Quick question, which is better for -6dB?

[bluebird]

Is there any advantage to an inverting style op amp with -6dB gain, say with 5K feedback and 10K into the inverting input over just a 5K/5K voltage divider into a unity gain buffer (voltage follower)? Is one noisier than the other?

 

[abbey road d enfer]

Is there any advantage to an inverting style op amp with -6dB gain, say with 5K feedback and 10K into the inverting input over just a 5K/5K voltage divider into a unity gain buffer (voltage follower)? Is one noisier than the other?

There is actually a tiny difference in noise performance, in favour of the non-inverting circuit, but the inverting circuit may have a subtly better transient response. The point is moot considering the performance of current opamps. The practicalities of the circuit are dominant in this respect.

 

[JohnRoberts]

There are some subtle trade offs pro and con favoring either topology. Since the ein of a modern opamp can be microvolts, the difference between a noise gain of 1x or 2x is inconsequential.

I will suggest a third option, using 4 resistors with a good quality opamp we can configure a differential stage. This can either be inverting or non-inverting depending upon which input you feed. The gain of -6dB can easily be accommodated with resistor ratios. 

The differential stage allows you to reference between where the signal is coming from and where it is going, a major benefit to maintaining signal integrity IMO.

JR

 

[bluebird]

I suspected there wasn't much of a difference. I'm playing with a M/S circuit similar to Wayne Kirkwood's famous M/S board. He used an inverting amp to knock off 6dB but then had to reverse the polarity on the output driver. in my circuit there's some EQ in the M/S loop and bypassing that needs consistent polarity. so I will just use the voltage divider and buffer.

Thank you gentlemen, much appreciated!

 

[emrr]

Hmm, I'm sure you've checked his site, there've been several updates and expansions to Wayne's M/S project, some covering your needs.  I don't recall what the changes have been offhand. 

 

[bluebird]

Yes! The new version is rad. I didn't analyze it fully. I'll go back and take a look. I'm basically transplanting the basic circuit into what I'm doing. This is part of a limiter I'm working on with M/S EQ function. One band with six selectable frequencies, cut and boost. Mainly for dipping the high mids that get annoying when pushing things. Leaving the sides alone when mellowing out the 2K-8K range seems to work best. Its one of those things that just keeps getting more complicated. There are too many stages already so I'm trying to optimize everything for lowest noise. I'm using only NE5534's because they are a bit quieter than 5532's or even LME4562's. At least in the circuits I've created. I've found the On Semi SMD 5534's to be just as good as the old Signetics versions so I've saved some space as its grows. Its still a mess and I'm over doing it but having fun...

I really feel like the 5534 still hasn't been beat by the new opamp's. There are circumstances where LME4562's specs may win, but in everyday audio circuits, the 5534 is still king in my book. Just have to have a bunch of 22p caps on hand... ;D

 

[john12ax7]

Have you noticed an audible and /or measurable difference in TI vs ON Semi for the 5534?

 

[bluebird]

No I haven't done tests between those. I would imagine the differences would be very slim. I did do tests on different brands of 5532's. The TI versions were the worst, the best was old Fairchild's. Very close (as good) are the currently available JRC's. That is the brand I would recommend. It was a while ago but I think I set up a bunch of unity gain circuits with 5K resistors in a row. 6 stages IIRC then measured THD+N on an AP. I also tested the LME4562 and another low noise JRC dual amp I can't remember. Neither were better than the Fairchild 5532. At the end of it I did the signetics NE5534's and they had significantly less distortion than the 5532's in the same circuit. I remember it was close to the limit of the AP. Its an old system one so it probably used 5543's! I'm only making these claims about the particular circuit I setup so it may not be true for all circumstances. I was also very surprised at how well the 5543 held up under 150 ohm load which the 5532's were not so happy about. But I did come away with a fondness for the NE5534. I also felt there really wasn't much use for newer dual op amps over the 5532 in most situations after the testing. If the circuit requires a BJT amp that is.

I really wish the rare NE5533 was still made or more available. Or even better, a SMD version. Those were true dual NE5534's.

 

[john12ax7]

I've only used the TI version so far,  interesting that there are differences.  Will have to try some other versions down the road.

 

[user 37518]

No I haven't done tests between those. I would imagine the differences would be very slim. I did do tests on different brands of 5532's. The TI versions were the worst, the best was old Fairchild's. Very close (as good) are the currently available JRC's. That is the brand I would recommend. It was a while ago but I think I set up a bunch of unity gain circuits with 5K resistors in a row. 6 stages IIRC then measured THD+N on an AP. I also tested the LME4562 and another low noise JRC dual amp I can't remember. Neither were better than the Fairchild 5532. At the end of it I did the signetics NE5534's and they had significantly less distortion than the 5532's in the same circuit. I remember it was close to the limit of the AP. Its an old system one so it probably used 5543's! I'm only making these claims about the particular circuit I setup so it may not be true for all circumstances. I was also very surprised at how well the 5543 held up under 150 ohm load which the 5532's were not so happy about. But I did come away with a fondness for the NE5534. I also felt there really wasn't much use for newer dual op amps over the 5532 in most situations after the testing. If the circuit requires a BJT amp that is.

I really wish the rare NE5533 was still made or more available. Or even better, a SMD version. Those were true dual NE5534's.

In what way was the Fairchild 5532 better than the LM4562? the only thing that could be better is that the 5532 has lower input current noise density, but aside from that the LM4562 is better in every way, Douglas Self has experimented quite a bit with the 4562 and 5532, and the 4562 is clearly a better opamp.

 

[bluebird]

Use the 4562! I'm just telling people what I found with the AP in the circuit i was using. Since I can't remember the exact setup, take my experience with a grain of salt. I expect people here to to be skeptical of all claims, as we should. Unless of course I can produce test results, which I can't.
Regardless I wouldn't want to lead anyone down the wrong path hence the disclaimers. I read the Douglas Self stuff and why I tried the amp in the first place. But it had higher THD+N readings. Op amp data sheets show the results of the test circuit. That's it. Different circuits different data.

 

[bluebird]

I was able to find the sweeps I did but I don't have the key to the colors. It does show the differences in different brands. I believe The very bottom yellow line is the 5534. But again this was a while ago and I don't have all the info. I mean, no one would be able to hear these differences anyhow. Again not saying anything concrete, just that I did do the tests and not trying to mislead anyone... Just trying to share some loose info.