces in our setups that still matter! Thanks for sharing with everyone and being so open about it!

Cheers,
Michael

 

Here are the results I spoke about for the 502DAC. Please be aware that testing and measurement are not the final word on sound. Something may not test and/or measure all that well and still be very enjoyable and pleasing sounding. I still believe music can be delivered in many ways and still be powerful enough to stir the heart and soul!

If this is not too much info then I will post the 502DAC3 (which has another set of tests for the HP amp out). Also let me know if these are readable enough.

Cheers,
Michael

24-bit@96khz, 1Khz, 0db:


24-bit@96khz, 1Khz, -5db:


24-bit@96khz, 1Khz, -60db:


24-bit@96khz, 600Hz and 1700Hz, 0db:


[email protected] Jitter Test:


24-Bit@48Khz Jitter Test:

 

Maybe I'm reading it wrong, but if that's the standard Jtest for jitter... it's very bad... it should be a thin 1 pixel spike, not fat like that

 

Actually, you are reading it wrong! The thickness of the line is related to the sampling rate of the analyzer, or the resolution chosen. I forget exactly what it is but what you are seeing is very normal and truly represents represents a single frequency.

Remember this is in the analog [edit] domain [edit] not digital.

By the way, I did not bother showing the digital results since they show a noise floor of -150db which is the limit of the analyzer. So in other words, bit perfect.

Cheers,
Michael

 

Thanks for posting Michael - definitely NOT too much info. Looking forward to the 502DAC3 results also.

 

After thinking about it I realized that Scott Kramer was both wrong and right! The plots are correct and show the proper results, but I was using too low a resolution setting. So I increased it four times and the results look more like what folks expect to see. Mind you it took way more time tor each capture and I had to actually increase the length of some waveform files so the system could capture and settle out! A big benefit was that the low frequency "noise" was an artifact of my low resolution.

I will not re-post them all here, but they are on our web site at:

http://www.pi2design.com/store/p15/502DAC_-_Pro_Audio_Shield.html

Just scroll to the bottom. I also changed the color to bright green and that seems to make them much clearer, to my old eyes anyway!

Now on to the 502DAC3.
Cheers,
Michael

 

502DAC3 results are temporarily located at:

http://www.pi2design.com/private-info.html

Note that at first glance the results might actually seem worse than the 502DAC, with all the crud at the bottom, especially at the lower frequencies. Until that is, you notice the noise floor of -140db. Some Raspberry Pi generated noise now shows up due to the ultra low noise of the DAC and it's circuitry (though it is all well below -120db). The PCM5122 in the 502DAC has a higher noise floor that hides this. It also shows there is some room for improvement. Although I used a low noise bench top supply, it was by no means the lowest. However, it also shows that much of any improvement would be inaudible!

If there are other tests you guys would like let me know.

Cheers,
Michael

 

Thanks again Michael. Is the ETA for the 502DAC3 still September?

And is it planned to run on the HiFiBerry code , like the 502DAC?

Cheers

 

Yes, we are on track for initial production in September. Unlike the 502DAC, the 502DAC3 will run with the Hifiberry Digi+ Pro driver. In this setup the WM8804 is the master and a pair of Pi GPIO's are used to select the clock. This is the same setup as our first board, the 503SPD, but using different GPIO.

Cheers,
Michael

 

Hello Michael,

Sorry for the derail, if this becomes too off-topic, please feel free to move it.

What is progress on your 502MFC board? and expected MSRP?

It looks like it would fit between a Pi and the 502DAC. If it is stackable like that, would it be possible to power all three boards the DC Barrel/Terminal on the 502DAC or would it be better to provide a separate power source for the 502MFC?

I think the MFC board would be a more elegant way of adding an SSD instead of just adding an external 2.5" enclosure. However, I am a bit worried about attempting to power those 1A USB ports if I chose to...need a big transformer or have to go switch-mode.

 

Hi, Michael,

My friend tested this layout in a complete Audio Note system: AN-J speakers, AN DAC 2.1, AN OTO SE. I also gave him the Digione board that I purchased.
By musicality, the 502dac board killed the digione. Digione is a dead audiophile sound, divided into atoms. 502dac correctly arranges accents in music, gives more pleasure from listening. Synergy AN.

What I do not like about this combination is the Raspberry Pi, it's hard to imagine the worst-sounding arm-platform. Very noisy and sensitive to power.
I had to lower the frequency to 700 mhz and reduce the voltage for an acceptable result. Will your board work with the Hummingboard platform or Odroid?

 

Hi Matts,

very interesting to me as I my DAC and Amp are from AN too.

In order to keep noise out of the digital section you might consider to add Allo Isolator between RPi and 502DAC .
Until now I heard of one person comparing DigiOne vs. 502DAC + Allo Isolator – while the gap got closer he still found DigiOne superior, but that wasn't in an Audio Note system.

Best regards
Stefan

 

We are tno sure about what we are doing with the MFC at this time. There is another board that was highlighted by t

I do not know about the Hummingboard, but the Odroid uses a separate connector for I2S, so it will not work.

Cheers,
Michael

 

We are not really sure where we are going with that board right now. Newark electronics came out with a raspberry pi enclosure that included an SSD adapter, case and a power controller, all for $56!

Although their board does not allow stacking, the low price makes it much harder for us to sell ours in the broader market.

We will make a decision within the next few weeks, because we plan to have a campaign by September for the audio boards. And we may keep the MFC alive in order to offer it bundled with a raspberry pi and one of the audio boards.

Cheers,
Michael

 

Hey scott did you tried feeding each board with a different LPS?