A Differential Deboo Integrator to Servo Mic and Low-Level Preamps

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

A Differential Deboo Integrator to Servo Mic and Low-Level Preamps

Post by mediatechnology »

I'm looking at using a THAT1570 or THAT1583 in a design without the digital gain controller and needed a servo.

...So I came up with a fully-differential form of the Deboo Integrator we used to servo the THAT1510.

You may recall that the THAT1510 provides a single-ended output to feed the servo but requires a differential output (at least in this topology) to feed the preamp inputs: viewtopic.php?f=6&t=14&start=114

The THAT1570 and THAT1583 have differential outputs and also require a fully-differential servo to feed the 1570 inputs.

Deboo describes a differential input for his grounded capacitor integrator in this 1969 patent drawing. This topology provides a single-ended output.

Image
Deboo Differential Input Single-Ended Output Grounded Capacitor Integrator, 1969, US Patent 3470495.

Read the 1969 patent by Gordon J. Deboo here: http://www.waynekirkwood.com/images/pdf ... 470495.pdf

What I needed was a floating capacitor, differential input and output, single-capacitor, Deboo integrator.

Image
A Fully-Differential Floating Capacitor Deboo Integrator to Servo the THAT1570 Differential In/Differential Out Mic Preamp

Unlike Deboo's differential input topology, the capacitor in the "fully-differential Deboo" floats.

Deboo's topology is modified to become fully differential using an instrumentation amp that provides positive feedback into both input legs.
The INA DC differential gain (without positive feedback) is "2." [1+{(10K+10K)/20K}]
The common mode gain is "1."

Though it does not remove the 1570's "-1" Vbe common mode offset, the differential offset error is held very low over a ~60dB gain range and eliminates the requirement for a Cgain capacitor.

I'm typically seeing <15mV output Vos at low gain and <3mV at high gain in a non-optimized test circuit.
I'm temporarily using 10K Rfb (x2) for the 1570 (which is on the high side) in order to re-use the stepped gain switch I built for the 1510.

Note that the non-inverting response of the requires the outputs to be cross-coupled to opposing inputs i.e. the integrator half for Out 1 connects to the In 2 current injection resistor. (100K Rinj, 1K1 Rbias both not shown.)
For it to have the same response as the 1510 circuit I think the 1uF needs to be lowered to 0.47uF since there are now 1M resistors in both legs.

I'm going to keep playing with this because I think it has a lot of potential.
THAT does not show a servo circuit for the 1570 or 1583 and this might fill the bill.

Edit: I think I also see a simple transformation to the "inverting" configuration.
User avatar
JR.
Posts: 3697
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: A Differential Deboo Integrator to Servo the THAT1570/15

Post by JR. »

Interesting.. I seem to recall another inspection of Deboo here a while back...

One concern is that the noise of the following differential section is not LP filtered or attenuated by the integration capacitor. This is probably already attenuated by the pad betweenn it and the input of the gain stage, since you don't need +/- 15v of correction range. At least the noise of all the 1M resistors is LPF.

Good luck... at some point you may need to name this a "Kirkwood servo"... :D

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A Differential Deboo Integrator to Servo the THAT1570/15

Post by mediatechnology »

Thanks JR.

I just re-built the "differential Deboo" using 1% resistors and the overall offset correction is far better than I thought. (Helps to practice what one preaches. :lol:)
I'm seeing the differential output offset pegged at <<1 mV at all gain ranges with either a 10K or 150R input termination.
One concern is that the noise of the following differential section is not LP filtered or attenuated by the integration capacitor. This is probably already attenuated by the pad between it and the input of the gain stage...
True - but we've got that all-important passive pole at the input in return.
See this post for a discussion on the limitations of the simple integrator using slow DC-precise op amps: viewtopic.php?f=6&t=419&start=19


I've been looking in the 1510 circuit for the servo noise penalty and it appears to be attenuated significantly by the source impedance to the point of not being a factor. With 100Kx2 injection resistors (which seems low but read on) and a 150R source the servo noise is attenuated by >60 dB. And the OPA2277 is pretty quiet.
With input capacitors 1M/leg Rinj would be plenty.
...since you don't need +/- 15v of correction range.
DC-coupled input-capacitorless I found that with the Behringer ECM8000 I do. It uses with 100K x2 Rinj about half that much: +/-7.

I did try the "inverting" differential Deboo (by rearrangement) and it's not that workable. Don't know why but it also lacks the passive input pole.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A Differential Deboo Integrator to Servo the THAT1570/15

Post by mediatechnology »

Just did a +60dB gain noise comparison with a 150R source and ~100 kHz measurement BW.

Servo vs. Cgain the difference was perhaps 0.25 dB in favor of the 4700uF/6.3V.

EDIT: Found an error in referencing one of the input bias resistors: There's no measurable difference with or without the servo.

Also added 100K "T-bias" and there is no measurable noise difference with or without.
User avatar
JR.
Posts: 3697
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: A Differential Deboo Integrator to Servo the THAT1570/15

Post by JR. »

Sweet, yes I like real poles early and often,,,

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A Differential Deboo Integrator to Servo the THAT1570/1583

Post by mediatechnology »

mediatechnology wrote: Sun Mar 05, 2017 3:21 pm "Consider The Deboo Integrator For Unipolar Noninverting Designs," Elliot Simons, Maxim Integrated Products, Electronic Design, March 18, 2002.

The non-inverting Deboo integrator has a passive input pole which is important for audio servos.
Other advantages of the Deboo Integrator are its non-inverting single capacitor topology.


https://www.proaudiodesignforum.com/ima ... esigns.pdf

Image
"Consider The Deboo Integrator For Unipolar Noninverting Designs," Elliot Simons, Maxim Integrated Products, Electronic Design, March 18, 2002.

Related reading:

A Differential Deboo Integrator: viewtopic.php?f=6&t=559
A Direct-Coupled Input-Capacitorless Active Mic Preamp using a DIfferential Deboo Integrator: viewtopic.php?f=6&t=598
Let's Discuss Servos (Previously OPA188 Thread) viewtopic.php?f=6&t=419

search.php?keywords=deboo
search.php?keywords=servo
In addition to the Simons article quoted above I recently found this TI citation:

Image
Howland Deboo Integrator, TI AN-1515.

Image

"AN-1515 A Comprehensive Study of the Howland Current Pump" http://www.ti.com/general/docs/lit/getl ... leType=pdf

Note resistors KR1 which can be scaled to increase gain.
User avatar
terkio
Posts: 304
Joined: Thu Feb 07, 2019 3:17 am
Contact:

Re: A Differential Deboo Integrator to Servo Mic and Low-Level Preamps

Post by terkio »

There is a catch in the Deboo integrator, differential or single.
It is sensitive to resistor mismatch.
This can be important or not, depending of the application.
When in need for near perfect integrators, one needs capacitors with very high insulation resistors. Using a Deboo configuration, the unavoidable resistor mismatch results in an additional parasitic resistor across the capacitor of a positive or negative high resistance.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A Differential Deboo Integrator to Servo Mic and Low-Level Preamps

Post by mediatechnology »

Using a Deboo configuration, the unavoidable resistor mismatch results in an additional parasitic resistor across the capacitor of a positive or negative high resistance.
Use precision resistors or a trim if its that critical.
I recall that the resistor mismatch error results in a small static DC offset.

In fairness we should point out the short-comings of other popular integrator topologies.

The simple integrator (R, C, Op Amp) has an "inductive" response when the op amp runs out of bandwidth.
A differential output simple integrator requires two matched resistors to avoid a small static DC offset.

The "2 C" "true" differential integrator requires capacitor and resistor matching to avoid non-monotonic response to step changes.

In most cases I would prefer a small DC offset and the passive input pole of a single-capacitor Deboo to the non-monotonic response of a two capacitor true integrator with active poles.
User avatar
terkio
Posts: 304
Joined: Thu Feb 07, 2019 3:17 am
Contact:

Re: A Differential Deboo Integrator to Servo Mic and Low-Level Preamps

Post by terkio »

I compared using a differential Deboo integrator, and a non inverting two caps integrator.
Integrating capacitor(s) insulation resistor does result in DC offset error in both cases.
With the Deboo the resistor mismatch brings a similar effect. With the non inverting two caps, there is no such resistor mismatch consequence and no need for matching the two caps.

In both cases, the DC offset error, not only depends of the capacitor leak, but depends of the op amps Vos, Ios, Ibias too.

By no mean I wanted to make an unfair point onto the Deboo. The point I raised is only a part of a design decision where all aspects must be considered.
I am not aware of bandwith and monotony issue about the non inverting integrator, I will look into these.
My point was to point at the resistor matching required in the Deboo.
I found a lot mentioned the formula R1/R2 = R3 R4 giving a perfect integrator but nothing about what goes on with real world resistor accuracy where it is unavoidably, about equal.
All one can find on the internet is the same incomplete presentation.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: A Differential Deboo Integrator to Servo Mic and Low-Level Preamps

Post by mediatechnology »

Take a look here at some servo bandwidth tests I did with an OP07: https://proaudiodesignforum.com/forum/p ... t=20#p4999

You might want to read the full thread.
Post Reply