Building a Pi WiFi radio receiver controlled by vintage aircraft instruments. Some pointers, please ?

Question

Hello, new to this forum, I am seeking some help to get started in the right direction with this project.

My goal is to build a WiFi radio receiver with speakers operated by RP and controlled by vintage aircraft cockpit instruments, probably a Pi3+ (?). It needs to look nice and be original.

Here's an example of the radio part, with credits to Bob Rathbone

https://bobrathbone.com/raspberrypi/pi_internet_radio.html

There are exquisite instructions and manuals by him on how to build the radio part of this project so that is not my question.

Here's the twist and where I need to help getting started. I have 5 of these 80s analog, precise, unused instrument controllers. I would like to incorporate these to replace the stock rotary dials and control channel and volume. It would look cool, be original, but.... I don't know how to use the controller to feed inputs to the RP.

Idea1: In the back there is a 37 pinout D-bus terminal. The most elegant would be to buy some female 37 pin board converter and find a way to feed the input to the RP, perhaps via the GPIO of the RP. I don't have a mapping of each pin's purpose, the closest I could find is in the pic below.

Does anyone have an idea how to discover the use of each pin and how to turn that into a listening variable input into a program on the RP ?

Idea2: If all else fails, there may be a brute force method of removing the backplate and gluing some rotary knob controller directly onto the disc. Might function but not elegant.

Expanding the project, I could even add some electric dial powered instruments/gauges to show volume level setting by feeding the right input back to the instrument.

I would greatly appreciate if anyone could give me some general steps and ideas to follow especially for idea 1 above. THANKS!

michaelkellett · Answer

Do you have a schematic for the actual controllers you own ? The schematic in the picture doesn't look right - it shows a different front panel. Then you need to apply power to the controller and monitor what happens on the pins while you play with the controls (informed by the schematic). You'll need an interface to the Pi and it may be necessary to run the controller with Pi unfriendly voltages. First step is to get the right schematic (and manual if possible) and post it complete in fully legible form here. If you can't get a schematic it may be possible to reverse engineer the controller or maybe easier to rewire it . It always seems wrong to wreck these ancient parts by re-wiring - I'd go to some lengths to preserve it in original state. Where are you (if in the UK I might be interested in having a look at one of the controllers). MK

marco&#39;polo · Answer

Thank you, Michael, for the prompt response! I am in Michigan USA but not opposed to shipping you a unit to the UK if you're willing to help out. I confess that when you say " you need to apply power to the controller and monitor what happens on the pins while you play with the controls (informed by the schematic)" , I have no idea as to how to do this. Your comment is correct, the schematic looks like it's for a different unit, the King KFS 580B while mine is the King KFS 585B. However they seem to be sisters since they share the same maintenance manual so I would like to hope that important info like power source and voltage are the same (someone is selling this manual for 160$ which I find not reasonable and not in the budget of this project). A 580B unit for sale suggests that there are 5, 14, and 28 Volt option to power the back panel lighting, which would be really cool to keep. I will post shortly a more readable format table of the pin table. Does this provide more clues? Thanks, m. edited with dsub pin diagram for kfs 580b. Same file attached as xls.

phoenixcomm · Answer

jc2048 neat redraw in the original UAD. This is what you normally end up with. having BCD outputs help a lot. Most of the logic has one diode drop but some have two, so the expanders are out, unless each output was pulled up by its own resistor. From what I understand he wants multiple heads, that is why I suggested the mega, it handles all of the i/o on one board and they can talk fairly easily of CAN bus, or RS-485 but this you have to poll and I think it is a drag on a CPU, while CAN is not a polling system is bidirectional. Depending on how the Pi wants the frequency presented most likely in decimal or ASCI? ether way the BCD converted whatever and then added where the dial is a 7 you will end up with three values (int) High, medium and low. ie decimalfreq = (high * 1000) + (medium * 100) + (low * 10 ) or just put them in to an array freq and use itoa() function or toascii() singly.

michaelkellett · Answer

I saw the rip off artists on the web too ! I've asked a UK air service co for a freebie - probably won't get anywhere but if you don't ask ......... It's really hard to do without a manual, there must be loads out there. You could check if the wire colours on the D connector on the 585B are the same as the 580 - if they are it would give a bit of comfort. Where did you get the schematic in the picture ? What happens when you rotate the knobs ? Is it just three separate rotary switches each worked by one of the concentric knobs or is there some clever linking between them and the display ? MK

DAB · Answer

Interesting project, but very challenging. Just looking over the basic schematic you showed, I am pretty sure the control units are just the switch housing. These are not the actual radio sets. That actually works to your advantage. You can interface to the wires on the connectors to generate switch connections you can pass on to the RPi, using suitable level controls, and then use an SDR stick with the RPI to do the actual radio bits. Please be aware that this is a challenging project. It will take a fair amount of hardware and a lot of software knowledge to bring off. DAB

phoenixcomm · Answer

marco First, Let me assure you this is not rocket science! You have the pin-out, or you can generate one your self, DVM to ground. and then probe. The good news is your print. above looks ok, So. Take a deep breath. I believe your pinouts show the tuning frequencies. What you have is only a Radio Head, ie each head has its own black box, radio. So for your indicated frequency "820Khz pins 12 and 6 will be connected via ground. now with that said the BFO switch is SPST. Then the function switch test/ADF/ant/off is just a rotary switch. All of this can be handled by Arduino. Now you have to send you selected frequency to your radio (pi whatever). since this is a Fixed system your and your Radio Heading will not work. unless you just want to point to the radio station from where you are. I am working on a project Nexgen. I have a motto: 1) Do not harm to the equipment. 2) see rule 1, rules 3 -9 see rule 1. rule 10 ) for vacuum drivin stuff you will have to convert to small stepper! Another one is my fuel gauge there are no docs and everybody says property. Which leads me to rule 11) Abuse it if you must! https://community.element14.com/resized-image/__size/620x300/__key/communityserver-discussions-components-files/91/contentimage_5F00_205930.html NexGen Flight Simulator

phoenixcomm · Answer

BTW only the panel lamps run on 28vdc. The rest use TTL levels and use pull resistors or the internal ones. and here is a link to my https://community.element14.com/resized-image/__size/620x300/__key/communityserver-discussions-components-files/91/5460.contentimage_5F00_205930.html RMI page

phoenixcomm · Answer

marco Marco just one small note: What have are dumb radio HEADS. You must have a controller to decode the frequency and the position of the switches. This can be done with an Arduino Mega and its output can go to your PI. outer 1000 800 400 200 100 Center 40 20 10 Iner 1 2 4 8 1 2 3 4 5 6 7 8 9 0 7The easiest way to build a small spreadsheet and you are in the NYC area and you know that you can get a good fix on WABC 770KHz. Therefore when the outer band is set to 7 signals 400, 200, 100 with the inner band which is also tuned to 7 400, 200, 100 will be low via common to the ground and all the outputs pulled up. I hope I have helped. Cris

michaelkellett · Answer

Hello John, You mentioned it earlier but your enhanced schematic (amazing app that PencilCAD ! ) really brings the point home - the diodes mean that the switches won't pull very low. For correct operation of the MCP23017 used in most I2C expanders you need to get below 0.2Vcc (0.66V on 3.3V systems, 1V on 5V systems). The two diodes in series mean that 20k and 10k will pulls down to about 1.2V (if silicon diodes). You could still use the MCP23017 but the 10k and 20k lines would need some additional components for it to work. Alternatively you could use a micro (an Arduino would do, a 3.3V one will interface mores easily with the RPi) and use its built in ADC to measure the 10k and 20k signals. MK

phoenixcomm · Answer

BTW please look at my blogs about the Audio challenge The formulas are not really correct as they are for RADAR, not radio. The formulas in my next blog will have radio distance over the sea and that works for DME. Now saying that ADF is very different it's mostly in AM radio band 19Khz to 1.75Mhz so the stations have bigger power and the antennas are much taller ie ABC was in Lodi, Nj but has 50kw output. At one point they transmitted from the top of the Empire State Building (1,250' + 200') and World Trade Towers. therefore height and power matter.

phoenixcomm · Answer

One last BTW if you use multiple heads you will have to devise a switching mechanism of sorts. One way is to use I/O expanders (I 2 C) or use separate Arduino per head (best as each radio head requirements are a little different. One last thing Audio is a mess!! you can switch them with CMOS switches. I aircraft you have a radio selector box with isolation amps one per channel and them mixed together (that is the selected ones with there own volume control). Since you can't do this you will be limited to one control head at one time. so you will have to fake it. Pi audio output can go and switch them with CMOS switches. that way each radio volume will work. FINS

marco&#39;polo · Answer

Many thanks for your perspective DAB. What was I thinking when I thought I would post this in the beginner project section :-) Yes you are correct this device (the 585b controller) is only the UI to select the frequency. For those interested to know more on this, an ADF (automatic direction finder) is a flight navigation instrument that will point an arrow to the source of an emitting beam. These types of ADF systems have 3 parts : 1- Input frequency selector, where the pilot chose the frequency of the beam to be located 2- The actual beam receiver which is a big black box connected to an antenna on the fuselage 3- and an indicator that looks very much like this In fact if I am lucky enough to move this project ahead and look for more unnecessary complication, I might want to an indicator in to show something else associated to the WiFi radio. But that's one for the next step...For right now it seems I am over my head, but I'll persist with the help of the great contributors on this forum.

jc2048 · Answer

Thanks phoenixcomm I know you and Micheal have basically answered the question. I was just emphasising to Marco that the switches do the encoding, so if he's careful with how he arranges his IO, he can have the numbers by simply inverting the code that comes in. Here's how it looks if you redraw the two switch poles so that it's a bit easier to understand [I find the original a bit confusing, even though it's a very nice piece of draughting work]. That's for the 10s of kHz switch, but the same applies to the other two. I've shown the current flows in red for the case where '7' is selected. The extra diodes are necessary for that case because there are only two switch poles, but three lines need to be pulled down. The truth table you get if you work through all the switch positions is this that's negative-logic BCD. Invert all the levels, so that it's positive-true, and you're back to binary. I think I would have suggested IO expanders on a plug-in board on the back of each head unit, with power and I2C running along to the Pi, but Michael's suggestion of a Pi shield makes more sense [I didn't realise such things were available, though it's a fairly obvious thing to make and sell].

michaelkellett · Answer

Hello Marco & CH, You don't need to bother with an Arduino unless you want. You can use port extenders on your RPi. Here is an example: https://www.abelectronics.co.uk/p/54/io-pi-plus This is a 32 bit IO board, you might be OK with a 16 bit board but this is the first one I found. There are lots available, choose on that uses the MCP3017 chip and has its own software libraries. You can connect this to your switch box (CH calls them HEADS). If the schematic you have is anything like your switch box then there are some diodes in the box so the polarity of the common is important. You'll need pull up resistors on the inputs and to connect the frequency common to 0V. If you go this route the RPi can do all the work in Python. MK

phoenixcomm · Answer

marco Back to your idea #1. DO NOT CUT UP YOUR HEADS! I will direct you to one of my older blogs Audio Panels Ins and Outs cutting up your heads is a disaster!! Lots of Soldering for nothing! You can buy a DB-37F on eBay for 4-5 buck hey don't want to build cables buy a National Instruments Compact FieldPoint Cables, DB37M - DB37F, L-com E119932-J and cut the male end off, they are like 9 bucks. Me most of my connectors are circular Bendix connectors and then can cost more than $50 each. Personally I use old printer cables! They're really cheap! DB-25 <> Pertec Connector, so I cut off the Pertec Connector this gives me about 6' of cable between my instrument and its controller (Arduino) terminated with the DB-25F. So to terminate to the instrument or radio head I use the original connectors mate (solder-cup). #2 Your Pi should just do the heavy lifting. You can use Arduino Megas for tunning and then run CAN bus between them! Just remember that you will have to write a protocol (software) to translate the Radio Head to Arduino for decoding. And then put that into a format that the Pi can understand.

jc2048 · Answer

In case you're not clear about the schematic, the frequency switches encode negative-logic BCD [Binary Coded Decimal]. One slight disadvantage with the scheme that the original designer used is that, on the setting for '7', there are two diodes in series on a couple of the outputs, so the drop across them is going to be a fair bit more than where there is just one diode, and the corresponding low-level will then be higher. You might want to measure that voltage level with a meter and then take a view as to whether you were happy with it for whatever logic it was going to be connected to.