• 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!

Resolving Measurement Issues with SONCOZ SGD1 DAC

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,394
Location
Seattle Area
About a week ago I landed on a thread on reddit where a few folks were making all kinds of wild accusations about me. I got engaged with one of them and probed into their claims. After a few back and forth, the poster says this:
Reddit post about AudioScienceReview.png


I challenged him to tell me what specific measurements he was talking about and he linked to this thread on measurements of the SONCOS SGD1 DAC: https://www.superbestaudiofriends.o...coz-sgd1-asr-measurement-discrepancies.11115/

As a way of background, I tested the original SGD1 and found it to have the famous ESS IMD Hump. The designer @Ben1987 immediately jumped on the issue and not only solved it in his DAC but also documented it for the world to see and use. I did not get another sample to test so my review was limited to the version I had. So clearly on this front the reddit NFTOxaile is wrong. What I tested is not the commercial samples that others had measured. Still, I wanted to investigate and get to the bottom of the problem.

Sonoz SGD1 Measurements.jpg


Here is my measurements of the pre-IMD-Fix:

index.php


Excellent measurements by any measure. Here is recent measurement posted of new version (I have cropped it for easier viewing; use the above link if you want to see the full snapshot):

Sonoz SGD1 Measurements theirs.png


Let's deal with a simple measurement thing: the FFT shown has no averaging. This causes all that "grass" at the bottom of the FFT. I suggest folks there use averaging of 3 to make it easier to see the spurious tones and more comparable to what I show. I hate for people to think that means there was more noise.

There were some other differences in measurements but none explained the higher harmonic content they were measuring versus my prototype. I went to buy one to test but none were available quickly. Fortunately member @CrimeInPartner had a unit he was not using and rushed it to me. Upon measuring it, I confirmed that the results were not nearly as good as what I had measured in my version:

Sonoz SGD1 Measurements Old Firmware.png


One clue that something was different was the output voltage. We are seeing 4.3 volt versus 4.1 before. Meanwhile I had reached out to the designer, Ben, asking him what was up. He picked up a unit at random from his supply and got much better measurements. Moreover the output voltage was lower, measuring my previous measurement. He said this was a firmware fix that was pushed out.

The SGD1 has a hidden Wifi feature that is used strictly for firmware update. I did not know that the owner had changed the Wifi SSID so spent a day going nowhere trying to upgrade. Once he told me that was the case, I reset the unit and it automatically upgraded itself. Here is the new measurements:

Sonoz SGD1 Measurements Latest Version Firmware.png


As noted, the output voltage is back to where it was and performance while not quite as stellar as it was in my original measurements, it is within shouting distance. And certainly lot better than what others have mentioned.

So to anyone who has a unit they purchased a while back, or regardless, create a mobile hotspot on your phone. Name it ANIG as the SSID/wifi name. Give it the password 12345678. Power cycle the SGD1. If it is all working, it will wait a few seconds and then light up all of its icons. It will then sit there for a minute or two and power cycle. You now have the new firmware and improved performance.

Going back to the poster, it is incredible what stories people make up and how little it takes for them to try to soil other people's reputation. Not only did he have his facts wrong on what I had tested versus what others had tested, he is of course totally wrong about me gaming the measurements to favor companies. As I have said a million times, I have no commercial relationship with any company whose product I am reviewing. If there is ever a conflict of interest, I will be sure to note it and clearly there is none with respect to this DAC. We had a sequence of explainable cause and effects which took a contact with the company to resolve.
 

Haruko

Active Member
Joined
Sep 14, 2021
Messages
105
Likes
116
This unit has wifi ? That's some scary shii bro all I want it to do is transfer signal from digital (usb cable) to analog domain
 
Joined
Apr 20, 2019
Messages
40
Likes
41
@Ben1987 Back when I first got my unit, I used the esptouch app to join my SGD1 to my wireless network, and I know at least at one point it was able to get firmware updates. Apparently that is no longer the case as Amir had to reset the firmware and use his phone as a mobile hotspot to get it updated. When I get it back, will I need to periodically do the hotspot method going forward, or is there a way I can rejoin it to my wireless and have it pull firmware updates successfully like it used to?
 

tmtomh

Major Contributor
Forum Donor
Joined
Aug 14, 2018
Messages
2,636
Likes
7,493
I’m confident in only two things:

1. @amirm ’s measurements are accurate and made in good faith; and

2. the Reddit user who made the accusation against Amir will not show a shred of accountability or retract their claim based on this new information.
 

Langston Holland

Member
Forum Donor
Joined
Aug 5, 2021
Messages
58
Likes
119
Location
Pensacola FL
Let's deal with a simple measurement thing: the FFT shown has no averaging. This causes all that "grass" at the bottom of the FFT. I suggest folks there use averaging of 3 to make it easier to see the spurious tones and more comparable to what I show. I hate for people to think that means there was more noise.

Amir is dead right (what's new?) - the "grass" in the FFT Spectrum plot has little to do with the actual noise floor of the DUT. This apparent noise floor will go up or down depending on number of averages applied and the selected FFT length. The FFT plot will reliably show correlated signal levels (fundamentals, harmonics, etc.). These signals will retain their same peak levels regardless of averaging or FFT length, the "grass" will simply go down as you increase either or both, improving signal visibility.

AP provides (2) other plots that look just like the FFT Spectrum, but show the "grass" as a genuine noise level regardless of averaging or FFT length. These are called Spectral Density plots and they present data in consistent increments of Hz. Interestingly, the correlated signal levels cannot be trusted with these plots because they change with averaging and FFT length.

You need both plot types to accurately observe both signals and noise floor. Fascinating stuff! : )

This is AP's excellent note on the issue if you'd like to know more.

You might need to register for a free account to get at it.

God bless you and your precious family - Langston
 
Last edited:
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,394
Location
Seattle Area
@Ben1987 Back when I first got my unit, I used the esptouch app to join my SGD1 to my wireless network, and I know at least at one point it was able to get firmware updates.
No, no. That should work. It didn't for me because I was setting my wireless access point to SSID of hardwired name, not the one you had set it to. As long as SSIDs/passwords match, it should work.
 

EJ3

Major Contributor
Forum Donor
Joined
Dec 10, 2019
Messages
2,154
Likes
1,663
Location
James Island, SC
[B]tmtomh[/B] gave a couple of good reasons. I had looked to other sites before but found them rather snarky & self-aggrandizing So I don't utilize them. At All.
 
Joined
Apr 20, 2019
Messages
40
Likes
41
No, no. That should work. It didn't for me because I was setting my wireless access point to SSID of hardwired name, not the one you had set it to. As long as SSIDs/passwords match, it should work.
Right but if that was the case, why was the firmware on my unit out of date? I know it was successfully connecting to my wireless, but it apparently wasn't getting firmware updates or you wouldn't have had to beat your head for a day trying to get the firmware updated in the first place.
 

da Choge

Active Member
Forum Donor
Joined
Dec 31, 2018
Messages
231
Likes
220
Location
DFW
I admire that you countered the claims and accusations in that Reddit thread . . . Although sometimes amusing and entertaining, Reddit has to be one of the least credible sources out there (IMO). I like the poster who said something about "alternate facts" -- if there is data that has been truly researched, tested, and scientifically documented, there are no such thing as alternative "facts" -- we live in some crazy times. Bravo, Amir, for putting in the extra work, debunking the naysayers, and maintaining the integrity of ASR !!!
 
Last edited:

phoenixdogfan

Major Contributor
Forum Donor
Joined
Nov 6, 2018
Messages
3,297
Likes
5,075
Location
Nashville
I stopped posting to Reddit a few months back as it became a complete snake pit. I don't even check in on what's going on there because it consists for the most part of stupid memes (which it's members no doubt consider "dank"), goodwill finds, guys posting their dead grandfather's passed on 70's vintage Pioneer receiver they just inherited, and random attacks on objectivist, evidence based audio reviewing.

SBAF? Nothing but trolly trolls doing their trolly thing. No doubt thinking they're "owning the objectivists."
 

Doodski

Grand Contributor
Forum Donor
Joined
Dec 9, 2019
Messages
20,750
Likes
20,760
Location
Canada
I stopped posting to Reddit a few months back as it became a complete snake pit. I don't even check in on what's going on there because it consists for the most part of stupid memes (which it's members no doubt consider "dank"), goodwill finds, guys posting their dead grandfather's passed on 70's vintage Pioneer receiver they just inherited, and random attacks on objectivist, evidence based audio reviewing.

SBAF? Nothing but trolly trolls doing their troll thing.
I don't even know what Reddit is. Never been there. :D Prolly never going there from the sounds of it.
 

GimeDsp

Senior Member
Joined
Oct 30, 2020
Messages
418
Likes
362
Location
Earth
Alternative facts. Lol.
Of all the people that have asked for Amir to chase rabbits to find the hidden performance one finally succeded. I hope he doesn't share his magic tactics!
 

Doodski

Grand Contributor
Forum Donor
Joined
Dec 9, 2019
Messages
20,750
Likes
20,760
Location
Canada
Reddit has a similar SNR to Twitter... i.e. unbearable. :facepalm:



JSmith
All that Twitter, Facebook and like stuff does not interest me. For some years I enjoyed OPing @ the unregulated IRC immensely when that was a thing and obviously enjoy ASR but those others are not my thang.
 
Top Bottom