Audio baby monitor
For a few months I was using an IP camera as a baby monitoring device.
While the video was ok, the sound quality was quite poor. It had a high noise level and a serious lag.
The high noise was very inconvenient especially during the night when it was quiet in the house.
Rather than looking for another model I decided that this is a good excuse to design a new circuit.
The plan was to mute the camera and use instead a small active speaker connected to a microphone in the baby’s room.
An analog wired solution seems to be the best option. I did not need mobility and there was already an unused
UTP cable running from baby room to our bedroom. I also already had a quite nice Phonic DM680 microphone with balanced output.
Obviously it was required a preamplifier since there are around 10 meters of cable in between.
A fully differential amplifier looks like the best choice. There are plenty of operational amplifiers that could do the job.
But I wanted to use discrete transistors (BC550C,low noise). Also, I just got a bag of nice colored Russian resistors from the soviet era, and I was wondering what to do with them.
So, I came up with the next circuit. The Q1 to Q4 form a classic differential but simplified amplifier. Because there are no resistors in emitter of Q1 and Q2. By omitting them I sacrificed CMMR for extra gain. So now the gain is determined only by R3, R4 and the transistors re.
As a result, the CMMR dropped to 20dB at 1kHz. But this is not an issue. Even with 2 meter of microphone cable. The gain is around 212 times. Just for curiosity I added a 500ohm balancing trimmer between the emitter of Q1 and Q2. The trimmer’s cursor was connected to Q4. After some adjustments the CMMR jumped to more than 50dB. But the gain dropped a few times.
The C1 capacitor was added after I noticed a barely visible 1053kHz sine wave at the output. The origin of that signal was a powerful broadcast band transmitter close to the city. R7 trimmer allow a minimal adjustment of CMMR. Since the current consumption of this stage is low, around 1.4mA, it is worth to increase the efficiency of supply voltage filtering by adding the R9 resistor.
Now the signal is amplified enough. But the output impedance is high. This is not the best choice for driving a long cable. Because of this I added a differential emitter follower with Q5-Q8. The current of this stage is set to 10mA by R10.
I did not want to use a switching mode power supply to avoid possible noise interferences. I also happen to have a 14Vac AC/AC adapter. This is the reason why there is a rectifying bridge on the right.
At the end of the 10-meter cable coming from preamplifier I added a 100ohm to 10kohm transformer. This takes advantage of the high input impedance of the active speaker and is increasing the voltage level 4 times. It also preserves the common mode noise rejection when switching from balanced to unbalance.
The overall noise and sensitivity are small enough to hear a whisper from 5-6 meters away from the microphone. Perfect!
And this is the masterpiece:
Noise level at output (no power)
Noise level at output (input short circuited)