What's inside this old military telephone? | TA-1042 Part 2

welcome to Nick snacks I'm Nick and today's Knack is yet again the ta-1042 a military Telephone from the 90s that uses a proprietary military digital telephone standard have you ever woken up looked outside to see the birds chirping and the sun shining and said to yourself I wonder what's inside this old military telephone no back buttons up there my friends but for the rest of you today we're going to crack this open and see what's inside this is one of a class of several phones using the digital non-secure voice terminal standard or dnvts if you haven't seen it part one of this series gives a good project overview and what we've done here to date but for Cliff Notes version we have reverse engineered the military switching protocol implemented a hardware interface which is this box which takes the phone from a paperweight into the 21st century now with USB support these boxes are available on Tindy for any of you crazy enough to want to buy one right now you'll be able to plug in several phones and make calls between them and we are working on getting them to interface with voice over IP today we begin the more in-depth overview of the reverse engineering process although this is a rather Niche application the process of reverse engineering itself is quite an interesting one and I'd like to provide some detail on what we did here today's video covers the first steps of reverse engineering taking a peek at the hardware and the basic protocol inputs outputs on these phones and working backwards to design a compatible phone switch for them the next video will be a deep dive on writing real-time firmware on the Raspberry Pi Pico to enable some basic functionality like being able to call between four phones connected to a single switch and in the last installment of the reverse engineering Deep dive we will talk about how we can connect these switches to a computer with a simple USB protocol and Bridge the data over the internet and thus attain true nerd status my partnering crime on this project is Rob Roark master of all things electrical I bear no responsibility for the design of this particular switch Rob basically started from poking and prodding at the phone itself and was able to design our own interface for four phones that is almost pocket-sized for reference the original switching equipment was trailer sized and eventually shrunk down to Van sized and had an incredible power draw so I've received a few comments from people who used to operate this equipment that they're Blown Away by the size and efficiency of this design hi I'm Rob I live down the street from Nick who happens to be the only other person probably in Seattle interested in these phones I've wanted to do something with them since college but at the time didn't have the time and also knew I didn't know what I needed to to program a microcontroller to decode everything in addition these phones have quite an aesthetic which was a big motivator in investigating these further [Music] but unfortunately we can't just plug them into a phone line sip adapter or something like that and talk is they use a proprietary four wire digital interface if we want to revive these phones we'll need to understand how the phone itself works so we can learn how to talk to them this process is called reverse engineering where we take a device like this crazy phone where we don't have specifications or a support number we can call and try to ascertain how it works the first thing we typically do in reverse engineering is begin to explore the inputs and outputs of the device to get a sense of what it's doing so these fonts uh when I as soon as I got one off of eBay I hooked one up to see you know what it was outputting so when you take the phone out of the Cradle you immediately see this differential Manchester code which if we go ahead and take a single capture of despite all of the horrible switching noise you can walk through and start to decode that so the second thing I did when I got these phones was hook this output of the phone to an isolation Transformer I happen to have one for injecting signals for doing Loop gain and phase measurements on power supplies so I hooked that up and actually got quite a nice signal that right there is an excellent segue too tangent time [Music] consider the following Transformers what are Transformers they're magic well technically they're electromagnetic Wizardry they use all those Maxwell's equations I've long since forgotten and what they do is they transfer AC power from one side to the other to construct one of these doodads we wrap some wire around a magnetic core on both sides power flows between the two windings the electromagnetism this right here is the Transformer we are using in the switches obviously it is a tiny dude they make larger ones but in order to understand how much of the rest of this video and reverse engineering process worked we're gonna have to explain what this is and what it does for us because I don't really remember my graduate ee degree nor does anyone else let's be real I'll explain how this is used in context a simple use of a transformer is for a DC power supply like one of those wall warts from back in the day that we had like 10 pounds and plugged into the wall to power your Marantz tape recorder we have ac voltage from the wall outlet and a rectifier or a couple of diodes that convert that AC input to DC output the only problem is the voltage from our split phase us Outlet is 120 volts ac which if you recall is a root mean squared value so the peak voltage of the sine wave coming from that outlet is actually 170 volts so if you feed that straight through a rectifier and into your tape recorder so how does a Transformer help us well a Transformer is a device that's going to pass AC signals through from one side to the other there's a lot more Nuance but let's just completely ignore that for now we feed the line voltage AC power to this Transformer on the primary or input side and then we connect the output of this circuit in this case a rectifier to the secondary or output side if we vary the ratio of turns on each side so for example we could have 10 turns on the input side for every one term on the output side we can effectively take that ac voltage and step it down to our desired voltage or up so if we needed let's say 17 volts Peak Peak we could use a 10 to 1 ratio to step the 170 volt input sine wave down to 17 volts because we're already in Tangent time the existence of the Transformer is why AC power one out against DC power back in the early 1900s we can use a very simple passive device this Transformer to modify the voltage up or down without transistors and switch mode power supplies and all that fun stuff this is important because the loss in a power line is proportional to current not power so if we is Knick snacks Power Company Incorporated need to send power a great distance we can use a Transformer on one side to step up the voltage very high say 500 000 volts from a much lower voltage like 10 000 volts what this does for us is reduce the current to carry the equivalent Power by a factor of 50 and in doing so reduces the current in our long transmission wire and the power loss to resistance by the square of that these Transformers cannot be used to pass DC we use crazy Power Electronics called buck boost converters using transistors which didn't exist 120 years ago to modify DC voltage and as such AC power proved much more manageable at scale the Transformer also accomplishes something else called isolation we're delving into what I call EE feng shui here but I'll do my best to explain the central theme which is the idea of relative voltage you might think of voltage as an absolute value but a voltage is always referenced to something because it implies a difference between two points that's why your voltmeter has two leads to compare with a Transformer the two wires on the output of the Transformer are not directly connected to the wires on the input they are connected through the magic of Maxwell's equations and not via the physical connection if they were directly connected it would explode because the voltage might be 500 times greater on the output side than on the input what does a Transformer guarantee it guarantees that the ac voltage between the output leads in this case will be 17 volts but the Transformer actually doesn't care about the absolute voltage of that output there's two main advantages to this isolation number one ground voltage is not always the same between arbitrary different points and if we stick a transformer in the middle we allow the signal to pass through on the AC side without the system actually caring about what the ground potential is we can imagine that if the ground potential was different on one side than the other your little ground wire in your circuit would be sinking current to try to equalize the two over your signal line the easiest way to illustrate this portion of ee feng shui is with a battery which shows you this concept of relative voltage in DC land so this is a 9 volt battery and it guarantees 9 volts between its two terminals so you might think that this could only ever be 9 volts but if we stack two of them together and we consider this bottom terminal to be the ground potential the output of the battery on the top is now 18 volts relative to the bottom so the battery voltage is still 9 volts but the output is now 18 volts because the negative terminal of the battery is now set to 9 volts by the first battery we can do the same thing in Reverse this positive of this third battery is now at zero volts and the bottom is at negative 9 volts it just depends upon what point you reference the Transformer maintains the same concept of the battery but this time the Transformer blocks DC voltages from passing through and transfers AC signals practically speaking that means we can use these two input wires for two different things we can use a common DC voltage to actually send power to a device and the differences between those wires would convey the actual signal the Transformer blocks DC power which we can pilfer off and send to a power supply and then this AC component the much smaller signal will be passed through the fact that the receive side is at 48 volts offset makes no difference to the digital logic when only the small signal is passed through and the logic only sees that small difference between the wires so in a roundabout way in the case of Rob's oscilloscope this isolation Transformer allowed Rob to measure the signal from this dnvt without any interference from the measurement equipment when you have an oscilloscope the ground clip on that oscilloscope is actually connected to Earth ground this will inevitably bite you at some point because if you clipped that onto your circuit at an arbitrary point it will pull that point to ground as well which can interfere with your circuit and therefore measurements an isolation Transformer like we have here is simply a Transformer with a one-to-one ratio of winding so it doesn't change the voltage up or down but it's going to prevent a DC bias from interfering with your logic on the other side and your final tangent time fun fact this isolation Transformer business is how power over ethernet Works in addition to providing the power and signal for the dnvt so with that out of the way let's take a look inside the phone and see what's cooking so knowing that sort of signaling then I pop the phones open to see what was going on inside this is the phone exploded apart it's just bolted together and it's a pretty easy disassembly aside from there's a sticker here you need to pull off the because it's a metal sticker it's pretty hard to tell it's blocking a screw hole but you take the phone off you've got receiving transmit on these binding posts you've got power on the side so the next step of reverse engineering is understanding the circuit inside the phone by understanding how the phone interfaces with the outside world we can understand how to replicate this in our own switch broadly speaking a circuit like this will be broken up into several parts we of course need some way to power the phone in this case we accept power from either the phone lines or from a battery or local DC Source attached to the side that power needs to be conditioned and then regulated down to the necessary voltages then we have the input and output of the phone which are connected to the four binding posts on top we expect some sort of protection circuitry which minimizes the risk from the outside world on the more sensitive Electronics down below next the circuit performs the conversion of the signal from whatever format is used to convey the data over the long wires to a logic compatible format often some sort of 3.3 Volt or 5 volt level lastly we have the central processing unit of the phone itself which receives the incoming data and crafts the outgoing data we will call this Logic the brain box and in the case of this phone it uses an ace or application specific integrated circuit basically a custom computer created to do a single task on our switch though we've been able to take advantage of modern microcontrollers and replace that Asic with a Raspberry Pi Pico I'm going to walk through the various circuits of the dnbt at the input you can see your six of these turrets coming up those turrets are for the wire connections on this side the little spring wire connections so starting at the right here we have the receive path to the left of that we have the transmit path to the left of that we have the auxiliary voltage input here we just have some protection components the common mode choke and some filtering this is for when you're powering it in local battery mode that power goes down the side of the board to this flyback power supply here which provides an isolated power supply for the phones there's a diode from the received circuit that takes the center tap of the primary of the received Transformer and passes the power from the lines over to the flyback as well ground return for that is through this inner tap to the primary the transmit Transformer interestingly the phone has no idea whether or not it is getting power from a battery or from the phone lines themselves the local battery switch on the front simply controls the mode selection for the brain box so it knows what to do so the power regardless of where it came from gets conditioned and is turned into stable voltages at the necessary levels down below by the is flyback power supply this section here is for the csvd audio these two chips to the left are the csvd encoder decoder chips there are two of them because one is used for transmit and one is used for receive either chip can do either but you can only use them in One Direction at a time as the phones are full duplex there are two of them uh these chips to the right of that are some op-amps and discrete transistors I didn't Trace out exactly what all this is doing but it's uh appears to just be analog audio for uh interfacing to the encoder and decoder we briefly touched on the cvsd codec in the last video but it was a codec designed in the 80s to have a good cost to complexity trade-off so it's not super efficient but it sounds adequate at 16 kilobits per second and it can be implemented with an integrated circuit with a relatively small number of components it is similar to a direct stream digital codec the codec used in super audio CDs in that it contains a single bit stream and a one means go up and a zero means go down the main feature of cvsd is that if you get a couple of ones or a couple of zeros in a row it actually increases the step size so that for example if you have to go very quickly upwards it won't take as many ones to get there this allows it to reproduce higher amplitude signals and lower amplitude with similar frequency content and similar Fidelity we were able to secure a cbsd codec implementation for my Dad's friend and we will cover how we can implement this codec on a computer in a future episode the final circuit to the right here is the custom Asic that does all of the differential Manchester encoding logic as well as transmitting the code words based on what's received on the keypad and receiving code words from the switch uh so now I'll go into the receive and transmit circuits uh the first component that we see is this white bipolar gas discharge tube it's a three terminal device the center terminal is ground and the terminal to the left and right are the positive and negative sides of the receive input so the purpose of this component is to shunt any large transients immediately to ground so this device and gas discharge troops in general present only a few picofarads of capacitance to the line but if the voltage is high enough it strikes over the gas inside ionizes it's typically a pinning mixture of argon and neon though I don't know exactly what's in this particular tube and then it becomes a short circuit that can handle hundreds or thousands of amps following this component we have two inductors these two inductors are for additional filtering and for shunting that current into the gas discharge tube as opposed to letting it come into the rest of the circuit as the inductors resist fast changes in current if you have trouble remembering which does what just remember the capacitor blocks DC because when you draw the DC on here it runs into the plate whereas if you draw AC you'll notice that it can Meander through the opening in the plates following the inductors there are some filter capacitors to ground and then that enters the Transformer there is also a protective diode so a TVs diode I was not sure if it's voltage but I assumed something in the realm of eight volts so why do we have this crazy array of protection circuitry we have a gpio output pin of our microcontroller can we just connect that over a very long wire to the input pane on the other side well if you just try to connect a gpio pin of a microcontroller to another one over like a four kilometer long wire a whole host of problems will appear to name a few the driver inside the microcontroller may not have enough oomph to take a long length of wire and phone on the other side and move it from 5 volts to zero and back 64 000 times a second these transmission lines have characteristics like inductance and capacitance that make them not ideal so just because you tell a microcontroller to set its upward pin high that doesn't mean the other side of a long length of wire will see that high voltage instantaneously in addition you may start to get attenuation of the signal over the length of wire that reduces the available signal levels you can amplify that signal but you'll find that over a very long length of wire it functions a bit like an antenna and can introduce a decent amount of noise if you've ever used a CB radio under power lines you're no doubt familiar with this and last you find that if you get a lightning bolt strike anywhere near the device or the length of wire that connects them it can induce a ginormous spike in potential that will often vaporize or let the magic smoke out of your poor little computer so for these reasons we employ an electrical interface after this brain box and before the output of the phone to provide a more robust signal and give us protection from any excessive voltages that appear on the wires after the Transformer on the receive side there is an op amp circuit used as effectively a comparator that is able to decide if it's a positive or negative on the differential side that signal then passes on to the Asic the transmit side is pretty similar I'll trace it from the Asic Direction there are two signals that come out of the Asic the drive push pull the secondary side or you could call it primary since that's where it's being driven from of the transmit Transformer those two signals drive to npn transistors that are connected through a resistance to the winding of the Transformer the center tap of that Transformer on the secondary side so that the side of the Asic is connected to 5.8 volts so by shorting one side or the other to ground you can drive current One Direction or the other through the Transformer after that we have the same series of protection components there are some filter capacitors inductors and a gas discharge tube for those of us that don't entirely remember our electrical engineering degrees I'll just do a quick explanation of what differential signaling is a classic example of this is an XLR microphone connector which carries what is called balanced audio so if we take a look here at oh that that was sub-optimal so if we take a look here at this other XLR connector you can see that we have three pins in this connector even though we only have one signal coming through inside the microphone there's a little Transformer and it takes the output of the microphone transducer and converts it into two signals that both carry the same audio signal but one of the signals is inverted they have opposite phase it's back the twisted pair Arrangement has a couple advantages the biggest here is noise resistance where if any noise is induced through interference it will be induced in the same proportion in each wire The Twist helps ensure that the average distance from each wire is the same from the noise Source when the noise gets to the receiving end the receiving equipment flips the phase of one channel and sums them this has the effect of canceling out the noise in the signal that came over our long run of wire for an audio application this enables the audio signal to be transferred a great distance without worrying as much about interference this applies in spite of the fact that the levels can be very low like in a dynamic microphone where the entire signal is generated by the physical motion of the pickup element for this reason and for some impedance benefits this Arrangement is used all over the place like audio ethernet HDMI and the dnvt so we've now explained differential signaling where the signal is sent over two wires in opposite phase and the basic functionality of the Transformer we can now put this together to understand the dnvt communication interface if you recall from tangent time these Transformers accomplish isolation so the phone is receiving 48 volts DC on the receive side of the Transformer so both wires and the entire input winding of the received side is nominally 48 volts this power is siphoned off through the center tap of the Transformer which is basically exactly what it sounds it's a tap halfway through the winding since the current flows half this way and half that way in other words in opposite directions this should in theory create a no voltage on the output because they counteract each other so while the power for these phones comes over DC or the offset of the two wires from the ground potential the signal component comes over a much smaller AC signal or the difference between the two wires so we are getting 48 volts on average but we might find when sending a positive pulse that this lower wire gets 47.5 volts and the top wire gets 48.5 volts thus inducing a current from the higher potential to the lower potential this AC component passes through the Transformer to the output and induces a small change the dnvt uses a comparator circuit which basically says which wire is higher thus the difference signal on the wires I.E which wire is greater voltage gets translated into a single output of this comparator which can be connected to the brain box the brain does the opposite on the send side we have a center tap just like on the side where we got power but on in this case the center tap has a voltage applied to it by closing a switch to ground on either the top side or the bottom side we can get current to flow up or down this obviously only causes current on one half of the Transformer windings but the output effectively cares only about the net current so the fact that it moves only on one side or the other doesn't break anything this allows the computer to generate a differential signal a positive or A negative pulse through the Transformer and that's basically the electrical interface for these phones you've got your 48 volts up here for power you've got your ground down here for return and then your signals come across these wires and then are transmitted through these so knowing how these phones were implemented I went ahead and implemented my own version of the circuit the dmvt switch it's a very similar circuit we have first on we'll Trace through receive path again a gas discharge tube inductors I matched the component values of the dnvt as closely as I could so these are 33 micro Henry inductors these are 33 micro Henry inductors they're followed by capacitors and following that there is also a protective diode so I have a 8.2 volt protection diode after the inductors on my board as well that's to clean up any additional energy that is not snubbed by the gas discharge tube and a handle transients of less than the flashover point of the gas discharge tube that then goes into the Transformer which well on the phone it's pulling power out of the Transformer so on the switch it's injecting power into the Transformers so the secondary winding here the center of the transmit side is connected to 48 volts through a PTC fuse so a positive temperature coefficient poly fuse this component if you short it out a lot of heat will be dissipated in it and it will melt the pass element when the pass element melts it's a polymer material its resistance goes up very high so it will maintain itself in that open State without dissipating so much power as to damage itself even if it has 48 volts across it until the short is removed so you could take any of the outputs of the dnvt switch and short them indefinitely without damaging the switch and without cutting off service to any of the other lines so after that poly fuse it goes to the transmit Transformer on the receive side that is connected directly to the ground Return of the 48 volt input here which is the same as the ground of The Shield of the RJ45 Jacks and the protective Earth ground here which is uh connected through this exposed metal to the case you know the case is anodized so it's not a good connection but it is at least trying then instead of having the op amp on the receive side and the bipolar uh the push pull on the transmit side we're using integrated rs-422 line drivers which can handle the differential inputs directly and transmit signals with levels that are equivalent to what the phone expects many of you have probably heard of rs-232 the old IBM standard for serial ports this isn't used as much anymore but there are two other common serial Port standards called rs-485 which is still used a lot in industrial applications and rs-422 which isn't used as much anymore but has several advantages over the older rs-232 what Rob has done here is repurposed drivers typically used for rs-422 in these phones we could for example use discrete transistors to drive the Transformers and a comparator to receive just like the phone but by using rs-422 drivers we can get quite compact parts that combine both the send and receive functionality into a single chip and since rs-422 is already a differential protocol the transceiver supports differential signaling out of the box so this single component has allowed us to go from the isolation Transformer all the way to the 3.3 volt logic levels of the Raspberry Pi Pico if you connected rs-422 driver to Transformer ordinarily if it was uh just passing uart data it would not work the uart data does not have a balanced waveform that is equal amounts of time low as high the way that this is able to work with the dnvts is that all of the data is differential Manchester encoded differential Manchester and Manchester code in general ensures that there's a DC balance so the Transformer does not saturate as you have an equal amount of time high as you do low the other cool part about this design is the size difference obviously the unit on the left is one phone and the unit on the right can switch four phones part of this is the rs422 transceivers and part of this is the smaller Transformers but Rob has leveraged a lot of other improvements to make this work so beyond just the Transformers there's other components that are smaller uh this phone I'm sure is more rugged to indirect lightning strikes and other line transients than the dnbt switches but part of it is just that we're using surface mount technology that is inherently smaller you can also obviously see that all of the digital processing circuitry has been tremendously reduced in size the Raspberry Pi Pico can handle everything and more than what this PGA packaged Asic can do it could replace all of the audio circuitry the csvd encode and decode quite handily we could do analog output of a single phone from a Pico without too much work and for this flyback power supply do the limitations of through-hole Technology it is quite a bit larger than any modern uh flyback power supply would be so let's recap where we are now we have a board here that can supply power to the phones through the center Taps of the Transformers can take the differential signal and convert it to the standard 3.3 volt logic levels the board includes the relevant protection resettable thermal fuses and gas ditch charge tubes to protect against transients at this point we now have 3.3 volt logic levels presented to the Pico for each of the four phones we have a transmit pin which enables the Pico to send data to the phone and a receive pin that sends the decoded data from the phone to the Pico all routed through the magic of the PCB this means we can receive data from the phones and transmit back but uh what do we send it and how do we interpret the data that comes back from the phone so the next step here is doing a bit of research and a bit of testing and hopefully we can figure out the communication methods so I've been interested in doing something with these phones since I was in college and when I was in grad school I downloaded the IEEE paper that describes the operation of these dnbts and in that paper it talks about the control codes and mentions that there are 20 of them out of as it says 255 possibilities for 8-bit code words but because these are cyclically permutable there aren't 255 distinct code words for example a code word that has a single one in it doesn't matter where in the eight bits that one appears those are all cyclic permutations of the same code word um so to determine which code words were actually distinct I wrote a quick python script that ran through and removed any code word that if you had a bit shift was any other word and then used the minimal form for each of them at kind of a normalized form so that code word that has a single one we just call it code word one instead of code word uh you know 128. so with that I came out to 36 code words that's more than 20. so then I had the idea well what if they're trying to select for code words that have better error resistance code words that have greater Hamming distance so one of the fundamental characteristics of digital protocols is error tolerance if your AM radio is off by a bit you'll hear some static but with digital encoding if you ask your bank to pay 100 to your credit card company you don't want a bit error to accidentally cause the bank to send nine hundred dollars instead to handle this we use a variety of different methods for relatively simple example we can create a checksum where we sum all the numbers in a packet and then add an extra byte at the end to ensure they all sum to zero for example then we could discard an entire packet if we knew it was corrupt and asked for a re-transmit because the checksum wasn't valid but in this case we want to be a bit more simple we don't have many different messages to send and we are dealing with a bit stream in which there are no packets to correct Rob has inferred that the designers excluded codewords which can be created through a bit error of a single other code word so the code words are both cyclically distinct and have a minimum of two bit differences from other code words many of the military codecs and protocols from this era exhibit a high tolerance to bit error rate so for cvsd and LPC codecs they can both maintain intelligibility over a very high bit error rate say five percent whereas if you apply a similar bit error rate to the PCM codec from an audio CD you'll hear disastrous crackling so I ran through and removed all the codewords that are Hamming distance 1 from other code words and came out with 20 which is you know those are the code words that I used for all of my Brute Force tests and those were all the code words that mattered so after you've set up a means of interfacing with the phones to take this differential signal and bring it down to logic level you can hook that into a microcontroller and start decoding you can also decode the code visually which I'll do here on the oscilloscope so when I take this phone off hook the switch places it into dial mode and once it's in dial mode and you hold down a key it will transmit that code word until you let go of the key if you briefly tap a key the phone will transmit some minimum number of code words so that you can't just tap it so shortly it only transfers a single code word we'll start with the flash override here so I'll hold that down we've got this scope here in normal triggering mode so I'll be able to probe the pen and then quickly let go of it so that we capture that on the screen so holding that down we've got that on the screen I'm gonna go ahead and pull it off and there's our captured signal so with differential Manchester encoding each bit has kind of two half periods with a zero you'll have continuous phase to the half periods so if it's high to low it will then be high to low again when there is a one you get a phase change in the middle of that half period so if you're starting out high to low instead of going high to low again you'll have a low to high so it stays in that low state for another half period which is why it looks like you've got two frequencies going on here at the scope so as this is a cyclically permutable code there isn't really a stop or an end but what I've come up with for this is just to minimize the value of the word so we'll start here with zero zero we know that's going to be the minimum value of it so go ahead and adjust those cursors find our zero zero and we'll go ahead and walk through so when we have a low and then a high of this short duration pulse we know that that's a zero because there wasn't a phase change in the middle um so I'll walk through another two of those so one two those are both the short pulses we know that's another zero now we've got a long pulse so that is a transition from high to low then a phase change so then it is a transition low to high which results in that long pulse there low so that is a one then we have another long pulse this is also a one because we have a low to high transition phase change high to low transition so that's our first four bits we've got zero zero one one we can see another zero after that we've got the long so we know that's a one we've got the two shorts that's a zero and we've got the long which is another one and this is using the terminology where that long is a one the default terminology used by the Raspberry Pi Peak is actually the opposite of that it's arbitrary this is just how I learned it in like my old textbooks so that is our code word here and we can look it up in our table of code words this is actually how I developed this table of code words was just manually decoding them because this is before I had the Raspberry Pi Pico running with everything to do this automatically so that for Flash override is the right code word zero zero one one zero one oh one which comes out to 35 in a hex and 53 in decimal so here are the four generations of the dnbt switch uh starting from the development boards using an stm32 f-103 blue pill breakout board on through the Raspberry Pi Picos so this first breakout board uh using the stm32 made a couple mistakes with its design forgot to break out serial as I really haven't had success with the USB boot loaders for these guys so here's our obligatory ftdi chip breakout soldered on had some fun with the silk screen on this you can trace through the circuit starting with what pins this actually connects to in silk screen on the top as all the components are on the inside on the inside of the box we have everything that actually has to get soldered so this is a pretty fun construction method for quick prototypes that look decent this is a Hammond 1590 BBK box so it's the 1590b series painted black in this particular size I've used these in other projects before so for example here's a geiger counter that is based in these boxes this guy we got working but with the limitations of the stm32 exclusively in transmit or receive mode we didn't attempt to implement both transmit and receive simultaneously on the same as gm32 the second version of this uses the Raspberry Pi Pico after we discovered that it has great support for differential Manchester code I included the ability to have four phones run off of the same Pico the way that this is accomplished is phone one is wired in parallel with the outputs and inputs from the rs-422 transceiver as well as to the Pico and the rest of these phone Connections phone two three four those all connect to other boards so what you'll do is you have a master board that has the Pico populated and it handles the phone one and then on the rest of the boards so for example the board that handles phone 2 you'll connect to this connector and connect the other end of this connector to the phone one port of that other board and that other board won't have a Pico and it also doesn't need to have any power input since this connector here passes the 48 volt power 48 volt return ground 3.3 volts RX and TX so we did test this with one slave board to run two phones at the same time but only about a week later I moved on to the first version of the four port switch very little changed from the first version to the final version here the main changes were changing from these led displays which you know looked pretty cool but actually are more expensive than including an entire all LED display the OLED display also gives you a lot more flexibility in what you're able to display on the status of the phone but we had to give up a couple of these gpio pins here just due to space on the front panel but those are just for expansion and not really needed anyway the third difference is just layout of some of these transceiver components that moved around just as I needed to install this dip switch here the dip switch is just for setting uh configurations one of the switches will be used for 16 versus 32 and there's a number of other options that can be encoded in the dip switch that are software selectable and haven't been finalized at the moment the final difference between this version of the switch and what we have now is the change from these wired board connectors to RJ45 connectors this was done because the wire board connectors went out of stock immediately after I bought the assembled boards so for these versions of the boards we had jlc PCB assemble everything except for the Transformers and all the through-hole parts it was only later that I figured out how to get jlc PCB to assemble all the through-hole parts but they still can't assemble through whole Parts they don't have in stock so this particular Barrel Jack connector and the Transformers we still have to solder ourselves so for both of these versions of the switch the the four port switch I designed it around this black aluminum Extrusion box it comes together pretty nicely it's uh this feature in the box is called card rail so the um the card well the PCB slots into that rail so you have to be careful about your clearances there and that comes all the way back and we have a nice flush front plate here the front plates are also made out of pcbs because pcbs are very cheap um these from jlc in quantity are on the order of 50 cents each and they can be nicely made with beveled Corners with uh the silk screen showing the function of everything and with pretty precise cutouts for all of your connectors on the other side of the Box we also have one of these PCB end plates and all of that is dimensioned around the uh the screws on the box to let that just screw together using I believe M3 screws now brought you up to speed on how we went from initial reverse engineering all the way through to a unit that I can manufacture in my extra bedroom right here All That Remains now is to stretch our fingers and program some firmware so we can call between multiple phones connected to a single switch so stay tuned for episode three after that episode 4 we'll cover how we got USB working and how we can connect this to a computer and enable voice over IP and at some point we will also demo the world's craziest fax machine that has an adapter cable to plug into the data port on these dnpts if you have any thoughts or questions feel free to leave a comment below and I'll do my best to respond to them all if you'd like to stay involved we have a Discord channel for a chat about the project and upcoming ones and at long last we finally have inventory of both the RJ45 and terminal block switches as well as the ta-1042 adapter boards but unfortunately this right here is the entire remaining stock of Transformers we need for this board that are available in the US with the exception of the part scalpers so we have enough for about 50 more boards but thereafter the lead times are into 2024 so uh get them while they're hot I've also debuted the Knicks next extras Channel where we will feature some more informal updates on our projects and show off some demos as well as general tomfoolery I've got a video demo of the current firmware version uploaded there as well as the soldering tutorial for those that ordered kits with some other fun stuff planned till next time Nick with Nick snacks have a great day thank you [Music] thank you [Music]

Channel: Nick's Knacks

Views: 67,484

Keywords:

Id: FZ7V1lvCrfY

Length: 42min 5sec (2525 seconds)

Published: Wed Jun 14 2023