This is a tear down of the recently reviewed Schiit Magni Heresy Headphone Amplifier. This version of Magni is uses integrated circuits ("OpAmps") as opposed to discrete output stage which in my measurements, garnered it far lower distortion and won my strong recommendation.
The screws were a bit less tight than the Magni 3+ I disassembled:
The architecture here could not be simpler. AC input is rectified, filtered with good quality capacitors (Nichicon) and regulated to + and - 17 volts. Input signal is amplified by the OP1662 opamp/buffer.
A cluster of four OPA1688 headphone amplifier/opamps are used per channel. OPA1688 has exceptionally low distortion+noise:
As you see, SINAD is 118 dB using 128 ohms. In my dashboard I use 300 ohms which is even a less stringent load. My measurements showed the same 118 dB SINAD confirming Schiit's implementation to be as good as the lab spec. Using four of these in parallel allows much more drive than a single one.
OPA1688 rated to operate at max voltage of +- 18 volts. So at 17 volts, it is near its max capacity so heat dissipation is going to be on the high side. To see how much, I put a 32 ohm load on the Heresy and drove it to 9+ volts output (about 2.5 watts). After just 3 to 4 minutes, the OPA1688s were cooking:
The highest temp is 63 degrees. Paralleling opamps is a crap-shoot in that you can't guarantee they all carry the same load. So some were hotter than others.
There is no heatsinking here whatsoever. If you are going to do crazy things like driving speakers with them, I suggest sticking a heatsink across all of them. Thermal efficiency will be low due to plastic packaging but hopefully better than none.
The back of the PCB was cleaner than the Magni 3+:
The LED is where the mess was in the other one but I see no issues there. There is however residual solder as indicated. Given the through-hole parts, this is likely gone through wave soldering with some residue remaining. I suspect that is where the board is being held so solder is sticking to it.
Conclusions
This is a very clean and simple headphone design. Which means it will get copied. Fortunately Schiit has priced this amplifier quite low so the business opportunity is not that good for the clones. Ultra low distortion comes in the form of purpose built audio opamp meant for lower power headphone drive. By using four in parallel, much higher output level is available. Thermal management is fine for headphone listening.
A bit of solder residue could lead to failures in the field and should be cleaned up post inspection.
--------
As always, questions, comments, corrections, etc. are welcome.
Images in these teardowns are much larger due to complexity of images. This means soon I have to pay for more server storage costs. And you know me, I hate paying for things out of my pocket. So I hope you do your civic duty by filling my bank account as much as you can using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
The screws were a bit less tight than the Magni 3+ I disassembled:
The architecture here could not be simpler. AC input is rectified, filtered with good quality capacitors (Nichicon) and regulated to + and - 17 volts. Input signal is amplified by the OP1662 opamp/buffer.
A cluster of four OPA1688 headphone amplifier/opamps are used per channel. OPA1688 has exceptionally low distortion+noise:
As you see, SINAD is 118 dB using 128 ohms. In my dashboard I use 300 ohms which is even a less stringent load. My measurements showed the same 118 dB SINAD confirming Schiit's implementation to be as good as the lab spec. Using four of these in parallel allows much more drive than a single one.
OPA1688 rated to operate at max voltage of +- 18 volts. So at 17 volts, it is near its max capacity so heat dissipation is going to be on the high side. To see how much, I put a 32 ohm load on the Heresy and drove it to 9+ volts output (about 2.5 watts). After just 3 to 4 minutes, the OPA1688s were cooking:
The highest temp is 63 degrees. Paralleling opamps is a crap-shoot in that you can't guarantee they all carry the same load. So some were hotter than others.
There is no heatsinking here whatsoever. If you are going to do crazy things like driving speakers with them, I suggest sticking a heatsink across all of them. Thermal efficiency will be low due to plastic packaging but hopefully better than none.
The back of the PCB was cleaner than the Magni 3+:
The LED is where the mess was in the other one but I see no issues there. There is however residual solder as indicated. Given the through-hole parts, this is likely gone through wave soldering with some residue remaining. I suspect that is where the board is being held so solder is sticking to it.
Conclusions
This is a very clean and simple headphone design. Which means it will get copied. Fortunately Schiit has priced this amplifier quite low so the business opportunity is not that good for the clones. Ultra low distortion comes in the form of purpose built audio opamp meant for lower power headphone drive. By using four in parallel, much higher output level is available. Thermal management is fine for headphone listening.
A bit of solder residue could lead to failures in the field and should be cleaned up post inspection.
--------
As always, questions, comments, corrections, etc. are welcome.
Images in these teardowns are much larger due to complexity of images. This means soon I have to pay for more server storage costs. And you know me, I hate paying for things out of my pocket. So I hope you do your civic duty by filling my bank account as much as you can using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/