This stereo amplifier is based on the LM3886 Audio Amplifier IC from Texas Instruments.
There are provisions for 3 switched audio inputs. 115VAC power is provided at the fused IEC connector.
Inside, the amp is configured in a stacked configuration with the transformer (Antek AS-2225) centered on the bottom plate of the enclosure. The power supply filter and amplifier PC boards are mounted on a shelf just above the transformer.
The source inputs are carried from the back panel to the selector switch by twisted pairs. Those long leads running from back to front look trouble prone and I was concerned they would pick up noise. So far I’ve not noticed anything.
The input signal for each amp is fed by a twisted pair from the dual gang, volume control, potentiometer on the front panel.
The amplifier PC board circuit closely follows the examples shown in the LM3886 data sheet.
A PDF of the LM3886 Amplifier PCB schematic is available here: LM3886 Amplifier Schematic
The values chosen for R1, L1, and C3 in the output networks are different than what is typically specified. The values here were arrived at experimentally and were chosen to provide what was, subjectively, determined to sound best through my speakers (a pair of Klipsch KG4s). I plan to do some more experimenting. The roll-off of the Zobel network might be too low.
The sharp eyed reader may notice that the LM3886 specified in the schematic is the ‘T’ package but the photo of the amp’s insides show the “TF” isolated package. Originally I had the “T” style packages installed along with shoulder washers and insulators. I replaced those with the “TF” package shortly before taking the above photo.
The same perceptive reader has probably already noticed that the construction of the output inductor, L1, as seen in the above photo, doesn’t match the description for L1 given on the schematic. L1 was modified after the photo was taken. I’m always tinkering with the things I build.
For the amplifier PC board layout I opted for two ground planes (one for the signal common and one for the power common) instead of the more typical practice of snaking traces around to a common star point. The two planes converge at the COM terminal on the PCB (I suppose that’s a 2 point star). I’ve not experienced any issues with this configuration (yet).
The Power Supply Filter PC board provides the smoothing capacitors and a fuse for each power supply rail. It has a total of 20,000uF per rail. The rectified DC is provided by a 35A bridge bolted to the base plate of the enclosure.
A PDF of the Power Supply Filter schematic is here: Power Supply Filter Schematic
The Power Supply Filter PCB:
The 25VRMS secondary of the power transformer provides just over +/- 34VDC supply rails after rectification and filtering. This value can vary up or down depending on the exact mains voltage available. Household mains voltage is nominally 110VAC in the U.S.
For an 8 Ohm load, +/- 34VDC is OK per the LM3886 data sheet. Although I’m using this amplifier with my 6 Ohm impedance KG4s, I believe a saner option would be to use a power transformer with a lower secondary voltage. Something such as the Antek AS-2218 might be a better choice. I will probably make the change the next time I’m inside the amp.
The custom enclosure is constructed from individual parts. The overall look and size was arrived at with the desire to have a unique appearance in a compact envelope while still allowing enough room inside to make the necessary connections. The heat sinks on either side are from Heatsinkusa.com. They are much larger than they need to be but they provide the side walls of the enclosure and reduce the number of parts needed.
The front panel and most of the other parts were fabricated at a local machine shop. The Walnut panel pieces on the front were laser cut at ponoko.com.
I’ve been happy with this amplifier. It is dead quiet with no inputs connected. There is no hum or any other perceptible noise when my Raspberry Pi running Moode Audio is connected. I haven’t tried it with any other sources.