single opamp making a difference??

[Svart]

Ok, so we all say and *know* that a single opamp can't possibly make a difference in sound right? I've always been skeptical of this but I think for the most part it's true until you get into certain situations..

here's my story..

so I was working on some new material in the studio and needed to use a number of channels in my console. I used 4 channels for cymbals because i needed to mute and change EQ( i only have mute automation.. :sad: ) but when i record all the tracks I do them dry, without compression, without EQ and so on. When listening back I noticed something strange.. two of the tracks sounded normal and good as usual.. the other two sounded lifeless and limited in headroom. oddly enough I have modified both of these channels in the past with high(er) quality IC opamps, removing coupling caps out of the audio path, upgrading supply decoupling and so on. FYI, the channels have 1% metal films throughout also.

Of course I couldn't stand to have such a noticeable difference between tracks of the same material, so i removed a blade of the *good* channel and a blade of the *bad* panel.

I traced everything and tested parts to find no problems with the actual parts themselves but I did find ONE difference between the channels...

on the *bad* channels, I had somehow left a TL082 in the audio path, It was either the booze or the sleep dep but it happened... the *good* channels have an OPA2604 in place. everything else was the same. I removed the TL082s and put sockets and OPA2604s in their places. I returned the channels back to the console and juiced it back up.

I now hear NO difference between the channels.

Now, I think if you have a ton of TL082s in your audio path and replace a few with something better, you won't hear much difference.. but reverse the situation and leave a few TL082s in a console of high(er) quality parts and you will bottleneck your audio. this is what i believe to have happened to my channels. they just sounded strange and choked until the odd part was replaced.

any thoughts or opinions?

am i nuts?

should i get drunk and forget about it?

should i get drunk and forget about work?

should i just get drunk?

:green:

 

[Samuel Groner]

I don't have an answer at hand, but I would greatly appreciate if you could post a few audio samples so that we could listen to your observation.

Samuel

 

[soundguy]

[quote author="Svart"]Ok, so we all say and *know* that a single opamp can't possibly make a difference in sound right? [/quote]

who ever said this? that doesnt reflect my experience at all. sounds like Im not alone. If that opamp is in the audio path I would be suprised to see the case where it didnt make a difference.

dave

 

[chrissugar]

Each opamp has a diferent sound.
The TL082 is an old opamp and it can sound ok if it is not stressed too much, but if you load it with low Z it starts to distort bad. Even in ideal conditions ( i mean ideal for the TL082 ) it will not sound stellar.
I tried diferent discrete opamps with my preamps and they all have a diferent sonic signature.

chrissugar

 

[Svart]

I'll try to get some samples of the differences this coming weekend, it's tonally subtle but the difference seems to be in the headroom.. the top end seems squished ( not like a compressor.. but not good either) and the higher frequencies seem dull, whereas the completely modified channels sound clear and clean without artifacts.

is this what you guys are seeing?

 

[alk509]

Things can only sound as good as the worse-sounding element in the audio chain. A single op-amp can make A WORLD of difference, as you just found out.

Peace,
Al.

 

[peterc]

Be intersting to compare a 5532 in there. In my experience the 5532 sounded slightly warmer, while the 2604 had a more extended top end.

The 5532 may also sound crap tho....... (not a FET input opamp) :wink:

Peter

 

[Svart]

PeterC, I've removed the coupling caps between stages, so i'd have to put the cap(s) back in before i could use the BJT input opamps. but I do indeed like the opa2604 better for it's color, at least for the circuits that it's in.


:thumb:

 

[bcarso]

FWIW here's my TL082 story: So in 1990 I have started to work for H*rm*n, coming from a failed family business with a few years of ~audio experience, prior to that scientific instruments for 17 years. I'm working for guys with lots of experience: Rich May, ex of Sumo et al., Brad Plunkett, major UREI history, etc.

I inherit a couple of car audio products not quite in production, one of which is a power amp that has a switchable lowpass filter for use with a woofer. It uses TL082's and seems to be non-controversial. Problem is distortion when the filter is switched out---serious (> 0.1%, well out of spec) distortion---not sure of the harmonic content but unacceptable. With the filter in there is no obvious problem.

So I look at what is being done and see that it's a three-pole Sallen-Key unity-gain structure. There are three 18k R's in series to the n.i. input of the op amp, and the three caps are being switched in and out for the lowpass function.

You look at datasheets for the bifet amps and you can find one showing quite low THD, not much to worry about for a mid-fi consumer product. But here was all this distortion.

I conjectured that the problem was common-mode distortion arising from a variation in the input capacitance of the amp. I was roundly disbelieved and told to keep looking for the real cause. But I demonstrated that if the R's were shorted out everything cleaned up, and went on to install an approximately compensating R-C feedback network from output to inverting input. It was an easy board relayout. It made the lowpass distortion a bit worse, but brought the bypass-mode distortion down significantly and fixed the problem.

Brad

 

[soundguy]

when a/b'ing stuff like this, onesy twosey there might not seem to be a big difference between different opamps. The real test is how the things stack up cumulatively. For instance, opamp A may appear to have the ever bit of a softer difference around 4K or 5K compared to opamp B. You note this, but doesnt seem like a difference to be concerned about. However, after you record or monitor 24 tracks, that ever slightest bit of softness cumulatively can amount to a total lack of clarity and mushiness. The ultimate test for this kind of thing is record a whole bunch of tracks quickly with option A and option B to see how they really stack up.

It's sorta like the logic Ive extended towards recapping stuff. If you have one or two channels of some old thing, you can change out old caps and maybe get increased performance, but maybe its not a big deal if the bottom is a little rolled off or the top isnt so clear, the circuit might not be performing to its optimum spec, but that doesnt mean it doesnt sound cool either. Over the scope of a whole console, however, the cumulative effects of old caps can really really work against you together.

dave

 

[bcarso]

Good points Dave.

Sounds like a fixture might be in order with a bunch of sockets that would chain a number of candidate op amps. Some non-inverting configuration, maybe the standard four equal R diff amp, or maybe some gain with attentuation, the latter providing some output loading as well.

You have the possibility of cancelling some of the artifacts when in the inverting mode and cascading---I've done some open-loop cancellation this way. But for the most part I'd think cascading would bring out the sonic signature more obviously.

Reminds me of a demo I did in a fancy sound room where the tweako guy insisted he couldn't stand sand state in his signal path. But his current favorite speakers had an elaborate line-level EQ which the mfg. came out and tweaked for your room, and of course under the hood was a sea of op amps. I decided not to rub it in.

 

[Svart]

i had noticed the top end sounding a bit strange once or twice but thought nothing of it really until i went totally mad and replaced everything on the channels. now the unmodded channels seem totally unusable to me and I notice when things are *wrong*. In a pair of modded channels I changed every single opamp in the audio path to an OPA2604 for the channel/EQ and opa404 in the EQ (around 10 opamps with EQ and filtes IN, about 5 without..).

on another set of channels I used OPA627s with adaptors for the summing portions and OPA2604/opa404 for everything else. I did notice a slight difference in that too but only when comparing to a completely unmodded channel. trying to compare to a half modded or fully modded channel using a good opamp for all was very hard to hear a difference.


I'm all for testing! I wonder if a certain opamp is good in a specific use than another is.. a comprehensive test and subsequent database would be very useful...

I think most of the bad rep of OPAMPs are the misuse and misdesign that people overlook when basing judgements. much like Bcarso's story, something overlooked in the (mis)use of the part can cause terrible problems and if one engineer was stumped, then i would bet that there are more where that came from..





:thumb: :thumb:

 

[soundguy]

[quote author="Svart"]
I'm all for testing! I wonder if a certain opamp is good in a specific use than another is.. a comprehensive test and subsequent database would be very useful...
[/quote]

I would certainly think so. For my taste, if you just wanted to look at slew distortion only, I would be inclined to use the fastest opamp possible in an active summing stage. A slower opamp could very well add a cooler color as in input buffer or in a EQ. The reverse of course could also be true, but your mix is always going to be defined by the limitations of the opamp on the summing buss so I'd want to go for the color on the channels and the speed on the buss... The biggest problem I encounter on low range but decent sounding consoles is the lack of attention to the summing buss. The input channels sound cool, but the summing buss typically will use similar topology and really limit the console as a whole. Id say this was the case on just about every mid level table top console Ive used which was chip based. Going with a different flavor in the summing buss than in the channel strip can also open up a different feel when you start combining things.

dave

 

[Svart]

Hey soundguy,

Right now i'm using mostly opa2604 in the channnels and using the dual OPA627s on the summing areas. this setup works very well! before this console had the tl082 in the summing/buffering areas and the 5532 in the gain areas. just switching out the summing areas was an improvement... I tried the opa2134 and others but keep coming back to the opa2604 for most everything.

I've noticed the diffference between the buss channels in my console, between the modded and unmodded ones. I can go on and on about the differences but it's similar to everything else i've said, the highs are dull and seem limited. this is especially true with cymbals. One buss track would sound like you were banging a trashcan lid when you were really hitting a 400$ handhammered cymbal. the other would actually sound almost realistic and detailed. the difference is night and day.

However I will say this about swapping parts like crazy and slew rates etc..

I have a couple pairs of channels that i keep stock, same for one buss blade. there is something about the mushyness of the original channelstrips that sounds great on vocals and keyboards. I tried the super-built channels on those two and they sound strange. strange because you can hear all the little-bitty raspy-spitty sounds in the voice, and you can hear all the sizzle and crackly-crap from the keyboard/synth. stuff that's impossible to EQ out or impossible to find an AFFORDABLE mic for..

:green:

but i digress.. i also keep a few channels of nothing but opa2134 which i find good for guitars.

 

[soundguy]

what kind of console did you start out with?

dave

 

[Svart]

An Alesis X2 is what I started with. Now with an empty wallet it is much more!

:green:

really it's a GREAT console to modify with the channel strips and straightforward design. you can get one for under a grand these days.. in fact I saw a broken one go for around 300$ on ebay a few months ago but just didn't have the $$ at the time. i was really bummed.

here's one now:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=23785&item=7313486067&rd=1&ssPageName=WDVW

 

[soundguy]

haha, I have the best story. A few weeks ago I was mixing something on one of those and the songwriter is sitting next to me and with a sigh he kinda hung his head and exhaled and goes "man, you know, this thing sounds exactly what it looks like..." everyone had a good laugh at that. What a plastic-y sounding thing that is. Glad you were able to make some improvements on that box. Isnt the stock summing horrid on that thing? Add a decent pair of output transformers to that for the cherry on top. What are you using for the main LR line driving chips?

dave

 

[Svart]

yeah it's not THAT great stock, it may be plastic-y but it's at least a decent base to start with.. no mackie single board crap here.. and mine came with the older A rev channels with all 1% resistors,Nichicon VZ caps and Signetics 5532s(most have been changed though..). The newer rev B cards used crap caps, njm5532's and 5% carbon resistors.. so you might have been using a Rev B X2 depending on it's year of manufacture.

I'm using the OPA627s on little adapter boards inplace of the main output driver ICs and most of the summing ICs as well(mainly in the bussing channels). I don't use the monitoring portions that much, except for cueing so I didn't bother to recap, replace or upgrade those sections. the output for the Studio Out jacks comes from there so you will hear the 50 or so tl082s in that path. you will also see Tl084's on the buss channels. you HAVE to replace those damn things in this thing. I used opa404s with good results.

Instead of running it through trafos only, I'm working on a pair of 1272's that I might be using strictly between the console and the computer.. As for now, I've been using a PRR vari-mu between the console and the computer. gets a nice warm sound out of it.

but again, since you've used one you now know what I was talking about. the thing is peppered with tl0XX's all over every channel. once you replaced them the thing really livens up!

:thumb:

 

[Svart]

anyone else have comments? additions?

 

[Svart]

ok i have a short and dirty mp3 of the sound of the console with the *good* opamps and mods done on the channels being used. anyone host it for me? thanks!

around 3.6 mb..