This was an article that I published on a website a friend of mine was involved with running back in 2014. My friend is not involved with it any more and the website is not what it once was. For that reason I'm reprinting it here.
A while ago I wanted to get a Bluetooth speaker with a built in AM/FM/SW radio. Not happy with what was out there for various reasons, I decided to see if I could build something to do the job. After about 10 seconds I realised that would be too much work, so I thought about hacking something that might be close. An AM/FM/SW boom box with a broken tape deck seemed like a good way to go. After trawling eBay for a bit this is what I ended up with.
Annoyingly the person who sent it to me did not pack it very well, so there were a few cracks in the case, and more importantly in the main circuit board. This meant that none of the radios worked any more. Fortunately the tape decks did still function (enough), so I could still go ahead with my plan.
Internals of the boom box
The internals of the boom box were actually quite modular. They fit together something like this:
The boombox internally is made up of 4 modules. There is a power control board that converts the incoming 110V or 240V power (this unit has a switch to go between) to 9V and controls power supply to the other modules. This is connected to the am/radio board via 2 sets of wires. One set acted as a switch that would cause the power module to turn on and start supplying power. My theory is that this must have been connected to the switch that goes between am/fm/sw/tape. The other connector was for the power itself. Similar connectors also existed between the tape modules and the power module, except when the switches from the tape deck were turned on they also turned the amp module on. And finally the audio from the tape decks was fed up to the amp/radio module so that it could be sent to the speakers.
So the way the bluetooth circuit is wired in is to take the audio line from deck B and wire it into the output of the circuit. The switch that used to run to the tape deck has been re-routed to a switch on the front of the stereo that used to control how it would switch between tape deck A and tape deck B when one side stopped playing. Originally I wanted to use the power that was used to power the tape deck as well, but at one point this stopped working, so I used the same power lines that went from the power module to the amp/radio section. As a result of this there is a bug where if you press play on tape deck A the bluetooth section also turns on. Maybe I will eventually get this fixed.
Modding the power and sound
At this point it there are two issues left. The first problem is the power. The bluetooth module I chose needed a power supply of between 4V and 8V. However the power coming off the boom box's power line was 9V. The module may have been able to tolerate more than this, but I thought it best to bring it down to something in the middle of that range. To do this I used a simple L7805 voltage regulator with a 100uF capacitor across both the input and the output to "clean" the power. Although the power coming out of the battery would be quite steady, the current draw of the bluetooth module might not be, and if I wanted to ever try and record something onto tape on deck A, then the motor could also cause fluctuations in the power provided.
The next problem I face was the sound. Initially I thought this would not be a problem at all, as there are quite a few tutorials out there that say all you need to do to convert a boom box to have an aux input is to cut off the read head and put an aux input in its place. Since the BT module was just like an external device, I thought the procedure would be the same.
However, this was not to be. Plugging the BT module straight into the amp section of the boom box resulted in a lot of bass distortion. I assume what was going on was that the tape read heads on the boom box had very poor bass response. To compensate for this the amp is designed to amplify the base more than the higher frequencies. So in order to make the sound coming out of the BT module work with the amp I would need to filter the signal that came out of it to make it more like the signal that would normally come out of the read head.
The circuit I initially tried was a simple high pass filter. Like the name suggests, a high pass filter lets high frequency signals pass, while attenuating lower frequency signals (http://en.wikipedia.org/wiki/High-pass_filter). After some experimenting I found some values that worked. At this point I was powering the circuit through a 9V battery, so I decided to try powering it from the boom box. This is where I hit the next hurdle.
This time the problem was that whenever I connected both the power supply and the audio connectors of the bluetooth circuit the dominant sound that came out of the boom box was the RF signal of the bluetooth module itself. This sounded a lot like the sound you hear in an FM radio when a cell phone talks to a tower, and since whenever music is going through it will be receiving data, this distortion showed up whenever I tried to play any music through the module. The most interesting part about this is that this would still show up even if only the ground wire of the audio connection from the amp section was connected.
I figured I needed some way to isolate the amp from the bluetooth circuit so that the audio input of the amp was not directly connected to the audio output of the BT module. Initially, I tried using opto-couplers, for the simple reason that I had some. This resulted in the sound being very "click". This is probably due to the fact that the optocouplers are digital, and therefore were unable to transmit an analog signal properly.
After a couple of other daft ideas (like literally trying to build a better optocoupler) I tried using small audio transformers, one on each of the left and right channels. I grabbed two 42TL transformers from Mouser. Putting these in between the BT circuit and the amp seemed to fix the interference problem, but the base problem came back, so I put the high pass filters from before in place after the transformers. After a little bit of tweaking with values I go this to a point where I was happy with the sound and soldered it all onto the board.
So, the complete circuit looks like this.
The section labled "BT module" in the middle represents the bluetooth module I used (and OVC3960). Note the use of the potentiometers on the filter. This was done to allow the cuttoff frequency of the high pass filter to be adjusted.
Putting it together
Initially I had thought of bolting the circuit board to the back of the case behind where deck B used to be. However, when I actually came to look inside the case I saw that this was not such a good idea.
That circled bit there is where the 110/240V power comes into the boom box. If I ever wanted to use this while plugged into the mains, this seemed like a very bad place to put the circuit. This is because if it came loose it could quite easily cause a short circuit across this part of the circuit, which would result in probably blown fuses in the fuse box, the nasty smell of melting plastic, possible electrocution, and a bad time for all involved. The solution was to rip out all of the mechanical parts off the back off tape deck B and mount the circuit there.
As I mentioned there is a problem with the resulting sound being a bit treble heavy. This could be resolved by lowering the cut-off frequency of the high pass filter, but using a two pass version rather than the simple single pass version used in this circuit. Using a two pass filter (which is simply two high pass filters one after the other) results in a smaller 'drop off range' and better control of exactly which frequencies get cut off. The other alternative is to put a low pass filter after the high pass filter to attenuate the higher frequencies.
The other thing that would be to make the buttons on the removed tape deck work again, so that the play button would pause and play the music and fast forward and rewind would allow next and previous track selection. The OVC3860 bluetooth module I used has a type of serial interface called a UART interface, which can be used to send in basic commands like this and something like an ATTiny85 (a type of microcontroller) might be able to be programmed to send these commands based on button presses. As for the buttons, none of those buttons mentioned have any electrical connections. All they did was turn the motor on via the switch on the back of the tape section, with the direction of the tape and speed being controlled by a mechanical process of which gears got lifted into place with the put-on was pressed. So I would need to mount some little switches behind where the buttons are to make this work.
As for the audio quality, it's pretty good considering that this thing is only a few years younger than me. I really wish the radio was still working because then I would be able to have something to compare the audio coming out from bluetooth to. I would really like to know if my circuit is adding in any interference. I guess I will need to get a pre-recorded tape from somewhere to try it out. All in all, I'm happy with the result and hope to get a few years use of my new creation.