Pro Audio Design Forum

hi

I notice c64 (unpopulated) has a nice large pad, just happens to connect to Vocm!

and the 2.7nf that I think you mention is ' c61 ' right up close to the PCM4222 so that is out of the way and post any potential modification.

I could scratch the traces (if the traces are only on the top layer) after r3 r4 and insert a signal relay (pic attached), to switch between the :

relay state 1= Figure 52. external home made, Noninverting Differential Input buffer ( I may use OPA1611's if I remember they are quieter than suggested) and have it mounted as close as possible to the EVM.
relay state 2 = The already on board Differential Input Buffer.

by switching/toggling between the two, that would stop any additional noise from which ever is not in use, perhaps if the relay is shielded well also.

I may just enjoy the EVM using 150uf cap's before I do this; but this might be a way to go?

thanks for the inspiration.


The diagram shows c61 (2.7nf) is pre to the Test points but it is not. it is very close to the PCM4222

Glad you found a good place to grab Vcom.

I would avoid putting too much stuff between the driving op amp pins and the modulator capacitor.
I'm not sure why I'm bothered by the potential for added inductance.

The modulator capacitor will shunt the differential path but I have to wonder what differences would occur in the common mode signal if the lead lengths to the op amp output were too long. I'm not looking at the EVM schematic but IIRC there's also a place for common mode caps.

I would listen to what you have with AC-coupling and see if you like it.

Hi thank you for you reply.

I have had some success with the circuit attached. inserted between the preamp and the EVM.

I had to use 47uf caps on the input of the EVM. (very little dc now. I have to zoom right in to see the dc, using a dc offset plugin is still beneficial after a recording. I have yet to try higher uf caps. (my tone generation is not accurate enough to determine the response of the EVM)

I am still interested in the circuit (diagram 52 Noninverting Differential Input buffer) I wonder if that would be a better option to eliminate the dc and allow for removal of the caps? As r23, r3 and r4 can all be removed to allow inserting of the dia.52 circuit. But as you say, keeping the wires short might be tricky.


thanks again. : )

Yesterday I decided to use a linear power supply with the PCM4222EVM and discovered that with this particular supply I have to reset the EVM when I power cycle it. It runs, but unless the /ADCRST SW4 and /DITRST SW5 pushbuttons are pressed the noise floor is very high.

i also have to reset almost every time i turn the evm on ( a fair amount of noise can be seen on the channels)

i am using a toroidals and regulated pcbs

I was considering to have a relay come on maybe 1 sec after the main power switch is turned on to do this for me, or, as this is for me only, a little button on the front panel until a relay circuit is eventually built.

I had been using a Meanwell P25A14E 25W switcher and never saw this issue.
Being a SMPS I'm sure it has a fast setup time.

The linear supply is a 5V 12A +/-15V 2.5A and it powers up much more slowly I suppose.

It took me quite awhile to remember to reset the EVM.

I'm seeing about -115 to -116 dBFS noise floor with single-ended inputs; around -121 dBFS balanced and -124 A-weighted.
Without a hard reset it degrades to -84.

C71 and C72 across the reset switches are only 10 nF. The time constant with the board's 10K pull-ups is only 100 us.
I'm wondering if C71/72 can be made much larger.