Go Back   Home > Forums > Design & Build > Parts

Parts Where to get, and how to make the best bits. PCB's, caps, transformers, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Thread Tools Search this Thread
Old 2nd March 2013, 03:03 PM   #1
diyAudio Member
Join Date: Jun 2003
Location: Hamilton, was Ottawa (Canada)
Default Gilbert's chips, AFA (active feedback amplifier): AD830, LT6552, TDA856X

These are some very interesting parts!

In a sentence, Barrie Gilbert stated that his AFA (active feedback amp) "unlike the OPA (vfb opamp), has high open-loop linearity and excellent closed-loop linearity due to its distortion-canceling topology, and a very high degree of versatility arising from its dual fully-differential inputs, has the potential to eclipse the OPA in all applications involving the manipulation of purely voltage-mode signals."

The purpose of this thread is to get people to try out this topology.
Actual, working designs.

These chips can be used in line-level applications (AD830, LT6552) and head amp/ power amp (TDA856X). abraxalito even said there's one power amp which has a line driver mode selectable by a pin! So I didn't want to put the thread in amp or line level ... 'parts' seemed a good compromise but I hope it gets enough attention ...

Here are some of the chips: AD830, LT6552, TDA856X
(all these are still available)

Let's start with the AD830

AD830 High Speed, Video Difference Amplifier

from the spec sheet:

AD830 connection diagram.png


The AD830 is a wideband, differencing amplifier designed for use at video frequencies but also useful in many other applications. It accurately amplifies a fully differential signal at the input and produces an output voltage referred to a user-chosen level. The undesired common-mode signal is rejected, even at high frequencies. High impedance inputs ease interfacing to finite source impedances and, thus, preserve the excellent common-mode rejection. In many respects, it offers significant improvements over discrete difference amplifier approaches, in particular in high frequency common-mode rejection.


High common-mode rejection ratio (CMRR)
High impedance inputs
Symmetrical dynamic response for +1 and −1 Gain
Low sensitivity to the value of source R
Equal input impedance for the + and − input
Excellent high frequency CMRR
No halving of the bandwidth
Constant power distortion versus common-mode voltage
Highly matched resistors not needed


The AD830 represents Analog Devices first amplifier product to embody a powerful alternative amplifier topology. Referred to as active feedback, the topology used in the AD830 provides inherent advantages in the handling of differential signals, differing system commons, level shifting, and low distortion, high frequency amplification. In addition, it makes possible the implementation of many functions not realizable with single op amp circuits or superior to op amp based equivalent circuits. With this in mind, it is important to understand the internal structure of the AD830.

AD830 fig 26 topology diagram.png

The topology, reduced to its elemental form, is shown in Figure 26.

Nonideal effects, such as nonlinearity, bias currents, and limited full scale, are omitted from this model for simplicity but are discussed later. The key feature of this topology is the use of two, identical voltage-to-current converters, GM, that make up input and feedback signal interfaces. They are labeled with inputs VX and VY, respectively. These voltage-to-current converters possess fully differential inputs, high linearity, high input impedance, and wide voltage range operation. This enables the part to handle large amplitude differential signals; it also provides high common-mode rejection, low distortion, and negligible loading on the source.

The label, GM, is meant to convey that the transconductance is a large signal quantity, unlike in the front end of most op amps. The two GM stage current outputs, IX and IY, sum together at a high impedance node, which is characterized by an equivalent resistance and capacitance connected to an ac common. A unity voltage gain stage follows the high impedance node to provide buffering from loads. Relative to either input, the open-loop gain, AOL, is set by the transconductance, GM, working into the resistance, RP; AOL = GM RP. The unity gain frequency, ω0 dB, for the open-loop gain is established by the transconductance, GM, working into the capacitance, CC; ω0 dB = GM/CC. The open-loop description of the AD830 is shown below for completeness.

Precise amplification is accomplished through closed-loop operation of this topology. Voltage feedback is implemented via the Y GM stage where the output is connected to the −Y input for negative feedback, as shown in Figure 27.

AD830 fig 27 closed-loop connection.png

An input signal is applied across the X GM stage, either fully differential or single-ended referred to common. It produces a current signal that is summed at the high impedance node with the output current from the Y GM stage. Negative feedback nulls this sum to a small error current necessary to develop the output voltage at the high impedance node. The error current is usually negligible, so the null condition essentially forces the Y GM output stage current to equal the exact X GM output current. Because the two transconductances are identical, the differential voltage across the Y inputs equals the negative of the differential voltage across the X input; VY = −VX or, more precisely, VY2 − VY1 = VX1 − VX2. This simple relation provides the basis to easily analyze any function possible to synthesize with the AD830, including any feedback situation.

The bandwidth of the circuit is defined by the GM and the capacitor, CC. The highly linear GM stages give the amplifier a single-pole response, excluding the output amplifier and loading effects. It is important to note that the bandwidth and general dynamic behavior is symmetrical (identical) for the noninverting and the inverting connections of the AD830. In addition, the input impedance and CMRR are the same for either connection. This is very advantageous and unlike in a voltage or current feedback amplifier where there is a distinct difference in performance between the inverting and noninverting gain. The practical importance of this cannot be overemphasized and is a key feature offered by the AD830 amplifier topology.

And here are some of the circuits from the spec sheet:

Connection diagram

AD830 connection diagram.png

Gain of 2

AD830 gain 2.png

AC coupled line receiver

AD830 ac coupled line receiver.png

Diff amp level shift

AD830 diff amp level shift.png

Much of this technical discussion is a bit beyond me so I'm hoping to attract some heavy-hitters to help summarize, explain, apply, and design some actual circuits. Now, if you want someone to poorly implement the design and then wax on about how good it sounds ... well I'm your man!

Anyone have any ideas? Ways to apply these circuits? New circuits that use this topology/chip?

Here's abraxalito with an application:

Opamps tend not to like the high levels of RF coming out of DACs, I recommend passive (LC) filtering before the opamp - OPA627 is one of the better ones at dealing with RF. Better still though an AFA like AD830, this one is highly RF resistant and sounded great when fed with TDA1545.



P.S. And here are Gilbert's two articles:

Are Opamps Really Linear?

Are Op Amps Really Linear.pdf

Nonlinear Effects of Radio-Frequency Interference in Opamps

Nonlinear Effects of Radio-Frequency Interference in Opamps opampsusc_01.pdf

Next up, LT6552 and TDA856X.

P.S.S. Here are some threads referencing the AFA that I'll draw from (if you know of any more, please add them):

Active Feedback Amplifier

AD1865 schematic

Powering a high-end USB DAC


Possibly the most frugal high-end sounding amp?


Last edited by AudioLapDance; 2nd March 2013 at 05:06 PM.
  Reply With Quote
Old 2nd March 2013, 04:54 PM   #2
diyAudio Member
Join Date: Jun 2003
Location: Hamilton, was Ottawa (Canada)
A similar chip (but this spec sheet has a simplified schematic!):


The LT6552 is a video difference amplifier optimized for low voltage single supply operation. This versatile amplifier features uncommitted high input impedance (+) and (–) inputs and can be used in differential or single-ended configurations. A second set of inputs gives gain adjustment and DC control to the differential amplifier.


The LT6552 is a video difference amplifier with two pairs of high impedance inputs. The primary purpose of the LT6552 is to convert high frequency differential signals into a single-ended output, while rejecting any common mode noise. In the simplest configuration, one pair of inputs is connected to the incoming differential signal, while the other pair of inputs is used to set amplifier gain and DC level. The device will operate on either single or dual supplies and has an input common mode range which includes the negative supply. The common mode rejection ratio is greater than 60dB at 10MHz. Feedback is applied to Pin 8 and the LT6552’s transient response is optimized for gains of 2 or greater.

LT6552 fig 3 single supply, fig 4 dual supply.png

Figure 3 shows the single supply connection. The amplifier gain is set by a feedback network from the output to Pin 8 (FB). A DC signal applied to Pin 1 (REF) establishes the output quiescent voltage and the differential signal is applied to Pins 2 and 3. Figure 4 shows several other connections using dual supplies. In each case, the amplifier gain is set by a feedback network from the output to Pin 8 (FB).


LT6552 simplified schematic.png

Figure 5 shows a simplified schematic of the LT6552. There are two input stages; the first one consists of transistors Q1 to Q8 for the (+) and (–) inputs while the second input stage consists of transistors Q9 to Q16 for the reference and feedback inputs. This topology provides high slew rates at low supply voltages. The input common mode range extends from ground to typically 1.75V from VCC, and is limited by 2VBE’s plus a saturation voltage of current sources I1-I4. Each input stage drives the degeneration resistors of PNP and NPN current mirrors, Q17 to Q20, that convert the differential signals into a single-ended output. The complementary drive generator supplies current to the output transistors that swing from rail-to-rail.

The current generated through R1 or R2, divided by the capacitor CM, determines the slew rate. Note that this current, and hence the slew rate, are proportional to the magnitude of the input step. The input step equals the output step divided by the closed-loop gain. The highest slew rates are therefore obtained in the lowest gain configurations. The Typical Performance Characteristic Curve of Slew Rate vs Closed-Loop Gain shows the details.
SHDN Pin (shutdown)
The LT6552 includes a shutdown feature that disables the part, reducing quiescent current and making the output high impedance. The part can be shutdown by bringing the SHDN pin within 0.5V of V–. When shutdown the supply current is typically 400mA and the output leakage current is 0.25mA (V– VOUT V+). In normal operation the SHDN can be tied to V+ or left floating; if the pin is left unconnected, an internal FET pull-up will keep the LT6552 fully operational.

Same thing here, anyone have any ideas? Ways to apply these circuits? New circuits that use this topology/chip?


Last edited by AudioLapDance; 2nd March 2013 at 05:07 PM.
  Reply With Quote
Old 2nd March 2013, 05:29 PM   #3
diyAudio Member
Join Date: Jun 2003
Location: Hamilton, was Ottawa (Canada)
Next is the TDA856x

Here's the selection chart:

TDA856x selection.png

What is DDD?
DDD = Dynamic Distortion Detection

(From TDA8561Q datasheet, we'll see later)
Diagnostic output (pin 16)
At the onset of clipping of one or more output stages, the dynamic distortion detector becomes active and pin 16 goes LOW. This information can be used to drive a sound processor or DC volume control to attenuate the input signal and thus limit the distortion. The output level of pin 16 is independent of the number of channels that are clipping (see Figs 3 and 4).

TDA856x selection AN10327.pdf

abraxalito is playing with the TDA8561 but also references the TDA8566 because it's still available at Mouser.


General description

The TDA8566 is an integrated class-B output amplifier which is available in several packages. TDA8566TH is contained in a 20-lead small outline plastic package. The TDA8566TH1 is a 24-lead small outline plastic package which is pin compatible with the I2C-bus controlled amplifier TDA1566TH for one board layout. TDA8566Q is a 17-pin DIL-bent-SIL package.

The device contains 2 amplifiers in a Bridge-Tied Load (BTL) configuration. The output power is 2 25 W in a 4 W load or 2 40 W in a 2 W load. It has a differential input stage and 2 diagnostic outputs. The device is primarily developed for car radio applications.


Differential inputs
Very high Common Mode Rejection Ratio (CMRR)
High common mode input signal handling
Requires very few external components
High output power
4 W and 2 W load driving capability
Low offset voltage at output
Fixed gain
Diagnostic facility (distortion, short-circuit and temperature pre-warning)
Good ripple rejection
Mode select switch (operating, mute and standby)
Load dump protection
Short-circuit proof to ground, to VP and across the load
Low power dissipation in any short-circuit condition
Thermally protected
Reverse polarity safe
Protected against electrostatic discharge
No switch-on/switch-off plops
Low thermal resistance
TDA8566TH1 is pin compatible with TDA1566TH


The TDA8566 contains 2 identical amplifiers and can be used for BTL applications. The gain of each amplifier is fixed at 26 dB. Special features of this device are:

Mode select switch
Clip detection
Short-circuit diagnostic
Temperature pre-warning
Open-collector diagnostic outputs
Differential inputs


Mode select switch (pin MODE)

Standby: low supply current
Mute: input signal suppressed
Operating: normal on condition

Since this pin has a very low input current (< 40 mA), a low-cost supply switch can be applied. To avoid switch-on plops, it is advisable to keep the amplifier in the mute mode for a period of 150 ms (charging the input capacitors at pins IN1+, IN1-, IN2+ and IN2-). This can be realized by using a microcontroller or by using an external timing circuit as illustrated in Figure 8.


Clip detection (pin CLIP)

When clipping occurs at one or more output stages, the dynamic distortion detector becomes active and pin CLIP goes LOW. This information can be used to drive a sound processor or a DC volume control to attenuate the input signal and so limit the level of distortion. The output level of pin CLIP is independent of the number of channels that are being clipped. The clip detection circuit is disabled in a short-circuit condition, so if a fault condition occurs at the outputs, pin CLIP will remain at a HIGH level. The clip detection waveforms are illustrated in Figure 5.


Temperature pre-warning (pin DIAG)

When the virtual junction temperature (Tvj) reaches 145 C, pin DIAG will become continuously LOW.


Here's the Block Diagram

TDA8566 block diagram.png

And a stereo BTL circuit

TDA8566 stereo BTL circuit.png

As mentioned, abraxalito is playing with the TDA8561:


Here are the first measurements - the Nitro amp being fed by the Ozone DAC with a six-tone multiple, peak digital level -20dB. Peak output power around 100mW in an 8R speaker (connected for this test). It looks from this that the datasheet is a little pessimistic on THD, though here I'm measuring the TDA8566, not the 8561. No measurement graphs are presented in the 8566 DS and the 8561 only has plots for 2R and 4R loads, not 8R.


The TDA8561 schematic


Actual circuit!
Opus11 TDA856 amp Nitro1.jpg

And the mods for the TDA8566 (at Mouser)


Additional schematic showing how to use TDA8566Q in place of TDA8561Q as the latter isn't available from Mouser. Mouser only has fewer than 300 TDA8566Q in stock so best to get in quick - I doubt they'll be getting more as this chip is going obsolete. The pins shown unconnected in the mod schematic aren't unconnected - it means they're connected the same way as for the TDA8561Q.

Here's some of the reasoning behind the choice of the TDA856x from abraxalito (opus11):
Possibly the most frugal high-end sounding amp?

Having built chipamp based amps for a number of years, they produce great but not stellar sound in the main, when issues like grounding and decoupling are addressed fully. I have the hypothesis that what holds them back is the input stage, being as it is typically a LTP and a bipolar one at that. Added to that, they do tend to have inadequate PSRR by virtue of not having power supply separation between the signal and output stages. So they pollute their own patch, so to speak and filters can't be included because the necessary wires aren't brought off-chip.

I have up until fairly recently, played with the TDA7293 because its fairly unique in allowing separation between the voltage and current amplifying stages, especially so when the amplification function is split between two chips. This device is the only one I know of that permits that so the power supply issue can be fully fixed. It does however still have the ubiquitous LTP input stage which has given me so many problems in opamps in terms of excluding HF interference.

The search for a chipamp sans LTP has been running for a few weeks now and recently I've come upon some very interesting parts from Philips - the TDA856X series

Just focusing on the lack of LTP rather than AFA topology ...

Last edited by AudioLapDance; 2nd March 2013 at 05:59 PM.
  Reply With Quote
Old 2nd March 2013, 06:22 PM   #4
diyAudio Member
Join Date: Jun 2003
Location: Hamilton, was Ottawa (Canada)
Default Gilbert's chips, AFA (active feedback amplifier): AD8129, AD8130

Knew I'd miss one!

AD8129 AD8130.pdf

AD8129 AD8130 connection diagram.png

AD8129 AD8130 circuit.png


The AD8129/AD8130 use an architecture called active feedback, which differs from that of conventional op amps. The most obvious differentiating feature is the presence of two separate pairs of differential inputs compared with a conventional op amp’s single pair. Typically, for the active feedback architecture, one of these input pairs is driven by a differential input signal, while the other is used for the feedback. This active stage in the feedback path is where the term active feedback is derived.

The active feedback architecture offers several advantages over a conventional op amp in many types of applications. Among these are excellent common-mode rejection, wide input common-mode range, and a pair of inputs that are high impedance and completely balanced in a typical application. In addition, while an external feedback network establishes the gain response as in a conventional op amp, its separate path makes it completely independent of the signal input. This eliminates any interaction between the feedback and input circuits, which traditionally causes problems with CMRR in conventional differential-input op amp circuits.

Another advantage is the ability to change the polarity of the gain merely by switching the differential inputs. A high input-impedance inverting amplifier can be made. Besides a high input impedance, a unity-gain inverter with the AD8130 has a noise gain of unity. This produces lower output noise and higher bandwidth than op amps that have noise gain equal to 2 for a unity-gain inverter.

The two differential input stages of the AD8129/AD8130 are each transconductance stages that are well matched. These stages convert the respective differential input voltages to internal currents. The currents are then summed and converted to a voltage, which is buffered to drive the output. The compensation capacitor is in the summing circuit.
When the feedback path is closed around the part, the output drives the feedback input to the voltage that causes the internal currents to sum to 0. This occurs when the two differential inputs are equal and opposite; that is, their algebraic sum is 0.

In a closed-loop application, a conventional op amp has its differential input voltage driven to near 0 under nontransient conditions. The AD8129/AD8130 generally has differential input voltages at each of its input pairs, even under equilibrium conditions. As a practical consideration, it is necessary to limit the differential input voltage internally with a clamp circuit.
Therefore, the input dynamic ranges are limited to about 2.5 V for the AD8130 and 0.5 V for the AD8129 (see the AD8129/AD8130 Specifications section for more detail). For this and other reasons, it is not recommended to reverse the input and feedback stages of the AD8129/AD8130, even though some apparently normal functionality may be observed under some conditions. A few simple circuits can illustrate how the active feedback architecture of the AD8129/AD8130 operates.
  Reply With Quote
Old 2nd March 2013, 06:40 PM   #5
diyAudio Member
Join Date: Jun 2003
Location: Hamilton, was Ottawa (Canada)
Default Gilbert's Pin, AD844 pin5 TZ

While not an AFA chip, this one I'll call:

Gilbert's Pin, AD844 pin5 TZ

From an old thread (2004!) with ... Holy Heavy Hitters Batman! Charles Hansen, John Curl, Walt Jung, ... and the old gang! Jocko, PMA, CarlosFM, Pedja, Elso

DIYHiFi.org &bull; View topic - "Op-amps unsuitable for audio DAC I/V" - myth?


AD844 simplified schematic:

AD844 simplified schematic.png

CG describes the issue:

"The AD844 is a special case. It's a current feedback type amplifier. The transimpedance node is connected to a pin on the package, which allows the part to be used open loop (actually, as the AD797 can be) if you want to run it that way. There may be good reasons to do so... Thank Barrie Gilbert for following Wyn Palmer's lead on making this pin available. (Wyn was a serious DIY audio guy at one time.) ..."

Gilbert's Pin!

CG continues with some great description:

"Back to the AD844 thing for those who didn't read the previous thread. Or threads.

The AD844, like the AD846 before it, use a classic CFB topology. The non-inverting input is a complementary pair of bipolar devices driven at their bases. The collectors are connected to current mirrors that ultimately connect to the Tz node. This Tz node is available at one of the pins. From the Tz node to the output is a complementary buffer stage. Or stages. The inverting input is at the connection of the emitters for the complementary input devices. It's shown in the data sheet.

So... If you connect the inverting input to ground through a resistor and add a resistor to ground from the Tz pin, you can build a pretty nice open loop amplifier. The devices inside are nicely matched because of the IC fabrication process. You can use the buffer as is or add your own externally..."

Georgehifi has been trying the AD844 in this exotic configuration with his PCM1704
Using the AD844 as an I/V

AD844 iv PCM1704 tina circuit pin 5.png

and I'm about to try it with my TDA1541 using Pedja's old circuit

Pedja 1541 NOS AD844IV.png

Pedja 1541 NOS AD844IV wBUFFER.png

Last edited by AudioLapDance; 2nd March 2013 at 06:59 PM.
  Reply With Quote
Old 2nd March 2013, 07:21 PM   #6
diyAudio Member
jan.didden's Avatar
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
I agree that the AD844 is a very interesting chip. I used it in my error correction power amp as published by Elektor and AudioXpress. Basically, when you run it open loop with a load at the Tz node, you can have a defined gain with no loop feedback.

As to the AD830 and company, I believe that history shows us that Barrie's enthousiasm was not universially appreciated, to be diplomatic.
Maybe that is because it really is not THAT linear - you always need a certain voltage difference on the input pins and those input stages are not very linear.

At one point when I was in contact with Barrie, I asked for a super-duper AD844 successor. He answered that they had one 'on the shelf' but that demand probably would not be high enough to take it in production.
Don't you wish you had a $1M to spare to spend on IC's

Never explain - your friends don't need it and your enemies won't believe you anyway - E. Hubbart
Check out Linear Audio Vol 7
  Reply With Quote
Old 2nd March 2013, 07:41 PM   #7
bcarso is offline bcarso  United States
diyAudio Member
Join Date: May 2005
Location: Canoga Park, California
Originally Posted by jan.didden View Post
Don't you wish you had a $1M to spare to spend on IC's

  Reply With Quote
Old 2nd March 2013, 07:52 PM   #8
diyAudio Member
Join Date: Jun 2003
Location: Hamilton, was Ottawa (Canada)
Originally Posted by jan.didden View Post
...As to the AD830 and company, I believe that history shows us that Barrie's enthousiasm was not universially appreciated, to be diplomatic.
Maybe that is because it really is not THAT linear - you always need a certain voltage difference on the input pins and those input stages are not very linear..
This is true, however as abraxalito points out they may be linear enough (diminishing returns) and they shine where many conventional topologies start to stumble--high freq performance, in our cellular/wifi world.
  Reply With Quote
Old 2nd March 2013, 08:30 PM   #9
diyAudio Member
Join Date: Jun 2003
Location: Hamilton, was Ottawa (Canada)


"...the concept I am running hard here with is noise modulation. When I have a hypothesis I try my damndest to break it, flog it to death, test it in many ways to see how strong it is. This is one such test of the 'noise modulation is the missing SQ metric' hypothesis. So the high THD of this chip works in my favour.


THD is the metric that Scott's using, he, like Doug Self pays little or no attention to what (IME) really matters for SQ which is low level IMD performance. So you're of the school that -100dB THD is a non-negotiable requirement for high-end sound? If so then there are a whole lot of high-end amps (considered subjectively) which violate that by huge margins!"

Jan, can you say more about your AD844 circuit?
  Reply With Quote
Old 2nd March 2013, 09:06 PM   #10
diyAudio Member
jan.didden's Avatar
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Well, the '844 circuit in isolation is simple - but it is embedded in an output stage (and there is also a Vas stage) error correction scheme.

Never explain - your friends don't need it and your enemies won't believe you anyway - E. Hubbart
Check out Linear Audio Vol 7
  Reply With Quote


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off

Similar Threads
Thread Thread Starter Forum Replies Last Post
Active Feedback Amplifier kyrgeo Solid State 9 2nd March 2013 01:37 PM
Chips for active cross overs ChrisA Chip Amps 10 5th April 2010 07:55 AM
Barrie Gilbert's articles 00940 Digital Source 4 16th April 2005 03:35 AM
Multiple chips, one feedback resistor xplod1236 Chip Amps 12 24th July 2004 04:55 PM
Active Feedback on Parallel Chips Stocker Chip Amps 8 4th May 2004 09:40 PM

New To Site? Need Help?

All times are GMT. The time now is 04:11 PM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2