OPA627 + BUF634 Preamp PCB Design

[carlosfm]

Referring to the original cct of the thread - OPA627 and Buffer. This is a line stage with GAIN, right? Why on earth wouldn't you use a OPA637 for the same price it's 5 times the GBW and has better +PSRR by 14 dB (error on datasheet)!

By all means, use the OPA637.
If you can live with a line preamp with a gain of 5...
I can't. Too much.

You don't need to compromise it with 200Pf and you don't need the buffer.

Hold on, here you show that you don't know what you're talking about.
It actually sounds better with the cap.

No need to subject the signal to any more abuse then, is there?

No, there isn't.
The OPA627/37 sounds better biased to class-A at around 10ma, or with the BUF634 in wide BW mode.
Just pick one.

 

[amplifierguru]

Carlosfm,

I've been using the OPA627/637 in pre- and power amplifiers since it's release in 1990! At one stage I was being rationed because they're also used in Patriot Missiles and they were being expended in the Gulf.

The Buffer with the OPA627 needs the C = 200pF which rolls out the overall loop feedback locally around the 627 so likely creates more distortion from the buffer into the output. You like this sound, good for you!

The 637 has more performance and you don't need the buffer in the application. Too much gain - you can always run it inverting with a lower gain and summing point resistor to GND to waste some GBW. But smaller Ccomp ensures it still has better slew rate and PSRR ( and undocumented CMRR).

Bias in Class A - gilding the lilly but no cost!
Cheers.

 

[padamiecki]

Re: High Speed Amplifier Design Application Note

Hi, An explanation for all this (driving a buffer with a high speed opamp) is beautifully explained in this application note by Jim Williams of Linear Technology:
http://www.linear.com/pub/document.html?pub_type=app&document=50
Even features a discrete and better sounding buffer than the BUF634........

Hi Elso
this link is dead now,
maybe do you have another link or could send this article at pawela@statoil.com ?
Thank you in advance

 

[carlosfm]

The Buffer with the OPA627 needs the C = 200pF which rolls out the overall loop feedback locally around the 627 so likely creates more distortion from the buffer into the output. You like this sound, good for you!

The treble lacks precision, it's somewhat "spitty" without the cap.
BB/TI recommends 200pf, but lower values can be used, at around 47pf.

The 637 has more performance and you don't need the buffer in the application. Too much gain - you can always run it inverting with a lower gain and summing point resistor to GND to waste some GBW. But smaller Ccomp ensures it still has better slew rate and PSRR ( and undocumented CMRR).

With the BUF634 the pre will drive long cables much better.
For those (like me) who prefer to have the power amp(s) as close as possible to the speakers, with short speaker cables, this is the way to go.

Anyway, I'm not using the OPA627+BUF634 as a line preamp anymore.
The AD815 is (sounds) superior.

Bias in Class A - gilding the lilly but no cost!

 

[Elso Kwak]

Re: Re: High Speed Amplifier Design Application Note

Hi Elso
this link is dead now,
maybe do you have another link or could send this article at pawela@statoil.com ?
Thank you in advance

Hi New link:
http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1154,C1009,C1028,P1219,D4138
I am going bananas from all these changes!
(Or search for AN47 in technical support on that site.)

 

[Fixup]

Alcaid, it is a feedback amplifier, if you use ground plane or ground fill be careful with strayed capacitances, at least from the inverting inputs and from the outputs. Though everything still could be well with a few tens of pF at both points... And though you can always stick with Zobel... My last experience tells me that even without feedback the things with analog circuits may (more info soon) be better without ground plane.

Pedja

Hi Pedia, and others:

Why analog circuits may be better without ground plane? I tend to agree, but would like to hear more thoughts/experience from others.

 

[PMA]

Hi Pedia, and others:

Why analog circuits may be better without ground plane? I tend to agree, but would like to hear more thoughts/experience from others.

They are not, just a myth. Groundplane is more than useful for link level circuits.

 

[peranders]

Why analog circuits may be better without ground plane? I tend to agree, but would like to hear more thoughts/experience from others.

Groundplane is in most cases good but if a node is very sensitive against stray capacitance to ground you may be forced to remove some.

I don't know about the TPA6120 but Texas recommends to have a groundplane free area under the inputs. See the pictures.

Remember also that 1.6 mm FR4 PCB has 3.3 pF per cm2 so what does it mean to have a SMD pad over a groundplane? 50-100 fF?

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

 

[sonnya]

The only reason i could come up with, is the very very low distortion will be compromised.

But then again, the ground plane should be removed in the complete feedback path to.

- Sonny

 

[peranders]

My last experience tells me that even without feedback the things with analog circuits may (more info soon) be better without ground plane.

Did you come up with something scientific or practical about using extremely fast circuits without groundplane?