I/V stage for balanced current out DAC (pcm1792/94/98)

[00940]

I've been researching a lot on the board the last weeks output stage for a pcm1794. What comes everytime are : the passlabs D1 output stage or the per datasheet stage with 6 opamps.

From time to time however, appears the idea to use something like the BB ths4131 or the TI opa1632. It seems that the THS4131 would be used in the Creek CD53 and the Belcanto dac2. After all, according to the datasheet :

The functionality of a fully differential amplifier can be imagined as two inverting amplifiers that share a common noninverting terminal

The common approach is the fig 1; the I/V based on the opa1632 would be the fig 2.




Sadly, the discussion I found about such possible use of fully-differential opamps are very short and about useless. The idea is thrown around and dies without any further comments. Why so ? Is the idea stupid, not fun enough or what else ? The specs of the opa1632 are pretty nice at first sight.

[niles]

We have apparently followed the same path. I did my searches in August. What I determined was that the OPA1632 will work, and will reduce the parts count, but it does not adress the primary reasons for discrete designs because it is still a voltage feedback circuit and therfore it is still limited by slew rate considerations. This is why I finally settled on Jocko's simple circuit. I havent gotten far enough to draw up anything, but it has been sugested to use matched devices for a balanced version. The other option is to take signal from between the positive and negative outputs of the 1794 into a single-ended circuit.

The 4131 is current feedback, and would address at least that one concern, but doesn't an open loop I/V statge just sound more elegant to you? It does to me. One thing I am certain of is that the experts are not going to offer up a ready-made solution. These guys like to help folks find thier own solutions, and really dislike solicitations for a functioning circuit. Have you read the Hawksford article yet? It offers at least one solution to the slewing problems in the form of a passive I/V pre-filter. That might be one way to use the 1632 with less concern for possible slewing issues. There is also plently of current out from the 1794 to shunt a little.

[00940]

Originally posted by niles
The 4131 is current feedback, and would address at least that one concern, but doesn't an open loop I/V statge just sound more elegant to you? It does to me. One thing I am certain of is that the experts are not going to offer up a ready-made solution. These guys like to help folks find thier own solutions, and really dislike solicitations for a functioning circuit.

Actually, I'm weighting elegance, simplicity and cost before taking any decision

The fact is, I'm not asking for a ready solution. I'm trying to understand why any discussion on those differential opamps died quickly while people can argue for ages on the merits of the NE5534 compared to the OPA134

[Jocko Homo]

Arguing the merits of any op-amp always seemed a waste of time. At least to me, it did. If Barrie Gilbert says that they make lousy I/V stages, then that is good enough for me.

I've seen a post that uses a "current-feedback" op-amp in a non-traditional configuration. Works in a manner similar to my discrete design, except that it saves lots of room.

Jocko

[Bricolo]

Originally posted by Jocko Homo
I've seen a post that uses a "current-feedback" op-amp in a non-traditional configuration. Works in a manner similar to my discrete design, except that it saves lots of room.

Jocko

Thorsten's OPA660 I/V?

[00940]

or Pedja's ad844 I/V stage ?

AD844 as a common base stage in the I/V converter

Problem with the OPA660 is that it is discontinued.

[Bricolo]

Originally posted by 00940

Problem with the OPA660 is that it is discontinued.

[00940]

you can still make stocks most suppliers still have those.

[ojg]

I say go for it!

What I would do is make room on the board for both the OPA1632 and two traditional I/V op-amps. That way you can compare both and have a safety fall-back if the OPA1632 fails. (no reason why it should though, except Murphy's law)

[pooge]

You may want to consider the simple passive I/V stage provided by K & K Audio using a transformer and resistor:

Passive I/V stage

There's a white paper on its use on the Raleigh Audio site:

Raleigh Audio white paper