Only if one limits what one attempts: many circuits are vastly superior or only possible with opamps, of course.
Discrete is the only way when you need higher voltage - the highest voltage amp IC I know of is LME498xx and those have just been EOLd. Below that chipamps typically peg out under 100V.
At the other extreme, my latest design headphone buffer draws considerably less juice from a battery than an O2, which is IC-based.
At the other extreme, my latest design headphone buffer draws considerably less juice from a battery than an O2, which is IC-based.
Last edited:
Advantages of integration: exquisite matching, amazing bandwidths. Disadvantages: as mentioned, limited voltage (and current) swings, limited dissipation, lack of flexibility, noise floors limited by limited areas of devices (although one can contrive compound circuits from a plurality of integrated devices that can push down the noise, increase the current limit).
But integration requires substantial volumes to be feasible economically. For limited-volume apps we have to take what we can get. Fortunately there are a lot of choices.
But integration requires substantial volumes to be feasible economically. For limited-volume apps we have to take what we can get. Fortunately there are a lot of choices.
Considering 100GHz ft complimentary discretes don't exist and/or there is no discrete packaging technology to make them usable, I would say that is a statement of fact in some cases. Take a buffer for a 90dB SFDR A/D at 500MHz sampling for instance.
Or any modern Cellular tech. 3G was made possible by the amazing advances in converter tech during the mid 90s. Much to the chagrin of the old school RF designers 🙂
(<= sign of approval, article 1, paragraph 1, Cal Clause)
Ah thanks, I had come across Apex some time ago and found that their idea of 'low cost' ($10 for a 20W-peak chip in volume) doesn't quite fit my own. At least they provide the internal schematic so a comparison with a discrete version is possible 😛
Must say the computer to vacuum tube one for 50 dollars sounds interesting for that direct drive tube buffer.
I thought I recently seen either a L.T. or A.D. offering in that category, but can't find it now .....
There is a new amp for envelope tracking power supplies in base stations, forgot the number. Heroic for an IC. There will be a +-200V driver but that will be for 1% or so duty cycles.
Note that Apex alleges no secondary breakdown, but that's not to say that DMOS parts can be used in linear mode with no concerns about localized self heating.
Apex used to do mostly hybrids, so it is interesting to see their monolithic offerings. And of course as one pointed out, they are hardly cheap. But if you need high voltage they have been the leaders for many years afaik.
Apex used to do mostly hybrids, so it is interesting to see their monolithic offerings. And of course as one pointed out, they are hardly cheap. But if you need high voltage they have been the leaders for many years afaik.
There is still a difference between discrete and IC's for audio. For the record, I do both. In fact, I recently asked everybody about what IC might be better than the NE5532, and I got a lot of good answers. However, even in that limited case: lowish supply voltages of +/-15V or less, unity compensated dual, etc, I found some contenders at a slightly higher price, perhaps 4 or more times, and I think a worthwhile substitution. I won't be designing a discrete gain block to replace the 5532. However, my REAL effort is in my Parasound JC-2 preamp design which is all discrete (except for servos) and it will still win out in an all out listening comparison, except for double blind tests, where little seems to matter.
Not quite, I do belive you meant to say ic's, not opamps. I haven't seen anyone here really disparage opamps.
Alan
But hang on the JC-2 is a CTC effort. It says so on the back. You can't claim all the glory for that?
- Status
- Not open for further replies.