Hi all,
It's been about a year since I announced the OPA1622 on this forum, so it's about time I released something new! I'm proud to announce that after much hard work we are releasing the INA1650 (INA1650 | Professional Audio Line Receiver | Audio | Description & parametrics, a 2-channel line receiver IC with some interesting features:
- Buffered, high-impedance inputs. We added buffer amplifiers in front of the classic 4-resistor difference amplifier to improve performance with mismatched source impedances. We also integrate two 500k resistors matched to 0.01% on the front end for bias current return.
- Integrated buffered mid-supply output allows the whole circuit to be biased to a mid-supply point for single-supply systems.
- Typical CMRR of 91 dB and with a specified minimum of 85 dB, tested in production on every single unit. We also specify a minimum CMRR over the full temperature range (-40C to 125C) of 82 dB.
- Noise is -104.7 dBu in a 20kHz bandwidth, no weighting applied.
- THD+N of -120 dB for a +22 dBu signal level, 22kHz measurement bandwidth, 1kHz fundamental. -106 dB for +4 dBu signals.
The product folder still says "preview" but that will switch to "active" this coming Monday or Tuesday. There is also an evaluation module which will sell for $49 on ti.com.
It's been about a year since I announced the OPA1622 on this forum, so it's about time I released something new! I'm proud to announce that after much hard work we are releasing the INA1650 (INA1650 | Professional Audio Line Receiver | Audio | Description & parametrics, a 2-channel line receiver IC with some interesting features:
- Buffered, high-impedance inputs. We added buffer amplifiers in front of the classic 4-resistor difference amplifier to improve performance with mismatched source impedances. We also integrate two 500k resistors matched to 0.01% on the front end for bias current return.
- Integrated buffered mid-supply output allows the whole circuit to be biased to a mid-supply point for single-supply systems.
- Typical CMRR of 91 dB and with a specified minimum of 85 dB, tested in production on every single unit. We also specify a minimum CMRR over the full temperature range (-40C to 125C) of 82 dB.
- Noise is -104.7 dBu in a 20kHz bandwidth, no weighting applied.
- THD+N of -120 dB for a +22 dBu signal level, 22kHz measurement bandwidth, 1kHz fundamental. -106 dB for +4 dBu signals.
The product folder still says "preview" but that will switch to "active" this coming Monday or Tuesday. There is also an evaluation module which will sell for $49 on ti.com.
Great that TI actually make new audio parts, but it's one noisy thing, has to be with those high value internal resistors....
Has to be? Pray tell why?
Looking at it, nice to see THAT have some competition now, although not had a change to compare the performances in detail. One comment I will make is that in sections 9.2 and 11.2 people will be forever telling you that you have a pin 1 problem there unless you take the time to explain in detail why you haven't!
Looking at it, nice to see THAT have some competition now, although not had a change to compare the performances in detail. One comment I will make is that in sections 9.2 and 11.2 people will be forever telling you that you have a pin 1 problem there unless you take the time to explain in detail why you haven't!
"For a 4–dBu signal level, the second harmonic is barely visible above the noise floor at –140 dBu. Increasing the signal level to 22 dBu produces distortion harmonics above the noise floor. The largest harmonic in this case is the second at –111.2 dBu, or –133.2 dB relative to the fundamental."
?
//
?
//
It's significantly lower noise than the older INA134/2134. About 31 nV/rtHz at 1kHz.
There is a tradeoff between CMRR and noise. Achieving high CMRR requires extremely close matching of the 4 resistors in the difference amplifier, this is easier when the resistors are larger values. The 10k resistors in the INA1650 are matched to around 0.001%
There is a tradeoff between CMRR and noise. Achieving high CMRR requires extremely close matching of the 4 resistors in the difference amplifier, this is easier when the resistors are larger values. The 10k resistors in the INA1650 are matched to around 0.001%
Yes! It was more a comment to Sören. Can -111.2 dBu /-133.2 dBc really be considered "noisy"?
//
Yeah, the price is good. But with one of the applications listed as "Audio Analyzers", it's just too noisy for that.... The front page also say "Ultra-Low Noise: –104.7 dBu, unweighted", in a world of 1nV/√Hz opamps I wouldn't call 31nV/√Hz "Ultra-Low Noise"....
It should also have been a little faster, THD start to rise quickly at 10Khz, even the old INA134/2134 is faster....
You might want to reconsider the THD+N curves, Fig 13 have around -100 dB using 500 Khz Bandwidth, Fig 14 around -110 dB at 3V RMS using 80 Khz bandwidth and the frontpage THD -120 dB number use 22kHz Bandwidth, the difference are due to different bandwidth used, might be better to use same bandwidth as the curves are mostly +N curves only....
Last edited:
The INA1650 has been moved to "active" from "preview" so you should be able to order samples and the EVM now (just in time for the holidays!) 😉 .
Yes, I could see a person unwittingly connecting the shield circuit incorrectly.
It seems that even two decades later, shields still get hooked-up incorrectly.
Pin 1 is shown connected to the chassis ground in all schematics. I do also show the bleeder resistors (and the COM resistor where used) connected to the same ground, which is perhaps up for debate and I'm open to opinions on that.
You are correct that maybe the datasheet should include a conversion on grounding in these systems. That leaves room for application notes in the future.
May I make an educated guess that the internal opamps in the INA1650 are based heavily on the OPA1688?
If I simulate the circuit with the same 8nV/1.8fA noise spec as the OPA1688 I get 4.64uV or -104.5dBu, which is spot on.
Some datasheet-comparisons show that the short-circuit output current is 75mA for both, and the output swing with 2kohm load is almost exactly the same. The FPBW and SR is also very close.
Since the INA1650 datasheet doesn't say that it is bipolar or FET based, I assume that means it's CMOS based. Thats probably also why the price can be kept affordable.
Am I close?
If I simulate the circuit with the same 8nV/1.8fA noise spec as the OPA1688 I get 4.64uV or -104.5dBu, which is spot on.
Some datasheet-comparisons show that the short-circuit output current is 75mA for both, and the output swing with 2kohm load is almost exactly the same. The FPBW and SR is also very close.
Since the INA1650 datasheet doesn't say that it is bipolar or FET based, I assume that means it's CMOS based. Thats probably also why the price can be kept affordable.
Am I close?
Yep, there's no reason to reinvent the wheel on every product. Using an amplifier core which has been thoroughly tested and meets the products needs can help shorten development timeline and ensure that you get it right on the first try. I also much preferred that our development team focus on the internal resistor matching rather than worry about debugging an amplifier core which had never been tried.
- Status
- Not open for further replies.