use NE5532 as the DC servo for the INA217 pre?

Just getting back into DIY, and wanting to build a 2 channel Pre with some INA217's that I have on hand.
I've gone through my boxes and I have everything I need for this except for the OPA137's for the DC servo.
(also minus 1M resistors, but I figure I can wire up a few in series just while i experiment)

circuit for reference: http://www.ti.com/lit/ds/symlink/ina217.pdf 

Can I simply use  NE5532 for the servo? I have plenty of those.  If so,  any problem in using one NE5532 for servo's for both channels?

Actually it's not ideal because of the input offset current of the bipolar input. You really need a fet input. But the servo isn't carrying signal so it can be a cheapo TL071 or similar. Or a TL072 with one half bypassed. The input offset current of the NE5532 with the 1M input Z is going to cause a significant offset on the output which is going to largely defeat the purpose of having a servo in the first place.

Honestly I really don't understand the importance of a servo on a mic pre. /Maybe/ if you're DC coupling two stages like in the Jensen dual servo circuit. But otherwise, does it really matter that your offset is perfectly 0V? As long as you have coupling caps where necessary, it should have no affect on fidelity. I tried a servo on a mic pre once and it actually added about 3dB to the noise floor.

Not a good idea. For starters, the OPA137 is a single amplifier, and the 5532 is a dual. So, the 5532 will pop if you plug one in.

At the very least, you want a FET input op amp, like the OPA137. If you want to throw something in there for the time being, a single BiFET like a TL071 or LF351 will be tolerable, but will still result in about 3-5x more offset voltage than the OPA137 spec'ed.

You can try a few different units of TL071 or whichever BiFETs you have, and maybe you can find one that will have a magically low offset voltage by sheer luck. The offset voltage range of those chips is very wide, but there will be some really nice ones if you search through a big enough pile.

If you can handle an SO8 surface mount package, perhaps with an adapter board, there are far better servo amps out there. My favorite is the Analog Devices ADA4522, available in singles, as well as the ADA4522-2 dual. It has extremely low offset, low noise, low distortion, and extremely low 1/f noise, everything you want in a servo amp. It's also reasonably low cost. If you use slightly optimal passive components, you can easily get a real world residual offset of around 5-8µV, instead of the handful of mV you'll get with the garden variety BiFETs recommended in the past. It's 2018 and we have nicer amplifiers than the TL07x series... why not use them.

squarewave said:

Honestly I really don't understand the importance of a servo on a mic pre. /Maybe/ if you're DC coupling two stages like in the Jensen dual servo circuit. But otherwise, does it really matter that your offset is perfectly 0V? As long as you have coupling caps where necessary, it should have no affect on fidelity. I tried a servo on a mic pre once and it actually added about 3dB to the noise floor.

Click to expand...

If you want to drive a nice output transformer without a coupling cap, then a quality servo will work, and never wear out. A sloppy servo will not remove enough DC, and will probably foul the signal as well. Just design a good one, provide attenuation at the servo output, use a quality chopper stabilized servo amp, and not some 3mV BiFET from the '70s, and you don't need coupling caps. Otherwise, if you're using coupling caps, then yes - servos are silly.

Monte McGuire said:

Not a good idea. For starters, the OPA137 is a single amplifier, and the 5532 is a dual. So, the 5532 will pop if you plug one in.

Click to expand...

But as the OP is DIYing it I'm assuming there isn't a pcb (at least not yet) so is free to use whatever single / dual / package preferred ?

Thanks for the answers everyone!

I am DIYing, so I can design for the servos to be in a dual package.

As far as using a servo or just using a cap, the board layout I started had provisions for both, so I could experiment and see if I found any difference, or at least see which I preferred.

the ina217's I have are surface mount, and was mostly considering other surface mount chips that I have on hand.  Turns out I actually have a few 8-DIP OPA2132's - which seems actually better than the 137's right? I think I'll just use those.

JamesHE said:

Thanks for the answers everyone!

I am DIYing, so I can design for the servos to be in a dual package.

Click to expand...

Cool - wasn't sure, since a lot of DIY I see is based on an existing PCB. I'd call that design, not DIY! ;-)

the ina217's I have are surface mount, and was mostly considering other surface mount chips that I have on hand.  Turns out I actually have a few 8-DIP OPA2132's - which seems actually better than the 137's right? I think I'll just use those.

Click to expand...

The OPA2132 is definitely better - not a bad choice . If you're starting from scratch, the Analog ADA4522 is ideal, far better than any of the audio grade BiFETs. It's actually designed for DC uses like servo integrators. So, don't ignore that either. For a current project, I decided to go deep and figure out servos completely, since I can't use coupling caps. The ADA4522-2 ended up being part of my two servo circuits, and the result works extremely well.

Monte McGuire said:

If you're starting from scratch, the Analog ADA4522 is ideal, far better than any of the audio grade BiFETs. It's actually designed for DC uses like servo integrators. So, don't ignore that either. For a current project, I decided to go deep and figure out servos completely, since I can't use coupling caps. The ADA4522-2 ended up being part of my two servo circuits, and the result works extremely well.

Click to expand...

I may go that route. Looks like I'm going to need to order some parts before I get this thing together anyway. I've determined the Power Supply I partially built years ago isn't going to work for this after all in it's current state. Guess I'll have to ask a few questions about that too

What about OP27, quite cheap these days and +- 22V rails. Very nice for DC.

The OP27 voltage offset is pretty low, and the noise, especially 0.1Hz - 10Hz, is also very low, but the input bias current is a lot higher than with FET input amplifiers. So, the net offset will still be a lot higher than with other amplifiers. Using 100kΩ input resistance, you'll typically get 1mV of net offset from the A grade, mostly due to the input bias current.

I stumbled upon something interesting.

http://www.diyaudio.com/forums/instruments-and-amps/309312-ina217-mic-pre-board-layout-analysis-enclosure-templates.html

this design uses half a NE5532 as the servo, and the other half as a voltage follower to balance the output.

After seeing this, my first thought was that it might be possible to do the same thing, but with the OPA2132's that I have.  But - this circuit has caps on the output - doesn't this defeat the purpose of the servo completely? In this particular circuit - is there any benefit of having the servo on the output of the INA217 feeding that voltage follower?

Other balancing circuits I've seen use both halves of a dual opamp - is this really simple voltage follower (which i assume just inverts the output) even a decent way of achieving a balanced output?

Is this idea just rubbish and not something I should follow?  

Another Pro Audio forum uses 0P07's w the that Corp.  1510 & 1512 xfmrlesss preamps with good success.

JamesHE said:

Just getting back into DIY, and wanting to build a 2 channel Pre with some INA217's that I have on hand.
I've gone through my boxes and I have everything I need for this except for the OPA137's for the DC servo.
(also minus 1M resistors, but I figure I can wire up a few in series just while i experiment)

circuit for reference: http://www.ti.com/lit/ds/symlink/ina217.pdf 

Can I simply use  NE5532 for the servo? I have plenty of those.  If so,  any problem in using one NE5532 for servo's for both channels?

Click to expand...

You can do anything you want BUT if you inspect the general concept of servos 553x is not the optimal choice.  Servos generally use an op amp to shift a high impedance RC to lower effective impedance.  A good op amp for this will generally have low input current (noise, bias, and offset) suggesting the use of a FET input stage. Further you want good DC characteristics for the servo to actually hit close to 0V so old early bifets are not the first choice either (while I have used them in servos last century). 

There are probably several good choices from the modern FET input op amps.

JR

<<If you can handle an SO8 surface mount package, perhaps with an adapter board, there are far better servo amps out there. My favorite is the Analog Devices ADA4522, available in singles, as well as the ADA4522-2 dual. It has extremely low offset, low noise, low distortion, and extremely low 1/f noise, everything you want in a servo amp>>

All things equal, can you take a wild guess as to whether this is an audible improvement over the opa137?

Low offset = hit song.

Ah ..... I was just getting used to my PRR signature line and now this?