What I meant is this:
I assume that the noise is dominated by the input stage, like it is in any well-designed low-noise amplifier. The base current of a bipolar transistor has two kinds of noise associated with it, shot noise √(2qIB) and 1/f noise. There are other contributions to the equivalent input noise current of a bipolar transistor, but those are negligible at frequencies below fT/√hFE. Looking at the typical specifications of the NE5532 and NE5534:
NE5532 and NE5532A:
input bias current, which is simply the base current of the input transistors, 200 nA typ.
equivalent input noise voltage 5 nV/√Hz at 1 kHz, 8 nV/√Hz at 30 Hz
equivalent input noise current 0.7 pA/√Hz at 1 kHz, 2.7 pA/√Hz at 30 Hz
NE5534 and NE5534A:
input bias current 500 nA typ.
equivalent input noise voltage NE5534: 4 nV/√Hz at 1 kHz, 7 nV/√Hz at 30 Hz
equivalent input noise voltage NE5534A: 3.5 nV/√Hz at 1 kHz, 5.5 nV/√Hz at 30 Hz
equivalent input noise current NE5534: 0.6 pA/√Hz at 1 kHz, 2.5 pA/√Hz at 30 Hz
equivalent input noise current NE5534A: 0.4 pA/√Hz at 1 kHz, 1.5 pA/√Hz at 30 Hz
√(2qIB) ≈ 0.25316 pA/√Hz when IB is 200 nA
√(2qIB) ≈ 0.40027 pA/√Hz when IB is 500 nA
(Each of the two inputs has a white noise current source with this value associated with it. Some op-amp manufacturers (such as the late Linear Technology) would split the current noise into a common-mode and a differential component and specify the differential value, to get √2 times better datasheet numbers. The difference can be much larger than √2 for op-amps with base current compensation.)
It looks to me like 1/f noise dominates the equivalent input noise current at 1 kHz for the NE5532, while (white) shot noise dominates for the NE5534A.
Last edited:
OPA1611 is an ultralow voltage noise, ultrahigh current noise op-amp, 1.7 pA/√Hz at 1 kHz. Far worse than an NE5534A for moving-magnet phono amplifiers. It would be pretty good for a dynamic microphone preamplifier, though.
Last edited:
Yes, that would be a very good choice.
Out of modern bipolar op-amps, OPA227 and OPA228 have a reasonably low noise current, low enough to be suitable for moving magnet amplifiers but of course not as low as a FET op-amp. With 0.4 pA/√Hz and 3 nV/√Hz at 1 kHz, they just outperform the NE5534A. Their dual versions are OPA2227 and OPA2228. I know from my own measurements that that 0.4 pA/√Hz also holds when the impedances driving the inputs are very different, that is, they don't use specification tricks to make the noise current look smaller.
TI revised their datasheet this month, which is good, because figure 7-4 in the previous version was all wrong - now they only have to change the unit from V/√Hz to nV/√Hz.
Out of modern bipolar op-amps, OPA227 and OPA228 have a reasonably low noise current, low enough to be suitable for moving magnet amplifiers but of course not as low as a FET op-amp. With 0.4 pA/√Hz and 3 nV/√Hz at 1 kHz, they just outperform the NE5534A. Their dual versions are OPA2227 and OPA2228. I know from my own measurements that that 0.4 pA/√Hz also holds when the impedances driving the inputs are very different, that is, they don't use specification tricks to make the noise current look smaller.
TI revised their datasheet this month, which is good, because figure 7-4 in the previous version was all wrong - now they only have to change the unit from V/√Hz to nV/√Hz.
I forgot to see their respective datasheet.
https://www.ti.com/product/OPA1612
I mean these are some highly regarded opamps out there, too much HOT to touch. Lol
https://www.ti.com/product/OPA1612
I mean these are some highly regarded opamps out there, too much HOT to touch. Lol
A BMW M3 is a highly regarded car. Too hot to touch. But it's not very useful if you need to move four bales of hay. Same with opamps. The OPA1612 is a great opamp as long as the source impedance is low. With a higher source impedance (such as inductive MM phono cartridges) the current noise becomes the dominant noise source. The NE5534 ($), OPA1642 ($$), or OPA628 ($$$$$) would be better options.
It's pretty easy to compare the noise of opamps in a circuit simulator.
Tom
It's pretty easy to compare the noise of opamps in a circuit simulator.
Tom
Burkhard Vogel recently wrote a book about simulating op-amp noise, Slopes and levels, Spice models to simulate vintage op-amp noise. (I tried to post a link, but it didn't work.)
Some of the manufacturer-supplied models produce results that are complete nonsense (he found such models for TL071, NE5534A, LT1128). Unfortunately, the same holds for a few of Vogel's models.
Some of the manufacturer-supplied models produce results that are complete nonsense (he found such models for TL071, NE5534A, LT1128). Unfortunately, the same holds for a few of Vogel's models.
Last edited:
That's pretty funny because the math is quite simple. It does go to show that you should always validate the simulation models before using them.
Tom
Tom
The problem with the LT1128 model was that it only models the differential part of the input noise current, not the much larger common-mode part. It's probably part of LT's effort not to emphasize that the chip has a huge common-mode current noise. The other two op-amp models had wrong voltage noise levels, and the TL071 model had no 1/f noise.