E1DA Cosmos ADC

[nanook]

Does this DC shift also appear when you just short both XLR inputs of the ADC(iso) to ground?
Pin2 and pin3 connected to pin1

 

[Jbenda]

Yes, I have a voltmeter and an oscilloscope. I will measure it with the unit powered on and report back.
One more observation: when both XLR outputs are connected, Audacity shows a signal on the VU meters at approximately –20 dB, which is inaudible.

 

[Jbenda]

Does this DC shift also appear when you just short both XLR inputs of the ADC(iso) to ground?
Pin2 and pin3 connected to pin1

How should I simulate it? (What exactly should I connect?) pin 2(+) to pin1. Pin3(-) to pin 1?
Won’t this damage the Cosmos?

 

[nanook]

How should I simulate it? (What exactly should I connect?)

Connect pin2 and pin3 to pin1 of the XLR inputs of the ADCiso.

It is strange however, that it seems to need both inputs connected to show the fault.

@IVX: are there resistors from the inputs to gnd to keep the internal AC coupling caps precharged?

 

[nanook]

Converting RCA to XLR and back is not the way to go.
There's something faulty that needs to be fixed.

 

[Jbenda]

Converting RCA to XLR and back is not the way to go.
There's something faulty that needs to be fixed.

Ok. Need both inputs connected to show the fault.

 

[nanook]

When you have this DC level with the pre-amp not powered, points towards a fault in the ADCiso.

Do you have XLR plugs to short the inputs?
Some wire stuck into the three poles of the XLR jack may do as well.

 

[nanook]

The inputs of the ADCiso buffer are at +2.5V. There are electrolytic caps in series to the inputs to make sure no DC current can flow into the signal source.

If one of these cap has a high leakage current, there's current flowing out of this pin and the ADC chips input sees a DC signal.

 

[nanook]

When you have a multimeter that can measure current, you can connect it to the inputs of the powered ADCiso to directly measure the leakage current, which should be in the order of micro Ampere.
You can measure the inputs one after the other.
Connection for each of the XLR inputs is
Pin1 - Pin2
Pin1 - pin3

 

[nanook]

@Jbenda: Did you find the root cause of this miracle?

 

[Jbenda]

Sorry for not replying earlier. I only got to the measurements today. I took the Gold Note PH-10 and the external PSU-10 power supply out of the rack. I disconnected the PH-10 from the external power supply as well. I wanted to test the PH-10 without the PSU-10. However, the PH-10 initially refused to turn on. After a bit of tinkering and checking the fuse, it finally powered up. The curve in Audacity looked fine. Then I reconnected the PSU-10 and everything was still OK.


I’m still testing, but it seems the issue was in the PH-10 software, and after restarting it, everything works properly. Hopefully this will help others who run into the same problem. The settings for the C1DA are 10 V (the highest possible), and the PH-10 is set to -3 dB.

 

[Jbenda]

The upper channel has a slightly wider signal line, but I wouldn’t consider that a problem

 

[Music1969]

@IVX the freq response (FR) calibration file - it is correcting FR for which ADC ESS digital filter? Or applies to all the filters?

I just bought Grade 0 ADCiso

 

[IVX]

Default filter Linear phase APO, CosmosADCiso_FR prepared for that.

 

[IVX]

and at 768k!

 

[Music1969]

and at 768k!

so the .cal file is correcting ADCiso frequency response up to ~384kHz ?

 

[IVX]

rather up to 380k.

 

[horias2000]

OK, I went and damaged the Cosmos ADC as well. I was testing a power amplifier and I forgot to unplug the XLR input connectors before powering off the amplifier. As you can imagine, a rather high DC component was present on the output of the amp and on the input of the ADC. The thing is that I used the XLR inputs set to 10V as I wanted to see if there is a difference from the TRS input. I'm hoping that the issue is similar to the Scaler and that there is a one or several resistors that were damaged. @IVX can you please point me in the right direction regarding the input resistors and their correct values? Thanks!

This is the version I have:

LATER EDIT!!!
FALSE alarm! It must have been a REW glitch as the ADC seems to work just fine. Here is the ADC+Scaler on my D50s. All good!