FCA202 FireWire Analog Frontend

If you do not have much more than a scoop and a tone generator it gets difficult to measure the behaviour of audio equipment, being it amplifiers, speakers or simple opamps. 

That is the reason I started to look for software on my PC that could help me to generate different test tones and do the processing of the signals in the PC to extract SNR, frequency curves and alike.
I decided for a PC program called "AudioTester" from German origin  www.audiotester.de that provided all functionality I was looking for. Over time several neat features have been added.

The next step was to find a quality audio input-output system to hook up or insert in my PC.

Nowadays there are plenty for sale at moderate prices but back then in 2008 they were hard to find and I decided to buy a Behringer FCA202 IEEE1394 or FireWire interface box.
I had a FireWire interface on my PC, it promessed to be a defacto standard and has sufficient bandwidth to support stereo audio streaming.
For the analog interfacing I wanted some more flexibility than the FCA202 box provided.
The FCA202 line input- and output did not have a volume adjustment capability, only the headphone output is adjustable.

What I also wanted was an input scaler so that I could directly connect both the output of a power amplifier and a low level signal smaller than 100mV. This resulted in the build of the FCA202 FireWire Analog Frontend.

The line input has a 100KOhm input scaler for a x1 and a x0.1 attenuation and the third setting is a gain of x10 so a range of x0.1, x1 and x10 that gives a maximum input voltage of over 90Veff.

The lowest input voltage is limited by the noise of the amplifiers that lies at -100dB (at gain 1x) at full scale of the converters in the FCA202.

The input resistor divider is switched using small reed-relais.
As amplifiers I used the LME49710 known for its very low distortion and noise.
The line input with the attenuators has a non inverting amplifier connected to a 5K potmeter to adjust the line input level followed by a second non inverting amplifier as voltage follower to drive the input of the FCA202 for both left and right channels.

The output from the FCA202 is buffered with a non inverting amplifier driving a 5K potmeter and the output is buffered by a non inverting amplifier with a gain of 3x to drive the line outputs, again left and right channels.

In retrospect I should have chosen for inverting opamp topologies since there are always two amplifiers in the signal path from FCA202 to line and vice versa. Inverting opamp topologies have an inherent lower distortion at higher output amplitudes since the input stage is not voltage excited but at virtual ground.

At the front panel there are four separate potmeters for left and right line-in and line-out in case I want to measure the input and the output of an amplifier simultaneously.  With the left and right pushbutton I can separately set the left and right line input attenuation. Be carefull to first set the correct attenuation before applying any high voltages above 10V eff.
The pushbuttons, reed relais and LEDs are controlled by a MSP430F2012 that also has a 12 bit ADC on board with input mux.
The line input signals are buffered and peak rectified by an OPA2353 and their output signals are read into the ADC of the micro to monitor overrange.

The reed relais and LEDs are driven by two TPL9201 buffers with serial inputs to save pins on the MSP430F2012.

The powersupply is straight forward. A traditional PCB transformer and rectifier with a TPS7A4901 LDO for the positive 15V regulation and a TPS7A3001 LDO for the negative 15V regulation.
A power module PTN78000WAH is used to generate a 7.5V rail to power the reed relais and the LEDs whilst a third LDO, a LP2951 provides the 3V3 supply for the MSP430F2012.

TPD2E007 devices are used at all inputs to protect against ESD discharges.

There are in total 8 Cinch connectors at the back side of the interface box, four for the FCA202 inputs and outputs and four for the line inputs and outputs.

The PCB is a 2-layer board with a bottom side groundplane and power and interconnect routed as much as possible on the top side.

Later I decided to route such analog boards with additional power planes to keep a massive groundplane. They are more expensive but make routing a lot easier and provide better noise suppression.

I also found out that when the FCA202 is driven at higher input and output levels the distortion increases to a -70dB. Most likely because no dual supply opamps were used.

Since there is no galvanic isolation between the PC, FCA202 box and my analog frontend there is a 50Hz mains component plus its harmonics measurable at -85dB.

A next project I am thinking of is to give the design a facelift support optical digital audio links and keep the ADC and DAC conversions under my own control. More on that in the map "Product Ideas".