I promise the package didn't attack any engineers during the development process. But I have personally been hurt on numerous occasions trying to remove stubborn DIP packages from sockets!
We did a lot of work to get the resistor ratio matching very tight, which should make it a useful device for building circuits that need high CMRR. I've been meaning to mock-up a few mic preamp circuits with it, but alas my job is mostly PowerPoint now.
Ah, right, I hadn't really considered how the ΔΣ architecture would push those components up much higher than the nominal sampling frequency. Thanks!
INA1620 is a very cool part - are you at liberty to divulge which OPA16xx core it's related to? I probably shouldn't wander off and compare the datasheets while I'm at work...
I don't think it's a secret that it's the OPA1622 inside, at least no one told me I had to keep it a secret. We did add some EMI filtering at the inputs of the OPA1622 to make it more robust against EMI/RFI but otherwise the core is the same, just with 4 matched resistor arrays around it.
Rad. Maybe I'll hold off on this DAC+amp design until I can get INA1620 in quantities <3000...
(Now that I know what I'm looking for, the specs do make it very obvious than INA1620 is OPA1622 - same quiescent current, same voltage noise, same rated output current)
Although as my plan was to parallel the opamps in the OPA1622 package, INA1620 might be a bit of a waste.
EDIT: Actually, they're pretty affordable direct from TI. Interesting.
Last edited:
Something tells me John might not entirely agree with that!
It is, of course, OK to have favourites.
Since most audio DACs are delta-sigma architectures, there is significant out of band noise above the audio frequency. High order delta sigma modulators have this useful "noise shaping effect". That noise is usually well above the audible frequency range, but if you want to implement a steeper filter slope to remove the out of band noise that's definitely an option. There are some examples in the PCM1792/3/4 datasheets of this. But be aware that there are usually some tradeoffs with THD+N performance due to the higher noise gain in an active filter circuit.
One a different topic, if you don't want to spend your hard-earned money on decent thin film resistors, check out the INA1620 we just released. We added matched resistor arrays to the die, pinned out and unconnected to the op amp, that you can use to build circuits to your heart's content. This was one of the last parts I worked on before taking a new role and moving to another group. My swan song product should hit the market later in the year...stay tuned
INA1620 looks great! Any hints on even the product category of the swan song?
Yeah, that's not reality, sorry. It is very good and has a good blend of characteristics but it is hardly the best in every position. In most circuits that I would use it, it's outperformed by OPA1611.
The National guys made a great op amp for sure, but I don't feel that it's specs are optimal for every application.
The OPA1622 development had the goal of making an excellent headphone amplifier for portable applications. So the focus was on high output current and extremely low distortion with low impedance loads, while consuming less power supply current and having a very small package. That's a very different set of priorities than drove the LM4562 development, which focused on being a high performance amplifier for line level audio signals.
For the intended application you should compare OPA1622 to LME49720 + 2 * LME49600, which is about the same price - sure, it's not a fair comparison (LME49600 has vastly higher output power, but OPA1622 takes up only a tiny fraction of the board space) but at the end of the day both are building blocks in a headphone amp.
I was wondering what the resistors were for on these DIP8 adapted OPA1622, after some inspection it turns out to be a virtual ground for the GND pin on OPA1622. you could remove the resistors and use the through-hole provided to connect the chip to true ground... would there be any point to that?
Not in a cell phone, no room. . . its a great opamp and focused on where real customers are. The entire world of audiophiles could use maybe 1-2 wafers. To pay for the $1,000,000 + cost of the mask set you need to sell a lot of them. Thats why cell phones are a primary target for new audio chips.
Well said.
But a set of reticles for an opamp will cost a lot less than $1M also.
I think you're underestimating the real cost of product development. The amount of development time, test development time, applications engineering, marketing, etc. that goes into launching a product is not free. At least I haven't found anyone willing to do all that work for free yet. The raw materials aren't free either. Neither is the mask set or the fab operating cost.
Have you actually looked at the specs for the OPA1622? Or played with the evaluation kit? I have. It's a quite nice opamp and headphone driver. It's no match for an LME49720+LME49600 (or OPA1611+2xLME49600 that I use in my HP-1) but it can certainly hold its own.
Tom