This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

THS6222: Composite applications for audio

Part Number: THS6222
Other Parts Discussed in Thread: THS6012, TPA6120A2, OPA1602, OPA1612, OPA2210

The TPA6120A2 and THS6222 are comparable in terms of cost for in-loop composite applications. The former appears appropriate for a gain of 10, though examples are thin on the ground. The latter is clearly designed to be “high” gain capable, but due to its market segment has some quirks that I need to dig into here. The THS6222 has the significant advantage of lower quiescent current, and I’m guessing there are a lot of other improvements that TI has made since the THS6012 came out.

So, regarding the THS6222:

1) Other than the mild bandwidth limiting associated with Low Bias mode, what are the other results? There is a reference to increased distortion, but no data to shed light on the topic. If you set up the amp for Low Bias and add resistance to ground at the Iadj point, you can significantly decrease Iq, but what happens as you dial from 0R to ~6K? Are there further bandwidth and distortion tradeoffs? To what degree?

2) Clearly, the amps must be used differentially due to the T-bias network. But what else defines this part’s differential-only requirement? The data sheet does not clarify. If the amps are given gain via the expected Rg/Rf network, a single-ended input should balance to a degree, but not fully (within 6dB IIRC). Does this part exhibit that typical behavior? If the amps receive a differential signal that is somewhat imbalanced (again within 6dB) is there a specific performance degradation or limitation? What happens?

3) Pardon my inexperience with IMD in high speed environments: Does the egregious ramp in IMD below 1MHz matter if the amps are used in-loop? What else does the IMD graph indicate about this part at audio frequencies? Are we simply looking at the results of two tones around 1MHz and if we saw the results of two tones conventionally used within the audio band we would have a completely different graph? Can you make that graph, if deemed necessary?

Dual 15 or 16V supplies, by the way. Lead op amp is anything in the OPA1612, OPA1602, OPA2210 genre. Output of 16dBu to 28dBu, differential, depending on application. The part might not quite hit that top mark but comes close enough IIRC.

  • Hi Jonah,

    1: Unfortunately, the devices' bandwidth and distortion were not fully characterized as resistance is added to the Iadj pin. However, as the Iq continues to drop off by adjusting the Iadj pin, we can expect the devices' bandwidth and distortion performance to decrease as a result.

    2: The device's architecture is designed to reject common-mode signals at both inputs and does not allow the devices' internal amplifiers to work independently. Setting the device in a single input configuration will degrade the performance of the device and the highlighted performance in the datasheet cannot be expected in this configuration.

    3. We reached out to our team for more information on the ramp below 1MHz for the highlighted IMD performance and we will get back to you as soon as possible on your questions regarding the devices' IMD performance. However, the tones are set around the given fundamental frequency like you mentioned. 

    Best Regards,

    Ignacio

  • 1. At all frequencies, or just in the PLC range? At what point does bandwidth drop below (say) 100MHz, and at what point does distortion increase below (say) 800kHz? Any other data along these lines would be fruitful.

    2. I get it. It’s not an FDA. The partial conversion from SE to diff in a conventional buffer-with-gain situation is not ideal. The non-driven side has the additional noise gain of the feedback network. But the stage benefits from having both inputs present the same impedance to the line, which is expected to have the same common mode voltage regardless of whether both or one side is driven. I’m definitely not talking about grounding one side locally.

    In this scenario, the op amps doing the driving are immediately in front of the THS6222, and have simply reproduced whatever CM voltage they have been presented with, and have brought the signal into a differential-but-unbalanced state. Make sense? Maybe it would help if i mentioned that Vcm is grounded and the CM signal is in the mV range for DC and uV for AC. There’s no level shifting here.

    I can accept part of your answer as “Don’t do that,” but I think it’s reasonable to ask again what happens if the signal seen at the inputs is A) dual driven, B) not balanced but close, and C) at the same CM voltage per side.

    3. Thanks for the reality check! I will look forward to their reply. Maybe they can look deeper into the questions above. Cheers.

  • PS re: #2

    I’m talking about an instrument signal. Literally a guitar, instead of the typical microphone.

  • Hi Jonah,

    Unfortunately, the only data that was characterized for this part is highlighted in the datasheet. As for your second question, in the scenario you mentioned it is safe to assume the imbalanced input signal would cause non-linearity. You will be driving both inputs but due to this mismatch you can expect irregularity to some extent at the output. We will also continue investigating with our design team to understand the IMD behavior, so I thank you for your patience.

    Best Regards,

    Ignacio

  • Please provide the graph that this 1kHz data point came from. I would like to see noise performance in the audio spectrum. Clearly it has been characterized but was not included in the data sheet due to market segment. Thank you.

  • Hi Jonah,

    The value you are seeing seems to be a typo on our end. The graph in the datasheet is the best source for the noise performance of this device. 

    Best Regards,

    Ignacio