Retronics

In my workshop, a Huldra 11 is used daily, along with a pair of Tandberg 2520 speakers, and until recently a separate DAC (Cambridge DacMagic 100) for playing music from a computer. The DAC, however, is periodically sounding distorted in high- and midrange, and I haven’t been able to find the cause. Therefore a new DAC was needed, and I decided to make a custom DAC board to be placed inside the receiver.

The receiver originally had two Tape in-/outputs with DIN5 connectors, and TAPE2 had to go, to give room for the DAC.

One important task when designing a DAC board is finding a suitable DAC chip; there are very many to choose between. After considering different alternatives, I went with PCM2900 from the Burr-Brown division of Texas Instruments. This a 16-bit DAC with 44.1kHz or 48kHz sampling rate (equivalent to CD- and DVD quality), and integrated USB interface. The chip also has an ADC, which I’m not using for this project.

With most of the needed functionality integrated in one chip, the circuit doesn’t need to be very complicated; it’s mostly about following the datasheet recommendations for voltage regulation and filtering, decoupling etc. This is a low-voltage device, which gets its power from USB.

Delta-Sigma DACs create a significant amount of noise above the audible range, and though the IC contains some low pass filtering, the datasheet recommends additional filtering on the audio outputs. The outputs also need buffers/amplifiers for driving the audio inputs of the receiver.

Mostly following datasheet guidelines, but also getting some ideas from other people’s implementations with similar DACs, I ended up with this circuit:

On the USB input, there is some filtering for the 5V supply, and ESD protection for the differential data lines.

On the audio outputs there are 3rd order low pass filters, first passive single-pole RC-filters, followed by Sallen-Key filters (active 2nd order LPF), and the cut-off frequency is around 28kHz. If symmetrical voltage supply of e.g. +/- 15V was available, I’d choose NE5532 op amps, which are both cheap and high performance. In this case only 5V is available, and inside a radio receiver I didn’t want to use a switched boost converter, as such things tend to be noisy. For low voltage, single-supply audio design, the OPA2353 (Burr-Brown/TI) is known to perform well, so I went with that one. I also gave the active output filters gain of about 3.6dB.

To minimize distortion, thin film resistors and NP0/C0G capacitors are used in the audio signal path. Some may argue about the use of electrolytic capacitors on the audio outputs. Electrolytics will cause distortion when there is AC voltage across them, typically when used in filter circuits. In this case, they are used for DC blocking on the audio outputs, and form high pass filters in conjunction with the audio inputs of the receiver. With the large capacitance of 100µF, and a receiver input impedance of 15kΩ, it causes a cut-off frequency of 0.1Hz. Therefore, these capacitors will not be a limiting factor for the overall frequency range, and there will be practically no AC across them, and hence very little distortion. If I was to use, say 10µF polypropylene caps for DC blocking, these would be physically larger than the whole board with the rest of the circuit. Recommended reading: Small Signal Audio Design by Douglas Self.
In Huldra 11, electrolytic caps are used for DC blocking between amplifier stages, and polyester caps in filter circuits like the tone control.

PCB design in KiCad:

The board was designed to fit in the space of TAPE2 input of Huldra 11, but will probably fit in other receivers as well.

The finished DAC board, installed where the Huldra 11 TAPE2 input used to be.

Rear panel of the receiver:

Connecting it to a computer is simply plug-and-play. The PC was already playing music to an other receiver, and once I connected the USB cable, the PC identified the new USB DAC and started playing through the Huldra 11 DAC. All I did is rename the device for easy recognition.

How does it sound? To my ears it sounds neutral and clear. No audible noise, no hum. Just as I want it.

How about high frequency noise? The DAC is after all placed right below the FM tuner and beside the AM tuner. I tuned through the FM band and all AM bands. There were three spots on the shortwave band where the receiver picked up a little noise which could be traced back to the DAC, i.e. the noise disappeared when I disconnected the USB cable: about 6, 12 and 24MHz. That’s all. The noise level was however so low that it wouldn’t be audible if the receiver had picked up actual radio stations at those frequencies. From that, I concluded that there is no need for EMI shielding around the DAC board.

Design files (KiCad) and production files (Gerber) is shared on Github: https://github.com/donpivo/Huldra-11/tree/main/USB_DAC