Vector TV

The thing that kicked off this project for me was this project I found on Hackaday. Lucy had designed almost the exact circuit I needed, which greatly simplified the work for me. I used their schematic as a start for the current feedback amplifier, and did my own design for the rest. I am inexperienced with analog circuit design, so not having to worry about the analog section is very freeing. The link to that project is here. The circuit that this project is based on is an amazing project by Jürgen Müller, linked here. This project provided the amplifier circuit and a great explanation of how it functions. 

After taking apart the 5 inch black and white CRT, I found a board that I could remove and replace with my own design. This board was most likely the radio receiver board, as this small TV doubled as an AM/FM radio. Using my calipers, I recreated the outline of the board in Onshape and converted it into a DXF file so I could import it as a board outline into EasyEDA. 

After rearranging all the components to fit onto the new board outline, I decided that I should make some changes to my schematic. My first change is that I was worried that the ESP32-C3 wasn't going to be powerful enough for all the analog math that I might require, as well as multiple functions at once. I decided that I would use the ESP32-S3 instead, as this version is dual core. This allows me to program 2 different functions to run simultaneously, sharing data with one another using FreeRTOS. This is a little complex, but I experimented with this approach in my light clock project, so I know a little about it. 

After more thinking about it, I decided that I would need to add another feature. I want this device to have 2 main functions: Playing back oscilloscope music, and emulating Vectrex/vector arcade games. In the original schematic, the speaker amplifier is connected directly to the inputs of the deflection amplifiers, playing back the audio sent to the deflection coils. This is great for oscilloscope music, but if I wanted to use this device for any games with separate audio, I would need to be able to switch the audio feed from the deflection amplifiers to a separate source of audio for the Vectex games. Using an integrated circuit called an analog multiplexer, I was able to design in a way to perform this function. I love experimenting with electronics like this, because it teaches me about components that I would never be able to learn about from classes. I would never have known that the perfect component existed for this if I didn't search through all the data sheets and figure out what all these things do. It's very rewarding :)

After I completed this design, I checked the cost for manufacturing. Unfortunately JLCPCB removed the ability to only have 2 boards assembled in economy assembly, so I will have to buy 5 assembled boards for around $160. I am currently in the process of increasing the cost effectiveness by removing unnecessary components and finding more affordable alternatives. My current area of research is whether I can use the ESP-32 built in touch sensitive pins as a capacitor-to-digital converter, so I can use the original AM radio tuner as an analog input device. This would save me money and complexity, as that potentiometer that I have on the board is an expensive and very unique part. Removing it would help save money, and I wouldn't have to redesign the gear that connects it to the TV knob. Once I decide I have the board finalized, I will buy them and begin the programming phase. I have some lovely code to begin with if I follow the Scopeman's code. 

After experimenting with the capacitive touch option built into the ESP32, I was able to consistently detect motion of the variable capacitor. This is great news, as it lets me save money on the PCB and also maintain the mechanical function of the tuning knob on the outside of the TV. This input will allow me to control the firmware, such as rotate rendered 3D objects, control volume, or maybe function as a game input. The possibilities are exciting, and I am glad it worked out this way. Here is the next iteration of the design:

In my never ending quest to make this PCB design more affordable, I actually added more features. I removed lots of parts, opting for manual soldering of the connectors. The most interesting addition is the snap-off sideboard, allowing remote connection of the USB, reset and boot buttons, and the external audio input. This board will connect with a simple soldered ribbon cable, saving lots of effort using breadboard wires to connect the 2 boards. I also added LED indicators for the negative and positive power supplies, so I can easily see any issues. When the board is powered over 5V, only the positive voltage will be present. This will disable the deflection driver, allowing only the ESP32 to be enabled. I did this mostly to protect the USB supply, as I expect this board to draw around 2A of power at times. Hopefully the power consumption of the CRT itself will go down when the deflection coils are removed, but there is a chance that the flyback transformer requires the deflection coil as a resonator to maintain function. In this case, I will need to supply about 4-5A 12V through USB-C. I am getting very close to ordering this board, but I want to make sure everything is checked well and perfect. I'm going over the datasheets for all the items to make sure all the components are correct. I already make many corrections to the power supply passives. 

It has been a bit since I've updated this log, and for good reason. Last update involved being sure that I had a finished product ready for ordering, so I did. That was an expensive mistake, as I had not realized that I had missed an incredibly subtle but important detail on one of the power supplies. A very subtle difference in my schematic crippled the entire project and I was not able to use micro soldering to fix it. I will explain the mistake I made in detail further on, but first I would like to explain my next steps after that lovely situation. I decided that I needed to double check EVERYTHING on my schematic, and I found more than one issue that needed addressing. First issue was that I didn't fully understand what my amplifier circuit was even doing. I spent hours studying the original inspiration for the circuit, and finally realized that the circuit that this one was based on was not powerful enough for my purposes. The circuit I was basing this on was for a small 90 degree neck CRT, mean that the beam angle only changes a few degrees, as opposed to my standard CRT, which has a beam deflection angle of 90 degrees. This means that it requires significantly higher voltage and current, and the circuit I was using to provide negative voltage was not enough. This led me to search all over for a specialized "high-current negative voltage supply". I was unsuccessful, but came across Texas instrument documentation on how to use a standard positive stepdown converter as a negative voltage supply. This allowed me to mirror the positive and negative supplies, and I also fully reverse-engineered the amplifier circuit to fully understand what it is doing. After all those changes, we have my new "final" design. I am hesitant to order it just yet, as I make small changes each time I see the circuit.

When I first began this project, I did not have a complete understanding of the stepdown converter ICs available on the market. I assumed that there was no way to use a standard converter as an inverter, which forced me to look for specific ICs designed for negative voltage. These ICs are expensive and do not provide enough current for my application. After finding an application document for one of the TI stepdown converters as a negative voltage regulator, I realized that I had been doing this project all wrong. I started from scratch with the power supply, using very simple circuits for the positive and negative voltage required. These stepdown converters are not only standard parts for JLCPCB, but they are able to provide much more current. 

After studying the amplifier circuit for a while, I was able to piece together a much clearer understanding of its function. The amplifier has a filter that provides some "pillow effect" to the signal, counteracting some distortion that can be present in this type of setup. Once this distortion is applied, the amplifier enters an operational amplifier set up for current feedback. This mean that the circuit provided current proportional to the input voltage, allowing the input voltage to linearly control the current in the deflection coils. Applying voltage directly to the coils does work, but because of the physics of inductors, the voltage will not be directly proportional to the actual deflection of the electron beam, causing lots of distortion in the final image.

As I am currently between jobs and unsure how current tariffs will affect this project, I am not able to complete this project just yet. The next steps for this project will be rewinding the yoke to reduce the inductance (for faster deflection), programming firmware, and finalizing circuit design. I still need to study the CRT circuit, as the G1 grid circuit behaves differently than expected. If I want to be able to control the beam intensity, I will need to understand how this circuit functions. Traditionally, the drive circuit uses negative voltage to deflect electrons and lower the brightness of the beam. After checking this, I discovered that this specific circuit does not seem to provide negative voltage, but rather positive voltage. I plan to use an oscilloscope to confirm this. Once this project is tested and my circuit is confirmed to work, I will redesign the board to be a more standard shape and opensource the project for other people to try this.