Mic Preamp DC-Servo: The "backwards" way

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Mic Preamp DC-Servo: The "backwards" way

Post by mediatechnology »

JR. wrote:Yes I do not appreciate caps in series with Rg, but if you put two caps between the feedback network and gain stage - inputs they don't need to be as large as in series with Rg . You will need another pair of caps in series with the gain stage outputs and a separate feedback resistor for DC feedback. Since the entire feedback net is no longer in the DC path, no more scratch or clicks.

I've played with such circuits in Jurassic times. But the extra caps in the feedback network make the circuit a higher order LF filter and its difficult to avoid serious peaking on some gain settings.

If you've successfully used this, please post a sketch or some clue on how you did it.
Wouldn't there be an example of that technique here in the P10 preamp?
viewtopic.php?f=6&t=423&p=7924

Going back to what Olaf is trying to accomplish is a way to make the bias resistors very large for common mode signals using bootstrap while also injecting servo correction and having those resistors not lower Rcm.
My hunch is the only capacitors desired are the 2u2 films at the input with a very high Rcm so they don't degrade CMRR.
I'm still thinking about this...
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Mic Preamp DC-Servo: The "backwards" way

Post by JR. »

ricardo wrote:
JR. wrote:Yes I do not appreciate caps in series with Rg, but if you put two caps between the feedback network and gain stage - inputs they don't need to be as large as in series with Rg . You will need another pair of caps in series with the gain stage outputs and a separate feedback resistor for DC feedback. Since the entire feedback net is no longer in the DC path, no more scratch or clicks.
I've played with such circuits in Jurassic times. But the extra caps in the feedback network make the circuit a higher order LF filter and its difficult to avoid serious peaking on some gain settings.

If you've successfully used this, please post a sketch or some clue on how you did it.
I have not used it myself, but I described it to a mic preamp chip manufacturer and he said he knows of two companies using his chip sets that way.

The caps need to be decent sized to keep current noise down, but they don't need to be as large as a cap in series with Rg that can drop to single digit ohms at high gain. The pole of interest is the cap working into the added DC feedback resistor.
I've used this trick to get higher order rumble filters into RIAA preamps.
Don't get me started on phono preamps... I am familiar with underdamped HPF to hit 20 Hz RIAA numbers and still get infra-sonic roll off below 20 hz where RIAA doesn't care.. I kind of liked the IEC approach to add a real pole at 30Hz but the audiophools want their 20Hz. :lol: I was selling tape noise reduction back then and when you compress a bunch of <20Hz signal then send it to tape... it doesn't come back, which causes a phantom modulation in the playback expander caused by the missing LF signal.

JR
Argue with something I actually believe... The customers punish clicks and pops by not buying the products.

Note: Even though I do not appreciate caps in series with Rg I have done it in many value designs because the customers dislike clicks and pops more.
I too am guilty of this EVIL and they weren't even 'value designs'. Mea culpa, mea culpa, mea maxima culpa. :o

It's really just my 'Broadcast' snobbery vis a vis da 'Music' guys. The unwashed masses rarely switch gain when tracking or mastering with the 'tape' running .. while the Broadcast guys often have to.

In da 60's & 70's, if the Studio Manager (BBC-speak for sound engineer) moved a fader during a Proms broadcast, there would be letters to the Times (OK HiFi News) from Colonel Auric Pinnae & Co.

"Sir, I distinctly heard you touch a fader just before the drum-roll on last night's Haydn Toy Symphony broadcast. And why are you using those evil new fangled transistorised AKG414s instead of C12s?"
______________

Olaf, you should really read Wayne's huge thread on simple servo on THAT 1510/12. It covers practically all the important points and has many 'real life' measurements.

I think its viewtopic.php?f=6&t=256 though there are other threads which expand on certain points.
______________

It appears THAT have incorporated many of these points in the system around THAT5171
Cancel the "cancel culture", do not support mob hatred.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Mic Preamp DC-Servo: The "backwards" way

Post by JR. »

mediatechnology wrote:
JR. wrote:Yes I do not appreciate caps in series with Rg, but if you put two caps between the feedback network and gain stage - inputs they don't need to be as large as in series with Rg . You will need another pair of caps in series with the gain stage outputs and a separate feedback resistor for DC feedback. Since the entire feedback net is no longer in the DC path, no more scratch or clicks.

I've played with such circuits in Jurassic times. But the extra caps in the feedback network make the circuit a higher order LF filter and its difficult to avoid serious peaking on some gain settings.

If you've successfully used this, please post a sketch or some clue on how you did it.
Wouldn't there be an example of that technique here in the P10 preamp?
viewtopic.php?f=6&t=423&p=7924
Similar in concept but not exactly... In the P10 the collector feed from the inout transistors (or drain from JFETs are cap coupled to the opamp. This way the DC output is independent of device Vbe. But that does not obviate scratchy pots or clicky gain switches. Now I am talking about caps between the emitters and feedback resistors. The transistors would need to get their operating current elsewhere and the whole feedback network is floating.

Maybe I'll draw something up this weekend but I don't feel like making one.

JR
Going back to what Olaf is trying to accomplish is a way to make the bias resistors very large for common mode signals using bootstrap while also injecting servo correction and having those resistors not lower Rcm.
My hunch is the only capacitors desired are the 2u2 films at the input with a very high Rcm so they don't degrade CMRR.
I'm still thinking about this...
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Mic Preamp DC-Servo: The "backwards" way

Post by mediatechnology »

Now I am talking about caps between the emitters and feedback resistors. The transistors would need to get their operating current elsewhere and the whole feedback network is floating.
OK, I understand now.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Mic Preamp DC-Servo: The "backwards" way

Post by JR. »

micpre.PNG
micpre.PNG (43.38 KiB) Viewed 22250 times
OK I have roughed up a cut and paste using an app note.

The servo is not used.

Caveat, this is proforma so may not work as drawn...and values are arbitrary. Added DC feedback resistors should be large so they don't step on gain.

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Mic Preamp DC-Servo: The "backwards" way

Post by mediatechnology »

Thanks for the clarification John.
So are the values of the DC feedback resistors relatively large compared to Ra and Rb?

I ask because I can't see the values clearly...
I had envisioned RaDC and RbDC being a low-ish value and split with a capacitor from the middle to ground forming an R-C-R "T" to make the loop through that path DC-only.
I have no clue how high Ra and Rb can be at DC.
IIRC the 10K upper limit was due to an AC parameter.

Providing a separate AC feedback loop - putting the coupling capacitors between the emitters and Rg - makes sense and doesn't seem to have too much of a penalty performance-wise.
I'm sure there's a small noise penalty but to put it in perspective the 1570/1573 IIRC has about 30 Ohms internally in series with its Rga and Rgb pins.
Anything to eliminate that big honkin' electrolytic in series with Rg (that works) is worthy.
olafmatt
Posts: 44
Joined: Wed Oct 22, 2014 11:33 am
Location: Europe

Re: Mic Preamp DC-Servo: The "backwards" way

Post by olafmatt »

Making RaDC and RbDC large has an upper limit since there is the collector current across them. So at some point you'll have too much voltage drop across them (i.e. the inst. amps output sitting at an extreme DC offset eating your headroom).
But the THAT1570 in JRs example shows internal constant current sources anyway, so it seems there is no need for an extra DC path to set the collector current (in contrast to the circuit I posted in the first post, which does set the collector current through the feedback resistors). Of course this is assuming that the internals of the chip are what it says in this simplified drawing.

Olaf
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Mic Preamp DC-Servo: The "backwards" way

Post by JR. »

mediatechnology wrote:Thanks for the clarification John.
So are the values of the DC feedback resistors relatively large compared to Ra and Rb?
Yes as large as we can get away with.
I ask because I can't see the values clearly...
Because there are no values there... TBD
I had envisioned RaDC and RbDC being a low-ish value and split with a capacitor from the middle to ground forming an R-C-R "T" to make the loop through that path DC-only.
Except any AC R to ground at the emitters would increase the minimum AC gain, so perhaps an application specific decision.. DC feedback in parallel decreases maximum gain, RCR network increases minimum gain... pick your poison.

[edit- oops.. my over-night engineering team was reviewing this while I was out riding my bike, and they observed that the response to common mode signals would be dramatically different between the two approaches. So RCR would at a minimum require precision tolerance for the resistors connected to the emitters. [/edit]
I have no clue how high Ra and Rb can be at DC.
IIRC the 10K upper limit was due to an AC parameter.
The DC voltage at that stage is not critical.. I would grab the audio from the top of the floating switched resistor network. A simple one op amp diff amp would provide a DC path to ground for the switched resistor network.
Providing a separate AC feedback loop - putting the coupling capacitors between the emitters and Rg - makes sense and doesn't seem to have too much of a penalty performance-wise.
It probably wants to be low Z wrt hundreds of ohms but that is better than low Z wrt tens of ohms.
I'm sure there's a small noise penalty but to put it in perspective the 1570/1573 IIRC has about 30 Ohms internally in series with its Rga and Rgb pins.
Anything to eliminate that big honkin' electrolytic in series with Rg (that works) is worthy.
Modern electrolytic caps are better thanks to process improvements related to use in switching PS (Low ESR low ESL) but in a perfect world they shouldn't have to carry audio into 10 ohms...

JR
Last edited by JR. on Fri Oct 24, 2014 2:28 pm, edited 1 time in total.
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Mic Preamp DC-Servo: The "backwards" way

Post by mediatechnology »

I bet that the RaDC and RbDC DC feedback resistors can be fairly high compared to the AC Ra and Rb.
ricardo
Posts: 262
Joined: Sun Apr 10, 2011 7:24 am

Re: Mic Preamp DC-Servo: The "backwards" way

Post by ricardo »

JR. wrote:OK I have roughed up a cut and paste using an app note.
I'm not sure 1570/1573 is the best demo of your floating feedback network.
  • The 1570/1573 has its own anti-click strategies which allow direct connection between the i/p emitters to Rg
  • The 2 separate feedback scheme is very similar to how I used to implement high order rumble filters so you would need to check LF response carefully .. both for cut off as well as peaking at various gains. You may need my 1mF CGO/NPOs in 0603 size after all :D
_____________________

Wayne, I seem to remember, your final Servo used 2 x OP07 instead of the single you show on this thread.
viewtopic.php?f=6&t=256

Where did you put the final circuit?
Post Reply