Bringing an antique wooden hand crank telephone from the 1890's into the app age
To make the experience fit your profile, pick a username and tell us what interests you.
The only major (and irreversible) modification I had to make to the body of the phone was drilling a hole for the power adapter to plug in to. Unfortunately I hit one of the screws that was used to hold the frame of the phone onto the back, but it still worked out and the power adapter fits fine. The two other holes I believe were used to hold the touch tone phone that was installed at one point; I can't think of a reason they would have been used for the original batteries.
I undersized the hole by a fraction and put the smallest dab of Gorilla Glue onto the powerjack threads before screwing it into the hole because I couldn't fit the nut.
The power cable comes out of the base of the phone at the back and will be hidden by the fake batteries.
I installed the door open/close sensor switch in the bottom right of the phone body, this ensures it stays pressed even if the door opens a few MM as it tends to want to do.
The door switch, crank switch, and earpiece switch were all connected to the breakout board which I ended up sticking to the back of the phone body to keep it from rattling around.
I connected everything to the batteries and finagled them into the cavity. I had to disconnect and reconnect a few cables to move them in front of/behind each other to get everything to fit nicely.
I made sure everything fit, and closed the door! The hardware was complete and ready for testing.
The final schematic of everything in the phone.
The schematic does show a neopixel on the switch connector; I originally intended to install neopixels in each of the 3 sections of the phone so I could do some cool color effects. In the end I couldn't find a clean way to install them without them being too obvious so I didn't end up using them. The wires for them are connected to a TRS jack in the switch breakout as a sort of 'expansion' port - it has 3v3, GND, and a GPIO connection so if I think of anything to add in the future it can just be plugged in there.
The ring generator battery houses the self contained ring generator circuit, as well as the power regulator that steps down the 12v input to 5v to power the Raspberry Pi and other components in the other battery and installed in the phone (microphone, speaker, LEDs, etc). The circuit board in this side came together much more organically as I worked around unexpected needs for different things like the relay and L298N board.
When I originally designed and printed the ring generator battery, I thought that I was going to be able to fit everything onto the circuit board, not anticipating the need for the L298N board. I designed a carrier for the L298N board that attached to the main housing using the top screw that holds the transformer in place.
I wanted to avoid as many non-permanent connections as possible to ensure there were no possibilities for things to come loose after the phone is installed. I removed the screw terminals and headers on the L298N board and directly soldered wires to the board.
This was connected directly to the circuit board, which had become quite cramped and components were installed on both sides.
The Raspberry Pi power regulator was installed on the back of the board along with the connections for the USB cable and button (ignore the resistor being used as a twist tie to hold the wires out of the way).
This board then slots into the fake battery housing.
The L298N is attached to its carrier and then fits into the open space below the circuit board.
The high voltage output wires from the transformer were then soldered to the wires connected to the jack being used for the bells. The wires were tucked into position and then the cover was screwed on. Staples didn't have any brown cardstock to print the label on, so I'm thinking I'll try to weather the label with coffee or something as it looks too new to me.
Ringing the 100+ year old bells turned out to be one of the more challenging things to accomplish. Due to the vast distances that signals had to travel to reach rural subscribers in the late 1890's, the ring signal is actually 90 volts AC at ~20Hz. Generating this from 5-12v DC is not the easiest thing to do.
There are a number of varyingly complicated 'ring generator' circuits that are available online including one developed by Sparkfun engineers for their Bluetooth "Port-o-Rotary" phone, and one from "Bowden's Hobby Circuits" using a 556 timer + 120v -> 12v power transformer in reverse to step up the voltage.
I initially avoided the Sparkfun circuit due to the number of required components that I didn't already have on hand, and attempted the Bowden Circuit with a transformer I ordered from Amazon. Unfortunately, the Bowden Circuit did not work correctly; it did generate a pulse of some sort, but it seemed to be 20Hz DC and not AC. I wasn't quite sure how the circuit was supposed to be generating the negative portion of the AC wave anyway, which I think may have been the issue. Unfortunately I don't have an Oscilloscope to dig any deeper. Only one magnet coil on the phone seemed to activate with this circuit which may have been designed for a later phone with only one coil.
I had initially planned to replace the D1 Mini with a single chip or something smaller, but it ended up fitting just fine into the fake battery so I left it as is. This entire 'ring generator' circuit is self contained and powered by a relay controlled by the Raspberry Pi in the other fake battery. This ensures that the high voltage circuit is always powered off unless turned on by the computer and actively ringing, making the phone safe to open while not ringing. There is a door sensor connected to the RPi that will not allow it to ring when the door is open as the high voltage wires are exposed and not insulated.
I wanted to be able to summon the Google Assistant by turning the crank of the phone, which historically is how you would have summoned the operator. Originally turning the crank would generate an AC voltage of approximately 90 volts, which would cause the bells on the operator's end to ring signaling that you wanted to place a call. I actually considered half a dozen different ways of detecting the handle being turned including hall effect sensors on the gear teeth, but eventually settled on what I think is the simplest. The way the crank is designed, when you start turning it it converts the first half rotation into horizontal movement to slide the shaft sideways and make contact with a terminal. This completes the circuit and allows subsequent turns of the generator to send voltage down the telephone line. When you stop cranking, a spring slides the shaft back away from the contact and breaks the circuit so that the lines can be used for the telephone call.
I removed the contactors and replaced them with a nano switch that would be pressed when the crank was turned.
This worked great to detect cranks but I had to adjust the mount to hold the switch at an angle due to space constraints in the telephone housing.
This works perfectly and fits in the telephone housing with a few mm to spare.
I added wires and the generator was ready to go back into the housing.
I soldered the cables onto the connectors I'm using for the main battery, which includes the 8 pin RJ45 jack for the Microphone, 6 pin Mini Din jack for the door, crank, and earpiece switches, and earpiece jack, as well as the micro usb power input/ring signal output. The micro usb jack is wired so that the main battery can be powered by itself, but when it's connected to the ring generator battery the D+ and D- wires are used to send the ring signal to the ring generator.
I started assembling the circuit board by attaching the cables to the board.
After 2 out of the 3 connectors were attached I started wiring the connectors to their respective pins on the raspberry pi header.
The 6 pin mini din connector had to be soldered in last after it was installed in the face plate due to the fact that it inserts from the front of the face plate and not the back like the rest of the connectors.
I installed the components into the housing, everything screws in directly except the RJ45 jack which is held in place by a set screw.
I attached the ribbon cable for testing everything before installing it permanently.
The raspberry pi is installed so that I can still access the micro USB ports by just removing the shell of the fake battery. The SD card slot is also externally accessible.
All of the jacks after being installed in the face plate and connected.
The interior housing is threaded on one end so the outside just screws on. There was some weird banding in the print for the housing but I left it because it will give more surface area for the glue to grab onto when I install the decals. The faceplate is intentionally red because the old batteries used to be sealed with red epoxy on the top.
After everything was installed I double checked everything worked as expected when connected to the phone.
I pulled the fabric covered earpiece cable through the hole that the original cable came from.
A ziptie around the cable will prevent it from getting pulled out or straining any of the connections if th earpiece is dropped. I added a zip tie pad to prevent pushing the cable in to the phone from causing any issues with the rotating generator next next to where the hole is.
I printed the fake batteries and shells to make sure they fit. The shells screw on and will be covered by ~2mm of plastic.
I wired the socket for the microphone plus the neopixels.
The microphone bracket fit perfectly into the top part of the mouth piece.
The middle piece screw onto the top piece and holds the bracket in place.
I reattached the top/middle piece to the main phone part and everything fits perfectly.
With the mouthpiece cone screwed on you can't even tell there's anything special inside.
I replaced the speaker in the earpiece with a modern 40mm 3W driver. The original speaker may have worked but I wanted modern audio quality.
I opened the microphone compartment on the antique phone to see what I was working with. It looks like this phone has been repaired before and the installed microphone probably isn't the original. Based on other repairs/modifications I'm seeing, it looks like this phone was converted to use a modern tone based system at some point.
I 3D printed a bracket to hold the digital I2S microphone in the space where the original went, and then decided to add space for some neopixels on the sides. I think these will be visible through the mouthpiece and be able to provide some basic visual feedback without being able to see them when they're off.
I need 7 wires for the digital microphone plus the neopixels, so I ran a slim cat6 cable through to the mouthpiece. I'll add an RJ45 jack to the enclosure so that this can be easily plugged in or unplugged if i need to remove the computer or other circuits.
The microphone and DAC share the only I2S interface on the Raspberry Pi, so I wanted to verify that both could work (be installed at least) at the same time. I've connected everything up and verified that I can record audio with the microphone and play it back with the DAC. Unfortunately the DAC is line level and not loud enough, so it will be replaced with a 3W amplifier + DAC that is on order.
Become a member to follow this project and never miss any updates