I remember trying to explain to Charlie Hansen (RIP) that semi companies didn’t give a **** about audio and weren't going to do anything special in any fab for a business that was perhaps worth 0.1% or less of the then $500 billion market. I don’t think he was convinced. Most semi companies even see auto as a PITA, so only a few specialists operate in it successfully (NXP and Infineon for example). Today the big bucks is in processors for computing, mobile and memory.
Niche companies like That and Linear Systems and won't interest the big guns because the potential for growth isn't there.
The market for opamps is healthy but there are too many of them. I don't see TI giving them up as they seem dominant currently. Innovation will be needed for consolidation of attributes. Mixed signal operation is another reason for not dumping analog. Microchip are embedding opamps into processors and for all the real world stuff to make it in/out to/from digital, you still need analog.
The market for opamps is healthy but there are too many of them. I don't see TI giving them up as they seem dominant currently. Innovation will be needed for consolidation of attributes. Mixed signal operation is another reason for not dumping analog. Microchip are embedding opamps into processors and for all the real world stuff to make it in/out to/from digital, you still need analog.
Ulrich Rohde seems to find time to support experimenters in the ham radio sphere...in fact he just wrote a letter to QEX !
So who came up with class-D chipsets?
While I would agree that the market for high end home audio is small enough that major semiconductor makers don't really care too much, the car audio and consumer device markets for audio amps is pretty big. Likewise for CODECs.
Analog Devices still participates in the pro-audio market. They are probably the largest semi maker there. Their line of Sigma DSPs also dominate audio DSP controller space
Car audio is a big market (class D). Consumer audio in mobile devices might be big, but the hifi market is small. Codecs are mostly done in software now. DSP in consumer is almost always embedded. Philips supplied the DSP IP to a lot of Japanese companies for consumer applications.
I'm not arguing with you, man. [rant on] But....
The 2021 version of the TL07xx datasheet is out to Ninety Pages!
30 pages are dimensions which SHOULD be in their own document as company or industry Standards.
There's real klunkers in here. Page 33 Fig 6-45: how does Vom drop to zero so quick? 2V or 3V at 0.1k seems more like it. That puzzled me in 1979, because it is just wrong-drawn.
Page 40: Equation (1) has numeric values which just are not right. Later RI is selected at 10k (not 1k) "because the amplifier circuit uses currents in the milliamp range." But can the source supply milliamps? (And if so, why use a JFET part?) Is +/-12V really ample for a +/-1.8V output? (Where is the loss data?) Or is it gross overkill (running up product costs)?
Fig 9-6: the 3.3k parts can't all be needed. Arguably the one on the cap, but surely not the other.
What is page 42 fig 9-9 teaching us? The 50Ω input termination suggests >1MHz signals. But the TL0s at gain=100 won't top 50kHz. The 0.1uFd on the bias divider won't touch the real world's most popular noise (100/120Hz buzz). A phono preamp would surely be a better example of... whatever.
And--- Where/when did page 42 Fig 9-4 9-5 come in? This is an intuitive insight into the BiFET's phase inversion mis-feature and should have been published 40 years ago. And why does page 30 fig fig 6-28 show "No Phase Reversal"?
[rant off] I know this is asking "why is your boss an idiot?" And from my idiot bosses, I know they may be brilliant in some way, but organizationally each has his/her head in a different patch of sand and any consensus is illusory.
How many new vacuum tubes have been designed and successfully introduced to the market since 1970? My guess is not many the big scheme of things.
Most of the innovation in semi’s is now in processes and packaging, and less in circuit design (as in analog) or system functions.
Seriously, how many more opamps do we need?
There is still quite a bit of innovation in the analogue circuits inside mixed-signal chips. The supply voltages are getting lower and lower as the processes are scaled down to make the digital parts cheaper. Somehow you have to get the analogue stuff to perform acceptably despite the low supply voltages (and leaky gate oxides and other peculiarities of deep submicron CMOS). Depending on the application, you can sometimes get away with running everything off the I/O supply and using only thick oxide transistors, sometimes not.