I haven’t had time to update the blog because I’ve been busy designing a new headphone amplifier. There’s been a lot of development when it comes to operational amplifiers and I was curious about how well I can get an amplifier to measure using the latest technology. The op amp I’m using still hasn’t got a development board, so I make one myself. The op amps are 5x5mm in PVQFN-20, so you need to use an SMD soldering gun or oven to be able to solder them. I simulated the circuit with both LTSpice and TINA90, but left biasing resistors and some other stuff in good old fashion though-hole components for easy testing.
Trying to test the unit with programs like RMAA with a EMU 0404 USB Audio card isn’t good enough. I’m essentially measuring the noise floor of the EMU. So next step is to test it on an Audio Precision Audio test equipment and hopefully get some better data.
I’m going to develop this in two steps. First step was getting the amplifier prototype up and running. During testing I used my RIGOL DP832 3 Outputs Programmable DC Power Supply, feeding the amp ±15 V to the board. I then used a wall-wart AC adapter, making it a voltage doubler and splitting the supply into positive an negative rail. I already had a PCB made for another project, so I just cut out the parts I didn’t need. When measuring it using my LNMP using jellybean parts like the classic 7815/7915 the noise floor on the power supply is around 1 mV. The final power supply are going to use Texas Instruments 1 amp DC converters with 4µV/15µV noise floor. The opamp has very high PSRR and CMRR, so it’s kinda overkill using these components, but it’s an experiment how well you can measure it, without thinking about cost. But even with the “bad” power supply, it sounds insanely great!
So now I’m waiting for the new revision of the amplifier board and will send the new power supply design to China. Looking forward to get the new PCBs.