OPA2210: Audio applications

Hi Jonah,

1. With a 600Ω load, is 7V the maximum output before THD+N takes off into the unusable region (as shown in figure 6-7), or is it 600mV from either rail (as stated in the Voltage Output Swing section)?

Figure 6-7 which shows the minimum THD+N with an output level of ~7 VRMS (~19.8 VPP)agrees with the THD+N information in the Electrical Characteristics table, Total harmonic distortion + noise (THD+N) G = +1, f = 1 kHz, VO = 20 VPP, 600 Ω, 0.000025 %. 

2. Can the amp put out 10V into a 2K load at its lowest THD+N? The actual load in question is a non-negotiable 1476Ω. The system's maximum level before significant noise increase has to be ≥22dBu.

We don't have THD+N characterization for a 2 kΩ, or 1.476 kΩ loads, because the emphasis was placed on the much more common 600 Ω load impedance. The one parameter that should make a difference in the THD+N is the open-loop gain (Aol) which will be higher with these lighter output loads compared to 600 Ω. Exactly what voltage the lowest THD+N will occur with these lighter loads would have to be determined by measurement.

3a. What is the minimum feedback resistance possible when the amp is used as a voltage follower (Av=1) at frequencies below 200kHz, but most importantly below 20kHz? Put another way, what is the lowest frequency transient (at, say, 10V) that might trigger latch-up in such an application? The amp is slower than a 5532, so unlike the LM4562, it would seem to be a good candidate for traditional 0Ω feedback buffer applications at audio frequencies.

3b. 3b. If feedback resistance is required, would ~500Ω be adequate, or is 1KΩ preferred? Could something smaller, such as 100Ω, be used? Would a feedback capacitor in parallel be required for stability, or rendered unnecessary due to the relatively low slew rate?

If the load is resistive with very little capacitance the preference is to directly connect the output back to the inverting input, without a feedback resistance. Doing so eliminates the thermal noise contributed by a feedback resistor. Additionally, feedback resistance in conjunction with the OPA2210 input capacitances introduces a pole in the feedback path which reduces the phase margin. If a feedback resistor is used adding a small feedback capacitance across the resistor can be used to compensate the circuit and increase the phase margin. 

Regards, Thomas

Precision Amplifiers Applications Engineering

Hi Jonah,

usually the distortion decreases when increasing the load resistance. So with a 2k load the distortion should not be higher than with a 600R load. The mechanism behind is explained by Thomas: A lighter output load (2k instead of 600R) results in a higher, linearizing open-loop gain :-)

Kai

Thanks Thomas & Kai. I understand: Look at the THD+N figure, not the maximum output figure, with the subtext being that you can get a higher swing out of the amp into a given load, but with a noise penalty.

Looks like the 2210 is quite the performer at 19dBu. I'm looking for a general purpose precision dual with four zeros in front of its THD+N figure at 22dBu into 600Ω. A maximum offset under 100uV would do it – the 2210 overachieves in that arena. Bias current doesn't have to be particularly low, as source resistances in this application are all 1K or less (well, there is one 2K source in there). Any suggestions? ±15V operation, which I forgot to mention earlier but (Thomas) you deduced. I had designed around the LM4562, but the strong power draw and popcorn noise issue with a small percentage of units started me off in a precision direction. Fewer capacitors are a good thing.

[EDIT:] I forgot to mention, this design incorporates the INA1620 as a balanced receiver and line driver, so basically I'm looking for something that I won't have to servo, which can match that amp's headroom. Even 20-30mA output would be fine.

Hi Jonah,

How about OPA1622 although it is not pin-to pin compatible to LM4562.
It has enough drivability for 22dBu/600ohm in good THD+N (about 0.00002%).

I suppose OPA1656 might be another alternative in the more standard package option(SOIC), but its offset seems to be higher than  you expected...

Best regards,
Iwata Etsuji

Same conclusion I keep coming back to, Iwata -- 1622. I had actually designed that in in a number of places, then wiped it, then put it back in tentatively when I finally looked at the offset population. It would be funny to use a headphone driver as an inverting line amp or buffer, but it does seem to outperform everything else.

The one thing I'm confused about regarding the OPA1622 and INA1620 is the sky-high input bias current. Looking at the input resistance charts, both seem to do just fine with sources under 10K, and even 10K is hardly worse than any audio FET, about a 5nV/rtHz difference. So maybe not explain-like-I'm-five, but explain-like-I'm-not-an-engineer: How does high Ib not equate to bonkers noise with these?

One thing I'm doing is using the 1620 as an instrumentation front end, so while pro audio line outputs down around 100Ω are the typical sources, I could also expect to see oddball '80s and '90s studio equipment with 1KΩ and 2KΩ output impedances -- and when you plug in an electric guitar directly, the source is 5KΩ-13KΩ. That's an unusual application but it will happen.

I'm also curious about the minimum feedback resistance with both of these. Seems like 499Ω would suffice (or the ~420Ω to ~510Ω of a 1620 resistor pair in parallel), based on some research, but I haven't found anything conclusive. Thoughts?

Hi Jonah,

AS the input stage of OPA1612 and INA1620 are constructed with the bipolar junction transistor (BJT), the input bias current and input current noise with BJT input are generally much higher than JFET (i.e. OPA1642, OPA828, etc.) and MOSFET(i.e. OPA1656, OPA1678, etc.).
Therefore, we recommend to use lower feedback resistor value for BJT input than 5k - 10kohm as you can see in the Fig.59 in OPA1612 datasheet.
While, I have never heard anomalous noise such as bonkers noise or popcorn noise for OPA1622/INA1620 so far. 

Regarding 500ohm feedback resistor for OPA1622 and INA1620 in unity gain configuration, it does not directly cause the degradation of THD+N since  both OPA1622 and INA1620 have the enough to drive 600ohm load in good THD+N as you can see in the FIg.16 of OPA1622 datasheet.
However, if the total value of the feedback resistors and output load resistor is much less than 600ohm, it should cause the THD+N degradation as you can see in the Fig.17 of OPA1622 datasheet.
In the case of 1.476kohm for the output load, the higher value as the total value of  the feedback resistors than 1.01kohm are recommendable.

If you would like to use the much lower feedback resistor value, how about the solution with opamp + buffer such as Fig.51 of OPA1656 datasheet.
This solution can be applied with OPA2210, but we have no data for THD+N evaluation results with OPA2210 + BUF634A unfortunately.

Best regards,
Iwata Etsuji