TLV320DAC3203: Need to know about the internal protection of the HP outputs

Hi Gustavo,

Referring to this previous thread, the OCP protection on HP outputs is about 200mA.
Connecting to 48V supply is not safe for the device, you would have to add AC coupling such that it doesn't reach the device pins.

Best regards,
-Ivan Salazar
Applications Engineer

Hi Ivan,

thanks for the answer!

Precisely those blocking caps are the problem. I have been following all the information I could gatter on this matter, but since my application is not a mic, nor a mixer providing the phantom power, but a intermediate device converting the 48V to 2.4V while letting Audio pass through, the thing is complicated. At least for me.

I found this thread interesting: https://e2e.ti.com/support/audio-group/audio/f/audio-forum/829315/pga2500-how-to-select-resistance-for-limiting-the-input-current-of-pga2500

There, Luis Chioye is suggesting this solution for a mic preamp using a PGA2500:

This is not really my case, since I am not using negative supply, but the principle og using TVS diodes together with Schottky diodes seem a very good option.

He says too:
"The 48-V phantom supply voltage will charge the 47uF DC blocking capacitors, and when the inputs are shorted, this will produce the large surge current.
You may consider reducing the value of the blocking capacitors.  The PGA2500 recommends a capacitor value in the range of 10uF to 47uF.  The DC blocking capacitors interact with the input impedance of the PGA2500 (4600-Ohm typical per pin), forming a high-pass filter, blocking the DC signals.  When using 47-uF DC blocking capacitors at the inputs, the corner frequency (f-3dB) is ~0.747-Hz.
f(3dB) = 1/(2*PI()*Cin*4600-Ohm) = 1/(2*PI()*47-uF*4600-Ohm) = 0.747-Hz
Similarly, using 10-uF DC blocking capacitors will set the corner frequency to ~3.51-Hz. 
If your application allows for using a higher corner frequency on the high-pass filter, you could consider reducing the capacitor further to reduce the current surge/voltage spike.  The trade-off is the higher corner frequency on the high-pass  filter, that will result on the attenuation of  the low-frequency signals.
"

I coud live with a higher corner freq., but for this calculations I need to know more about the HP outputs of the TLV320DAC3203. This also because, in order to have enough power, at least on the first stage (the DC/DC) I need to use very small resistors (milliohms to few ohms), instead of the usual 6.8kOhm before the blocking caps.

This could also be a problem for the DC/DC itself, if a fault condition arises on the following section, then the surge current could reach the output of the device too. (?)

Would you mind to continue this thread outside the forum? I don't  feel comfortable disclosing this ideas on public.

Thanks a lot for your patience,

Best regards,

Gustavo

Hi Ivan,

haven't hear from you since I sent the email.
But do not worry, I realise now that this could be a dangerous endeavour and I am refraining of using ghostd to feed my sensitive Ti-Electronic ;)

Meanwhile I am looking for another way, which looks very promising.
Thanks, anyway, for the help till now.
Do not worry, I'll be back...

Best regards,

Gustavo