John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
Half the guesses should be right. ;)
Second try, i'm not so sure of the difference...
When i listened the first time, just after having powered my computer, i found an obvious difference. The second try, after 10 minutes, i listened again, and the two samples seemed very similar to me.
Are they, in fact, identical ?
 
Last edited:
Of course, this is where the mind "filling the gaps" kicks in - once you've heard both versions enough times your head starts to interpolate between the two, the more you listen the more they sound the same - the "reconstruction" that you will refer to, internally, is an average of the two - one's perspective on what's going on is lost ...
 
Dvv and Esperado, I think you are on the right track. SY was just politely dissing you for believing in you own choices and listening opinions. In a small way, he is right, we CAN be fooled, especially if we have never been proven wrong by others, but what you do to improve sound quality is about all you can do. You can't just rely on specs. i.e. that this op amp is much better than another, you can't even completely rely on measurements, even though they are usually a good indicator of quality sound. You have to listen for yourself, and hopefully others will give you a second opinion, to keep you on track.
I design the way that I design for several reasons: First, I like to create interesting topologies that are 'elegant' as well as well performing. Kind of like a sports car engine designer might want to do. This is MY hobby, and why I am so sensitive about topology, and who creates it.
However, I have been taught, mostly by rejection by others, of something that I designed was OK by measurements (at least with what was available and always worked with tube circuits) such as SMPTE IM distortion).
This is where I HAD to try new and sometimes cumbersome ideas, like Otala's, to see if I could truly be successful sonically, and not just be an 'also ran' with the best tube circuits held up as best, and solid state being always behind. That happened when I designed and got rejected by the Grateful Dead, and later when they gave me a second chance, I chose high slew rate, high open loop bandwith, and even rather simple, but elegant circuits to try. As I have said before, this ultimately became the Levinson JC-2 about 1 year later. So, I satisfied the Grateful Dead, then Mark Levinson, and ultimately the audio public.
But that was 40 years ago.
Today, discrete design is still better, but many situations come up where all we can do is change an IC. I am all for picking and choosing IC's for audio improvements. That is what I do, one half the time, these days.
 
Listened one more time, right now (third time). If there is any difference, i would not change a cap for it ;-)
That is in this kind of situation that i need blind verification, or, better, that i consider that there is nothing witch matters.
I imagine the first time (it was so obvious) my sound card was too cold during the first sample (y), and hotter enough for the second one.
 
Esperado, your experience mirrors my own - and that's exactly what happens ... the real world of audio reproduction is that there is a constant flux, a changing, of everything in the equation: how familiar are you with the track? how up to speed is the equipment? how much have you lost interest? how much has other chatter around you made you question your take? It's a complex mix, it's always dynamic, always ...

How to solve this "madness" ...?? By using carefully developed, over time, techniques, that suit your own situation, your personal perspective, that allow you to make the "right" decisions, for moving whatever you're dealing with forward. An impromptu something, from left field, aiming to "prove" that you don't know what you're talking about, proves nothing ... except, that anyone can be caught unawares, that a particular situation for myriads of reasons may not be particularly conduicive for finding any sort of "truth" ...
 
Last edited:
www.hifisonix.com
Joined 2003
Paid Member
You probably need to listen to Mooly's recordings about 50 times blind and on each occasion log which you think is which. Of course, to make the test valid, someone will have to operate the keyboard - no peeking! Once you have completed the test, apply some stats to see if you have something that is valid.

The brains ability to screw with itself is no doubt quite amazing.
 
Disabled Account
Joined 2004
How to solve this "madness" ...?? By using carefully developed, over time, techniques, that suit your own situation, your personal perspective, that allow you to make the "right" decisions, for moving whatever you're dealing with forward. .

Hey Frank how about a specific piece of equipment and some real physical thing that you have done. Please tell us.
 
Okay, I'll repeat a very specific test that I have used over the years, and refer it to the original audio setup that gave me a taste of the good sound, some 30 years ago. This was a nominally powerful Perreaux amplifier, 200W into 8R, driven by a high quality Yamaha CD player, and B&W DM10 speakers, decent bookshelfs. I used a track from Status Quo hits, "Caroline", and played that at various volume levels.

What I was listening to was the quality of the cymbal strikes, which occur regularly, very predictably, throughout that track. Specifically, the quality of the "shimmer", the complex, higher harmonics that particular instrument projects. And I found that at a certain, very specific volume level the quality of those harmonics started to degrade - in conventional terms, you would say the sound of that specific sound element was starting to go "flat" at that point. Now, I had done enough up to that point to know that the cause of that was not the speakers, which only left the amplifier. I reasoned that the most likely suspect was the power supply not being sufficiently stiff under higher load, and so focused on improving the supply's behaviour. Over many years, in different spurts of activity.

And each time I would go back to that recording, and play that: the Yamaha CD player had a display which allowed the precise volume setting to be known - and by virtue of the fact that the setting could be substantially increased before this particular degradation occurred, that gave me the "measure" of progress made.
 
Last edited:
The Perreaux amp - is that a bridged or a single-ended design? And presumably has several output devices in parallel. How to work out the optimum places for the capacitors to be added (assuming that's a major element of your PSU mods, adding more caps)?

Also was the subsequantly modded Perreaux able to be mass-produced in a form with the same SQ? In other words did the SQ depend on 'air sculpture' cap arrangements or would normal PCB construction methods have been sufficient to deliver a low enough impedance supply to the output transistors?
 
Nicely timed, Richard, I just found this ... Perreaux_Amps. It was the 2150B - can you pick the obvious problems ... ;) ?

That transformer got monstrously hot, even idling - the heatsinks ran at just under 60°C, all the time - in hot weather I was always probling temperatures, to be on the safe side, :).

Mine looks a total mess from the top, caps, wires everywhere - a dog's breakfast if ever there was one. In hindsight, I could do much better now, it's always a learning path ...

I wasn't as much into "air sculpturing" as now, and it showed, in that there were still quality issues that would have improved had I been even more fastidious. My chip amp beast, the next project, is where I really started to push that concept as far as I could ...

And this will throw some people here into a real tailspin, :D - at one stage, for quite some time, I had my DIY LM3875 driving the left channel, and the Perreaux doing the right !! And it worked beautifully, the gains happened to match, and it all made good music - the soundstage and imaging didn't disappear up some tailpipe ... :p
 
Perhaps of interest, this is the Yamaha - http://www.diyaudio.com/forums/digital-source/156478-yamaha-cdx-1110-worth-modding.html ...

That belt drive, and the tray loading cam were the weak spots of the mechanics design, I've had to do a bit of fiddling there over the years. Almost no modding of the electronics, just some tidying up - I focused on cleaning up the mains it was being fed mostly, and its physical environment.

From dead cold a very smooth sound, so easy to live with when not being fussy in the listening - trick was to condition it to the point such that high levels of detail emerged.
 
Disabled Account
Joined 2012
I tried with Foobar and doing an ABX and could fairly easily get 5/6 5/7 without to much effort and yet in absolute terms its hard to say one is obviously worse than the other.

You'll know soon enough.

:D

BTW - Easily heard differences on the very first comparison.... I was only using the speakers in my new portable HP computer! BeatsAudio inside!

Should be extreamly obvious there is a difference between low z and high z loading on the amp with a good HiFi system.

So loading does affect/change the sound with this amp. Now we knew the OPS is a weakness in IC amps.... so where do we go now or get back to that subject?

THx-RNMarsh
 
SY, quoting Benaek's Law is wrong on both counts. About as wrong as one can be.

Taking it at face value, the frequeny response at low and high power drive, for my speakers was the No.1 concern we had. Hence the tight tolerance of 40-18.000 Hz +/- 1.5 dB.

Taking it as applicable to my CD player, I have no idea at this time whether I will or will not change anything, as I haven't looked over the scematics in any depth. Exchanging a 1,000 pcs per dollar NJR op amp for an AD high speed amp is really no brainer, it's a rather obvious move which I suspect you also would make under the circumstances. It takes a few minutes to do and just as much to go back to original if you're not happy with the results.

In fine tune with scientific curiosity.
 
Administrator
Joined 2007
Paid Member
I preferred Y on my computer speakers.

Mooly
Y sounds a bit “brighter” to me
George

Thanks folks, appreciate you taking the time to listen.

When i listened the first time, just after having powered my computer, i found an obvious difference. The second try, after 10 minutes, i listened again, and the two samples seemed very similar to me.
Are they, in fact, identical ?

Definitely not identical :)

BTW - Easily heard differences on the very first comparison.... I was only using the speakers in my new portable HP computer! BeatsAudio inside!

Should be extreamly obvious there is a difference between low z and high z loading on the amp with a good HiFi system.

So loading does affect/change the sound with this amp. Now we knew the OPS is a weakness in IC amps.... so where do we go now or get back to that subject?

THx-RNMarsh

Well, 100 ohm loading is severe. Theoretically the opamp could have been asked to deliver just below -/+3 volts pk/pk had a 0db test signal been applied and in fact visually on a scope it passed that test with ease. I wonder if any transient limiting or compression is occurring.

No, they are not. But I won't reveal the differences until Mooly says something.

Perhaps a little later today if there is no more interest.
 
can you pick the obvious problems

The Perreaux models of the late '80s run 10% in class A.
As in 20W class A for the 200W PMF-2150, 30W for the PMF-3150, 50 watts class A for the PMF-5150/5550.

Means that your standard 200 watter idled at a 250W dissipation level.
Stiffing up the rails by adding large electrolytics only pushes heat issues up.

Farmer John would have lowered the bias level

(or replaced the single EI by 2 toroidals with lower secondary voltages, plus a small transformer in series with the main transformer, and added a real driver stage)
 
Status
Not open for further replies.