Hi all,
Hopefully there are some op-amp experts here. I am making a mic. preamp board and will be using the PGA2505 IC for amplifying the input signal.
However, I am trying to decide on an op-amp for various audio roles: audio peak-detect circuit, low-pass filtering, buffering, and a balanced line driver. I could use different op-amps but prefer to consolidate to one.
I have +-5V rails available so will be using dual-supply and obviously rail-to-rail input and output is a necessity.
Initially I was going to use the NE5532 but saw the TI recommends the TLV9362 as a replacement.
Searching TI audio op-amp sections revealed the LME49721.
Then there are more general purpose options: AD8542 and TLV9002
Here are some stats.
Hopefully there are some op-amp experts here. I am making a mic. preamp board and will be using the PGA2505 IC for amplifying the input signal.
However, I am trying to decide on an op-amp for various audio roles: audio peak-detect circuit, low-pass filtering, buffering, and a balanced line driver. I could use different op-amps but prefer to consolidate to one.
I have +-5V rails available so will be using dual-supply and obviously rail-to-rail input and output is a necessity.
Initially I was going to use the NE5532 but saw the TI recommends the TLV9362 as a replacement.
Searching TI audio op-amp sections revealed the LME49721.
Then there are more general purpose options: AD8542 and TLV9002
Here are some stats.
| TLV9002 | AD8542 | TLV9362 | LME49721 | |
|---|---|---|---|---|
| Cost $ (USD) | $0.4 | $0.87 | $1.0 | $1.1 |
| Noise (nV√Hz) at 1kHz | 30 | 40 | 8.5 | 4 |
| Offset voltage (mV) | 1.5 | 6 | 1.7 | 1.5 |
| Slew rate (V/µs) | 2 | 0.75 | 25 | 8.5 |
- I like the TLV9002 for the cost but is 30 nV√Hz acceptable for audio?
- Maybe for peak-detect this would be OK, but probably not for anything in the signal path?
- Is there a compelling reason to pick the TLV9362 over the LME49721 (other than cost)?
LME49721 will go up in smoke at your supply voltage. It is meant for at most 5.5 V between the positive and the negative supply, while you have 10 V. The same applies to the TLV9002 and AD8542.
Hi MarcelvdG,
Did I miss something? Yes! These seem for total voltage 5V, was reading the info. Wrong!It seems alright to me, for +-5V rails, from the webpage (both TVL9002 and LME4972 say the same)
What about the AD8542? This seems to imply it’s OK?
Regards,
Dan
Did I miss something? Yes! These seem for total voltage 5V, was reading the info. Wrong!
What about the AD8542? This seems to imply it’s OK?
Regards,
Dan
Last edited:
No, the underlined part doesn't mean it can be powered with +/-6V just that being a differential input it can go to 6V magnitude both ways without either of the inputs going negative.
As for your 30nV/rtHz question - it depends how high fidelity you're looking for and the particular circuit configuration. The input noise would be around 4uV in the audio band or -107dBV.
As for your 30nV/rtHz question - it depends how high fidelity you're looking for and the particular circuit configuration. The input noise would be around 4uV in the audio band or -107dBV.
Hi Dan,
Like abraxalito, I'm afraid it doesn't mean it is OK.
Suppose you connect the positive supply pin to +5 V and the negative supply pin to 0 V. The voltage between the positive and negative supply pins is then 5 V, obviously.
Now connect the positive input to the positive supply pin and the negative input to the negative supply pin. The differential input voltage will then be +5 V - 0 V = +5 V.
When you swap the inputs, the differential input voltage becomes 0 V - (+5 V) = -5 V.
Best regards,
Marcel
Like abraxalito, I'm afraid it doesn't mean it is OK.
Suppose you connect the positive supply pin to +5 V and the negative supply pin to 0 V. The voltage between the positive and negative supply pins is then 5 V, obviously.
Now connect the positive input to the positive supply pin and the negative input to the negative supply pin. The differential input voltage will then be +5 V - 0 V = +5 V.
When you swap the inputs, the differential input voltage becomes 0 V - (+5 V) = -5 V.
Best regards,
Marcel
The one remaining op-amp has a rail-to-rail output, but no rail-to-rail input.
One thing you should know about rail-to-rail inputs: there are exceptions, but the traditional way to make them is by giving the amplifier two input stages, an N-type (for example, NMOS or NPN) stage for medium and high common-mode input voltages and a P-type (PMOS or PNP) input stage for medium and low common-mode input voltages. These stages normally have different offset voltages, as they have independent random mismatch. When you make a voltage follower and drive it through the voltage range where the op-amp transitions from one input stage to both or from one to the other, the offset changes with the signal, causing extra distortion.
One thing you should know about rail-to-rail inputs: there are exceptions, but the traditional way to make them is by giving the amplifier two input stages, an N-type (for example, NMOS or NPN) stage for medium and high common-mode input voltages and a P-type (PMOS or PNP) input stage for medium and low common-mode input voltages. These stages normally have different offset voltages, as they have independent random mismatch. When you make a voltage follower and drive it through the voltage range where the op-amp transitions from one input stage to both or from one to the other, the offset changes with the signal, causing extra distortion.
Not rail-to-rail though. Well specifically not rail-to-rail on the inputs, as there is 3.5V drop-out - the outputs have good swing, so as a gain stage it's fine, but as a follower it would be hopeless at +/-5V as the input only works to +1.5V...
Yes, but note that output swing is always load-depedent, often there's graphs to show the dependence of output swing on output load or current.