[mwhouston]

  Here is my latest HP amp: 6AS7 with Sophia Electric Mesh Plate rec.tube - 274B. This build was inspired by the Woo Audio HP amp which can be purchased with the same rec. tube.
 




 
Take a look at the write-up: Sanguine

Dropped the amp on the CRO and here are some specs. I think the most impressive is the Ultra low noise figures:
 
 
With 33ohm load:
-3db points -10Hz to 240KHz
max power with 2.03Vpp input = 1Vpp @ 1KHz
noise with input S/C = 0.45mV to 0.88mV RMS (that's really low) knob at 10:30 position my comfortable listening level
Xtalk - 1Khz at 2.03Vpp input in one channel = 1.6mV in the other channel (excellent)
 
With 180ohm load
-3db points -10Hz to 164KHz
max power with 2.03Vpp input = 6Vpp @ 1KHz
noise with input S/C = 0.45mV to 0.77mV RMS (that's really, really low)
Xtalk - 1Khz at 2.03Vpp input in one channel = 1.6mV in the other channel (same, excellent)
Channel matching was perfect 100%. Noise didn't change with attenuator fully open.

 

[diego]

After four months of work I finally finished my Buffalo IIISE . It is a very nice upgrade over my five year old Buffalo DAC.
 
 

 

 

 

 

 

 
Added higher resolution pictures.

 

[mwhouston]

Very neat pro looking build. Dual PSs?

 

[nippon]

Wow, first my pessimistic mind was thinking about the 2x switches near the volume controll... but no. It is very very close to the perfect DIY DAC.
I can identify the impressive powersupply, the input stage and the protection relay(!?), but can you tell us something about the signal path in detail?
Does the DAC oversample?
 
It looks just awesome

 

[diego]

Thank you guys.
 
For the power supply I used AMB Sigma 22 and Sigma 11 power supplies. The first has a +15 and -15 output and feeds the IVYIII amplification stage and the second has a +5 volt output and feeds the DAC and WaveIO  USB converter. Additionally, both the DAC and WaveIO have their own local regulators that improve performance and isolate them from each other. Both of these supplies have very low noise and output impedance. I also put an AC filter and SumR 50VA custom transformers with a metal band to reduce horizontal radiation from them.
 
The switches and the volume control do not carry signal. They just control a small voltage that orders the dac chip to digitally modify the volume or the input. I used ufl coaxial cables for the digital inputs to minimize noise. All output and power supply cables are twisted and go between two grounded planes that protect them from some noise.
 
It is not a relay, it´s a ground loop breaker. The signal ground is connected through the ground loop breaker and has a switch to float it so that I can choose what gives me the lowest noise.
 
It is a Twisted Pear Audio Buffalo IIISE. There is plenty of info on their website but yes it oversamples and reclocks the input.
 
The whole thing has 0 DC offset at the output and the noise is too low to measure with my Fluke 179, so the total output noise is probably less than 100 microvolts. I certainly can´t hear any at least.

 

[vixr]

simply stunning sir...

 

[RudeWolf]

Why did you opt for Sigma PSU's instead of shuntregulators?

 

[diego]

Mainly for better thermal efficiency. Shunt regulators waste a lot of current particularly in the case of the WaveIO. This card uses a lot less current when playing 44.1 khz material than it uses when playing 384khz but if you want it to be able to do 384khz you have to set the current for it and waste it most of the time.
 
Additionally, the buffalo IIISE has it´s own shunt regulators so even if I put a linear regulator behind them the dac still is going to see a shunt regulator. In the case of the WaveIO, it has on board linear regulators so even if I put a shunt behind them the card still is going to see a linear regulator.
 
Finally, Sigma PSUs have very good performance and I am not convinced that I can gain anything by using shunts.

 

[2leftears]

I built this headphone amp back in 1992 when I was a student.  It was very loosely based on the NAD 1020 pre-amplifier high output/ headphone section, but I made a lot of major modifications (Darlington output stage, switchable bias spreader etc.) so pretty much completely my own design.
 

1. Runs either on mains or 4 x 8.4V NiMH batteries (about 10hrs)
2. Fully discrete amplifier
3. Full class A for headphones down to 32Ohms at normal listening levels
4. Can be switched to class AB to preserve battery power
5. Output can be used as a pre-amplifier
6. Output impedance just over 5 Ohms

 
I recently treated it to a new Alps pot and a full new set of Nichicon caps.
 
Of course I am biased but I think it still holds up very well against my Lehmann Linear!
 

 

 

 

 

[RudeWolf]

What elements are in the TO-99 packages that I see heatsinked?

 

[2leftears]

  What elements are in the TO-99 packages that I see heatsinked?

BC516 and BC517 Darlingtons in the output stage, h_FE > 30,000.  They are no longer being made I think.
 
These are not meant to be heatsinked, but the heatsinking helps a bit. They are running 25mA at 12V when on mains supply, dissipating 0.3W each. They are rated at 625mW.  At a stretch the current could be set at 50mA, but with the high impedance headphones I have got I don't need more for full class A operation.
 
The other transistors are various BC546, BC556 and BC559 used in the input stage, voltage amplification stage, and also for short-citcuit protection.
 
I built this as a student, and although the amp sounds really great, there are some things I would do different now.

 

[Androb]

Dad just finished building this little badboy
 


 
A phono stage

 

[2leftears]

Neat!

 

[KimLaroux]

 
 
 

 

[mwhouston]

Looking very smart. Nice build. Some technical info please?