The Opamp thread

[murrays]

 
Quote:

your EF2 runs off a wall wart, LT1363/4 has a low PSRR...it will not sound good.
 
Try AD797B/LT1028AC/ADA4627-1B/OPA827/etc?

That's not so much of a problem if the amp has good local regulation (does it?). Bear in mind that the LT1363 has a PSRR is 90 to 100 dB.
A lot of the point of PSRR is not supply ripple rejection, but rejection of voltage changes due to reflection of the changing output current.
If you're having to rely on really high PSRR because of large levels of noise and fluctuations at the power inputs then your circuit design is lacking.  Good power supply filtering and local HF bypassing will eliminate most of that.
A good test is to measure (or even better - listen to) the AC voltages on the power supply pins of the op-amp. Ideally there should only be DC (zero hertz) but it can be a real eye-opener to find that you are feeding audio back into the power supply inputs.  Even a high PSRR won't block all of that.

 

[Charminbaer]

Hi, ive been reading your posts for a while now. Now i want to participate actively.
 
I´ve got a question regarding AD797BR in a Matrix m-stage.
I tried it and it sounded great but i got a feeling, that the longer i keep hearing the more distorted it sounds, especially in complex musical scenes.
Could this be due to heat? Or is it just not stable? What can i do to stabilize it?
 
I found this tutorial on stabilizing an AD797: http://www.acoustica.org.uk/t/naim/35clockreg.html
 
 
Quote:

Between pins 6 and 8 we can connect a small capcitor (47pF) which closes an internal feedback loop and has the effect of reducing distortion by reducing output impedance dramatically at high frequency

 
He also talks about some other tweaks but i dont know if they work for the m-stage circuit
 
Since i am no engineer i dont know what he exactly talks about. Could someone here with greater knowledge explain to me what to do?
 
Is there a way to get it stable without soldering something on the board itself? Something like just soldering a resistor to the legs of the opamp?
 
I would be happy if someone could give me some advice on this.

 

[dlaloum]

Quote:

Is there a way to get it stable without soldering something on the board itself? Something like just soldering a resistor to the legs of the opamp?
 

You can use the same method as used for the Class A mod - get a socket - do the soldering on the socket - plug the opamp into the socket and the socket into the board.
 
This way if you make a mess - you toss the socket - no risk to the opamp or board, and sockets are cheap

 

[Charminbaer]

hi,
 
yes i´ve  thought about that too , i also have some spare sockets.
Isn´t this, in the tutorial mentioned mod, a Class A Bias ?
 
 
PS:
 
I played around with the DIP-switches and it seems that almost all stability issues are gone when set to +20db.

 

[murrays]

 
Quote:

Hi, ive been reading your posts for a while now. Now i want to participate actively.
 
I´ve got a question regarding AD797BR in a Matrix m-stage.
I tried it and it sounded great but i got a feeling, that the longer i keep hearing the more distorted it sounds, especially in complex musical scenes.
Could this be due to heat? Or is it just not stable? What can i do to stabilize it?
 
I found this tutorial on stabilizing an AD797: http://www.acoustica.org.uk/t/naim/35clockreg.html
 
He also talks about some other tweaks but i dont know if they work for the m-stage circuit
 
Since i am no engineer i dont know what he exactly talks about. Could someone here with greater knowledge explain to me what to do?
 
Is there a way to get it stable without soldering something on the board itself? Something like just soldering a resistor to the legs of the opamp?
 
I would be happy if someone could give me some advice on this.

 
I recommend Tangent's website, which has a good write-up on dealing with "cranky" op-amps http://tangentsoft.net/audio/hs-opamp.html.  If you find it difficult to understand you may need to search out some op-amp tutorials to fill in some of the basics.
 

 

[murrays]

Quote:

I played around with the DIP-switches and it seems that almost all stability issues are gone when set to +20db.

 
I presume that was an increase in gain.  That is one established way of improving stability. 
The only real downside is that it may reduce the usable range of your volume control.
 

 

[qusp]

yep, compensation and bandwidth limiting are often needed to help the 797 cope with lower audio frequencies and lower gain. 797 works quite well with class A bias, but it doesnt always benefit many modern chips, as they are mostly designed to work in class A as it is.
 
loving THS4150 at the mome (fully differential chip) this range of chips contains much of the same circuitry as the discrete solid state Pass Labs super symmetry (SUSY) designs, so much so that they had to buy a license from Nelson in order to use it. handy little group of chips, the VCOM pin servos the output to the voltage applied to it, easy offset adjustment. they will also directly take a single ended input with the - signal pin tied to ground and output fully balanced audio
 
but most importantly, it sounds excellent!! no grain, slams like a mofo and dynamics for miles

 

[dlaloum]

Quote:

yep, compensation and bandwidth limiting are often needed to help the 797 cope with lower audio frequencies and lower gain. 797 works quite well with class A bias, but it doesnt always benefit many modern chips, as they are mostly designed to work in class A as it is.
 
loving THS4150 at the mome (fully differential chip) this range of chips contains much of the same circuitry as the discrete solid state Pass Labs super symmetry (SUSY) designs, so much so that they had to buy a license from Nelson in order to use it. handy little group of chips, the VCOM pin servos the output to the voltage applied to it, easy offset adjustment. they will also directly take a single ended input with the - signal pin tied to ground and output fully balanced audio
 
but most importantly, it sounds excellent!! no grain, slams like a mofo and dynamics for miles

What have you compared it to?
 
Would it work in fairly standard buffered opamp layout (ie in my M-Stage) or would it require additional work / mollycoddling (like the 797's apparently often do)?
 
(considering whether to add it to my extensive opamp trialing list)

 

[Avro_Arrow]

Be warned the THS4150 has a completely different pin out from a single op amp.

 

[dlaloum]

Quote:

Be warned the THS4150 has a completely different pin out from a single op amp.

As in the standard Dual Op amp pin out (such as OPA2134) - or something completely different?

 

[Avro_Arrow]

Most Dual op amps share a common pin out.
Most Single op amps also share a common pin out which is different from dual op amps.
Even the BUF634 has a pin out compatible with single op amps in some situations.
THS4150 does not share the pinout of any dual or single op amp.
You would not be able to roll this into a circuit which was not designed for it.
 
Quote:

As in the standard Dual Op amp pin out (such as OPA2134) - or something completely different?

 

[qusp]

Quote:

What have you compared it to?
 
Would it work in fairly standard buffered opamp layout (ie in my M-Stage) or would it require additional work / mollycoddling (like the 797's apparently often do)?
 
(considering whether to add it to my extensive opamp trialing list)

THS4131, PCM1632 and thats about it, I dont sit down rolling opamps so much these days, its really not the best way to gain good results IMO, but its fun for a while. this one I just knocked up a simple datasheet-o-like circuit on perf initially to test (powered with 4 x A123 lifepo4 cells) and i've just rolled it in in place of OPA1632. a couple of days ago. 
 
the guys are right, its not going to be much good unless you at least have some way of changing the power and signal pinout of your socket and are able to add different decoupling etc its actually pretty well behaved, given some bandwidth limiting and resistance to buffer it from any capacitive loading, yes it needs a totally different set up to what you will have installed. there arent all that many fully differential chips around, this one is more flexible than some too given what I said above. basically if you can bend the circuit to supply power (it will actually work with single supply as well) to the right pin and are able to pull the output from it, it will give you balanced output fro SE input with a single supply.needs at least 16-20v to come into its own though (+/-8-10v)

 

[Charminbaer]

 
Quote:

 
 
I recommend Tangent's website, which has a good write-up on dealing with "cranky" op-amps http://tangentsoft.net/audio/hs-opamp.html.  If you find it difficult to understand you may need to search out some op-amp tutorials to fill in some of the basics.
 

 
Thank you for this very informative link.
 
So in my case i have to connect a 4.7uf Tantalum and a 0.1uf ceramic capacitor in parallel as close as posible to pin 4 and from there to ground. The same for pin 7. As described on Page 13 in the AD797 Datasheet.
And i have to measure the resistance of R1 and R2 , then fill them into the formula provided from tangent to get the right capacitance. And this, preferably a polypropylene, is to be put between pin 2 and 6.
 
Is that correct ?
 
Ps: The pin-layout is for a single opamp, so whats about a Dip8 socket? what has to be modified in that case ?
      I found a schematic : http://cimarrontechnology.com/%5Cpdf%5C020302sch.pdf
      The polypropylene has to be put between pin 1/2 and another one between pin 6/7?  The tantalum&ceramic have to go from pin4 to ground and from pin8 to ground?
     
 

 

[murrays]

Quote:

Thank you for this very informative link.
 
So in my case i have to connect a 4.7uf Tantalum and a 0.1uf ceramic capacitor in parallel as close as posible to pin 4 and from there to ground. The same for pin 7. As described on Page 13 in the AD797 Datasheet.
And i have to measure the resistance of R1 and R2 , then fill them into the formula provided from tangent to get the right capacitance. And this, preferably a polypropylene, is to be put between pin 2 and 6.
 
Is that correct ?
 
Ps: The pin-layout is for a single opamp, so whats about a Dip8 socket? what has to be modified in that case ?
      I found a schematic : http://cimarrontechnology.com/%5Cpdf%5C020302sch.pdf
      The polypropylene has to be put between pin 1/2 and another one between pin 6/7?  The tantalum&ceramic have to go from pin4 to ground and from pin8 to ground?

 
The pinouts for the 8 pin package are on the front page of the AD797 datasheet. They are the same for DIP and SOIC.  Pins 4 and 7 are the power inputs.  There is no pin for ground - you will have to find the nearest ground (zero volts) point on the circuitboard.
I suggest starting with the power supply decoupling which are the 2 pairs of 4.7uf Tantalum and a 0.1uf ceramic capacitors in parallel that you mentioned from the datasheet.  That may be enough.  Otherwise, the capacitor from pin 2 to pin 6 will be for bandwidth limiting in the feedback loop.
This can be more difficult if you are using chip adapters, as the bypassing components should be as close to the chip as possible.  Some adapters have facility to mount bypassing caps on board.  I like to solder the bypass caps to the underside of the PCB beneath the op-amp socket, directly from the socket pin to the ground point.
Some design techniques work for all op-amps in general (e.g. power supply bypassing), while others are specific requirements of individual chip designs.  Pin 8 on the AD797 is one such individual feature which is outlined in the datasheet. The LT1028 has pin 5 for tuning.

 

[qusp]

sometimes you will find that the pins that are labelled NC or 'not connected' are tied to ground