ON Semiconductor 5534

[clintrubber]

--

 

[JohnRoberts]

I believe measurements tell more than enough.

My criticism of 5532/4 can be "measured" in open loop transfer function (specifically phase response), PSRR, input bias current, input offset current, ein, etc .

Good, and perhaps better, but not best and not great.

There are much better parts available today. IMO

JR

 

[CJ]

OP90

 

[Samuel Groner]

Perhaps some are interested in facts rather than guessings and opinions: opamp_measurements.zip

"Transfer" is a measurement in inverting mode and noise gain of 60 dB, "Common Mode" the same but noninverting and "Output" inverting with added resistive load. "Slew rate" is unity gain inverting and noninverting.

These are preliminare results--the "full edition" (which I hope to have ready in about a week) will include THD+N vs. amplitude measurements as well as FFT and time domain analysis of the residual. Note that there is something goofy about the OPA627 slew-rate plot.

What we can say: the AD797 is hard to beat.

Samuel

 

[clintrubber]

[quote author="mediatechnology"]What happened to Peter?[/quote]
Nothing really, just regretted it that I had contributed to making this thread go where it shouldn't have gone.

Regards,

Peter

 

[clintrubber]

[quote author="mediatechnology"]Don't expect measurements with equipment as good as Samuel's though...[/quote]
Wouldn't THAT have a spare & calibrated AP handy that they can lend to you ? For the benefit of all.

 

[magicchord]

I replaced the 5534s in the summing circuits of my ol' ADM console with LT1115s, to avail myself of the low noise and low noise impedance of that particular chip. It worked, too: the bus noise was reduced by around 6 - 8 dB. It looks like the AD797 would be "all-that" for that application as well. The AD797 is more than $9 while the LT1115 is a bit less than $3. I wonder if the AD797 would be more than 6-dollars'-worth of improvement, eh...?

 

[Rossi]

[quote author="clintrubber"] just regretted it that I had contributed to making this thread go where it shouldn't have gone.[/quote]

??? can you elaborate? I think this is an interesting thread, even though I have nothing substantial to contribute. I'm greatly impressed by Samuel's diligence.

I have to admit that I still like the 5532 a lot and usually find no reason to replace it in some piece of gear. But that's my limited experience. Maybe it is time to move on to more advanced opamps for new designs.

 

[JohnRoberts]

[quote author="magicchord"]I replaced the 5534s in the summing circuits of my ol' ADM console with LT1115s, to avail myself of the low noise and low noise impedance of that particular chip. It worked, too: the bus noise was reduced by around 6 - 8 dB. It looks like the AD797 would be "all-that" for that application as well. The AD797 is more than $9 while the LT1115 is a bit less than $3. I wonder if the AD797 would be more than 6-dollars'-worth of improvement, eh...?[/quote]

The AD797 and LT1115 look like roughly same ein so no I don't expect significant noise difference between them in practice.

Is that price comparison apples to apples or distorted by distribution channels?

JR

 

[magicchord]

Hey, John...

I was quoting the small-quantity prices from DigiKey for the 8-dip versions. I guess I was hinting that it would be nice to see how the LT1115 did on Samuel's distortion tests in comparison to the other chips he's tested :grin:

 

[clintrubber]

[quote author="Rossi"][quote author="clintrubber"] just regretted it that I had contributed to making this thread go where it shouldn't have gone.[/quote]

??? can you elaborate? I think this is an interesting thread, even though I have nothing substantial to contribute. I'm greatly impressed by Samuel's diligence.[/quote]

It's hard to comment without triggering the subject that I tried to avoid triggering... :roll: :wink:
Interesting thread indeed, but in my feeling it breathes the implication that (with all respect!), diligent & elaborate measuring tells it all (whether or not intended that implication, but that's at least how it felt for me) - just as the Douglas Self opamp-data I previously linked to.

I don't think this thread needed to be infected with yet a new listening vs measuring debate etc, nobody needs that. For that reason this message will self-destruct within 741 seconds.

Regards,

Peter

 

[clintrubber]

[quote author="mediatechnology"]Hmm....[/quote]
Should we start a poll to convince them ?

 

[Guest]

[quote author="Samuel Groner"]Perhaps some are interested in facts rather than guessings and opinions: opamp_measurements.zip

"Transfer" is a measurement in inverting mode and noise gain of 60 dB, "Common Mode" the same but noninverting and "Output" inverting with added resistive load. "Slew rate" is unity gain inverting and noninverting.

These are preliminare results--the "full edition" (which I hope to have ready in about a week) will include THD+N vs. amplitude measurements as well as FFT and time domain analysis of the residual. Note that there is something goofy about the OPA627 slew-rate plot.

What we can say: the AD797 is hard to beat.

Samuel[/quote]

Wow, thank you!!!

ARE YOU GOING TO TRY AS WELL 4560 AND 4562?

 

[clintrubber]

[quote author="mediatechnology"]Peter - You may find this story interesting. Back when I worked for SSL we, like many other manufacturers, had to use Raytheon 5534s. These parts had this sputtering 1/f popcorn whatever you want to call it noise. They sounded awful.[/quote]
Wayne,

After having replaced more than a few of them you'd most likely be qualified to answer this question: who all made 5532's, 5534's with the RC-prefix ?
As I understood at least Fairchild & Raytheon did, and probably others as well. Have a PCB full of RC..., but can't identify the brand.

Thanks/regards,

Peter

 

[clintrubber]

Thanks Wayne for responding, much appreciated.

Here's a thread with pics in which they're all together
( http://www.groupdiy.com/index.php?topic=26083 )
but here we have a better usable zoom:

And here's a datasheet:

I won't directly base my decision for gutting or re-using the analog section of this box on the brand of opamps in it presently (I'd be a pityful guy), but it'd at least good to know who we're dealing with.

Regards,

Peter

 

[clintrubber]

[quote author="mediatechnology"]Peter - those look like all the Raytheons I used to change from the package to the markings. I still remember a few that got stuck in the end of my thumb. Ouch![/quote]
Thanks for having a look at it. Hope you're thumb has healed by now :wink: Makes me remember a trainee here that ended up with the footprint of a DIL-16 on his forehead but that's another story...

Most likely be taking the route as MartyMart did in the linked thread above.

and I can't recall seeing Fairchild 553X in the wild between 1975-1995.

The PMI SSM2402-ICs in that box say it's from the early nineties ('9240'),
so then Fairchild's definitely out.

Just curious - what is the copyright date on the datasheet? This looks to be the pre-National Fairchild. (Fairchild now again being independent of NSC.)

It's from 1998.

I may cascade all those samples and do a null test. Or, maybe I'll just ship some to Sam....

Just read a review of the Neve 8803 EQ the other day, using 'the venerable NE5532s' and as stated embellishing the sound already with all EQ-bands disengaged...

http://www.soundonsound.com/sos/jun07/articles/amsneve8803.htm
Pulling them into class A ? Some TX in there ? :roll:
What's really different here than the signal struggling its way to a few dozens of ICs in a channelstrip ? The supply will be more tidy & less crosstalk etc for a separate box, but...

Can well imagine these ICs don't do too much harm when you have to put something into the signalpath, but while it'll be true, as HR says it over there is still a bit puzzling.

BTW A Google search for "ON semiconductor 5534" (with no quotes in search term) points to this thread as the first hit. We must have some clout with Google here.

Who's mad enough to talk opamp-brands ?! Just a few forums like this one I guess... 'real people' sit on the couch in front of the TV! :wink:

Regards,

Peter

 

[clintrubber]

dbl post

 

[clintrubber]

OK, then both AMS-Neve & the 'real' Neve (Rupert) must be doing the same thing... (except for RN adding iron to that)
...but what brand would they be using ?

Regards,

Peter

 

[Guest]

[quote author="Wavebourn"][quote author="Samuel Groner"]Perhaps some are interested in facts rather than guessings and opinions: opamp_measurements.zip

"Transfer" is a measurement in inverting mode and noise gain of 60 dB, "Common Mode" the same but noninverting and "Output" inverting with added resistive load. "Slew rate" is unity gain inverting and noninverting.

These are preliminare results--the "full edition" (which I hope to have ready in about a week) will include THD+N vs. amplitude measurements as well as FFT and time domain analysis of the residual. Note that there is something goofy about the OPA627 slew-rate plot.

What we can say: the AD797 is hard to beat.

Samuel[/quote]

Wow, thank you!!!

ARE YOU GOING TO TRY AS WELL 4560 AND 4562?[/quote]

Hi again Samuel;

are you here?

 

[Samuel Groner]

I don't have them on my priority list as I would expect the LM4562 to be similar to the LME49860 and because I'm more interested in the high-performance parts (which makes the NJM4560 loose priority). I'm happy to test them as well though once the rest is done. Unfortunately I had to delay the next measurement day again as I'm just too busy with recordings (nothing bad about that, actually!). New schedule is for mid February.

What I can offer right now is a preview on the final measurements for the AD797: opamp_comparison.pdf

The distortion residual is digitally (and phase-linear) high-pass filtered to remove hum and remaining fundamental as well as 3000x averaged to reduce noise by about 35 dB. The duplicated 1 kHz THD+N vs. level-plot will be replaced with one at 100 Hz. Note again that the FFT and time-domain measurements are done at 60 dB noise gain; typical applications will have 40-60 dB better actual figures.

The low-level common-mode distortion seen for the AD797 is pretty strange--anyone seen that before? Will need to verify with another sample.

Samuel