E1DA Cosmos tests

[hayabusaxps]

and the unit is on an isolator so its supposed to be equivalent to a battery connection

https://www.digikey.com/en/product-highlight/t/texas-instruments/isousb211-high-full-low-speed-isolated-usb-repeater

ISOUSB211 High/Full/Low-Speed Isolated USB Repeater​

Texas Instruments' galvanically-isolated repeater supports 1.5 Mbps low-speed, 12 Mbps full-speed, and 480 Mbps high-speed signaling rates​

ill do a 2nd test with tablet when charged just in case

 

[Music1969]

In REW distortion settings, you can specific to the distortions calcs to be 20Hz to 20kHz

Maybe your THD+N is being calculated for full measurement bandwidth ?

 

[hayabusaxps]

i had the 20hz high pass and 20khz lowpass turned on

In REW distortion settings, you can specific to the distortions calcs to be 20Hz to 20kHz

Maybe your THD+N is being calculated for full measurement bandwidth ?

 

[Music1969]

i had the 20hz high pass and 20khz lowpass turned on

View attachment 345888

I mean the OP @horias2000

 

[horias2000]

I mean the OP @horias2000

I tried both with and without the 20Hz and 20kHz filters. There is not much difference. As for you measurements, as someone else mention before, you might want to supply the Cosmos from from a battery pack. I had similar issues when powering from a mains brick. At the moment I'm busy with a speaker project and I left the Cosmos for a bit. I will come back to it. I will try to get the THD+N reading lower but for the moment it is OK as it is as I'm most interested in the graph and seeing the HD components.

 

[hayabusaxps]

* change your FFT length to 256k instead of 1M
* change your window to Blackman-Harris 7
* change your averages to 2 or none
* select “use 64 bit FFT”

and cosmos runoff a tablet not connected to the wall. and remote desktop controlled.
best i can do as a battery pack haha. there's no 60hz in the response anyway for noise.

also set the dips to 10v as cosmos instructed


this should have been the best settings, 2channel 48khz sampling 64-bit fft, and everything listed above.
but in stereo. cosmos volume at zero.

32bit 192khz sampling on cosmos

32bit 192khz , regular non 64bit sampling

no changes on the -126 dBFS


lastly added the scope and mono

im stuck at -126dBa FS

and this is supposed to be better than 129 correct.... on a grade 0

 

[BKDad]

View attachment 345855


and the unit is on an isolator so its supposed to be equivalent to a battery connection

https://www.digikey.com/en/product-highlight/t/texas-instruments/isousb211-high-full-low-speed-isolated-usb-repeater

ISOUSB211 High/Full/Low-Speed Isolated USB Repeater​

Texas Instruments' galvanically-isolated repeater supports 1.5 Mbps low-speed, 12 Mbps full-speed, and 480 Mbps high-speed signaling rates​

View attachment 345857
ill do a 2nd test with tablet when charged just in case

I think this is the isolation device Ivan added to the Cosmos ADCiso version.

I've used an external isolator on my old Cosmos ADC that didn't have an isolator. It made a substantial difference.

With either an external isolator or the one built into the ADC iso, I get identical results from using a "linear" wall wart AC powered supply as with a battery supply. But, every situation has the potential to be different.

 

[BKDad]

There's some subtleties associated with using a Topping and probably other DACs with either REW or Multitone. It has nothing to do with the software itself, both of which work really well, but with how the DACs use distortion correction.

Some very helpful forum members pointed me in the right direction to get that part right:

Using A D50 with an ADCiso

 

[hayabusaxps]

I think this is the isolation device Ivan added to the Cosmos ADCiso version.

I've used an external isolator on my old Cosmos ADC that didn't have an isolator. It made a substantial difference.

With either an external isolator or the one built into the ADC iso, I get identical results from using a "linear" wall wart AC powered supply as with a battery supply. But, every situation has the potential to be different.

he uses a different device, (different brand and model)
but same exact function/circuit

"from when i messaged him when i purchased my unit" i didnt buy the ADCiso because it would add to the complexity of the device increasing the probability of error or noise.

imo.
(https://www.audiosciencereview.com/...ds/e1da-cosmos-adc.27038/page-65#post-1745784) thread 1299
this verifies i was correct too, a small irrelevant issue.


my unit was 1 dB shy of the quoted rating, which miffed me because i paid extra for the grade 0 to get the 129+ ratings.

but thats kinda nitpicking when dBA rating throws everything in the toilet from 20-1000 and 10k-20k hz

 

[hayabusaxps]

* change your FFT length to 256k instead of 1M
* change your window to Blackman-Harris 7
* change your averages to 2 or none
* select “use 64 bit FFT”

and cosmos runoff a tablet not connected to the wall. and remote desktop controlled.
best i can do as a battery pack haha. there's no 60hz in the response anyway for noise.

also set the dips to 10v as cosmos instructed


this should have been the best settings, 2channel 48khz sampling 64-bit fft, and everything listed above.
but in stereo. cosmos volume at zero.
View attachment 346372


32bit 192khz sampling on cosmos
View attachment 346373View attachment 346371


32bit 192khz , regular non 64bit sampling
View attachment 346369View attachment 346368

no changes on the -126 dBFS


lastly added the scope and mono
View attachment 346376

im stuck at -126dBa FS

and this is supposed to be better than 129 correct.... on a grade 0

i also resolved my 126dBA issues as well, the 126 occurs because you have to manually bridge the XLR too, but mine only gets 128.5dbA

manual bridge the TRSS (Vaux if i recall) and manual bridge the XLR (L+ L- and R+ and R- and GND/shield) in the same way.
its not automatically done internally like i was inferring for XLR

also a not about anyone else that tries to use the volume slider.
mine only functions prolerly and 0 and 100% wierd stuff occurs at 18-25%

 

[hayabusaxps]

if i recall the cheaper usb isolators
i think the 60$ fi isolator unit, has a 6000hz hum introduced at like -40db drove me crazy!!!
and the toping isolator broke after 24hrs. (same hardware as the TI dev unit)
the Ti DEV unit has been functioning since day 1 though the device is already pretty good though i forget which combinations of isolator and power source on the dac still hold a small 60hz hum at like -110db though.

but seriously you can just ignore it if its at -110db its not going to be audible,

 

[BKDad]

he uses a different device, (different brand and model)
but same exact function/circuit

"from when i messaged him when i purchased my unit" i didnt buy the ADCiso because it would add to the complexity of the device increasing the probability of error or noise.

imo.
(https://www.audiosciencereview.com/...ds/e1da-cosmos-adc.27038/page-65#post-1745784) thread 1299
this verifies i was correct too, a small irrelevant issue.


my unit was 1 dB shy of the quoted rating, which miffed me because i paid extra for the grade 0 to get the 129+ ratings.

but thats kinda nitpicking when dBA rating throws everything in the toilet from 20-1000 and 10k-20k hz

I'll have to take your word on the device type. I can't read the printing on the one on my ADCiso, even with magnification.

I am not sure if the 129+ rating normalizes out the noise of the ADC and associated parts or not. That can contribute to the overall noise reading.

 

[BKDad]

if i recall the cheaper usb isolators
i think the 60$ fi isolator unit, has a 6000hz hum introduced at like -40db drove me crazy!!!
and the toping isolator broke after 24hrs. (same hardware as the TI dev unit)
the Ti DEV unit has been functioning since day 1 though the device is already pretty good though i forget which combinations of isolator and power source on the dac still hold a small 60hz hum at like -110db though.

but seriously you can just ignore it if its at -110db its not going to be audible,

I built my own isolator previously using a different device. Unlike the Topping, it has no switching converter in it to supply the downstream side from the computer USB power. Having the isolation built into the ADC is more convenient, of course.

 

[hayabusaxps]

I built my own isolator previously using a different device. Unlike the Topping, it has no switching converter in it to supply the downstream side from the computer USB power. Having the isolation built into the ADC is more convenient, of course.

yes!!!!!!
it much more convenient if your computer is the major cause from the noise.
think older cpu/ big GPU/ older psu leads to fluctuations.

if your not really fluent and have all the diagnostics hardware ADCiso is the way to go.


thats really awsome you built your own isolator with the usb 2.0 speeds.

 

[BKDad]

yes!!!!!!
it much more convenient if your computer is the major cause from the noise.
think older cpu/ big GPU/ older psu leads to fluctuations.

if your not really fluent and have all the diagnostics hardware ADCiso is the way to go.


thats really awsome you built your own isolator with the usb 2.0 speeds.

The ADCiso also has some improvements in it for better distortion at higher audio frequencies.

I use the isolator now between this computer and the DAC attached to it. Nice improvement. Didn't break after 24 hours, either. Gave a couple away to friends, too.

 

[hayabusaxps]

I'll have to take your word on the device type. I can't read the printing on the one on my ADCiso, even with magnification.

I am not sure if the 129+ rating normalizes out the noise of the ADC and associated parts or not. That can contribute to the overall noise reading.

normalize , it shouldnt, -129 should be considered when you have clean signals, and no major noise source leakage issues.

though with the iso because its built in, it should become part of the system so better than -129 should mostly be a requirement.
though you still have technicalities because the ADC still has a power source port. you have 2 loops here. 1 from data and 1 for power but the usb side should be isolated in ADCiso

i would consider it fair game to be below -129 if i wasnt running it off a (battery) tablet (low power + battery), though... depending on the tablet power circuits it may still require the isolator,
but i wasnt seeing frequency response characteristics of leakage and i don't recall if i also connected the isolator on my last measurements.
it was last month from the dates on my testing.

 

[hayabusaxps]

my opinion is that the only part holding back better capabilities in the ED1h ADC is going to be the actual ESS ES9822PRO its already impressive but the dBA rating leaves a bit left on the table if used as an oscilloscope .