Help on reconfiguring PCM1738 dual dif outputs please

[kevinkr]

Originally Posted by georgehifi

kevinkr, I'm going to try it next week, just making sure with you before I do with the attached diagram one channel only shown. (don't want to blow anything)


1: 2 x positive outputs of dac together go to + input of I/V

2: 2 x negative outputs of dac together go to - input of I/V

3 : vcom2 from dac goes to G of I/V link 4 NOT bridged.

Cheers George

Hi George,
I think what you propose is correct, but I would take a look at the datasheet for the PCM1738. The current does sum as you have indicated and if the transimpedance amplifier cannot handle 5mA FS then you do have a problem.

One of the things I am wondering is just how close to a virtual earth the transimpedance amplifier inputs are? If they really are that good you can offset the dac current with an opposing current of 1/2 the magnitude of the FS output current making the D/A appear as a bipolar dac to the transimpedance amplifier. In this scenario you would not connect Vcom2 to the transimpedance amplifier and the output would inherently have close to 0 dc in the output at the mid point. (0 signal) Note that there is some risk of damaging the dacs if the current outputs are forced below ground and internal substrate diodes are turned on. You have to assure that the supplies feeding the transimpedance are always up first.

There may be modifications to the transimpedance amplifier that will allow it to operate at higher currents, but I am not sure about that.

An alternative would be the THS4130 which I am running at much higher currents than this - FS with the PCM1794A is about 16mA (IIRC) in mono mode in my set up.

As always caveat emptor. Be really careful as you go - I'd hate for you to roach your dac. In my case it just costs me money as the parts are all current, in your case it could be a headache.

Incidentally the changes you propose are completely worthwhile regardless of the risk. There are a lot of op-amps in the signal path, and IMHO a more theoretically perfect amplifier device there isn't - but in practice a couple of them in cascade are often (if not always) worse sounding than a few very parametrically inferior discrete devices in a much simpler circuit.. (Just my opinion... )

[georgehifi]

Thanks Kevin, I will do it once I have time away from the Lightspeed Attenuator orders.
LC audio say the Transimpedance amp is a just a couple of ohms input impedance when used in current mode.
I think if I increase the 16kohm I-ref resistor that goes to ground on pin 21 of the PCM1738 to say 20kohm or so, I can reduce the current output of the dac, this may well see the Transimpedance amp just under clipping at 0dbF then.
Cheers george

[kevinkr]

Originally Posted by georgehifi

Thanks Kevin, I will do it once I have time away from the Lightspeed Attenuator orders.
LC audio say the Transimpedance amp is a just a couple of ohms input impedance when used in current mode.
I think if I increase the 16kohm I-ref resistor that goes to ground on pin 21 of the PCM1738 to say 20kohm or so, I can reduce the current output of the dac, this may well see the Transimpedance amp just under clipping at 0dbF then.
Cheers george

Hi George,
I think that should work just fine. This is commonly done with the PCM1794 as well. Note that you should not reduce this more than necessary to eliminate clipping at FS because it does have an implication on dynamic range and the lower current per bit moves you closer to the dac chip & transimpedance amplifier noise floors. This is probably not an issue here as most dacs have significantly more dynamic range than most systems can reproduce, and you are only trading off a few dB..