AMB Laboratories DIY Audio

[thomaspf]

Thank you. Compared to your reference design I switched the location of the β24 and σ22 boards relative to the connectors which seems to allow for shorter cables. I also went dual mono for the current stages. With the all the supplies for the DAC part there are no less than 7 power supplies in the case. I will post some more pictures how all the components fit inside the box.

DAC / poweramp combinations are apparently not that unusual anymore http://thewelltemperedcomputer.com/HW/USB_amp.htm. Especially with a high quality USB interface they are rather convenient when your main source is a computer.

Now it's time to get the Arduino hooked up for the DAC control and IR interface...


Cheers

Thomas

[amb]

DAC / poweramp combinations are apparently not that unusual anymore http://thewelltemperedcomputer.com/HW/USB_amp.htm

I would call that one an "integrated amplifier" with onboard DAC, because it's not just a DAC and a power amp on the same chassis. I've yet to see another DAC paired (sans volume control) only with a power amp like yours.

[thomaspf]

Actually, I do have volume control as well as source selection. That is what the Arduino is for. It's just not connected quite yet. I have two external DACs which are almost identical except for the volume control. One is using a relay ladder and the other one relies on the digital volume control in the ESS 9018 DAC chip. At the levels I am using it I could find no disadvantage of the digital volume so that is what I will be using here as well.

This build will not have a volume knob. I put one on the build for my wife and she ended up only ever using the remote control. Since this build even a bit larger than that I figured I cut the knob and make it remote only.

btw. I wired the Buffalo/β24 in parallel to my main system and did a bit of an unfair comparison. Speakers are Super-HL5 vs. Monitor 40 and I could only use the digital volume control in Jriver at 7% max because of the gain. My son who has a lot better ears than mine at his age commented that there is almost no difference. In fact I probably liked the Buffalo/β24 a bit better but I can't say whether that is the speakers. I am very impressed.

I do get a bit of mechanical transformer hum from the 6 toroids in the case and just ordered the parts to add a DC blocker.

Cheers

Thomas

[thomaspf]

I finished connecting the Arduino to the LCD, IR receiver, and the DAC i2c link. Firmware is all up and running.

I also received new metal film resistors to adjust the gain into a range where the digital volume control in the DAC works comfortably with a full digital input signal. This allowed me for the first time to hear this rig in full resolution with all the bits transferred into the DAC. Not unsurprisingly the result is a lot more detailed than setting the Jriver Media Center volume control to 7%...

Since the Buffalo DAC needs a 3.3V I2C link but the Arduino only provides a 5V output I took the opportunity to also galvanically isolate the Arduino from the Buffalo via an ADUM 1250 chip. The chip provides the isolation and allows for different levels on its two sides. The software is a derivative from the early Hifiduino code with a simple menu system to allow adjustment of the various settings.

Now I can start to do some real comparison to the Bryston 4B-SST2 in my main system.

Cheers

Thomas

arduino.jpg

[thomaspf]

To address the hum I was getting from my β24/Buffalo II build I have built a DC blocker according to the Audio Asylum discussion below.

http://www.audioasylum.com/audio/tweaks/messages/16/161953.html

If I plug the amplifier in using this blocker the hum gets quieter but is not completely gone. It is however lower in level than what I can hear from my Bryston 4B-SST2 which I believe also has a variation of this circuit built in. There is no hum on the speaker outputs.

What do other poeple hear for transformer hum?

Cheers

Thomas

[Shaman]

Hi Thomas,

Give the Swenson Snubber a try.

Good job on the project!

[dsolodov]

Nice built.

I also received new metal film resistors to adjust the gain into a range where the digital volume control in the DAC works comfortably with a full digital input signal. This allowed me for the first time to hear this rig in full resolution with all the bits transferred into the DAC.

What I/V stage is used (Legato or IVY-III or something else)?? What values and which R?

Since the Buffalo DAC needs a 3.3V I2C link but the Arduino only provides a 5V output I took the opportunity to also galvanically isolate the Arduino from the Buffalo via an ADUM 1250 chip.

Could you please share schematics and/or wiring diagram? I was going to follow the application note from AD and use two Adum chips with Arduino and DAC behind the left side, which is not I2C compliance, and run I2C on the right side of ICs.

[thomaspf]

What I/V stage is used (Legato or IVY-III or something else)?? What values and which R?

I am using an IVY-III with 100Ohm resistors for R13-R16. This in combination with the β24 and a pair of Harbeth Super-HL5 seems to work pretty well. If you really play it loud you might want to use 120 Ohms but I wanted to go for maximal number of bits.

Could you please share schematics and/or wiring diagram? I was going to follow the application note from AD and use two Adum chips with Arduino and DAC behind the left side, which is not I2C compliance, and run I2C on the right side of ICs.

I am not quite sure I follow your question. The app note for the ADUM1250 is very simple and that is exactly what I used. It is a bidirectional chip so you only need one for the whole setup. To start with I soldered it onto a $2.95 breakout board from Sparkfun http://www.sparkfun.com/products/494.

Then I simply followed the attached appliaction application example connecting one side to the Arduino GND, 5V, Analog 4 (SDA), Analog 5 (SCL) and the other to the matching connections on the Buffalo II that I am using. I am still waiting for the caps to get delivered but it works just fine without them.

adum1250.JPG

Hope that helps.

Thomas

[thomaspf]

I am still waiting for the USB interface and the S/PDIF mux but I ave been using the amp daily in my main system. The digital interface is far from optimal at this point with USB->Toslink converter providing the interface so I expect the final result to be "even" better. since that simple USB interface does not have an ASIO driver and is flaky with WASAPI. I am currently using DirectSound and have the interface set to 24/96 which means whatever I play is being resampled by the operating system to that sample rate and bitdepth.

As mentioned in my earlier posts the speakers are Harbeth Super-HL5 and this works out to be a nice match. At this point I prefer this combination over the Bryston. I am really bad at describing this but it feels like the β24 moves you up closer to the artists in a classical recording. For concerts with a larger number of intruments you seem to have more details of the individual instruments and that without adding any unnatural brightness.

I had a few friends visit and demonstrated this effect and it is not placebo and you can tell these amps apart in direct comparison with 100% certainty. It is actually surprising given that both are supposed to be very accurate. So far people have preferred the β24.

I also looked at a very low signal on my scope and the fully balanced circuit in the β24 seems to work quite well. If I connect my probe to the speaker posts and play the lowest possible volume so the DAC does not mute I get a noise floor that is identical to removing the probe cable. I interpret this to mean that the self noise of the scope is higher than the noise floor of the amp. If I repeat that experiment on the Bryston I see a higher noise floor when the probe is connected so the noise floor appears to be higher than that of the β24. I wish I had access to a Dscope or AP analyzer to look at this with a bit more accuracy.

Cheers

Thomas