As for the earlier comment about inability of IC's to reject incoming EMI, that's simply not correct. You have EXACTLY the same issues to deal with in discrete circuits. If you are referring to JFET vs bipolar ability to reject RFI, then again, this applies to discretes as much as it does to opamps.
So far wrt opamps we have debunked just about everything thrown at them that are supposed to make them 'inferior'. We've agreed that for some specific ultra low noise applications, high output drive current (although this is easily fixed in my view) and high (i.e. > +-15V), discretes offer advantages.
Anything else we can look into?
No EMI is often input stage current related. I can increase the current on discrete devices to levels that would destroy monolithic devices!
SY,
The absolute proof would be to listen to every contender from both camps in a reasonable test, as my time may actually be finite such a test is most unlikely. So for my personal use I will do a quick listen and rule things in or out based on my taste. That then provides my OPINION on what I prefer.
That seems like an extreme position, Ed. Here's something much simpler: Compare level-matched input to output of a good chip amplifier by ear (blind). If you hear no difference (and I'll put my money on that if the level matching is done tightly), you're done- the amplifying stage is as good as it needs to be for you. Anything else is either fashion, fun, or the need to perform under unusual conditions (e.g., having to swing 100V).
Actually that is most of the way I do it. Sometimes it is two identical devices (ex. CD players) where one is modded and the other not. Differences do show up and swapping the DUT will show if it is mismatch or actual difference.
The best example is the Burson opamp does work better in my CD player. Most likely due to reduced EMI issues than the stock bipolar I/V output converter.
On the bench I will plug in several different parts and often there is a measurable difference but no change to my ears. Then sometimes there is a difference but I have yet to encounter a difference that did not show up in measurements. The real issue is are you measuring the right thing!
Did someone say blind ?
YouTube - David Byrne Live - Blind (1992).mpg
(best is the one with background vocals by O-Sammy-B)
YouTube - David Byrne Live - Blind (1992).mpg
(best is the one with background vocals by O-Sammy-B)
You don't seem to get the "blind" bit!
Without it the rest is .......... silence.
Standard procedure is to use a six position switch. Pos 1 is A Pos 2 is B Pos 3-6 are mixed and unknown. When the difference is large enough it is easy to pick out the other 4 positions. When it is not... switch set number 1 has a noisy position 4, switch set #2 has a noisy position 2 and not so noisy position 4....
Last edited:
Well, OK, I'd take this discrete baby over my National opamps... Too bad you can't order free samples.
Shades of the JE990 inductor degenerated input.
Well, OK, I'd take this discrete baby over my National opamps... Too bad you can't order free samples.
Yup, that's what ultra-high feedback can give you!
jan didden
Then you need to look at you front end filter design. I'd say fiddling with the LTP current is trying to cure the symptom and not the cause.
And, is this the sole cause of the 'gritty, harsh' sound, when there are many discrete designs that also operate at low input currents?
You can add parts but it is simpler to raise internal levels to where it is not an issue, as is done by many talented designers.
Harsh and gritty were not the issue. EMI is more often perceived as an increase in noise level.
I looked at the Weiss data sheet a while back and some of the distortion figures are truly amazing. I can only assume they used distortion magnification techniques to get some of those readings, or used the trick NS did to get them.
If these are genuine numbers, then they are very good, no question. But, they are going to be expensive.
If these are genuine numbers, then they are very good, no question. But, they are going to be expensive.
Well, I don't entirely disagree with you, but leaving the input of any amp wide open is not a good thing in my view, so even if you did raise the input stage current levels (BTW on my designs I always run the front end 'hot', but for different reasons), its still wise to bandwidth limit using a simple filter. There's too much gunk floating around nowadays to take that risk.
Yes, truly amazing claims - but he seems to have the test results to back them up. And I don't see anything in the photo on his "OEM page" that makes me skeptical.
I can make out an AD MAT12 in a metal can - a US$20 part in its own right, though I have occasionally wondered if it's measurably better than the THAT300 series. The heatsink on the output devices supports his claim of being able to drive a back-terminated 50 ohm line. Not sure what all those things in the 6-pin SMT packages are, or whether the LED is there for benefit of the signal or the user.
His sales flyer says the "competitive price" is "Swiss Francs 120 in small quantities".
Dale
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.