doobooloo
Headphoneus Supremus
- Joined
- Nov 3, 2002
- Posts
- 2,544
- Likes
- 11
(Note: This is a continuation of a project started two and a half years ago, on this thread.)
First, some photos:
And the board layouts and pending front and rear panels:
All:
Top:
Bottom:
To recap, my three primary objectives of designing and building this amp were:
1. Create a transportable amp that will take an unbalanced source but drive both balanced and unbalanced headphones;
2. Design the amp section around the TPA6120 chip; and
3. Utilize the DS1802 digital pot.
To achieve this, I have one TPA6120 in noninverting configuration and the other in inverting configuration, both buffered by dual opamps for matched input impedance on the TPA6120 side. So, both headphone jacks are independently amped outputs, with opposite polarities. Hence, both jacks will function properly as single-ended headphone amps, but the real kicker is when you build a cable to connect both for a balanced output.
Since this was the first batch, I ran into a few issues. After everything was put together and powered on, I tested the amp for any DC offset before plugging my headphones in and found out that there was +/- 3V (!!) on all channels! I was a bit bummed, but remembered that n_maher had a similar issue with his build two years ago so I thought it's probably not the issue with my soldering or some other silly error. So, I went through the board with my multimeter checking where this could be coming from... and eventually I found out that the pads for the voltage reference for the DS1802 power supply were mixed up and hence was supplying all the wrong voltages to the chip! Fortunately all I had to do to get it in the right configuration was rotate the TO92 package a little (i.e. so that pins 2 and 3 go where 1 and 2 were supposed to go - see the photos above) and once this was fixed, everything was working fine without any issues.
So, how does it sound? I initially did all my listening tests in single-ended mode (primary reason being me not having made the adapter cable yet). First, it has that immense impact and sense of power that is characteristic of the TPA6120 chip. Compared to my Headamp GS-1, everything sounds much more forward and impactful. Bass is not the only thing that contributes to this sense of impact - all throughout the frequency spectrum, there's a sense of power and the effortless nature of how the amp drives headphones.
I tried it out with three opamps in the pre-TPA stage - AD8066, AD8620, and the LM6172, all three of which I have had fond experiences with and have stellar reputation among DIYers. Listening comparisons were pretty easy because I could install one chip in one channel and another on the other - I couldn't detect any sonic difference from polarity shift so for me that was good. My source was the Gigalab moon Non-OS DAC.
Previously, my favorite was the AD8066, which had the most forward and impactful presentation of all, so my first choice was to try the amp out with this chip. Oh boy, what a combination! So much bass impact, treble detail, and forwardness, it was real fun to listen to. However, it was too much and I felt that combination lacked fluidity, texture and a natural presentation needed for a more enjoyable, extended listening session. In other words, often the combo sounded artificial and a bit fatiguing. Now on to LM6172. After 30 seconds on this chip I promptly de-soldered it. It sounded thin, harsh, and just wrong to my ears. So my extended comparison at the end of the day comprised of AD8620 in one and AD8066 in the other. AD8620 sounded somewhat veiled compared to AD8066 at first after having been accustomed to the AD8066 sound. Also, there was less of that immediate sense of impact and deep bass response that came out of AD8066. Sondstage-wise, the AD8066 had the upper hand as well due to its highly resolving nature and the AD8620 sounded more "room-like" in comparison. However, I found that the sound was more balanced and easier to listen to with the 8620, and tracks just sounded more real. What really got me after an extended comparison was the really crunchy midrange texture that came out of AD8620 - how shall I put it, it was both lush and very textured, kind of "tasty" to the ears. (I'm running out of audiophile speak here!) This really got me, and at the end of the day, I decided to go with a pair of 8620s in the amp.
So, how do things change in balanced mode? To be honest, given that TPA6120 is such a marvelous and powerful chip, I expected very little. But the difference wasn't subtle, for the better. The relative "veil" of the AD8620 was lifted, and the amount of impact and sense of power had increased. The picture that came into my mind was the amp holding the headphone drivers by the collar and shaking it around! I tried the amp with two headphones - Grado HP-2 and Sennheiser HD580 - and it was a little funny how the two cans' differences had narrowed on this balanced amp. I thought the HD580 sounded smooth and laid back, all of a sudden it sounded smooth but quite forwardly articulated. I thought the HP-2 sounded forward but a bit airy, all of a sudden its sound gained mass and substance. It was quite a revelation, it seemed as if the amp really had full control over the cans' drivers and could dictate the diaphragm's motion to a degree I've never experienced before (and hence my imagery of amp holding cans by collar.
)
In other words, it sounds real good, and it makes me want to try all over again the HD650s and a bunch of other headphones that I had given up on in the past.
Ah, a few final comments - the DS1802 paired with the Alpha rotary switch works like a charm. I think with a proper knob in place the switch would feel very nice but for now it the clicks feel a little hard/sharp. That said, I feel having a stepped attenuator with 60 or so steps is a huge advantage over any conventional pot and even the expensive Elma stepped attenuators in my GS-1. Also, even with a humble WM080 the amp is pretty quiet, probably thanks to a beefy PSU rail caps. Even with the Etymotic ER-4S/hf2, the noise is almost nonexistent. This was a real surprise to me and now I am completely de-motivated to build a nice dedicated power supply.
Anyway. I think that sums it up nicely. Of course I'm emotionally biased about this amp, so take all my comments above with a grain of salt, but I really am enjoying this amp a lot. I still haven't ordered the front panels yet, need to wrap that piece up soon. I've also got some new ideas cooking for the next batch (including a PPA-style bass boost) and hopefully I can move to that phase soon and have a second build with a lead time of less that 2.5 years.
Any comments, thoughts, or feedback would be greatly appreciated!
First, some photos:
And the board layouts and pending front and rear panels:
All:
Top:
Bottom:
To recap, my three primary objectives of designing and building this amp were:
1. Create a transportable amp that will take an unbalanced source but drive both balanced and unbalanced headphones;
2. Design the amp section around the TPA6120 chip; and
3. Utilize the DS1802 digital pot.
To achieve this, I have one TPA6120 in noninverting configuration and the other in inverting configuration, both buffered by dual opamps for matched input impedance on the TPA6120 side. So, both headphone jacks are independently amped outputs, with opposite polarities. Hence, both jacks will function properly as single-ended headphone amps, but the real kicker is when you build a cable to connect both for a balanced output.
Since this was the first batch, I ran into a few issues. After everything was put together and powered on, I tested the amp for any DC offset before plugging my headphones in and found out that there was +/- 3V (!!) on all channels! I was a bit bummed, but remembered that n_maher had a similar issue with his build two years ago so I thought it's probably not the issue with my soldering or some other silly error. So, I went through the board with my multimeter checking where this could be coming from... and eventually I found out that the pads for the voltage reference for the DS1802 power supply were mixed up and hence was supplying all the wrong voltages to the chip! Fortunately all I had to do to get it in the right configuration was rotate the TO92 package a little (i.e. so that pins 2 and 3 go where 1 and 2 were supposed to go - see the photos above) and once this was fixed, everything was working fine without any issues.
So, how does it sound? I initially did all my listening tests in single-ended mode (primary reason being me not having made the adapter cable yet). First, it has that immense impact and sense of power that is characteristic of the TPA6120 chip. Compared to my Headamp GS-1, everything sounds much more forward and impactful. Bass is not the only thing that contributes to this sense of impact - all throughout the frequency spectrum, there's a sense of power and the effortless nature of how the amp drives headphones.
I tried it out with three opamps in the pre-TPA stage - AD8066, AD8620, and the LM6172, all three of which I have had fond experiences with and have stellar reputation among DIYers. Listening comparisons were pretty easy because I could install one chip in one channel and another on the other - I couldn't detect any sonic difference from polarity shift so for me that was good. My source was the Gigalab moon Non-OS DAC.
Previously, my favorite was the AD8066, which had the most forward and impactful presentation of all, so my first choice was to try the amp out with this chip. Oh boy, what a combination! So much bass impact, treble detail, and forwardness, it was real fun to listen to. However, it was too much and I felt that combination lacked fluidity, texture and a natural presentation needed for a more enjoyable, extended listening session. In other words, often the combo sounded artificial and a bit fatiguing. Now on to LM6172. After 30 seconds on this chip I promptly de-soldered it. It sounded thin, harsh, and just wrong to my ears. So my extended comparison at the end of the day comprised of AD8620 in one and AD8066 in the other. AD8620 sounded somewhat veiled compared to AD8066 at first after having been accustomed to the AD8066 sound. Also, there was less of that immediate sense of impact and deep bass response that came out of AD8066. Sondstage-wise, the AD8066 had the upper hand as well due to its highly resolving nature and the AD8620 sounded more "room-like" in comparison. However, I found that the sound was more balanced and easier to listen to with the 8620, and tracks just sounded more real. What really got me after an extended comparison was the really crunchy midrange texture that came out of AD8620 - how shall I put it, it was both lush and very textured, kind of "tasty" to the ears. (I'm running out of audiophile speak here!) This really got me, and at the end of the day, I decided to go with a pair of 8620s in the amp.
So, how do things change in balanced mode? To be honest, given that TPA6120 is such a marvelous and powerful chip, I expected very little. But the difference wasn't subtle, for the better. The relative "veil" of the AD8620 was lifted, and the amount of impact and sense of power had increased. The picture that came into my mind was the amp holding the headphone drivers by the collar and shaking it around! I tried the amp with two headphones - Grado HP-2 and Sennheiser HD580 - and it was a little funny how the two cans' differences had narrowed on this balanced amp. I thought the HD580 sounded smooth and laid back, all of a sudden it sounded smooth but quite forwardly articulated. I thought the HP-2 sounded forward but a bit airy, all of a sudden its sound gained mass and substance. It was quite a revelation, it seemed as if the amp really had full control over the cans' drivers and could dictate the diaphragm's motion to a degree I've never experienced before (and hence my imagery of amp holding cans by collar.
In other words, it sounds real good, and it makes me want to try all over again the HD650s and a bunch of other headphones that I had given up on in the past.
Ah, a few final comments - the DS1802 paired with the Alpha rotary switch works like a charm. I think with a proper knob in place the switch would feel very nice but for now it the clicks feel a little hard/sharp. That said, I feel having a stepped attenuator with 60 or so steps is a huge advantage over any conventional pot and even the expensive Elma stepped attenuators in my GS-1. Also, even with a humble WM080 the amp is pretty quiet, probably thanks to a beefy PSU rail caps. Even with the Etymotic ER-4S/hf2, the noise is almost nonexistent. This was a real surprise to me and now I am completely de-motivated to build a nice dedicated power supply.
Anyway. I think that sums it up nicely. Of course I'm emotionally biased about this amp, so take all my comments above with a grain of salt, but I really am enjoying this amp a lot. I still haven't ordered the front panels yet, need to wrap that piece up soon. I've also got some new ideas cooking for the next batch (including a PPA-style bass boost) and hopefully I can move to that phase soon and have a second build with a lead time of less that 2.5 years.
Any comments, thoughts, or feedback would be greatly appreciated!