THAT 1512 to 1646- ac or dc coupling?

Working on building a THATcorp 1512 based pre using a 1646 line driver to balance the output. Should the two be DC or AC coupled? The 1512 datasheet says you can get up to 250mV offset at output (without DC servo or ac coupling). Thanks for any help, I've been sifting through pages upon pages of datasheets tonight and not finding a real answer to this.

gemini86 said:

Working on building a THATcorp 1512 based pre using a 1646 line driver to balance the output. Should the two be DC or AC coupled? The 1512 datasheet says you can get up to 250mV offset at output (without DC servo or ac coupling). Thanks for any help, I've been sifting through pages upon pages of datasheets tonight and not finding a real answer to this.

Click to expand...

What do you want to do with this? If it's for audio, audio is intrinsically not DC. If you end up with some offset at the outputs, that will induce clicks and pops when switching or connecting/disconnecting. If this is a concern to you, you must take preventative measures.
Using output caps is very efficient at removing DC, and easy and unexpensive. The quality of most modern electrolytics is eons better than what they used to be (and gave them bad rep).
If you choose DC servo must be applied to both legs of the line driver, it's not as efficient at removing DC, unless you add trimmable offset compensation, it takes at least two opamps and a handful of components and PCB real-estate.
DC servo is often used under the false assumption that there are no more caps in the signal path; it's not true, the caps are actually in the signal path (although they generally are plastic film type) and the additional opamps are also in contributing to some performance degradation. And DC servo disqualifies the circuit for direct-coupling (FWIW).

The 1646 is commonly used with no caps on the output with no problems at all (you do need a cap between the force and sense lines, see THAT docs..) I think they show a servo for the output of the preamp chip on the app notes, very simple and easy to build. If you do not want to go there, use a cap to couple the pre chip out to the output driver, but it does need something there.
I would be very curious to have a blind listening test between the 2 approaches.... you are more likely to hear coupling caps when you have some 20 or more in a console path!

thanks to the both of you for your input. I know that dc offset can be a bad thing, but being that these two IC's will be living together on one board, with no switching or disconnecting between them, I thought a decoupling cap might be unnecessary. Both the 1512 and 1646 demo board schematics show no input or output caps, as well as the 1512 demo board only has the empty footprints for dc servo on the gain control, and not the output of the opamp. So that's what lead me to ask if there was a reason for decoupling the two, or can the 1646 get along fine with the possible 250mV offset (if even that). I WILL be decoupling the output of the 1646 for good measure, even though my ADA I/O has decoupling caps...

so I guess I'll just try it out with a jumper in the place of a decoupling cap and if that doesn't work, I'll just throw a cap in.

thanks again! GREAT forum!

gemini86 said:

thanks to the both of you for your input. I know that dc offset can be a bad thing, but being that these two IC's will be living together on one board, with no switching or disconnecting between them, I thought a decoupling cap might be unnecessary. Both the 1512 and 1646 demo board schematics show no input or output caps, as well as the 1512 demo board only has the empty footprints for dc servo on the gain control, and not the output of the opamp. So that's what lead me to ask if there was a reason for decoupling the two, or can the 1646 get along fine with the possible 250mV offset (if even that). I WILL be decoupling the output of the 1646 for good measure, even though my ADA I/O has decoupling caps...

so I guess I'll just try it out with a jumper in the place of a decoupling cap and if that doesn't work, I'll just throw a cap in.

thanks again! GREAT forum!

Click to expand...

That's exactly what I meant; no need to block DC between the two chips, but the output to the outside world should be DC-free. I had somehow not properly understood your question.

I have both IC's but no engineering knowledge . could you share your Schematic for this project? thanks

Go to thatcorp.com and check out the datasheets on both demo boards for the 1512 and the 1646, my design is pretty much pulled off those, with a standard +-18 and +48v phantom PSU on a separate board.

An Improved Servo for the THAT1510 and THAT1512 from Wayne Kirkwood can be found over at the Pico forum:

http://www.picocompressorforum.com/forum/php/viewtopic.php?f=6&t=256

I see a lot of questions here and elsewhere asking about servos for the THAT1510 and THAT1512. This post will discuss a new servo circuit which has the goal of reducing gain switch clicks and eliminating a large-value Cgain. An added benefit of this new circuit is that output offset is virtually eliminated.

Click to expand...

yeah, just read up on that, too bad I already built my pre's... (WITH a huge ass cap in the gain loop. :-[ )