E1DA Cosmos ADC

[Music1969]

For reference, the blue trace shows the same input recorded at 384 kHz.

Just so I understand: so if you measure at PCM 384kHz, then none of the 8 filter options make a difference ?

Everything will measure the same?

 

[IVX]

I clearly see the difference if changing filters at PCB 384k, 180-190kHz area changes a lot.

 

[Rja4000]

@IVX
HI ivan
Any clue wheb the scaler and APU will be available for order ?

 

[IVX]

Rja4000, you know, I keep improving that slowly with the hope $ilicon prices settle down.

 

[Grooved]

Hi,

I really didn't have enough time available lately to work on it sooner, but I made some progress on the digital outputs we talked about with @IVX @pkane @HiZ in this thread.

On my first try, I was able to get SPDIF output and get an audio interface record and to be synced to this source. I did a loopback recording, and after listening the results, I immediately heard some distorsion.
So I ran a REW test (not easiest case as there is DAC -> Cosmos ADC -> Digital Board -> Audio interface) and saw peaks of distorsion between odd and even harmonics (sadly, I made a screenshot of that, but I've overwritten it with the next screenshot )

Strangely, since this test, I can't get any audio interface to sync again on the SDPIF signal, but audio is still coming as I can record it on the SPDIF input, but without sync (internal clock of the audio interface).
It show a worst signal on the right side where there are cuts/jumps but the left is cleaner.

As I have not so much time, I didn't add a resistor on the MLCK output, so maybe the problem comes from here.
@IVX , can you confirm me please the value of the resistors you placed on SCLK, SDOUT and LRCK is 17 ohms?
(have that in mind as I'm writing, I think it's what I measured)

For information, I'm using one of the boards in the second page of these thread, with the inputs on the left and 5V power supply (tried several, even battery, didn't change the results)
Tested it with TOSLINK and SPDIF outputs only (will test it on AES as soon as possible), and cables have no problems as I measured a TOSLINK and SPDIF loopback on the audio interface and get perfect results on REW, Multitone and Deltawave (bit perfect match)

I also tried first with a 0.2mm diameter wire for the MCLK, then with a 0.1mm which was easier to solder but got worst results

 

[IVX]

20ohm actually.

 

[Grooved]

20ohm actually.

Thanks @IVX for the confirmation! I knew it was in that around that but certainly confused with the 17 on other things I measured the same day.

 

[Music1969]

I clearly see the difference if changing filters at PCB 384k, 180-190kHz area changes a lot.

Which is 'best' at PCM384k in your opinion/recommendation?

 

[IVX]

I didn't notice any THD difference, only different impulse responses, delays, and frequency responses.

 

[JimiPegorino]

Forgive me for asking but can the COSMOS ADC be used to record a effects/mastering chain? Looking for a cheap ADC for this purpose.

 

[staticV3]

Forgive me for asking but can the COSMOS ADC be used to record a effects/mastering chain? Looking for a cheap ADC for this purpose.

As long as the last link in that chain before the ADC doesn't mind the Cosmos ADC's rather low Zin and can produce high enough output voltage, then sure!

 

[JimiPegorino]

As long as the last link in that chain before the ADC doesn't mind the Cosmos ADC's rather low Zin and can produce high enough output voltage, then sure!

Thanks for the reply and answer! The Cosmos should have a very good frequency response for the price comparable to some DAC’s costing 5-10x as much right?

 

[staticV3]

Thanks for the reply and answer! The Cosmos should have a very good frequency response for the price comparable to some DAC’s costing 5-10x as much right?

Cosmos ADC frequency response is rated at +0.1/-3dB 5Hz-190kHz.
I don't know how that compares to other ADCs and interfaces, sorry.

 

[Rja4000]

Thanks for the reply and answer! The Cosmos should have a very good frequency response for the price comparable to some DAC’s costing 5-10x as much right?

Most ADCs have good enough frequency response.
The main difference vs other ADCs is elsewhere:
Very low noise and distortion.

But for optimal performance, the input level needs to match the ADC optimal.
On the Cosmos, you may adjust range from 1.7Vrms to 10Vrms for 0dBFS by dip switch.
The lower the range, the lower the impedance though.

 

[charliecola]

I am wondering if anyone can help with high frequency roll off on a Cosmos ADC on a particular computer.

Everything tests well apart from frequency response - when I first got the ADC everything was perfect when tested on this particular machine including the frequency response... now though, the high frequencies roll off substantially (please see image below).

It works fine on my laptop but on the desktop I get this high frequency roll-off.

I have tried removing driver, reinstalling driver even tried other filters in the tweak app, cables, different usb port etc... just can't work out what is causing is - It is strange that it works on laptop perfectly and also worked before on this machine perfectly but now has this single issue.

Any help would be appreciated.

 

[phofman]

charliecola: What is your audio software chain? Any chance your "now" case going through the windows audio engine?

 

[charliecola]

charliecola: What is your audio software chain? Any chance your "now" case going through the windows audio engine?

I got to the bottom of the problem - it was the word clock clocking the DAC - the other ADC I was testing with was producing the master word clock for the whole system (apart from the Cosmos) so didn't reveal the problem when testing through itself, when I switched the clocking of the audio interface to 'internal' the frequency response came back. The problem didn't show up when tested new because I used USB for the DAC rather than SPDIF and the audio interface (RME HDSPe RayDat) was out of the equation.

Thank you very much for your reply nonetheless - the software chain was Rightmark RMAA using MME drivers and also ASIO4ALL v2...

 

[IVX]

I am wondering if anyone can help with high frequency roll off on a Cosmos ADC on a particular computer.

Everything tests well apart from frequency response - when I first got the ADC everything was perfect when tested on this particular machine including the frequency response... now though, the high frequencies roll off substantially (please see image below).

It works fine on my laptop but on the desktop I get this high frequency roll-off.

I have tried removing driver, reinstalling driver even tried other filters in the tweak app, cables, different usb port etc... just can't work out what is causing is - It is strange that it works on laptop perfectly and also worked before on this machine perfectly but now has this single issue.

Any help would be appreciated.


View attachment 190421

looks like you've changed the filter setting in Cosmos_Tweak app.

 

[Grooved]

@IVX
I have a problem right now. After my last try with a 20ohm resistor added on the MCLK, it was still not perfect with my digital output board, but Cosmos ADC was still working great via USB (I even could record with ASIO driver in Wavelab).
I checked each time to be sure that changing the resistor or changing the length of cables going to the digital output board doesn't impact the measurements I got via USB.
I used it again another day via USB only, left it for two days, and now that I wanted to use it again (without having change anything since two days ago), I have no audio at all, using WASAPI, Comtrue ASIO or ASIO4ALL.

Cosmos ADC is seen by softwares like it was before, I can select it, I also have blue or green LED working when analog signal comes in (I can turn it to mono or stereo with Cosmos Tweak software), and I even tried a firmware update with CT7601_ISP_Tool and it worked and said it succeeded.
But REW, Multitone or Wavelab : no signal at all.

Any idea on something I could do, please?

 

[IVX]

Grooved, huh, that looks like you've killed I2S input/s of the CT7601 or output/s of ES9822. Probably ESD. Pls take a look with a scope is there normal I2S signals?