• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

xrk971 Desktop Class A (DCA) Headphone Amp

Finally finished my take on the DCA. A very nice headphone amp. I really like the ease of construction and the "frameable" circuit board!!

Thanks X and JP.
 

Attachments

  • DCA1.jpg
    DCA1.jpg
    207.9 KB · Views: 1,281
  • DCA2.jpg
    DCA2.jpg
    729.2 KB · Views: 474
  • DCA3.jpg
    DCA3.jpg
    743.4 KB · Views: 449
Founder of XSA-Labs
Joined 2012
Paid Member
Woohoo! This is great news :). How about some pics?

Here is the new board with silkscreen properly applied.
attachment.php


2mm thick FRP with 2oz copper traces (double thickness) and ENIG finish. Of course, covered with vias to connect both sides together for equivalent 4oz+ thick copper traces!
 

Attachments

  • 94D6BFE6-2A17-443D-BE7F-4B1219A3BA13.jpeg
    94D6BFE6-2A17-443D-BE7F-4B1219A3BA13.jpeg
    237.5 KB · Views: 583
Last edited:
In case anyone has a DAC or preamp with balanced outputs (+ve and -ve phase balanced line driver), and they have headphones modified to accept balanced drive amps, there is an easy way to make yourself a unique, and very cost effective balanced SE Class A HPA. Just use two DCA boards, one for each earpiece (driver). By doing this, you get 4x the power, and twice the slew rate. Best of all, because this is a SE Class A amp that is cap-coupled, you never need to worry about the amp going DC on you and blowing out your expensive cans. I have tried this with the PCA before and it sounds fantastic. So much dynamic reserve, and everything sounds so much more powerful.

Here is schematic:
681132d1526524167-xrk971-desktop-class-dca-headphone-amp-dca-balanced-drive-hpa-png


So, if you were to run each amp with Vcc of 25v, and 150mA bias current, with 42ohm cans you can get about 2.2wrms and +/-20vpp swing from an effective Vcc of 50v. This is useful for super inefficient 74dB planar headphones.


When the amp is operating in Single-Ended mode (where the Return lead from the Headphones goes to GND), the polarized output capacitors will always have their (+) terminal at some voltage above GND.
.
But, when the amp is operating in Balanced mode (where the Return lead from the Headphones goes to the other amp channel output), won't the polarized output capacitors be reverse biased by the AC audio signal, ~50% of the time? When using the DCA as a Balanced amp, should bi-polar capacitors be used for the output?
 
While I'm putting together my shopping list for capacitors to use in the 3 DCAs I'm assembling (2 connected as a Balanced amp, 1 as a Single-Ended amp), I'm trying to figure out the voltage requirement of the large (4700uF) non-polarized output capacitors.


If I understand the setup and operation of the amp correctly -
1.) R106 / R109 are adjusted/sized, so that the bias current is set to a level that causes the DC voltage drop across V111 / V112 to be one-half of V+.
2.) This allows the AC audio input signal, imposed on the Gate of V111 / V112, to cause them to turn further ON (raising the voltage going to the output capacitor), or turn further OFF (lowering the voltage going to the output capacitor).
3.) In Single-Ended mode, the voltage across the load can vary from a minimum of ~0V from GND, to a maximum of ~V+ from GND. This means the output capacitors need a voltage rating of just a bit more than V+.

4.) In Balanced mode, it seems like the same voltage ratings would apply, since neither output can go higher than V+, nor lower than GND. So, I'm trying to figure out your reference below that schematic, of a +/- 20Vpp swing - which would require output capacitors rated at 2x V+.


Am I misinterpreting how the circuit works, in Single-Ended and/or Balanced modes?
 
Founder of XSA-Labs
Joined 2012
Paid Member
The headphone or speaker load sits in between the two anti-phase outputs. One swings +Vcc, the other side swings +Vcc but at 180deg out of phase so appears as a -Vcc to the headphone voice coil. It’s similar to a BTL amp that has double the voltage swing for single polarity supply. The coupling cap is going to experience the same thing that voice coil experiences as it is in series. So you should size caps for more than Vcc and use two back to back for rating of 2Vcc and non-polar. It’s inportant to note that the output voltage is not referenced to GND as neither of the speaker wires are allowed to touch GND. The outputs are referenced to each other - hence the 2xVcc.
 
Founder of XSA-Labs
Joined 2012
Paid Member
No don’t do that. Just take two 4700uF caps and strap them together with their -ve leads lined up and solder the -ve to -ve. Then add some shrink tubing to the remaining +ve legs and bend a dog leg in them to get the two pins to fit the existing holes. You now have 2350uF bipolar cap. It should be plenty big, I don’t think you need to go with bigger caps unless you plan on driving 8ohm speakers.

I do this trick all the time. Use a good cap like Panasonic or Nichicon or Elna Silmic. But even budget caps (50 in a bag from China) can sound quite good to get started. Panasonic FC or FM are fine. Elna Silmic are very fat and may be harder to fit.

But to get started just use regular polar electrolytics rated 50v - it will work fine and not be an issue or blow up since the negative polarity is brief and AC not DC. It’s reverse DC for long periods that blows up caps.
 
Somehow in Balanced mode, I'm seeing the two 2350uF caps, one on the output of each amplifier, being in series with the load, and each other, which would mean the total capacitance in the circuit between V111 (Vout+) and V112 (Vout-) is halved again, to 1175uF.
.
That's compared to Single-Ended mode, where one 2200uF capacitor is between Channel A V111 (Vout+) and GND, and between Channel B V112 (Vout+) and GND.
 
Founder of XSA-Labs
Joined 2012
Paid Member
With a load of 50ohms in between the caps you no longer have caps in series equation. The load of resistor in this case acts as the virtual “GND” as it is dissipating the power being put through it to the other side which is in opposition. The lower your impedance goes, the closer it goes to being two caps in series. But, when we plan on a lower impedance load, we increase the capacitor so that it’s falloff for low bass notes is not noticeable. Try it and tell me if the bass doesn’t go low enough.

If you are still bothered by this seeming series use of caps, you can in fact remove one of the caps and DC couple one side. It will be fine as the blocking cap on the other side prevents DC from going through.

If you are still bothered by use of caps you can even do this: remove both caps and adjust the relative bias current so that they are perfectly balanced and thus the midpoint at the output is, say 10.00v for both the phases and hence, a net 0.00v across the balanced output. Your FETs are matched already so the bias currents should be quite close. You can put a trim pot at the resistor below the JFET that feeds back to the input and adjust for perfect balance.

I have been meaning to try this trick and it should result in an amp with much more transparency and slam. Plus it saves you from having to buy expensive big caps. But beware, any imbalance such as one amp going off first will cause DC to flow. A DC protection circuit would be needed.

Compromise is use just one cap.
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
Dbaldock9,
Jhofland told me that using a single (polarized) cap per amp should work fine and the value of the cap is not halved like in series. The way to think about it is this: inagine one amp driving half the load (25ohms for nominal 50ohm load) directly to ground. The load is a plane of symmetry so it’s like GND at its half point. The cap will never see reverse bias - which is what we want to avoid. What a single cap “sees” is a 25ohm load to a virtual ground. It never goes negative.
 
Just realized that I have been confusing / mixing (in my mind) the DC and AC states of the amplifier channel outputs. The DC voltage on the side of the output capacitors closest to the transistors, will always be ~ +10Vdc. So, while the AC component of audio output may cause the voltage going to the load to vary from ~V+ to ~GND, the DC voltage level and direction on the capacitors doesn't change.