Project Log
Hello all ! This is my first project log posted here. Hope you'll like it.
The goal of this project is to create a SCART switch which accepts up to 8 inputs and can support composite video, RGB and S-video (plus audio of course). The switch will be wall mounted for it to have stability and easy access.
It will NOT support audio in both ways, and thus cannot be used in the opposite direction.
Details of the steps in following updates, and discussion about the log is in the forum, as it seems to be the rule here
Here is my current status, that I will update while I go forward.
Update 4 : gotcha, PCB puzzle!!! (2018/06/15)
Finally!
Took me a really long time, but I finally managed to finish the PCB layout for the whole stuff.
There were a lot of things to do :
I first started to use the autorouter provided by the software, but the results were so messy I couldn't follow any line easily. This is why I routed all by hand, and it gave me several sessions of troubles, especially for the last part. The deal was to route 8 scart connectors to 7 analog multiplexer ICs, because each scart connector had to connect one pin to every IC, so all the signal lines had to cross each other!
I tried several methods, but finally got to this one : one layer per direction. Top layer (in red) for vertical lines, and Bottom layer (in blue) for horizontal lines (with a few exceptions on places there would be only one layer, to limit the number of vias on the PCB):
As you can see this looks like a bit messy, but I tried to do this meticulously, starting from the middle IC, and going to the sides gradually. This way I had the ability to use the same horizontal lines for 2 separated signals, when they were both on the opposite side. This way it took me less space between the scart connectors and the ICs, so I could optimize the board size (and save costs for the manufacture of it)
The rest had to be slowly made, feature after feature, but I finally got quite well. Several components moved to different places to limit the global number of crossing signals I had to do.
At the end, I am very proud of this, especially on the fact that I managed to do the whole relay control components on a single layer Many things may and will change so there is still some work to do, but the PCB should go to manufacture very soon.
Thanks for reading, and don't forget to check out the gallery for great PCB pr0n
(samples follow)
Update 3 : not much to say (2018/04/30)
A very quick update here, just to stay I'm still on it. PCB layout is veeeeeeeery long to do, as I have to wire a large number of things. But at least the components are in place. Also, the ground plane is here, and some basic stuff is wired (the power for the LEDs, and the switch monostable latches.
I'm still waiting for the components to start prototyping, they should arrive very soon.
Update 2 : circuit design, continuation & end (2018/04/02)
(Click on this thumb to have a full scale image of the schematics)
This is done, I finally added the missing items in the update 1.
The bistable relays + their control circuitry is in place. And oh god that makes a lot of components ! I wonder if I wouldn't better have used IGT transistors... anyway I'll keep the relay design for the reasons mentioned in update 1.
But if you read me, have a lot of consideration for place and cost, and are struggling with the same choice, then prefer the IGTs when you can
Also, for the relay thing, I had to add a demux chip (74HC238), so I could change my 3 line binary signal in 8 selection signals for each of the relays. Hopefully there is no need for decoupling capacitors for it, since the relays state won't change really often (only when switching the SCART source)
Speaking of them, the decoupling capacitors (1 polarized 0.1µF electrolytic, and 1 non-polarized 100pF ceramic) for each 4051 chip have been added, along with the 4 leds and their control circuit (basically 3 transistors).
By the way, I've read over the internet that for decoupling we better use tantalum capacitors than electrolytic ones, because they are smaller and last longer. But I still preferred to use catalytic capacitors, since tantalum capacitors have 2 major drawbacks for me :
for the PCB design... well the really impressive number of components of the finished schematics really do not help me at all... I'm on it but I'll need way more time than expected. News in the next update
Update 1 : circuit design (2018/03/29)
Schematics creation.
The first thing I did before starting is to break open a commercial scart switch, to see how it can work. And... surprise. No electronics at all, the switch is only done mechanically using long switches with a dozen of contacts each. Seems like this does not disturb much the signal using such switches. Anyway, this is a good thing : no need to do complex circuitry to decode & re-encode video signals. Would be a real pain to do.
So I had to find a way to switch electronically between each source, without having to bother much about the signal. And I finally came to use the 74HC4051 IC series, which are 8 to 1 analog multiplexer / demultiplexers. To control these 7 4051 (one for each signal, all grounds are linked), I need 3 bits, which happens to be perfect because I wanted to have my switch work with 3 lever switches (2-state switches coupled with 555s to avoid the bouncing signal, so I can have up to 8 combinations). The last main issue for me concerned the "function switching" signal of the SCART specification, which is a linear voltage signal used to determine the aspect ratio of the image sent to the TV. This voltage can go up to 12V, so this can not be handled by 4051s which accept up to 6V only.
I thought about many solutions for this, including weird and stupidly complex methods (like reducing the voltage and multiplying it again after the output of the 4051...), and finally chose to use low signal relays. I also had the solution of IGT MOSFETs, but I preferred the relays because:
So, next step :
Thanks for reading, and see you for the next update !
The goal of this project is to create a SCART switch which accepts up to 8 inputs and can support composite video, RGB and S-video (plus audio of course). The switch will be wall mounted for it to have stability and easy access.
It will NOT support audio in both ways, and thus cannot be used in the opposite direction.
Details of the steps in following updates, and discussion about the log is in the forum, as it seems to be the rule here
Here is my current status, that I will update while I go forward.
- Schematics : 100%, waiting for prototyping for validation
- PCB Layout : 100%!!!
- Case design : 0% (But I have my ideas)
- Prototyping : 5% (parts ordered)
- PCB manufacturing : 0% of course
- Soldering : 0%
- Construction of case : 0%
- Test & Validation : 0%
Update 4 : gotcha, PCB puzzle!!! (2018/06/15)
Finally!
Took me a really long time, but I finally managed to finish the PCB layout for the whole stuff.
There were a lot of things to do :
- Power and ground
- Wiring the 555s with the switches
- the relays circuitry (all on the top layer \o/ )
- the 3 to 8 demux to control the relays
- The output scart connector
- And the hardest one, by far : the 8 input connectors toward the 7 analog multiplexers
I first started to use the autorouter provided by the software, but the results were so messy I couldn't follow any line easily. This is why I routed all by hand, and it gave me several sessions of troubles, especially for the last part. The deal was to route 8 scart connectors to 7 analog multiplexer ICs, because each scart connector had to connect one pin to every IC, so all the signal lines had to cross each other!
I tried several methods, but finally got to this one : one layer per direction. Top layer (in red) for vertical lines, and Bottom layer (in blue) for horizontal lines (with a few exceptions on places there would be only one layer, to limit the number of vias on the PCB):
As you can see this looks like a bit messy, but I tried to do this meticulously, starting from the middle IC, and going to the sides gradually. This way I had the ability to use the same horizontal lines for 2 separated signals, when they were both on the opposite side. This way it took me less space between the scart connectors and the ICs, so I could optimize the board size (and save costs for the manufacture of it)
The rest had to be slowly made, feature after feature, but I finally got quite well. Several components moved to different places to limit the global number of crossing signals I had to do.
At the end, I am very proud of this, especially on the fact that I managed to do the whole relay control components on a single layer
Many things may and will change so there is still some work to do, but the PCB should go to manufacture very soon.Thanks for reading, and don't forget to check out the gallery for great PCB pr0n
(samples follow)
Update 3 : not much to say (2018/04/30)
A very quick update here, just to stay I'm still on it. PCB layout is veeeeeeeery long to do, as I have to wire a large number of things. But at least the components are in place. Also, the ground plane is here, and some basic stuff is wired (the power for the LEDs, and the switch monostable latches.
I'm still waiting for the components to start prototyping, they should arrive very soon.
Update 2 : circuit design, continuation & end (2018/04/02)
(Click on this thumb to have a full scale image of the schematics)
This is done, I finally added the missing items in the update 1.
The bistable relays + their control circuitry is in place. And oh god that makes a lot of components ! I wonder if I wouldn't better have used IGT transistors... anyway I'll keep the relay design for the reasons mentioned in update 1.
But if you read me, have a lot of consideration for place and cost, and are struggling with the same choice, then prefer the IGTs when you can
Also, for the relay thing, I had to add a demux chip (74HC238), so I could change my 3 line binary signal in 8 selection signals for each of the relays. Hopefully there is no need for decoupling capacitors for it, since the relays state won't change really often (only when switching the SCART source)
Speaking of them, the decoupling capacitors (1 polarized 0.1µF electrolytic, and 1 non-polarized 100pF ceramic) for each 4051 chip have been added, along with the 4 leds and their control circuit (basically 3 transistors).
By the way, I've read over the internet that for decoupling we better use tantalum capacitors than electrolytic ones, because they are smaller and last longer. But I still preferred to use catalytic capacitors, since tantalum capacitors have 2 major drawbacks for me :
- When they get destroyed for whatever reason (including age), they become completely conductive, and can cause serious damage to the circuit, whereas electrolytic capacitors become isolated when destroyed
- Still when getting destroyed, they have a tendency to burn instead of just breaking open and spitting smoke, which you guess can lead to more serious damage...
for the PCB design... well the really impressive number of components of the finished schematics really do not help me at all... I'm on it but I'll need way more time than expected. News in the next update
Update 1 : circuit design (2018/03/29)
Schematics creation.
The first thing I did before starting is to break open a commercial scart switch, to see how it can work. And... surprise. No electronics at all, the switch is only done mechanically using long switches with a dozen of contacts each. Seems like this does not disturb much the signal using such switches. Anyway, this is a good thing : no need to do complex circuitry to decode & re-encode video signals. Would be a real pain to do.
So I had to find a way to switch electronically between each source, without having to bother much about the signal. And I finally came to use the 74HC4051 IC series, which are 8 to 1 analog multiplexer / demultiplexers. To control these 7 4051 (one for each signal, all grounds are linked), I need 3 bits, which happens to be perfect because I wanted to have my switch work with 3 lever switches (2-state switches coupled with 555s to avoid the bouncing signal, so I can have up to 8 combinations). The last main issue for me concerned the "function switching" signal of the SCART specification, which is a linear voltage signal used to determine the aspect ratio of the image sent to the TV. This voltage can go up to 12V, so this can not be handled by 4051s which accept up to 6V only.
I thought about many solutions for this, including weird and stupidly complex methods (like reducing the voltage and multiplying it again after the output of the 4051...), and finally chose to use low signal relays. I also had the solution of IGT MOSFETs, but I preferred the relays because:
- I don't need to be very fast when I switch from one input to the other
- This is not bound to happen regularly
- The low signal relays are really small now (1x0.6 cm in size for the one I will probably use)
- I can use bistable relays, so I can save energy (even if the gain is ridiculous, I agree)
- This gives me a reason to work on the usage of relays, and also on capacitor charging / discharging circuits using transistors
So, next step :
- Create the eagle library for the low signal relay, as I didn't find one on the Internet
- Finish the schematics by adding my 8 relays with the circuitry I need on them
- A few tweaks (like decoupling capacitors on the +5V lines of the 4051s, the famous LEDs...)
- Advancement on the PCB layout
- First steps of prototyping, if I receive the components on time
Thanks for reading, and see you for the next update !
