Look at these two circuits. They do the same.
The difference is the single and the dual supply.
Another difference is the capacitor C1 and C3.
Because the negative voltage at input pin C3 should be put 'upside down'.
This goes for all bipolar opamps with NPN input transistors.
And when using dual voltage supply.
If a bipolar opamp has PNP transistors input the C3 should be turned the other way.
R3 will pass any 12 volt noise into the signal path.
It indeed is a very popular op-amp, and for good reason. It has up until and including today not really met a suitable match for the sheer versatility, cost-effectiveness and performance at that. At some point, good enough is good enough.
Nigel is right. Normally you use your divider with a capacitor to clean that point up, then a higher value resistor to your bias point, like an op amp input.
Yes, the NE5532 and NE5534 were both great designs that stood the test of time. It will be more than good enough for most applications.
Yes, the NE5532 and NE5534 were both great designs that stood the test of time. It will be more than good enough for most applications.
Nowadays, there's certainly better op amps for very little extra money. But what sets the NE5532 apart is its legacy, and how many designs use it that still work the same way today thanks to this same op-amp. Tried and true remains true.
You're unlikely to find an opamp with PNP input transistors. The PNPs in most semiconductor processes tend to be awful. I have worked in a process where the opposite was true, but that was an SOI process optimized for analog work.
The NE5532 and NE5534 are indeed pretty amazing. It's easy to find something that's a smidge better than the 5532, but finding something that performs better than the 5534 in a moving magnet phono stage is actually a bit challenging, especially if you take price into account too.
So, yeah. It is remarkable that a 1979 opamp wasn't really beat on specs until around 2005-6 when the LM4562 came out.
Tom
The original LM833 (but not the much later TI sourced P and D variants which does use NPN input stages) and the old NJM2068 come to mind. Both devices were 100% audio oriented and I believe the 2068 beats the 5532 on noise performance.
Hi Karl,
Yes, those are pretty good products too, although I didn't like the LM833 in some applications. When it comes right down to it, there isn't much point in pulling an NE5532A and putting one of those in. At that point you would install a newer product with specs more in tune with your application. There is a very real danger of the replacement not making any real difference in performance. Noise for example, comes from many sources.
I don't think you can beat the NE5532 in most applications for near the same money. It's still current production I think, and that counts for something (you can buy real parts).
Yes, those are pretty good products too, although I didn't like the LM833 in some applications. When it comes right down to it, there isn't much point in pulling an NE5532A and putting one of those in. At that point you would install a newer product with specs more in tune with your application. There is a very real danger of the replacement not making any real difference in performance. Noise for example, comes from many sources.
I don't think you can beat the NE5532 in most applications for near the same money. It's still current production I think, and that counts for something (you can buy real parts).
I just got 25 NJR2068 for 25 cents each.
This chip has great performance , open airy highs and very low noise. May hit 88db s/n on MM phono preamp.
This chip has great performance , open airy highs and very low noise. May hit 88db s/n on MM phono preamp.
Not offhand... but you did generalise in saying 'You're unlikely to find an opamp with PNP input transistors' when there are many, including highly regarded ones such as these two.
It sure is! Although TI does suggest a newer alternative, which is definitely more expensive. https://www.ti.com/product/NE5532
But if the source is low impedance the noise will be strongly attenuated as 100k is very high impedance for modern equipment. C2 will win over R3 from a few Hz upwards.
All the "single supply" bipolar opamps that accept inputs below the negative rail have to use PNP inputs... So any bipolar opamp described as "single supply" will have PNP input devices. The very early integrated PNP devices were poor, but they moved to buried one's without surface defects to worry about and they got a lot better, as I understand it.
... but the ESD protection in the opamp wouldn't allow for voltages below the negative rail. But if you want rail-to-rail inputs you'll likely have to use some sort of complimentary structure (either MOS or BJT).
Lateral PNPs tend to have really crappy beta. 1-10 on a good day. Vertical PNPs are much better but with the common P+ doped wafers in modern CMOS processes the collector often gets stuck at the most negative potential in the IC, which really limits the usefulness of the devices. You can isolate them with a silicon-on-insulator (SOI) process, but that's expensive and the insulator limits the power dissipation in the IC. I don't know if you can well-isolate them. Maybe that's a thing. But that's not to say that people haven't built PNPs in the past.
Tom
A properly designed +/- PS (read regulated) should out perform a conventional single pole power supply. All other relevant variables the same.
If nothing else, you get more headroom with a dual powder supply.
I understand choosing components and topologies because
But you might regret specing single pole supply. Down the road, you might want to add some circuit that requires a dual +/- supply.
I strongly recommend your dual supply design. If you can.
Plus greater voltage rail vales, like +/- 24 volts. If you can pull it off.
If nothing else, you get more headroom with a dual powder supply.
I understand choosing components and topologies because
- it is what you have on hand
- it is what you can get your hands on
- it is what you can afford
But you might regret specing single pole supply. Down the road, you might want to add some circuit that requires a dual +/- supply.
I strongly recommend your dual supply design. If you can.
Plus greater voltage rail vales, like +/- 24 volts. If you can pull it off.
Hi audiomot,
Well, a single 40 V supply is the same as +/- 20 V supplies. Biasing is just more difficult, that's all. It can be easily done.
24 V rails, now that is an issue unless you use specialty op amps or discrete designs. The worst thing you can do for a design is use special parts as one day there will not be a replacement (= dead product if needed). Many examples of this in the world. That and if one supplier runs into long lead times, you need an alternate source ASAP. Probably best to use parts that are common in requirements as for supply voltages and performance.
"Sometimes good enough is good enough." Well it is more recognizing the entire system performance and not wasting money rather than trying to be cheap.
Well, a single 40 V supply is the same as +/- 20 V supplies. Biasing is just more difficult, that's all. It can be easily done.
24 V rails, now that is an issue unless you use specialty op amps or discrete designs. The worst thing you can do for a design is use special parts as one day there will not be a replacement (= dead product if needed). Many examples of this in the world. That and if one supplier runs into long lead times, you need an alternate source ASAP. Probably best to use parts that are common in requirements as for supply voltages and performance.
"Sometimes good enough is good enough." Well it is more recognizing the entire system performance and not wasting money rather than trying to be cheap.
The LM317 is from 1976 ... and still selling. I wonder how many times it has been redesigned to be ported to the next new process.
Tom
Honestly, i'm not very surprised.
Maybe it's because i'm a big believer in vintage, but i also believe that something well made lasts over time.
Remember the TDA1541?
Not more than 0.3 V or so below the negative rail anyway. Maybe you could get away with a snapback device or a stack of diodes rather than a single diode to the negative rail, but that wouldn't help, because the PNP base-collector junction would still limit the input voltage range.