Glenn,
The HD970 you posted is a classic implementation of a DC servo.
Your DAC uses the second opamp in exactly the same way but it is also part of the filter, that is why there is no "integrator" function as such. It both sets the DC conditions and is a vital part of the filter implementation.
If you want to check just short the output of the second opamp to ground and check the DC conditions in that state and also check the filter response 🙂
The HD970 you posted is a classic implementation of a DC servo.
Your DAC uses the second opamp in exactly the same way but it is also part of the filter, that is why there is no "integrator" function as such. It both sets the DC conditions and is a vital part of the filter implementation.
If you want to check just short the output of the second opamp to ground and check the DC conditions in that state and also check the filter response 🙂
If the OPA2134 can't drive line level loads with minimal distortion, I'd just have used a better opamp...
And anyway there are high-end players that use the OPA(2)134 to drive their line outs without anything in between.
For driving headphones... there are better solutions! At least I'd want an usable volume control
At any rate, try the LME49723 as it's very good. I expect the LME49725 to be very good too, but we'll see.
Wrong. From what you say, all the OA are "DC servo" because all of them do work like you describe - keep the output at "zero" via negative feedback...
That's not what a servo does. If you model that schematic and apply 1V DC at the input you will see that the output will NOT stay at zero. A real servo, with intergrator, will keep the output at zero no matter what.
Hi SoNic_real_one
An apology, I have made a fundamental (daft) statement 🙂 suggesting that shorting the output of the second opamp will cause an offset problem. It won't as it gives pin 5 a zero reference.
However the circuit as seen does indeed need this second opamp to clamp (or more correctly set) the output to zero volts... we can't see what is to the left of the diagram... how is pin 5 voltage (DC) determined to be specific... it may well be AC coupled. As it stands, it is floating without the second IC and connection via R189.
An apology, I have made a fundamental (daft) statement 🙂 suggesting that shorting the output of the second opamp will cause an offset problem. It won't as it gives pin 5 a zero reference.
However the circuit as seen does indeed need this second opamp to clamp (or more correctly set) the output to zero volts... we can't see what is to the left of the diagram... how is pin 5 voltage (DC) determined to be specific... it may well be AC coupled. As it stands, it is floating without the second IC and connection via R189.
From a dc standpoint I don't think the second op-amp is required, grounding as shown in red will provide the required dc reference point, however from an ac standpoint the filter performance will be somewhat compromised. I haven't really thought this out so I can't say to what extent this would be true.
From a quick calculation (means I could be wrong) (asuming the caps dont do much in the audio band) the voltage at the output of the top OA (2) is not ground its -Vout of the first OA (1).
Iin-(2) = Vout(1)/2.2k, Vout(2)=-Iin-(2)x2.2k= -Vout(1) so it ballances the currents (and impedances) of the + and - inputs of the first OA assuming the voltages there are = but opposite. Putting a ground there will change the circuit operation.
Iin-(2) = Vout(1)/2.2k, Vout(2)=-Iin-(2)x2.2k= -Vout(1) so it ballances the currents (and impedances) of the + and - inputs of the first OA assuming the voltages there are = but opposite. Putting a ground there will change the circuit operation.
Last edited:
Of course that output it is not zero in AC - it is feedback on the positive input of the first OA. It makes all symetrical.
My point was that removing the feedback amp and putting in a ground will change the functioning of the circuit.
I got interested in this and turned this up,
Differential Multiple Feedback Bessel Low Pass Filter
which is the same configuration. As I suspected earlier that second opamp is vital to the filter operation. The opamp circuit in the above link has both inverted and non inverted outputs.
Differential Multiple Feedback Bessel Low Pass Filter
which is the same configuration. As I suspected earlier that second opamp is vital to the filter operation. The opamp circuit in the above link has both inverted and non inverted outputs.
I see Mooly finally has found an important piece of the "puzzle"
the circuit is a fully differential amplifier made with the dual op amp and a fully differential multiple feedback low pass filter
single fully differential op amps like THS4130 are popular for high speed signal processing like A/DSL
but you will find fully differential op amps also include a a output common mode servo on the chip
http://focus.ti.com/lit/an/slyt165/slyt165.pdf
the posted circuit fixes the outputs to be symetric with respect to gnd - which means the typical mid supply offset of the DAC outputs causes DC current to be drawn from the DAC +/- output buffers - not always a good thing
using just one side of the differential output you are relying on lots of components ratio matching for common mode rejection/differential-to-single ended conversion accuracy
there can be stability issues using the 2 op amps from a dual to build the fully differential amp - the phase margin may not be high enough - especially with the odd internal compensation of the 5532
the circuit is a fully differential amplifier made with the dual op amp and a fully differential multiple feedback low pass filter
single fully differential op amps like THS4130 are popular for high speed signal processing like A/DSL
but you will find fully differential op amps also include a a output common mode servo on the chip
http://focus.ti.com/lit/an/slyt165/slyt165.pdf
the posted circuit fixes the outputs to be symetric with respect to gnd - which means the typical mid supply offset of the DAC outputs causes DC current to be drawn from the DAC +/- output buffers - not always a good thing
using just one side of the differential output you are relying on lots of components ratio matching for common mode rejection/differential-to-single ended conversion accuracy
there can be stability issues using the 2 op amps from a dual to build the fully differential amp - the phase margin may not be high enough - especially with the odd internal compensation of the 5532
Last edited:
You may also find the below article informative w/ relation to the thread topic
"A Differential Op-Amp Circuit Collection" : http://www.ti.com/lit/an/sloa064/sloa064.pdf
..."Proper operation of high-speed op-amps requires proper decoupling techniques. That does not mean a shotgun approach of using inexpensive 0.1-µF capacitors. Decoupling component selection should be based on the frequencies that need to be rejected, and the characteristics of the capacitors used at those frequencies"
Last edited:
- Status
- Not open for further replies.