[sdiy] Op amp notes

[Paul Higgins]

On Saturday, September 4, 2004, at 01:55 AM, Cary Roberts wrote:

> I don't believe in monolithic opamps for critical audio stages.  I 
> prefer
> the API 2520, Hardy 990, Forssell JFET992, JLM99V, Quad Eight AM10, UA 
> 1108,
> etc.  There was a time not too long ago when I didn't know the 
> difference
> and thought a Mackie full of 4560s sounded just fine.  Then again, done
> right, designs employing monolithics can still sound fine.  My D&R 
> Orion is
> chock full of TL074 and TL072 with a bipolar transistor pair and 5532 
> for a
> mix buss and it still sounds great.

Aren't nearly all boards, regardless of price, chock full of 
monolithics?  I seem to recall that even SSLs have tons, likely 
thousands, of 5532s in them.  Not to mention what must be literally 
hundreds of VCAs.

> Fred Forssell has the schematic to his original JFET opamp (same as 
> Millenia
> still uses) on his website.  You can also find Melcor 1731, QE AM10, 
> and
> Jensen 990 (good luck making it work) schematics on the web.

I have some of that stuff archived, but haven't built any of it.  It 
looks interesting, though.  Nelson Pass has a good article on building 
discrete high-voltage opamps, too.

> See above.  SSM2017 weren't bad when available, although these days you
> gotta use the SSM2019 or INA217.  Neve/Focusrite used 5534s in various
> designs with great success.  I actually have a little 5534 circuit with
> follower pair that is a drop in replacement for API2520 and compatible
> discretes.  I can't hear much of a difference in EQ circuits but you 
> can
> tell with mic pre high gain circuits.

Well, after hearing so many list members' (justifiable) frustration 
over SSM's capricious policy of sudden-death discontinuation of their 
parts, I try to avoid using anything from SSM.  I'm kind of also to the 
point where if I can't get the part from Mouser or preferably Digi-Key 
(I'm in MN, where they're located), I'm not too inclined to use it.  
Especially with the high minimum orders many suppliers have.  Digi-Key 
carries lots of Burr-Brown/TI stuff, though, which I have used.  (Isn't 
the INA217 a Burr-Brown part?).  Interestingly, the OP-27s I've got are 
from both B-B and Analog Devices.  It must not be a patented part 
anymore.  It is quite high-performance when it comes to S/N; like I 
said, I have used it (along with a FET) as a on-board guitar 
pre-preamp.  That's a fairly critical application; the pre-preamp feeds 
a hot-rodded Mesa/Boogie Quad Preamp (among other things), which is 
about as high-gain as you're going to get!

I believe you on the mic preamp applications; nearly all well-received 
designs use discretes in the front-end differential stages, and it 
certainly makes sense from just a S/N perspective.  And then there are 
the input transformer zealots...

> Other than building some Mackie-like circuits with 4560s and some 
> generic
> TL0xx buffer circuits I don't have much experience using monolithics 
> for
> audio.  I haven't seen the need to spend time tweaking monolithics when
> building '70s vintage discrete audio designs are so much more fun.

I build a lot of stuff with tubes and hybrid circuitry, so I know what 
you mean.  Still, when you want to build a four-band fully parametric 
EQ, etc. monolithics make life a lot easier.  MHO is that monolithics 
are fine in line-level circuits (as long as you don't use total junk).  
Of course, I was mostly asking list members about line-level synth 
applications, though I'm of course interested in pro audio uses as well.

-PRH