Real Men Don't Use Opamps

[Wavebourn]

Everyone hears the same, but different people pay more attention to different errors. Also, some are satisfied by fast food that is "good enough" (C) John Curl, some are not... After folk festivals where I tested my equipment nobody said that sound was bad, everyone said that the sound was great. The problem is, performers did not get used to pay such many for sound quality. The paradox is, one audiophile pays more than they are ready to pay for couple of sounds listeners who listen to their performances!

 

[Salas]

I would absolutely agree to this view, provided that you start all your statements with "In my opinion...", "According to my taste...", "It is in my experience...", or any other introduction that excludes generalizations. As flawed as you may think, the ABX test results are at least statistically valid.

It goes without saying. Of course in my opinion. I would not think about fairness if I was oblivious to the subjective factor. I don't think that ABX is flawed. Just politically correct. Its my feeling that using a method that excluded the individual, its nice for the statistics, but defies the purpose of making the equipment for personal use. Must be much closer to public address purposes. There statistics are king. You see, I don't see a flaw, just an application error.
I mean if we show to an alien the average healthy good proportioned young woman of the same race and ask him to ABX her in bed double blind, he will be very right to say that she is as good as the next. But we don't pick our mates like that in real life.

 

[unclejed613]

Everyone hears the same, but different people pay more attention to different errors.

i beg to differ. the shape and length of your ear canal is as unique as your fingerprints.

 

[Wavebourn]

i beg to differ. the shape and length of your ear canal is as unique as your fingerprints.

...and how such differences impact on ability to recognize presence of errors?

 

[Juergen Knoop]

personally, I just find nailing together black boxes boring.
Discretes call for more design work and this is part of the fun in DIY.
It is a hobby, isn't it?
regards

 

[Bonsai]

I thyink there is an element of truth in wha t you say Juergen.

I also think that the IC vs discrte issue is used by some manufacturers to promote their brand. In other words, there is no conclusive proof (measured) or subjective proof (double blind test) that either is better - an certainly none that prove disrete is better.

This is marketing, pure and simple or plain opinion. Not fact.

Yes, discretes are more fun. but are they conclusively better. Probably not.

 

[hags]

I thyink there is an element of truth in wha t you say Juergen.

I also think that the IC vs discrte issue is used by some manufacturers to promote their brand. In other words, there is no conclusive proof (measured) or subjective proof (double blind test) that either is better - an certainly none that prove disrete is better.

This is marketing, pure and simple or plain opinion. Not fact.

Yes, discretes are more fun. but are they conclusively better. Probably not.

Wow, beautifully put and to the point. This has been my experience as well.
I would have to say I wholeheartedly agree.

 

[Wavebourn]

Hmmm. Blowtorch preamp made on IC opamp?
Or Hiraga 20W class A?
(from current threads on the forum)

 

[whubbard]

Hmm, This is all very interesting.

Is anybody willing to try 'copy' the Burson OpAmp? Or a DIY design like it. I wonder what the parts alone cost.

I feel like the estimate of $10 wasn't too far off.

This could be a very interesting diyaudio.com project if it gets started.

-West

 

[Bonsai]

Sometimes I spell like ****.

 

[Bonsai]

Heres my discrete op-amp effort. Will need the addition of a shunt regulator - something similar to Edmonds/Syn08 PGPA topology.

Distortion simulates at nearly -130db ref 0db output. Output flat to 10MHz (-3db) so in practical application it would require some input filtering. In ppm terms this distortion is around 0.7ppm. There is no discernable 2nd harmonic and just a hint of some third if you look at the FFT plot.

Yes, it does have feedback. However, you can change the loop gain by altering the value of R18 - also then you can lower the value of C1. The distortion does increase, but even with c. 15db of loop gain (so R18 set to 4k7), the 2nd and 3rd harmonics ref 0db out are below -100db or about 12ppm. Again, after th e 3rd harmonic there is nothing on the plot - its very clean. not bad for this level of low loop gain.

These figures are admitedly simulated. They still don't come close to a $4 LM4562 which can deliver better performance but with all the bugs ironed out (eg PSRR, offset etc). I guess a big challenge converting a paper design like this to a working discrete op-amp is in the physical execution. Syn08 can comment here!

(sorry, dont know how to post this up as a readable graphics picure - i.e. with enough resolution)

 

[jcx]

Hmmm. Blowtorch preamp made on IC opamp?
Or Hiraga 20W class A?
(from current threads on the forum)

you can do quite a lot with op amps if you care to:

http://www.diyaudio.com/forums/showthread.php?s=&postid=1392458&highlight=#post1392458

by rearranging my amp's output to a bridge/parallel scheme the 6 TPA6120 output op amps on the same heatsink would fine driving 8 Ohms 20 W in Class A - as monoblocks instead of the cascade/stereo Headphone drive they are set up for now

 

[unclejed613]

bonsai, the thing with taking a simulated design and actually building it is that your simulated components aren't imperfect. real components are imperfect. real transistors have different betas and Vbe. 1k resistors aren't always 1k, with a 5% tolerance a 1k can be anywhere between 950 ans 1050 ohms, components also have lead inductances and resistances. they hace pin-to-pin capacitances. electrolytic caps have ESR. all of these things (and more) will affect your circuit's performance.

 

[AndrewT]

Hi Bonzai,
what do we need to read your zipped file?

 

[lineup]

Looks like a Spice file
for some simulator.

Not LTSpice what I can see.

CFJ2.asc
CFJ2.plt

The plot file has got this for Fourier FFT

[FFT of time domain data]
{
   Npanes: 1
   {
      traces: 1 {524290,0,"V(vo)"}
      X: ('M',0,5000,0,1.63835e+008)
      Y[0]: (' ',0,3.16227766016838e-008,10,1)
      Y[1]: (' ',0,-120,30,180)
      Log: 1 2 0
      GridStyle: 1
      PltMag: 1
   }
}

 

[Salas]

It called up LT Spice in my case.

 

[Salas]

.

 

[lineup]

thanks salas
I now got Bonsai discrete op-amp in one schematics image

thanks! Bonsai, too

 

[Bonsai]

Thanks Salas, Thanks Lineup.

Unclejed, Yes, I know the components in a simulation are not perfect remeber the fully balanced current mirror loaded amps - simulate perfectly but cannot be built in practice) - but its pretty easy to change the values in th e simulation and then check the results - I have not done that yet.

Right now I just finished laying out my LM4562 based balanced pre-amp. I'll be looking into a discrete based design in the near future, but I still have a lot of work to do to decide exactly what topology to use, whether is bipolar or JFET based f/end, VFB or CFB etc.

 

[fotios]

Thanks Salas, Thanks Lineup.

Unclejed, Yes, I know the components in a simulation are not perfect remeber the fully balanced current mirror loaded amps - simulate perfectly but cannot be built in practice) - but its pretty easy to change the values in th e simulation and then check the results - I have not done that yet.

Right now I just finished laying out my LM4562 based balanced pre-amp. I'll be looking into a discrete based design in the near future, but I still have a lot of work to do to decide exactly what topology to use, whether is bipolar or JFET based f/end, VFB or CFB etc.

Hi Bonsai

What is this IC "THS4130" of T.I. ? 😉

Regards
Fotios