3 Transistor HP Amplifier with low dist

Thanks for the tips, X.

I'd been looking through various texts to learn how to set up an input properly, but most seem to skip that stage and go straight to the VAS. I can't even remember where I saw the 100n, but I must have copied it from other designs I'd seen.

It all makes sense now that I stop and think about it and the calculated frequency response does tally up with what I was using - 75ohm output impedance of the DAC into 100nF gives this:

attachment.php


Vs. real life:
925266d1613989906-3-transistor-hp-amplifier-low-dist-fr2-png


It should be simple enough to retrofit a board with the proper filter values and I'll give it another test and see how it compares.

Fun stuff and a learning process at the same time!
 

Attachments

  • 20210226_105751.jpg
    20210226_105751.jpg
    104 KB · Views: 1,173
Last edited:
Hello,
I also finished the stereo amplifier.
I used BC546, BC556 and BD139.
It sounds great.
I present you some pictures.
I will have to mention radiators on the BD139 transistors and the 7812 voltage regulator that I used.
 

Attachments

  • 5.jpg
    5.jpg
    64 KB · Views: 294
  • 4.jpg
    4.jpg
    89.6 KB · Views: 299
  • 3.jpg
    3.jpg
    64.7 KB · Views: 218
  • 2.jpg
    2.jpg
    85.1 KB · Views: 579
  • 1.jpg
    1.jpg
    84.9 KB · Views: 582
I added the filter to the second board I built as follows:

attachment.php


This has flattened out the frequency response:

attachment.php


And THD now 0.00362%:

attachment.php


As a reminder, I measured my O2 head amp using the same set up. The O2 is specified at 0.0016%, I measured 0.00490%. The T3 under the same conditions is 0.00362%, so it is in the same ball park.

These are the figures:
Noise: -98.4dBA
Dynamic Range: -99dBA
THD: 0.00362%
IMD+n: 0.00456%
Crosstalk: -70dB

The only issue I have is lots of buzzing when nothing is connected to the input, and the buzzing goes away when I touch the metal part of the case (which is connected to the negative). When there is an input connected (even an ungrounded source such as a mobile phone), the amp is dead silent.
 

Attachments

  • Filter.PNG
    Filter.PNG
    47 KB · Views: 1,111
  • FR.PNG
    FR.PNG
    40.7 KB · Views: 1,100
  • THD.PNG
    THD.PNG
    48.2 KB · Views: 1,100
Founder of XSA-Labs
Joined 2012
Paid Member
Add a 10ohm resistor and two anti parallel 1n400x diodes plus 22nF film cap (all in parallel) for a ground loop breaker (GLB) between your analog ground and wall plug “earth” ground). Tie the chassis to the earth ground on the dirty side of the GLB.

That might make that hum go away. But we never listen to an amp with no signal connected anyway.
 
I've used that with good effect before, but the power supply for this one is the double insulated SMPS type that has no earth connection.

I do have a 'signal ground's which is connected via star to the power ground, so I could try isolate it there.

But as you say, we dont listen without it connected to something,so it's a minor issue in the scheme of things.and probably just best left alone.
 
marin cornel, thank you!
Is it required to have 100K pot? I'm just wondering if it would be possible to use digital pot like this: Overview | Adafruit DS1841 I2C Logarithmic Resistor | Adafruit Learning System
which is 22K max? I'm planning to build the integrated amp which will also include the headphone amp (still not decided which one to use). There will be Raspberry Pi to control the whole system. It could also control volume over I2C using that digital pot.
Thanks!
 
Founder of XSA-Labs
Joined 2012
Paid Member
It does only have 3 transistors... but I kinda feel that using a line driver is cheating ;)

(Going to look at ordering some THAT1646s now!)

Nothing wrong with using a line driver - not cheating in my book :)

THAT1646 might be getting scarce. DRV134 from TI works too. Modify circuit per schematic note on bottom left.

If you want to go crazy, use DRV134 and some big fatty TO247 BJTs. Like TTC5200 and TTA1493. Use 0.22ohm 3W emitter resistors. Use +/-19v laptop bricks in series and set bias for 1.2A. Use 18v Zener diodes to ground on DRV134 rails.

Makes for a 10w into 8ohm Class A amp. It can drive ANY headphone to ear bleed levels :) Mediatechnology (Wayne Kirkwood) has some nice simple but far out designs! :cheers:

DCAO_10W_Schematic_small_bw.JPG


This was Kirkwood’s build:
DCAO-10W_Class-A_Amp_With_Switching_Power_Supplies_small.jpg


I would seriously think about using SSR DC protection circuits before connecting a nice set of headphones to it though.
 
Last edited: