New Posts  All Forums:Forum Nav:

Panda amp - Page 2

post #16 of 40

It's like a big opamp... jfet input ltp, followed by a bjt ltp for high gain, followed by a quite basic double emitter output stage.

 

I'm a bit puzzled because the gain stage is very good: good fast transistors, nice design, extra regulation. But the output stage is crude: it uses old, slow transistors and simply a zener for the output stage biasing. The slow output transistors probably don't help to compensate the whole thing. At the power required by headphones, they could have used the same small signal bjt and bd139-bd140 (or the like) for the output. In cfp fashion, it would have helped both linearity and thermal stability (see the ckIII).

 

Gain is 6 btw.

post #17 of 40

You can probably find info on the gain part in the link I posted earlier, what I can say is that without the outputresistors it cranks out about 3 watt, which will be able to drive anything headphone related..

The gain can be adjusted to be more or less..

 

I use one for powering my Ergo AMT and they are supposed to be driven by a small speaker amp.

 

In short it sounds awesome, I wont say that it is the best I ever heard, but you have to go pretty far up the foodchain to something on par (talking well above the 500-1000 £ mark) so in terms of value for money it is a clear winner.

 

Yellowjeep: plug in the soldering-iron and get to work, you will not regret building it

post #18 of 40
Quote:
Originally Posted by francisdemarte View Post

 

Using hardwood flooring is genius! :) Great job!


Thanks, I have been using it for years, my table has it my windowbenches got it, my diningroom table got it, so a lot of my home is covered with oakfloor, it is fun to work with, easy and it is very entertaining to come up with new patterns and variations, it is almost as addictive as Tetris............

It is in building the box, I find my pleasure, the soldering part is OK, but not that funny..

 

here are some of my other boxes

PICT0506.JPG

 

PICT0508.JPG

 

PICT0566.JPG

 

post #19 of 40

The spiral patten cube is just too cool...

post #20 of 40

I would have to spend more time  but on a quick look the basic circuit topology is basically OK

 

but any designer is going to have differing concerns - to me some of the parts choices seem odd - TIP41/2 are slooow, 3.9 V zeners are poor shunt regulators

post #21 of 40
Quote:
Originally Posted by jcx View Post

I would have to spend more time  but on a quick look the basic circuit topology is basically OK

 

but any designer is going to have differing concerns - to me some of the parts choices seem odd - TIP41/2 are slooow, 3.9 V zeners are poor shunt regulators


Looking forward to your ideas.. 

post #22 of 40

@mr slim: Just for kicks (it's all sims, so take it with a grain of salt), with similarly biased amps (100ma):

 

THD @10Khz, 2Vrms into 30r : 

 

Original:

 

Harmonic    Frequency     Fourier     Normalized     Phase      Normalized
 Number       [Hz]       Component     Component    [degree]    Phase [deg]
    1       1.000e+04    8.982e-02    1.000e+00      179.71°        0.00°
    2       2.000e+04    1.266e-06    1.410e-05      -20.56°     -200.27°
    3       3.000e+04    3.220e-07    3.585e-06      -86.45°     -266.16°
    4       4.000e+04    1.914e-08    2.131e-07      111.41°      -68.29°
    5       5.000e+04    2.963e-08    3.298e-07      102.45°      -77.26°

Total Harmonic Distortion: 0.001456%

 

"Improved":

 

Harmonic    Frequency     Fourier     Normalized     Phase      Normalized
 Number       [Hz]       Component     Component    [degree]    Phase [deg]
    1       1.000e+04    8.984e-02    1.000e+00      179.87°        0.00°
    2       2.000e+04    8.903e-07    9.910e-06      -33.48°     -213.35°
    3       3.000e+04    2.808e-08    3.125e-07     -135.43°     -315.31°
    4       4.000e+04    1.746e-08    1.944e-07      111.29°      -68.58°
    5       5.000e+04    1.628e-08    1.813e-07      107.32°      -72.56°

Total Harmonic Distortion: 0.000993%

 

Bandwith @-3db

 

Original: 3.5Mhz

"Improved": 11.5Mhz

 

 

Would it translate into any audible difference ? Who knows...

 

 

The circuit, with classical vbe and cfp outputs.

 

Sans titre.png

post #23 of 40


 

Quote:
Originally Posted by Avro_Arrow View Post

The spiral patten cube is just too cool...

 

It sure is.... Lovely woodwork.

post #24 of 40
Quote:
Originally Posted by 00940 View Post

@mr slim: Just for kicks (it's all sims, so take it with a grain of salt), with similarly biased amps (100ma):

 

THD @10Khz, 2Vrms into 30r : 

 

Original:

 

Harmonic    Frequency     Fourier     Normalized     Phase      Normalized
 Number       [Hz]       Component     Component    [degree]    Phase [deg]
    1       1.000e+04    8.982e-02    1.000e+00      179.71°        0.00°
    2       2.000e+04    1.266e-06    1.410e-05      -20.56°     -200.27°
    3       3.000e+04    3.220e-07    3.585e-06      -86.45°     -266.16°
    4       4.000e+04    1.914e-08    2.131e-07      111.41°      -68.29°
    5       5.000e+04    2.963e-08    3.298e-07      102.45°      -77.26°

Total Harmonic Distortion: 0.001456%

 

"Improved":

 

Harmonic    Frequency     Fourier     Normalized     Phase      Normalized
 Number       [Hz]       Component     Component    [degree]    Phase [deg]
    1       1.000e+04    8.984e-02    1.000e+00      179.87°        0.00°
    2       2.000e+04    8.903e-07    9.910e-06      -33.48°     -213.35°
    3       3.000e+04    2.808e-08    3.125e-07     -135.43°     -315.31°
    4       4.000e+04    1.746e-08    1.944e-07      111.29°      -68.58°
    5       5.000e+04    1.628e-08    1.813e-07      107.32°      -72.56°

Total Harmonic Distortion: 0.000993%

 

Bandwith @-3db

 

Original: 3.5Mhz

"Improved": 11.5Mhz

 

 

Would it translate into any audible difference ? Who knows...

 

 

The circuit, with classical vbe and cfp outputs.

 

Sans titre.png



I like the looks of your improved version much better. I think there probably would be an audible improvement and it looks more "right".

The power rails design on the original is pretty funky looking and I agree the output stage transistor choices are odd.

On the original, what is the unregulated +- voltage on the rails for the output stage ?

The original power rail arrangement looks more like it's intended for a speaker amp than a headphone design.

post #25 of 40
Quote:
Originally Posted by Avro_Arrow View Post

The spiral patten cube is just too cool...



It was a terrible job to do, there is more than one pattern in it, hard to see on the pictures, but it drove to the edge of insanity, I had to stow it away for a month and then, try again, there is two faults in it, nothing anybody would notice, but having build it it is the only things I see...

post #26 of 40

How can you judge an amp. you have not heard ??

 

I think it is sad to sit in ones chair listening to nothing and telling others how it sounds...

 

Build the thing listen to it, then you can comment on the sound....

post #27 of 40

Wow, from reading the first fifteen pages of the Rock Grotto thread this seems like the real deal. I think I'll bite, I've got £250 in scholarship funds that I'm meant to be spending on practical projects anyway. This almost seems too perfect...

I'd love to see what two of these rigged for balance could do.

Hurm, anybody know an easy DIY DAC for under £100 (assuming I'm left with about that much)? Or I could just lie and buy a uDac2...

post #28 of 40

I don't think either jcx or I meant to say anything about how it sounds. It's just that there are some "technical oddities" about that design.

 

Now, I would caution anyone about jumping to conclusions too fast. At normal listening levels the theoretical discrepancies in thd (as in sims) in between the stock amp and the "improved" one are nothing to go crazy about. While the slow output require more compensation and restrict the achieved bandwith, I'm not certain it matters that much for the reproduction of audio frequencies. And, finally, an unregulated output stage, while uncommon for headphones amplifiers, isn't  big concern in such a design: the amp has plenty of feedback, which offers it plenty of power supply ripple rejection.

 

Honnestly, I'm not certain that chasing the last % of performances will always and automatically translate into a better subjective sound (but it sure is an amusing intellectual exercise). Considering the input stage and the compensation scheme, the designer apparently knows what he's doing. If he had considered that a better output stage was necessary, he had the skills to put one onboard. Why didn't he ? No idea. Could be cost cutting but it could also be that he thought that it sounded better that way.

post #29 of 40
Quote:
Originally Posted by JoetheArachnid View Post

Hurm, anybody know an easy DIY DAC for under £100 (assuming I'm left with about that much)? Or I could just lie and buy a uDac2...


Gamma 1

 

I really want to build one of these. If anyone wants one I might be tempted to build one for you. With that said it looks ridiculously easy to build.

post #30 of 40

I understand the nerdfactor very well, after all I am one, and the amp, could prabably be even better with some tweaks here and there, but the thing is, it is already very very good, so why bother ??

If you end up "only" getting an amp that is better on paper and sonicly inferior, then what has been gained ??

 

By all means I love improvements, but I just think it is the wrong way to do things, build it and then start tweaking it..If we just tweak it before we build it we cannot know if it is better..

New Posts  All Forums:Forum Nav:
  Return Home