Mooly said:
Not much you can really say to thatKeep taking the tablets.
take the red pill, and i'll show you just how deep this rabbit hole goes, take the blue pill, and you'll wake up in the morning, and all the "no feedback", magic cable, unobtanium foil capacitor, 10% 2nd harmonic equipment you have will sound just fine............
i have seen excellent preamps use 4558's as headphone amps, with no complaints from anybody. not cheap preamps, either, one i'm familiar with was $700.00 US in 1979 (adjusted for inflation, about $2100.00 today...). the 4558's were being used at a gain of 2, so there wasn't a problem with input stage noise, and hardly any contribution to the THD figure, so 4558's were fine in this role.
the reason 4558's were used in a preamp where everything else was done with TL072's was because the 4558's could source a lot more output current than the 072's.
unclejed613 said:
You are right about 4558s.
I know those from Peavey mixing desks, equalizers, DJ mixers etc. I remember very well the nice sound of Peavey DJ mixers compared with many hi-fi preamplifiers - to my big surprise - their smoothness in highs and their warmness in lows. Peavey concretelly used the type RC4558.
Fotios
Good quality descrete sounds good, and also good quality op-amps sound good. Sometimes the biggest difference is what chip is being used.
IMO if the amp can play a bit wider than the whole audio range from 5hz-40khz accurately, with no crossover distortion, and has plenty of output-drive, then it sounds great, despite the minor inaudible differences some may claim.
On a large 350W Mono amp I built some time ago, I used an LM324 quad op-amp as a preamp. This amp was meant for subwoofers, but also for full-range use or PA, or whatever I want to use a big amp for.
Despite the VERY CLEAR bass & treble sound of the 350W amp itself directly from a line-input with no preamp, when I used the quad op-amp based preamp to drive the amp, the sound was HORRIBLE! The treble sounded like a low-quality MP3 but the bass was fine. Then to my amazing I read in the datasheet that the LM324 is CLASS B so no wonder it sounded like ****.
I switched to a much better On-Semi MC33079 with 0.002% thd and 7V/us. The datasheet mentions it for audio use and uses NPN-output stage to eliminate crossover distortion. It fit in the same IC socket, so the swap was easy, and it sounded like a new amplifier. The treble was accurate, and missing nothing with the instruments, and also the bass got better too. Now the preamp sounds great with no changes to the original sound !
Sure, you may be able to build a high quality descrete solution, but IMO most people's negative opinions about opamps comes from using OLD junk like uA741 and LM324.
Get a good opamp and design your circuit right, and you will get good sound.
IMO if the amp can play a bit wider than the whole audio range from 5hz-40khz accurately, with no crossover distortion, and has plenty of output-drive, then it sounds great, despite the minor inaudible differences some may claim.
On a large 350W Mono amp I built some time ago, I used an LM324 quad op-amp as a preamp.
This amp was meant for subwoofers, but also for full-range use or PA, or whatever I want to use a big amp for.Despite the VERY CLEAR bass & treble sound of the 350W amp itself directly from a line-input with no preamp, when I used the quad op-amp based preamp to drive the amp, the sound was HORRIBLE! The treble sounded like a low-quality MP3 but the bass was fine. Then to my amazing I read in the datasheet that the LM324 is CLASS B
so no wonder it sounded like ****.I switched to a much better On-Semi MC33079 with 0.002% thd and 7V/us. The datasheet mentions it for audio use and uses NPN-output stage to eliminate crossover distortion. It fit in the same IC socket, so the swap was easy, and it sounded like a new amplifier. The treble was accurate, and missing nothing with the instruments, and also the bass got better too. Now the preamp sounds great with no changes to the original sound !
Sure, you may be able to build a high quality descrete solution, but IMO most people's negative opinions about opamps comes from using OLD junk like uA741 and LM324.
Get a good opamp and design your circuit right, and you will get good sound.
i remember an early 6 channel mixer when i was in a band, and we wanted to record some rehearsals to get a more objective listen to our music. we had an 8 channel mixer for the instruments and vocals, and were going to use the 6 channel to mix down the drums and run the drum mix into the line level inputs on the 8 channel. only problem was that the 6 channel had a nasty hiss and had lousy high frequency response (the last pair of things you want in a drum mix). it turned out it has a 741 as it's mixer amp. pulled the 741 and replaced it with a TL071, and it was like a different mixer. the toms no longer went "thud" and the cymbals didn't sound like PFFFFF PFFFFFF anymore (they should sound like "BOOM" and TSSSSS TSSSSSS). i don't even remember the brand of mixer it was, but the upgrade made a night and day difference.
i'm of the opinion that if an op amp is a good solution for an application, use one. if a discrete is a good solution, use discrete. there are advantages to both methods. some have the opinion of "8 legs bad, three legs good. fine, that's their opinion. i have an article from 1953 or 54 from an engineering magazine where the debate was whether transistors would ever be able to come close to matching tubes for performance. back then op amps required dual 300V power supplies and were the size of a very large bread box, and had a unity gain bandwidth of a few hundred hertz. transistors changed that by reducing the size of that op amp to the size of a deck of playing cards, only requiring +/- 24V, and a bandwidth of 10 or 20khz. then the 709 and the 741 really changed everything. the debate hasn't changed much, except for the technologies being debated. the substance of the debate has stayed pretty much the same "new technology bad, old technology good". don't forget that if it hadn't been for monolithic op amp design methodology, you wouldn't have constant current sources and current mirrors to make discrete op amps perform much better than they did in 1965. these are the same spinoffs from IC technology that make audio power amps perform far better than could be imagined in the 60's.
i'm of the opinion that if an op amp is a good solution for an application, use one. if a discrete is a good solution, use discrete. there are advantages to both methods. some have the opinion of "8 legs bad, three legs good. fine, that's their opinion. i have an article from 1953 or 54 from an engineering magazine where the debate was whether transistors would ever be able to come close to matching tubes for performance. back then op amps required dual 300V power supplies and were the size of a very large bread box, and had a unity gain bandwidth of a few hundred hertz. transistors changed that by reducing the size of that op amp to the size of a deck of playing cards, only requiring +/- 24V, and a bandwidth of 10 or 20khz. then the 709 and the 741 really changed everything. the debate hasn't changed much, except for the technologies being debated. the substance of the debate has stayed pretty much the same "new technology bad, old technology good". don't forget that if it hadn't been for monolithic op amp design methodology, you wouldn't have constant current sources and current mirrors to make discrete op amps perform much better than they did in 1965. these are the same spinoffs from IC technology that make audio power amps perform far better than could be imagined in the 60's.
It's all relative. Swap the OpAmps in a top CD player for a 741 or 4558 and it's still "good". Ask someone with no real interest in audio to listen first to one then the other and I bet they say, well they sound the same to me.
Now that difference is worth all the tea in China to a music lover/audiophile. But can you honestly say it's 1000 or 2000 or 10,000 pounds/euros /dollars better.
I tried a £15 portable CD through my system a while back. In terms of "good HiFi" it came to 95% of the sound of my Micromega. But that missing 5% is EVERYTHING, it's what makes the system what it is, so totally and utterly magical to listen to -- but it's so elusive to pin down---to describe---to measure .
Now that difference is worth all the tea in China to a music lover/audiophile. But can you honestly say it's 1000 or 2000 or 10,000 pounds/euros /dollars better.
I tried a £15 portable CD through my system a while back. In terms of "good HiFi" it came to 95% of the sound of my Micromega. But that missing 5% is EVERYTHING, it's what makes the system what it is, so totally and utterly magical to listen to -- but it's so elusive to pin down---to describe---to measure .
Hi Anatoliy,
I don't know what to say without offending you. I always try and use the inverting configuration anyway with OpAmps, it does seem to sound better -no CMRR problems probably as was mentioned earlier. Can't use that configuration of course when swapping them in a CD player if it's non inverting.
What can I say ? the 4558 etc doesn't come close to other more modern designs, one of my favourites being the OPA604/2604.
It might only be that 5% but it's night and day
I don't know what to say without offending you
. I always try and use the inverting configuration anyway with OpAmps, it does seem to sound better -no CMRR problems probably as was mentioned earlier. Can't use that configuration of course when swapping them in a CD player if it's non inverting.What can I say ? the 4558 etc doesn't come close to other more modern designs, one of my favourites being the OPA604/2604.
It might only be that 5% but it's night and day
I like that Anatoliy advises the same procedures to get a good sound from 4558 etc that I would: low gains, inverting, class-A (via R to the local supply rail).... and I may add another very effective "tweak": use the second half of a dual to mirror the current of the first, but with negative sign (and with the same class-A bias). Point is, make the supply current not just a copy of the signal current (already a strong point for using class-a) but make it virtually constant. With known loads (say, inside a more complex circuit) this is easily accomplished, for example on can just use a gain=-1 inverter on the master amp's output and load it so that the sum of suppy currents (as they appear at the dual's supply pins) is constant at all times. The slave output is not used as a signal by itself, it only makes for a constant supply current class-A "bridge" circuit. To get the (dynamic) loads right needs a little thought, I prefer to check with a sim with a special simple opamp model I made which pefectly reflects output current to the rails (plus a constant "bias" amount of current).
- Klaus
- Klaus
Mooly said:Hi Anatoliy,
I don't know what to say without offending you
What can I say ? the 4558 etc doesn't come close to other more modern designs, one of my favourites being the OPA604/2604.
It might only be that 5% but it's night and day
I plan to design a new console where opamps will be used in servo only. Modern opamps are still opamps, so they are suboptimal for audio. They seems to be easy to use as novices believe, but in reality you have to know the specific character of each and every beast you use. It is like Windows OS: advertised as if it is the easiest OS for dummies, it requires more human resources to work properly.
So, you may say whatever you want, you can't offend me because I know well what I say.
Anatoly,
please help me to understand your objections to IC op-amps (I have no objections to either IC op-amps or discretes designs). I'd like to understand why an op-amp optimised for audio performance is inferior to a discrete solution. Lets do it like this - score 1 for better and 0 for worse
My reference op amps are LM4562, AD479 and venerable NE5532.
Spec/Feature Discrete Op-amp
CMRR 0 1
Distortion THD 20 0 1
Distortion IMD 0 1
Distortion PMD 0 1
Noise 1 (only just) 0
PSSR 0 1
Slew-rate 0 1
Output Drive 1 0
Bandwidth 0 1
DC performance 0 1
*Ease of use 0 1
*Cost/Performance 0 1
* optional and not really important if we are talking about cost no object
Feel free to modify as you see fit.
Once we have a few of these down on paper, maybe we can start to understand where to look as to the reasons why one is better than the other
please help me to understand your objections to IC op-amps (I have no objections to either IC op-amps or discretes designs). I'd like to understand why an op-amp optimised for audio performance is inferior to a discrete solution. Lets do it like this - score 1 for better and 0 for worse
My reference op amps are LM4562, AD479 and venerable NE5532.
Spec/Feature Discrete Op-amp
CMRR 0 1
Distortion THD 20 0 1
Distortion IMD 0 1
Distortion PMD 0 1
Noise 1 (only just) 0
PSSR 0 1
Slew-rate 0 1
Output Drive 1 0
Bandwidth 0 1
DC performance 0 1
*Ease of use 0 1
*Cost/Performance 0 1
* optional and not really important if we are talking about cost no object
Feel free to modify as you see fit.
Once we have a few of these down on paper, maybe we can start to understand where to look as to the reasons why one is better than the other
Bonsai said:
I do not know why people would use the lame NE5532 dual instead of the better & different single op-amp NE5534.
Actually, it is the people refering to the first mentioned
that many times drags down the reputation
of the excellent, for the price,
low noise op-amp and general very useful for audio NE5534.
I can get something like 50 (fifty) NE5534 for price one OPA627 / AD797 / LT1028 etc.
And I wont be disappointed with NE5534 any time in almost any audio application.
NE5534 popularity in all sorts of commercial gears and among hobbyists
is due to the very allround and good level of data figures.
It does not do really bad in hardly any parameters important in audio.
Take this fact and add The Low Price .. and you have a winner.
There are other good allround operational amplifiers chips, today.
But not too many in the same price range of NE5534.
It is not with NE5534 like it is with
OPA134 (single) or OPA604 ( single) vs. OPA2134 / OPA2604 duals.
Besides, NE5534 has got 2 pins for custom internal capacitance compensation trimming.
And so we can do real compensation .. instead of putting cap from output to inverting input.
=====================
Bonsai.
Your list idea is a good way to try make a good compare.
But as we will never be able to agree on
what op-amp & what discrete small signal amplifier to compare, it is useless to try
But nevertheless, you are right!!! we see too much opinions
without any explaning of the reasons for these opinions.
Opinion without reasoning may be just a fashionable breathe of air.
And such wont give us any better knowledge and background to form a valid and good personal opinion.
I, like most diyaudio members, have no problems to use Op-Amps, Transistors, Tubes and hybride mixings of them.
Whenever I like or find this may give me what I am after in a specific audio application.
To a fair DIY Price, of course
I wont let my hobby ruin my economy .....
Lineup
Andrew
Exactly. Without hard evidence either way, we should not dis the other option. Keeo an open mind on the issue.
Tico
you are correct. My concern abouyt the otherwise excellent AD797 is its unforgiveness to less than optimal layout - you have to know what you are doing to get the best out of it.
Lineup
Good comments!
Exactly. Without hard evidence either way, we should not dis the other option. Keeo an open mind on the issue.
Tico
you are correct. My concern abouyt the otherwise excellent AD797 is its unforgiveness to less than optimal layout - you have to know what you are doing to get the best out of it.
Lineup
Good comments!
Many audio circuits like crossovers etc are unity-gain. AFAIK the NE5534 is the non-unity-gain stable brother of the NE5532. I guess it will become as "lame" as the other one when being compensated forunity-gain stability.
According to Douglas Self's measurements the NE5532 is still among the best regarding THD & noise:
http://www.dself.dsl.pipex.com/ampins/webbop/5532.htm
In non-inverting circuit it even beats the LM4562:
http://www.dself.dsl.pipex.com/ampins/webbop/LM4562.htm
There was once an active studio-monitor by a famous Swiss company that used tons of the 5532 in its crossover. And this was a precision tool for audio professionals !
One would maybe not use this OP-AMP in a CD player's output stage but for crossovers etc it is still fast enough IMO.
Regards
Charles
According to Douglas Self's measurements the NE5532 is still among the best regarding THD & noise:
http://www.dself.dsl.pipex.com/ampins/webbop/5532.htm
In non-inverting circuit it even beats the LM4562:
http://www.dself.dsl.pipex.com/ampins/webbop/LM4562.htm
There was once an active studio-monitor by a famous Swiss company that used tons of the 5532 in its crossover. And this was a precision tool for audio professionals !
One would maybe not use this OP-AMP in a CD player's output stage but for crossovers etc it is still fast enough IMO.
Regards
Charles
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.