IC's get hammered

Bonsai

Member
Paid Member
2003-07-25 10:44 pm
Europe
www.hifisonix.com
I was doing some research as part of the work on my new pre-amp, which uses a modular gain stage design (both IC version and discrete version using LSK389).

My main reason for the modular gain stage approach is to be able to drive heavy loads (600 Ohms) fully in class A up to 10 V pk output, in amongst a few other things.

Anyway, I came across this and thought I'd share it.

Low Quality Components = Tasteless music
 

jcx

Member
2003-02-17 7:38 pm
..
read http://www.ti.com/lit/ds/symlink/tpa6120a2.pdf and laugh at Burson's pretension, ignorance

graphs show one channel driving 600 Ohms to 12 Vpk with < 2 ppm distortion, 64 Ohms, up to same level @ < 6 ppm - all with the nominal Class AB output


you can parallel the 2 channels in the package with a offset V, 1 Ohm current sharing resistors for Class A push-pull

I did it with >200 mA Class A bias

even a young college kid who swore he could tell op amp swaps "by ear" in blind tests couldn't tell when the Class A bias offset was turned on or off in my TPA6120 multiloop headphone amp
 
Last edited:
why is anyone surprised? it's all over the web nowadays.

IMO the real issue is the fact that some people believe them simply because they like the 'concept'. "pure parts", "custom design" etc. just look on this website, people hear night and day differences between resistor brands when even complete amps are many times not very different when compared side to side. so I guess the guys actually own some market...

and, again, it's the same symptom: pseudo-objective explanations by subjectivists. look at the comments about EMI. geez, I wonder how come my mobile phone works, being full of ICs and working at almost GHz ranges. but have you looked at their discrete stuff? I'll show it to you. no physically close parts or anything like that.
oh, and you can bet the IC makers have no idea what EMI/RFI or coupled noise is. and sure as hell they don't care about it, because none of them actually make ICs designed to be used in precision applications.
 

Attachments

  • 3714_BURSON-OPAMP-DOUBLE_2.jpg
    3714_BURSON-OPAMP-DOUBLE_2.jpg
    44.9 KB · Views: 506
In order to sell a solution, first you have to create the perception of a problem.

To sell medicine/vitamins etc., convince healthy people that they are ill.
To sell cosmetics, convince beautiful women that they are ugly.

Then, to keep them, convince your customers that they are wiser/better/richer people than the foolish/jealous/ignorant/poor critics who scoff.
 

FoMoCo

Member
2012-12-04 10:04 pm
In order to sell a solution, first you have to create the perception of a problem.

To sell medicine/vitamins etc., convince healthy people that they are ill.
To sell cosmetics, convince beautiful women that they are ugly.

Then, to keep them, convince your customers that they are wiser/better/richer people than the foolish/jealous/ignorant/poor critics who scoff.
You just hit the nail on the head. Same thoughts as mine, just said better.
 
the Burson discrete regulator:
Superpower Super Regulator by Belleson

worse than a 7805 at ripple rejection. and looks like for some reason the Burson NASA-level scientists reached the conclusion that horrendous transient performance is a must for good sound.

funniest thing of all is that in a completely subjective shootout done by some local audiophiles their DAC came out as one of the worst sounding.

what beats me is... why not simply market your products as good sounding and be done with it? why resort to such ridiculous statements like the ones about EMI? anyone wanna bet that their products are worse EMI-wise compared to standard ICs?

PS: the banner just above this edit box has a Burson ad LOL
 
its in the magazine, worth a read. the dac PCB layout is quite terrible (scary), its not surprising they managed to make an ESS dac that performed at a level able to be clearly distinguished as bad. funny they talk about EMI, the thing looks like an EMI nightmare. 2 layer PCB punctured with hundreds of PTH, because SMD parts, except for the dac itself of course, are so abhorrent and dont get me started on the evils of ground planes

the test result worse than 7805 was what I was talking about, quite funny, that takes some effort
 
dont get me started on the evils of ground planes

[IMGDEAD]http://bursonaudio.com/images/160DS_In2.jpg[/IMGDEAD]

and since it seems that the regulator shooutout is only available in the printed version of Linear Audio, here's a table and Thorsten L's opinion: DIYHiFi.org • View topic - Power Supply testing/debate at Linearaudio & Pub...

I'm rubbing my eyes in disbelief. I mean... is anyone surprised that a regulator which, well, seems to not regulate too much is imposing something on the signal? therefore changing the signal? (at which point we should remember that different sound = better sound). why is it even called a regulator when in fact it seems to be an effect box?
or why even consider the fact that the behavior of that thing is clearly circuit-dependent and it will likely sound different in other implementations? "but is has electrolytes!"
 
this is really too much. I mean, measurements don't tell everything, we know it etc.

but when one designs regulators which don't regulate in order to modify the signal, something is certainly wrong. ok, I get that with op amps there's the slightest trace of logic in it, because signal passes through them. but when one intentionally gets to alter the signal by altering the power supply, it's the Apocalypse of common sense.
"but it has electrolytes!"
 
this is really too much.
Meh, I've seen worse. Have you seen any of Chris Daly's creations? e.g. this: http://www.diyaudio.com/forums/analog-line-level/209054-light-dependant-resistor-current-control.html

Take a long hard look at the schematic in post 5. Nope, no mistakes, that really is what the circuit looks like. The explanations of how it works were hilarious.

And yes, it counts as a commercial product - He actually built those and sold them on eBay, along with opamp upgrade modules based on this idea: http://www.diyaudio.com/forums/solid-state/202003-powering-opamps.html
 
let me see if I get this right. guy said to himself "I'll draw a schematic so that everything suggest current instead of voltage, because current is good and voltage is bad. I'll put a LM317 in current source configuration and throw diode connected transistors all over the place. just throw them together and make it as weird as possible. maybe it even works".

PS: thanks for the Pease articles on Howland current sources though :) there's a good part in everything I guess.