E1DA Cosmos ADC teardown

[Blumlein 88]

View attachment 164577​

Above would be Topping D90 MQA on Cosmos ADC. Almost 2dB worse than ARTA, bunt not sure if it's due to the internal Multitone's generator or perhaps Multitone is measuring a bandpass bigger than 20...20kHz? @pkane pls hlp.

Later Edit: I redid same test with ARTA and now I got a THD+N of -118.416dB. So, like Ivan said, warming up Cosmos ADC prior to do measurements should help (it was calibrated for 27C). This is why sometimes I get 1-2 dB better.

So that looks pretty good. And if I understood, you think if it were warmed up it would match your ARTA results.

So what I had in mind actually was using one of the multi-tones. Maybe the 32 tone AP version like Amir uses or one of the others with more tones just to see how well this pairing does.

 

[dougi]

Does anyone know what the different grades mean? There is a large price difference between them? I would love one for the odd amp measurement, so as not to risk the RME ADI-2 PRO. With the 43V input they would be great for that, where the relatively low input impedance would not be an issue, hopefully! (in parallel with a load block of course)

 

[IVX]

Maybe the 32 tone AP

https://cdn.l7audiolab.com/wp-content/uploads/2021/10/Multitone.jpg

Regarding the grades and more please read the product description https://e1dashz.wixsite.com/index/cosmos-adc

 

[IVX]

I trying the "thermostat oven" to warm up ES9822 to stabilize a 3rd harmonic performance. Profit - distortions-harmonics are very stable over 20-30C range, cost - up to 100mA of a current draw is needed to feed the oven.

 

[pkane]

View attachment 164577​

Above would be Topping D90 MQA on Cosmos ADC. Almost 2dB worse than ARTA, bunt not sure if it's due to the internal Multitone's generator or perhaps Multitone is measuring a bandpass bigger than 20...20kHz? @pkane pls hlp.

Later Edit: I redid same test with ARTA and now I got a THD+N of -118.416dB. So, like Ivan said, warming up Cosmos ADC prior to do measurements should help (it was calibrated for 27C). This is why sometimes I get 1-2 dB better.

To clarify: the measurement bandwidth in Multitone is determined by the minimum and maximum frequency settings above the plot. In your screen shot, it's 20-20,000Hz. If you click Reset Axis button, you'll see all the data available, and which part is measured (white is the measured noise+distortion, blue is the captured test signal), something like this:

 

[wwenze]

Should load-bearing screws be soldered? Would the solder not break off?

 

[trl]

So what I had in mind actually was using one of the multi-tones. Maybe the 32 tone AP version like Amir uses or one of the others with more tones just to see how well this pairing does.

Already did here: #59

 

[H-713]

Nice to see good quality on this. I don't see anything I don't like.

Next step is going to be to design an audio analyzer front-end for it. Might be time for me to dig that project out of the mothballs...

 

[dougi]

Nice to see good quality on this. I don't see anything I don't like.

Next step is going to be to design an audio analyzer front-end for it. Might be time for me to dig that project out of the mothballs...

Mine is on it's way. Does anyone know if the 43V/TRRS input is DC protected? i.e. if I am testing a class D amp which is floating above ground, as some do, do I need an external block or filter?

 

[trl]

Per https://e1dashz.wixsite.com/index/cosmos-adc: "All inputs are AC coupled(-2V to +12v DC allowed)".

I don't know what exact capacitors are in the signal path that handle the input DC, might be those four small SMD non-polarised caps nearby the two XLR plugs or those nearby the OPA1612 or the four black 105C electrolytic caps in the middle, but maybe @IVX would like to detail this to us.

However, @dougi might need to measure the output DC of that Class-D amp prior to connect it to Cosmos ADC. Usually there should be tens of mV allowed there, but not more, unless the amp is defective or has some design problems. You could also use a laptop running on battery or with an isolation transformer in front (most laptops PSUs are already isolated inside, but not sure about yours).

 

[IVX]

trl, yes you are right these 4 TH black aluminum caps are DC blocking, and I have to correct the product page because of safe range for a DC is -7.5..+12.5VDC.

 

[trl]

-7.5...+12.5VDC is way better than -2...+12VDC, so great news Ivan, thanks for sharing.

 

[IVX]

huh, I don't think if 220uF/10V will not withstand +30% for a long time, so there is no problem at all, don't worry.

 

[777]

I am going to tease you with mine below then.

View attachment 164593
​

It would be interesting to find out what the first row says, perhaps @IVX may help here with translation.

maybe it's mine

I am going to tease you with mine below then.

View attachment 164593
​

It would be interesting to find out what the first row says, perhaps @IVX may help here with translation.

Romania ? My unit was send in Romania on 21.10.2021 but it has not arrived yet. My name start with "R" too . TRL, are you from Romania ?

 

[dougi]

Per https://e1dashz.wixsite.com/index/cosmos-adc: "All inputs are AC coupled(-2V to +12v DC allowed)".

I don't know what exact capacitors are in the signal path that handle the input DC, might be those four small SMD non-polarised caps nearby the two XLR plugs or those nearby the OPA1612 or the four black 105C electrolytic caps in the middle, but maybe @IVX would like to detail this to us.

However, @dougi might need to measure the output DC of that Class-D amp prior to connect it to Cosmos ADC. Usually there should be tens of mV allowed there, but not more, unless the amp is defective or has some design problems. You could also use a laptop running on battery or with an isolation transformer in front (most laptops PSUs are already isolated inside, but not sure about yours).

Thanks and @IVX as well. Certainly for my previous Lyngdorf sda2175, with 35v DC I would need an external block! At least it warned you.

 

[trl]

maybe it's mine

Romania ? My unit was send in Romania on 21.10.2021 but it has not arrived yet. My name start with "R" too . TRL, are you from Romania ?

No, it wasn't yours for sure.
Yep, I'm from RO as well.

 

[IVX]

huh, I don't think if 220uF/10V will not withstand +30% for a long time, so there is no problem at all, don't worry.

Oh, I have to apologize, the original specs were correct +12V/-2V because of the polar cap used with the 2.4VDC bias.
This morning I mixed up that with APU and Scaler inputs where are NP caps used.

 

[trl]

Well, I guess that swapping the polarised caps with non-polarised type, like anything between 6.8...10uF MKP will dramatically increase the max. DC input voltage range, right? However, size and price would increase a lot, so not sure it worth it, unless it's really-really required.

 

[NTK]

Should load-bearing screws be soldered? Would the solder not break off?

Soldering threaded brass standoffs to a PCB is a bit unusual. However, soldering standoffs to PCB's is not. There are ones made for that purpose. I suppose they are not as strong as threaded standoffs held down by nuts on the other side, relying on the adhesion strength of the copper to the PCB, but they don't easily loosen.

https://www.pemnet.com/fastening_products/pdf/kdata.pdf

 

[Lambda]

@IVX
So is there a I2S signal coming from the pins on the Board?
If so i would be very interested if it can run be synchronized to an external Clock.
And if some sort of AES / SPDIF or ADAT digital output can be Added.

This product is already amazing.
For "studio" uses the ability to have world clock to synchronize and digital output would make it a "killer" product.

For Measurement use (not only audio) the Ability to isolate at the i2s level would make it very flexible and idiot proof.
This would make it able to have floating/highside shunt to messue current to a speaker with high precision/resolution.