It's unfortunate that Linear Audio has created a bad vibe here. I would prefer that everything I publish simply be put into the public domain, but Jan does all the work. It's not like it's a staggering hardship to support this process.
No problem using premium parts everywhere. BTW - I do not recieve money from their sale. I have a patent on it and did get a small percentage before the royalty expired and then they cut me out of the income loop. Oh well.
LA hasn't created the bad vibe, and I think most folks understand the necessity of supporting high quality publications.
Yes patents and payments - I did market development work for a company from India - as soon as they were introduced to clients or large projects occoured they cut me out or don't answer my phone calls either.......................... Their main technology came from the Swedish parent company they convinced to take 40 % ownership in the indian operation and transfer manufacturing technology to be eligible for Indian government contracts and then bankrupted the first and sold all the imported tooling to a second Indian company for cents on the dollar and reopened.
I owe you a scotch or two then.
I owe you a scotch or two then.
Last edited:
Quite an amazing figure for an open-loop architecture (assuming your final design is an open-loop one: isn't it?) - 30 ohm is my design goal, too, and I've managed to get pretty low THD figures on resistive loads (well, not as low as < 0.001%, anyway - I'm really impressed: have to get this LA issue...), but when switching to real loads (ie genuine ~30 ohm Z headphones) THD raises by almost an order of magnitude or so at low frequencies - that's why I was wondering about output impedance of your amplifier.
L.
I agree wholeheartedly. LA hasn't created the bad vibe, its quite blameless (in the everyday meaning of that word!). I freely acknowledge my share of the blame for the bad vibe, and I apologise for that.
As RNMarsh seems to acknowledge in post #80, Linear Audio is a commercial concern - nothing wrong with that at all. What I do have a problem with is selective enforcement of rules, as per Note #4, wrt advertising in non-commercial sections.
Most of those who write for Linear Audio are also prominent members here. I think that quite accidentally, an unfortunate situation has arisen. I also think that it could be resolved quite easily by giving Linear Audio its own section, possibly with a simple header stating that it's a good friend of DIY Audio, which supports it wholeheartedly, and that many of its contributors post here.
Simple, up-front, clear and open.
LA does look good. Those involved in it presumably know it's good. The rest of us have to pay up to find that out. DIY Audio's clear and open endorsement would make that an easier decision to make.
The figures very quickly deperse the idea of open loop.
These figures probably have been achieved for a headphone amp in the late 1960s when JLH published his design, at the time there was no measuring instrument that could measure it. The quality of the parts of that time might have had a detrimental effect.
One thing i like of this design is that is NOT one of (too) many amps for high Z HPs.
Mmhhh... never say never. The buffer jackinnj assembled and tested is indeed working open-loop, and my open-loop prototype (which is by far more complex, anyway) seems capable of THD in the 0.001% range. As I have already pointed out, output impedance seems to play a crucial role in keeping THD low on heavy * real * loads (~30 ohm or so) at LF, so imho the real issue with OL architectures is how to keep output Z very low (< 1 ohm or so, I mean) without GNFB.
L.
Last edited:
It isn't easy without gnfb. At one point in time, I was up to a 12 transistor buffer circuit. I thought this was getting way too complex. How are you doing with your buffer circuit? Did you put it on diyaudio somewhere?
Not yet - there are still some issues to be addressed in the low Z out (< 1 ohm) prototype: I'm working on it, and it seems quite promising (nice figures and good sound, I mean, both closed and open loop), but I'm still not completely satisfied with what I've got so far. It is anyway a pretty standard (and rather complex: 18-20 bjts, not counting protection circuitry, optional DC servo and shunt ps...) discrete CF architecture: diamond buffer in - I/V gain stage - diamond buffer out with Sziklai pairs in the power stage. It was a let's try a discrete CFA and see what happens project; when I decided to go for an open-loop/low-Zout try it became a full (spare, of course
) time work...L.
Buffer as in no voltage gain ?? I thought it had 12db gain. From post 1, what is R7 ??
Nope - not the circuit shown on post #6, at least: it has unitary gain, but I think it's only the output/power buffer.
L.
There is both a buffer and a 12db gain stage. This is why you need the article. not to make money but to explain what is going on best.
If all is needed is a buffer, then what's wrong with the classic diamond buffer? I think it's really awesome.
Why in the world would anyone need 12dB of gain in a headphone amplifier? The max output of a CD player is 2Vrms, which is enough to blow out your eardrums with most headphones.