DOA virgin.

[12volts]

OK. I'm in my 50's, with a wealth of experience in electronics from a very early age. I cut my milk teeth on valves, O.K.
Now, with tons of experience, I have to confess, I've never used discrete operational amplifiers.

What do YOU recommended I should do to fix this gap in my experience and prove they are better than NE5534s?
I have 3 months of winter to devoted experimentation/building.


Frank

 

[shabtek]

stay with valves

 

[thermionic]

The famous Neve in Air Studios contains not the discrete output stages, but 5534s... I like discrete gain blocks as they give me (speaking personally) a multitude of options in terms of the resulting subjective sound. For example, I know of some very basic 3-transistor 'op-amps' which clients like because of their phenominal amount of 2nd harmonic distortion... Personally, I prefer a far more evolved op-amp which competes with ICs in terms of THD and noise - but has the added benefit of high-current driver transistors, enabling me to use it in numerous apps. However, the latter op-amp is deemed 'too clean' by certain clients... You can't win. I even know some people who like TL071s and the like (I certainly don't, speaking subjectively - I find them raw and unmusical).

So, what's my point? Unless you have some stonking discrete op-amp, probably utilising items such as 'supermatched' transistors (which, ironically, are from the IC family in terms of manufacturing technology), it's unlikely to compete with some of the uber-modern ICs from the likes of NSC and LT in terms of THD and noise. Therefore, how could anyone describe DOAs as 'better'? To do so would be purely on subjective merit (aside from output drive and high-voltage, which is where DOAs really score) - which is why I like discrete

(retires and waits for someone to put rocket up my proverbial)

Justin

 

[zebra50]

I haven't really mucked about with DOAs (dead on arrival?), having preferred to do tube projects for much of my DIY

But I do love my (non-DIY) API preamps, which are DOA based. They just sound good on most sources. Of course there could be other reasons why they sound good beyond the DOA, but there you have it.

 

[JohnRoberts]

While this may not be the popular opinion, modern opamps are remarkably linear and precise, so I see little reason to delve into DOA other than personal amusement, if that floats your boat.

JR

PS: The 553x was state of the art for modest cost IC opamps in the 1970s. Opamp technology has moved the goal line since then.

 

[thermionic]

But I do love my (non-DIY) API preamps, which are DOA based. They just sound good on most sources. Of course there could be other reasons why they sound good beyond the DOA, but there you have it.

Many of the pretty respected engineers I know like discrete gear simply because they claim to find IC-based gear, quote 'boring' in many cases. They all know an 8078 is blown away by a J-series in terms of spec, but the 8078 gives them the sound they're after. But as you say, you have to compare apples with apples. You can add transformers to IC gear and make it sound more 'interesting'. Likewise, I can think of quite a bit of unremarkable-sounding discrete gear. Properly designed gear will always win out, whether it's IC, BJT, FET or Valve.

None of what I'm saying will be news to most here - I'm just trying to give perspective on the concept of discrete sounding 'better'. Although I do have a penchant for discrete, if I want transparency, convenience and compactness, ICs are king and I use them all the time.

I haven't really mucked about with DOAs

I'm glad to hear it

 

[thermionic]

While this may not be the popular opinion, modern opamps are remarkably linear and precise, so I see little reason to delve into DOA other than personal amusement, if that floats your boat.

JR

Don't you think discrete blocks have a place on account of being able to run on high voltages, and have large O/P devices that'll drive anything? What about my clients who genuinely want loads of 2nd harmonic, without the hassle of maintaining tube gear? I'm sure the designers who conceived the high-distortion opamp would shudder to think people *want* that distortion today, but I can assure you they do... Not my thing, but what am I to do? Tell them they're wrong? I guess this falls under the scope of 'personal amusement'.

(NB - for those unaware, it's very easy to put driver transistors on the back of an IC et voila: instant headphone / low load ability)

 

[JohnRoberts]

While this may not be the popular opinion, modern opamps are remarkably linear and precise, so I see little reason to delve into DOA other than personal amusement, if that floats your boat.

JR

Don't you think discrete blocks have a place on account of being able to run on high voltages,

Once upon a time, it was useful to support large voltage swing, to deliver good dynamic range in a world of relatively high noise levels. However you can also deliver wide dynamic range by lowering the noise floor (lowering the water instead of raising the bridge).  Perhaps  for live sound reinforcement where you may need to send signals over hundreds of meters with good integrity, but even common IC processes can support +24 dBu without raising a sweat. 

and have large O/P devices that'll drive anything?

For power amps discrete devices are still indicated for higher power points, while modern class D is getting closer to an IC solution, but IMo still years off for decent output power.

What about my clients who genuinely want loads of 2nd harmonic, without the hassle of maintaining tube gear? I'm sure the designers who conceived the high-distortion opamp would shudder to think people *want* that distortion today, but I can assure you they do... Not my thing, but what am I to do? Tell them they're wrong? I guess this falls under the scope of 'personal amusement'.

Indeed.. and heaven forbid I suggest that there are plenty of low voltage ways top make all kinds of distortion. Another truism is that the customer is always right. If they need to see that heater glow to separate them from their geld, or lots of bitty parts hanging off small squares of FR4, do what it takes. 

Call it personal amusement, style, fashion... whatever... My argument is that DOA are not needed for pure performance reasons in a modern system. If you make a fashion decision to drive a bunch of 600ohm transformers in parallel, you will may need discrete buffers as a minimum, but again this is not pure performance decision, and some of the modern uber-opamps have robust output drive capability.

If you are making EFX, you have a blank page to work from, do what works for you. 

(NB - for those unaware, it's very easy to put driver transistors on the back of an IC et voila: instant headphone / low load ability)

JR

PS: I can appreciate the sense of accomplishment from discrete design, I've done a lot over the years. Lately I have been seduced by the power of microprocessors to embed some simple intelligence and decision making into designs. In fact for many applications the digital approach wins on cost by a healthy margin, while providing features impossible or impractical using just analog. 

 

[conleycd]

What about some hybrid designs like JLM's hybrid opamp?  There he uses an OPA2604 and hangs BD139 and BD140 transistors on the tail end of it all.  Of course... some parts to make it all work.  I figure just about any opamp would work in the above noted circuit - remembering not to blow up the IC opamp with too much voltage.

The advantage is the ability to drive low loads (like a transformer - especially a 1:2 step up or headphones even).

So... keep your NE5532 and throw some transistors on the end (keep it to +/- 18v of course).

CC

 

[PRR]

For perspective: discrete WAS very valid 1960s-1970s. First there were no chips, then expensive poor performers, then cheap poor performers.

5534 set a level that few discrete designs really beat for "quality". (Yes, a "bad" unbalanced pair has ear-tickle virtue.)

Chip economics do not favor low-low voltage noise and high output drive. Deane's 990 "squandered" Silicon area to get better in both areas than the chip-makers could print at a profit. (But note that most 990s use a "chip" front-end: the LM394 matched-pair is a 100-transistor integrated circuit for wealthy fussy customers.) Deane also threw in a couple inductors which is not possible in IC fabrication to get a compensation which is novel and useful and discretely simple, but can be done other ways if demand justifies the development cost. (Many snazzy new ICs have both the noise and bandwidth that Deane was hoping for.)

So as a practical matter.... if you got this far with 12AX7 and 741 and 5532 and OPA2604, then dinking with a dozen 3-leg parts on an inch of board is not essential.

> OPA2604 and hangs BD139

There are DSL chips which will make that kind of power (assuming the BDs are naked not heat-sunk).

They usually aim at lower voltage and maybe higher current than some of the old heroic discrete consoles used, so that particular combination may be a best-fit for some legacy sockets. In a clean-sheet design, instead of 1:2 step-up you could run 1:3 and get hot lines with +/-15V rails. With today's lower noise voltages, you don't need such whopping signal voltages except maybe outside the console.... long-lines will always be a special case, always too-good for 99% of venues and not-good-enough for the 1% nasty-gigs (welder's convention).

 

[strangeandbouncy]

HI,


  My comments here are without prejudice . . . They are also unqualified, except by my opinion, but then I am in the trenches under fire actually using the stuff every day . . .

  I have a lot of gear, discrete, ic and tube. I have some bits of gear that I have lashed 2520 or OA10 in instead of 5534, etc. I know which bits of gear get the most use! Anything where a lot of gain is made, it is DOA every time, and in eqs, there is a very perceivable difference, especially at the top end. Now I do accept that it may well be the Failings of said DOAs that add something special, and I am sure that the newer ics( National LMExxxx etc) measure better, but you can prise my 2520s( I have at least 3 different original types and some diy) and OA10s from my cold dead hand!

  Kindest regards,


    ANdyP

 

 

[conleycd]

HI,


   My comments here are without prejudice . . . They are also unqualified, except by my opinion, but then I am in the trenches under fire actually using the stuff every day . . .

   I have a lot of gear, discrete, ic and tube. I have some bits of gear that I have lashed 2520 or OA10 in instead of 5534, etc. I know which bits of gear get the most use! Anything where a lot of gain is made, it is DOA every time, and in eqs, there is a very perceivable difference, especially at the top end. Now I do accept that it may well be the Failings of said DOAs that add something special, and I am sure that the newer ics( National LMExxxx etc) measure better, but you can prise my 2520s( I have at least 3 different original types and some diy) and OA10s from my cold dead hand!

  Kindest regards,


     ANdyP

   

I don't mean to be facetious... but are there many occasions to use an opamp (or even a DOA opamp) where gain is not involved?  Most of my preamps that I value have DOAs (original 2520 and then my own grown ones).  I don't have any current eqs (some in the works) so I can't compare but eqs use opamps as gain as well. 

I agree that potential "flaws" or unintended things can sound 'good' or pleasing.  I almost always want 'character'.

Is the thought that using DOAs as line receivers is not audible or running them at unity gain at the end of a signal chain to drive a transformer is not really 'audible'?

With respect to "pure sound" discussion - this drove me nuts in my guitar circles - I just stopped discussing it.  No one buys a pedal to make their electric guitar sound the same.  I think what people are really trying to articulate with this "pure" discussion talk is character that is somewhat predictable, not too obtuse, and subjectively pleasant.  The same is true in the studio - FWIW.

CC

 

[analag]

OK. I'm in my 50's, with a wealth of experience in electronics from a very early age. I cut my milk teeth on valves, O.K.
Now, with tons of experience, I have to confess, I've never used discrete operational amplifiers.

What do YOU recommended I should do to fix this gap in my experience and prove they are better than NE5534s?
I have 3 months of winter to devoted experimentation/building.


Frank

Well Frank, so you got everybody's opinion. Want mine?
Fix your gap by building and testing/blowing up a few DOA's...I did and it was totally worth the experience. As far as DOA vs NE55xx or whatever flavor may be popular I offer you this. I use what's available and build what is not. So I will build DOA's to work with increased voltage rails/ increased current drive capability and anything else that may tickle my fancy. I too cut my teeth on tubes...they are such simple devices one must step into other things to learn more. Start your journey.

 

[analag]

Imagine a mic pre with DOA's swinging 60V or 70V peak to peak can you go get it at the local parts store. YOU have to make it if YOU can. So if you're not like the majority of DIYers who live inside the box I say experiment, experience and learn from it, like anything else in life.