10 console multi-cab - need help w controller switching

[rCadeGaming]

DISCLAIMER: The ideas presented here are currently in the prototype design stage; they are incomplete and untested. It is possible that this could fry every console connected to it, electrocute you, and burn your house down. DON'T TRY THIS AT HOME.

In my custom cabinet, a MAME PC will be inside, and several consoles will be on shelves next to it. Basically, I want all of them to be hooked up at the same time. I want to be able to select one to play with the flip of a switch, instead of changing around a bunch of connections. I have the audio/video side covered (I'll be using an XRGB-Mini and a few switches), the problem is switching the controller connections.

I've designed a circuit to accomplish this, but I'm hoping for some feedback as to whether it will work or not before I start building.

First, here is a "concept diagram." I made it to explain the project to a friend who's an electronics guru, but isn't familiar with video games. It's pretty oversimplified for anyone here, but it's a good starting point:

http://rcadegaming.com/images/concept_diagram.jpg

Here is the actual circuit:

http://rcadegaming.com/images/player_co ... _color.jpg

I didn't put the images directly in the post because they're very large.

Note that this circuit is for one player. It will have to be duplicated for the other player. The two circuits will be completely isolated from each other, except for both using (but not being connected in) the 4PDT switch (at the top left, marked as a "shared component"). The purpose of this switch is to connect the menu buttons to the consoles and lock them out from the PC, or vice versa.

At the top are the momentary contact switches representing the joystick and buttons. These are connected to all four controller PCB's, which all share power and ground at all times. Power, ground, and signal lines exit from the controller PCB's, and connect to their respective consoles, identified at the bottom.

On the right is a two deck, ten position rotary switch, which is used to select which console will be connected. At the bottom are groups of solid-state relays for each console, these connect the console's signal lines to its PCB when that console is selected.

Near the relays, the voltage line coming from each console splits into two lines. One (which I will still call the voltage line - V) runs straight to the top deck of the rotary switch; while the other (which I will call the control line - CON) runs through the console's relays in parallel before reaching the bottom deck of the rotary switch.

When is a console is selected with the rotary switch, the top deck connects that console's voltage line to the PCB's; while the bottom deck connects its control line to ground, powering its relays and connecting its signal lines to its PCB.

On the power line, connected between the rotary switch and the PCB's, there is another momentary contact switch. The idea is that the rotary switch will be underneath of a door which holds the momentary contact switch closed. To turn the rotary switch, the door must be opened, cutting power during switching.

Here is the information on the solid state-relays relays:
http://search.digikey.com/us/en/product ... ND/1212841

-

Anyone have any thoughts? Do you think this will work? Any potential problems or suggestions?

[matrigs]

though about using a kvm ?

i made something similar at home with my arcade stick using this box:

http://www.dealextreme.com/p/manual-4-p ... 151?item=5

for a vga dreamcast, xbox360, ps2 with ebay scaler and laptop.

the vga ports are connected normally to every console and tv. the controls are connected through the ps2 ports.
i need 11 lines routed (6 buttons, 4 for the joystick and ground) and both ps2 give you 12 together which is sufficient.

from my arcade stick i'm going directly into the kvm and then to a hacked gamepad of every console.

looks horrible and was a lot of work to do but hey - it works.

you could just use a bigger kvm or a few cascaded together.

[burgerkingdiamond]

seems a little complicated. Just get one of these

http://www.google.com/products/catalog? ... 3509427929#

some ribbon cable, and solder on DB-25 connectors and start wiring. It works great for me.

[matrigs]

wow mam before today i didn't even know about these.

fantastic - that'll manage two players with one switch. nice.

[rCadeGaming]

Not that I don't appreciate some comments, but the KVM switcher doesn't make sense because most of these consoles will not be using VGA.

Furthermore, I'm going to use at least 10 consoles. The DB25 is nice, but I would need array of at least 4 of them for that many systems, PER PLAYER.

I'm not going for the easiest way out during the build, I'm going for the most elegant final product.

Does anyone have any advice about the circuit I'm planning?

[viletim]

Does anyone have any advice about the circuit I'm planning?

Connecting the button inputs from each controller PCB in the way you have done might result in some unpredictable behavior or even damage. Some game controllers have voltage regulators inside them which drop the VCC for the microcontroller. The Sega Dreamcast controller is fed a +5v rail from the console but it drops this down to +3.3v for the micro. When you join it to a Megadrive controller running at +5v you know what might happen....

I think you would be better off completely isolating each controller except for the ground.

[rCadeGaming]

Villetim, I'm connecting the switching circuit's voltage line to the point on the Dreamcast controller PCB which normally connects directly to the controller cord, and the Dreamcast puts out 5 volts through its controller port. Whatever it does with it beyond that point is irrelevant, I'm giving it 5 volts where it expects it.

Sharing power and ground between the PCB's is not a problem, all of these PCB's run on 5 volts and all of the systems give it to them. Here is what I've measured personally from a few of my systems at the controller ports:

Nintendo - 5.0v
Super Nintendo - 5.06v
Sega Saturn - 5.04v
Playstation 3 (USB) - 4.94v
Sega Genesis - 5.03v
Sega Dreamcast - 5.04v
PC's (USB) - ~5v

This is done all the time with custom sticks. I made one that has an MC Cthulhu sharing power and ground with a Dreamcast PCB, and it's had no trouble for about two years running on all the systems the MC Cthulhu supports.

The PCB's sharing power is not a problem. What I'm more concerned with is that the circuit isolates the consoles themselves from each other, connecting only one console's power and signal lines at a time, and allowing that one console's signal lines to communicate freely.

I'm not an electronics expert, so I'm asking if the circuit presented will function that way, or if it will have problems. I appreciate that you took the time to respond, but the area you're talking about is the part that's already tested and proven to work. I'm more concerned with everything on the diagram below the PCB's which is all untested.

[viletim]

Villetim, I'm connecting the switching circuit's voltage line to the point on the Dreamcast controller PCB which normally connects directly to the controller cord, and the Dreamcast puts out 5 volts through its controller port. Whatever it does with it beyond that point is irrelevant, I'm giving it 5 volts where it expects it.

I think you have misunderstood. I am referring to the nodes inside the magenta circles:

http://etim.net.au/temp/player_controls ... edited.png

You can measure them with a multimeter easily enough. Dreamcast, Wii, and possibly others (check the x360, PS3 and PC) will pull up these points to 3.3v, the other consoles will pull them up to 5v. The 1.7 volt difference will cause current to flow from one controller PCB into another through the button inputs. Perhaps you could fix this with a low leakage Schottky diode on each button input of each controller PCB.

The PCB's sharing power is not a problem. What I'm more concerned with is that the circuit isolates the consoles themselves from each other, connecting only one console's power and signal lines at a time, and allowing that one console's signal lines to communicate freely.

I'm not an electronics expert, so I'm asking if the circuit presented will function that way, or if it will have problems. I appreciate that you took the time to respond, but the area you're talking about is the part that's already tested and proven to work. I'm more concerned with everything on the diagram below the PCB's which is all untested.

I can see problems with that part too.

With this setup it is dangerous to switch while any one of the game consoles is powered. There is no mechanism prevent delay turning the next V+ line on before the last one has fully switched off. The MOSFET switches do not switch instantaneously and this should be accounted for somehow. Also the rotary switch must be of the break-before-make variety. I recommend cutting power (AC mains) to everything before switching unless this is a particular reason why you can't do this.

Secondly, there is the matter the capacitance of each of the MOSFETs. The datasheet specifies 25pF as typical, which is quite low actually, but the lack of a maximum is a bit disturbing... This capacitance will cause a little bit of signal to get through when the MOSFET is switched off. I only mention it because you are using a lot of these devices and you are dealing with some fast rising edges on USB, maple bus, etc. It probably won't be a concern, but it's something to watch out for.

Finally, you need a separate current limiting resistor for the indicator LED and the opto-MOSFET's LED. As they are, either the indicator will light or the MOSFET will switch on, but not both because one will have a higher forward voltage than the other.

One last thing, for the sake of saving a bit of bandwidth, please use the PNG format for line art. JPEG is for photos.

[rCadeGaming]

First of all, I apologize for the delay in responding. Please don't take it for a lack of appreciation, the kind of things you've brought up are exactly why I made this thread. Your comments are HIGHLY appreciated. I did see your post the night of, and have been thinking about it, but hadn't been able to write out a full response yet because of a crazy work week and personal stuff. Anyhow...

I am referring to the nodes inside the magenta circles:

http://etim.net.au/temp/player_controls ... edited.png

You can measure them with a multimeter easily enough. Dreamcast, Wii, and possibly others (check the x360, PS3 and PC) will pull up these points to 3.3v, the other consoles will pull them up to 5v. The 1.7 volt difference will cause current to flow from one controller PCB into another through the button inputs.

I'll try to get out a multi-meter this weekend and take readings for the different systems at this point. What kind of problems could this cause? Not noticeable, dropped inputs, fried hardware?

With this setup it is dangerous to switch while any one of the game consoles is powered. There is no mechanism prevent delay turning the next V+ line on before the last one has fully switched off. The MOSFET switches do not switch instantaneously and this should be accounted for somehow. Also the rotary switch must be of the break-before-make variety. I recommend cutting power (AC mains) to everything before switching unless this is a particular reason why you can't do this.

Maybe you didn't notice this part:

On the power line, connected between the rotary switch and the PCB's, there is another momentary contact switch. The idea is that the rotary switch will be underneath of a door which holds the momentary contact switch closed. To turn the rotary switch, the door must be opened, cutting power during switching.

I'm realizing now that there needs to be a cutoff switch for the CON lines as well the V lines, but wouldn't that do the trick?

Finally, you need a separate current limiting resistor for the indicator LED and the opto-MOSFET's LED. As they are, either the indicator will light or the MOSFET will switch on, but not both because one will have a higher forward voltage than the other.

I thought I had read that the relays had the same forward voltage drop as a normal LED, but either way I will be removing the indicator LEDs entirely. They aren't necessary if there is a label around the rotary switch, and they wouldn't light up while the door over the rotary is open anyway.

Secondly, there is the matter the capacitance of each of the MOSFETs. The datasheet specifies 25pF as typical, which is quite low actually, but the lack of a maximum is a bit disturbing... This capacitance will cause a little bit of signal to get through when the MOSFET is switched off. I only mention it because you are using a lot of these devices and you are dealing with some fast rising edges on USB, maple bus, etc. It probably won't be a concern, but it's something to watch out for.

Not sure what to do about this, but once the design is finalized, I'll be building it in prototypes, adding one system at a time to check for problems. I'm planning to start with just the Super Nintendo and Genesis, as they will test the use of 2 PCB's, and they're cheap if they get fried.

One last thing, for the sake of saving a bit of bandwidth, please use the PNG format for line art. JPEG is for photos.

I will fix this when I upload the next corrected diagram. At least I converted from the original BMP which was about 4 times the size of the JPEG, haha.

Thank you.

[iammud]

If you're going to do this, do it right and go through a frontend.

[viletim]

I'll try to get out a multi-meter this weekend and take readings for the different systems at this point. What kind of problems could this cause? Not noticeable, dropped inputs, fried hardware?

Something between no problem at all and a broken controller board. Generally speaking, it's just bad practice to raise the power supply of a device beyond its normal operating voltage. For the sake of a packet of diodes and some time, it's not really worthwhile finding out.

On the power line, connected between the rotary switch and the PCB's, there is another momentary contact switch. The idea is that the rotary switch will be underneath of a door which holds the momentary contact switch closed. To turn the rotary switch, the door must be opened, cutting power during switching.

Does this mean it is possible to an unpowered controller connected to a powered console?

I'm realizing now that there needs to be a cutoff switch for the CON lines as well the V lines, but wouldn't that do the trick?

Yes, that sounds like a good plan.

I thought I had read that the relays had the same forward voltage drop as a normal LED, but either way I will be removing the indicator LEDs entirely. They aren't necessary if there is a label around the rotary switch, and they wouldn't light up while the door over the rotary is open anyway.

No, the voltage drop is lower than a visible LED. Even it was similar, if you place two diodes in parallel then one will always have a lower voltage drop than the other, even if only by a tiny amount. This one will get all the current and the other will get none.

I think you should leave the LEDs in, just add an extra resistor. They will be very helpful to troubleshoot if things don't work as expected.

Not sure what to do about this, but once the design is finalized, I'll be building it in prototypes, adding one system at a time to check for problems. I'm planning to start with just the Super Nintendo and Genesis, as they will test the use of 2 PCB's, and they're cheap if they get fried.

Perhaps try to build a simple switch between two USB devices out of these opto-MOSFET 'relays'. Unless you are completely hopeless the a failure should not result in any damaged equipment.

[iammud]

*Edited - Bad theory.

[rCadeGaming]

Something between no problem at all and a broken controller board. Generally speaking, it's just bad practice to raise the power supply of a device beyond its normal operating voltage. For the sake of a packet of diodes and some time, it's not really worthwhile finding out.

I just tested the Genesis PCB I'll be using. It reads 5.07v on the button lines, same as the input voltage. Same thing with a Cthulhu hooked to a PC, 4.57v on both.

I'm assuming that the MC Cthulhu won't do anything different when connected to other consoles. The original controllers might have stepped down the voltage for the button lines, but there's no reason for the Cthulhu would have to, and it doesn't seemed to be wired for it.

I can't test DC or 360 at the moment. The only DC PCB's of the type I'm using are already wired together with MC Cthulhu's in my custom sticks (obviously I can't get an accurate reading there), and I don't have any 360 PCB's on hand.

So far the trend seems that the PCB's simply use their full voltage for the button lines, meaning there's no problem necessitating the diodes. I will continue to watch for this though, and be sure to test this with all of the systems when I'm able.

Does this mean it is possible to an unpowered controller connected to a powered console?

Not sure what you mean here. The "controllers" here are the PCB's, and all four are powered if any console is on.

I think you should leave the LEDs in, just add an extra resistor. They will be very helpful to troubleshoot if things don't work as expected.

Unfortunately, the LED's make no sense with the final design. It does indeed seem necessary to cut both the V and CON lines during switching, and this means that the LEDs will not be illuminated during switching, rendering them useless. No point in a light that's supposed to indicate what you're switching to, that doesn't work until you're done switching and close the door.

It could be helpful for troubleshooting, but it's another thing to troubleshoot in itself, and I don't want it in the final design.

I think you've approached the PCB all wrong. There's no reason to switch anything aside from the power from the console to the Cthulhu/game controller board. When the controller's voltage is disconnected, it's basically the same as being unplugged from the system. The Cthulhu has inputs for each controller port, so sharing and switching those is totally unneeded. The relays on the button I/Os are totally unneeded, too, since you can easily run all the button wires to the same inputs on the Cthulhu and controller boards (daisy chain 'em). Unless you're just being lazy with stripboard, you're wasting a ton of money on relays you can do without.

I don't think you understand what's going on in the diagram. All the button lines are "daisy chained" to all the PCB's. The relays are on the signal lines in the controller cords, which is very necessary.

To switch controls I'm going to use an Arduino to send the appropriately leveled voltage to that controller board's power input, leaving all the controllers hooked up, but omitting their power from the console.

You can't just leaved all of the signal lines connected to all the consoles all the time. If you share button lines between the controller PCB's, you have to share power and ground as well. All of the PCB's will be powered no matter which console is selected, and that power will be leaking out through the signal lines to all the consoles which aren't supposed to be selected.

Aren't you running an arcade setup? You're going to want 15 or 31khz, which is VGA. Nonetheless, practically any analog video signal can be sent through VGA.

KVMs, serial, vga, parallel, pretty much any type of switch box makes sense for practically everything, since they're essentially 9-25 channel relay switches. For example, a 4 port VGA switch is basically a 15p4t relay switch. Add USB and it becomes a 19p4t, or PS/2 for 21p4t. Parallel only would be 25p4t. Then consider the fact that they make some of these with 10 outputs. Where else are you going to find a 25p10t switch?

You can also use a KVM to swap the a/v and controls. For controls just hack the controller's voltage line into the KVM's PS/2, DB-9, USB, or even headphone jack lines, using a new channel for each system. Then you use another hacked cable to complete the circuit to that system, only using the appropriate channel. Since you're running an individual channel for each controller's power line, there's no chance of a controller getting the wrong voltage from a bad switch.

Some of the consoles I'm using use HDMI, and I'll be using an XRGB-mini which doesn't even have VGA input. You can transmit any analog signal over 15 pins, but it won't be RGBHV unless you transcode it. These 10 port KVM's might be expecting RGBHV.

Do you realize that most KVM's with a lot of ports aren't using a straight connection through a physical switch? They're using microchips to handle video switching, as well as to emulate USB and PS/2 communication, meaning they won't work for controller connections in other formats either.

I know you can get smaller switches, 2 maybe 4 ports, that are straight rotary switches, but if you've found a large one that's straight mechanical connections, please share the link.

[iammud]

You can't just leaved all of the signal lines connected to all the consoles all the time. If you share button lines between the controller PCB's, you have to share power and ground as well. All of the PCB's will be powered no matter which console is selected, and that power will be leaking out through the signal lines to all the consoles which aren't supposed to be selected.

Ah yes, I see now.

I failed to realize that it'll still power the other boards. My setup is still in variable land as far as controllers go. I'm working on getting the frontend scripted to remotely launch games, and change video inputs (I have all the scripts, I just need to create themes for all the games I want). Controllers are the last on my list. Your setup might be my answer.

While you're right, no single switch is going to do exactly what you want for 10 systems. Will matrix switches not work with power and ground? Couldn't you use a matrix switch IC to do some of your switching?

Some of the consoles I'm using use HDMI, and I'll be using an XRGB-mini which doesn't even have VGA input. You can transmit any analog signal over 15 pins, but it won't be RGBHV unless you transcode it. These 10 port KVM's might be expecting RGBHV.

Those same consoles that use HDMI will also work with RGBHV. At least for video and sound, a VGA matrix switch would work perfect. What are you using to switch between the video? HDMI matrix switcher?

Also, upscaling from composite/s-video? Why not upscale from the consoles' native resolutions through RGB? Super easy for most of those systems. If you're willing to put all that effort into the controls, why not the video? *shakes fist*

Don't you think this setup screams for an XP29?

[rCadeGaming]

While you're right, no single switch is going to do exactly what you want for 10 systems. Will matrix switches not work with power and ground? Couldn't you use a matrix switch IC to do some of your switching?

Ground doesn't need to be switched, it is shared by all the systems all the time. For power, its best to have the complete mechanical severance of a mechanical switch. For signal lines, the optical relays I'm currently planning to use are great, because of their incredibly low on-state resistance.

You may not have seen the earlier versions of my circuit. It used 4066 chips to switch signals, which would be cheaper but brought several complications; one of the main ones being that they require power to the VCC pin and ground on the CON pins just to maintain an off state.

Those same consoles that use HDMI will also work with RGBHV. At least for video and sound, a VGA matrix switch would work perfect. What are you using to switch between the video? HDMI matrix switcher?

Also, upscaling from composite/s-video? Why not upscale from the consoles' native resolutions through RGB? Super easy for most of those systems. If you're willing to put all that effort into the controls, why not the video? *shakes fist*

Who ever said I was going to f*** up the video? Looks like I need to explain my video setup.

TV: LG 32LD450 LCD TV - only 16ms input lag in game mode, more info here: http://shoryuken.com/forum/index.php?th ... se.145141/

Switching/Scaling/Scanlines: XRGB-Mini - more info here: http://retrogaming.hazard-city.de/ (Fudoh's site)

The consoles will use these video connections:

Composite - Nintendo

RGB SCART - PC Engine, Sega Genesis, Super Nintendo, Sega Saturn

Component - Nintendo Wii

HDMI - Playstation 3, XBox 360

VGA - PC, Sega Dreamcast

First, you're probably thinking, why ruin it with an LCD TV? It's simply the best compromise to run all of these systems. A 15kHz CRT is obviously best for 15kHz sources, but Street Fighter IV is crap in 480i. A 15/24/31kHz tri-sync would be best for 480p stuff like Dreamcast and that era, but the scanlines aren't as bold as I like for 15kHz stuff, and its still not great for modern consoles. An LCD is best for new consoles, and with proper upscaling and scanlines and a bit of reduced sharpness, they can actually look pretty good for 15kHz sources. They may never look as good as a real 15kHz CRT, but it's close.

Another big factor is rotation. I'm working on a bracket that will allow the TV to physically rotate for vertical games in seconds with ease. Try doing that with a heavy CRT. It has been done before, but a 32" 16:9 LCD has a 4:3 letterboxed display area equivalent to 27". On-the-fly rotation of a CRT this big would require an unsightly cabinet of immense size.

I'm already thinking about making cabinets dedicated for specific purposes, like one with a vertically oriented CRT for 15kHz shmups only, but for my first cabinet I want it to play everything. After that, I'll decide if I still think its worth it to branch out into specialized builds.

An XRGB-Mini is the best way to get proper upscaling of the older consoles, with scanlines and low lag. It will need a few purely mechanical switches in front of it though. For instance, I'm using four RGB SCART sources but I think it only has one input for that. I'll run those through a four-way DB-9 switch.

The video formats used are the best possible connections for these systems, with the exception the Nintendo. It can output RGB, but it takes a lot of modding.

Using RGB does not mean a difference in resolution. Whether you output RF, composite, s-video, or RGB SCART from your Super Nintendo, it will always output the game's native resolution. The difference in RGB is that it provides separate lines for the red, green, and blue picture signals, as well as another line for composite sync. Composite compresses all three color spaces, as well as composite sync, into one line. Due to the loss in color quality, color bleeding can make the picture look blurrier, but it is still the same resolution.

Also note that the RGB SCART signal from old consoles is not the same as VGA. Both use RGB, but the consoles' SCART uses composite sync, whereas VGA uses separate horizontal and vertical sync (RGBS vs. RGBHV). It also lacks the monitor ID, bidirectional data, and data clock lines of VGA.

The 360 supports VGA output, but the PS3 does not. You would have to either convert HDMI, which may add lag, or transcode component. There is no point in using VGA or component for either system, when they can output HDMI and the XRGB-Mini accepts it. You have to realize that although all three can do 1080p, VGA or component only looks as good as HDMI when its working perfectly. Digital signals aren't susceptible to the things like noise and ground loops that can show up with VGA and component.

The PC on the other hand, will rarely use 1080p for gameplay, probably just for the Front-End and a very short list of games. It will be connected straight to the TV's VGA port through an SLG-3000. This will allow the Front-End to control the SLG through a microcontroller, turning scanlines on only when a 15kHz game is loaded up.

As for Dreamcast, I'm still undecided whether to have it share the TV's VGA port on a mechanical VGA switch, or run it through a VGA to component transcoder and run it through the XRGB-mini.

Don't you think this setup screams for an XP29?

A what?

[iammud]

I've been brainstorming this quite a bit for my setup.

Instead of using a pad hack for Genesis I'm using a Playstation->Genesis adapter. I also have PS->N64 (these are rare, but there's also Gamecube->N64), PS->Arcade, PS->Dreamcast (for VMU), and I'm using the Imp circuit to tie in a 360 PCB.

Now that I've got everything simplified to just the Cthulhu pcb, all I have to do is switch the ABCDEFV channels. To do so, my buddy is building me a 7 channel mux board. Six of the multiplexers are used for the signal channels, and the seventh is used to switch relays on the +5v channels; allowing for 16 consoles. Since 16 consoles isn't enough (heh), we're hacking multi-taps with relays to disconnect the controller ports, allowing for even more consoles. I can also use KVMs to combine all the USB capable systems to a single channel (this is how I'm currently using the Cthulhu).

For the monitors I'm using dual NEC MultiSync XM29s, and soon a ViewSonic CD4230. I'm also upgrading the video card to an ATI HD4670 that has VGA, DVI, & HDMI, and supports the Calamity driver. This way I can have a tri-desktop display on the XM29s and the CD4230, allowing for 4:3 3:4 & 16:9 frontends.
I'm then using Belkin Omni-Link KMVs to share all the consoles to one VGA input, which is shared using the NEC's RGBHV throughput ports. This is my current gaming setup, sans the 16:9 display. Currently waiting for the new Hyperspin to come out, which will support all those aspect ratios. Once that's finished I'll upscale the 15khz consoles so everything on the CD4230 is at least 31khz.

Finally, the PC frontends aren't just used to execute roms on emulators. I've been very particular about purchasing things that can be controlled from a computer, so the frontends are able to choose which console they want to play. The mux breakout and multi-taps will be Ardunio controlled; the NEC and ViewSonic have RS232 remote support; even my power strips are RS232 controlled. Even though the KVMs don't have RS232, they have hotkeys which can be controlled from RS232->ASCii.
Then I have methods for Xbox 1 & 360 to execute games over a network (execute, not transfer); some Naomis can flash roms over a network (transfer then execute); NES, SNES, Genesis, N64, and [soon] TG16 all have ways to flash & execute roms to carts over USB; PS3s, unmodified 360s, and various other consoles can be macro controlled with this nifty little RS232 to USB ASCii device. Plus, a SC-500N1 to record all the consoles and PCBs. Ultimately, instead of changing games/systems manually, the frontend will handle everything, including recording.

Right now the setup is looking like the following:
KVM 1: stock NTSC-J 360, stock 40gb PS3
KVM 2: 3.55 60GB PS3, Trusty 1.4ghz Xbox, JTAG 360, Dreamcast w/ serial SD, Naomi Net Dimm
KVM 3 (31khz): Gamecube w/ GB Player & flash cart, N64 w/ Everdrive, SNES w/ Everdrive, Wii
KVM 4 (15khz): 2x Saturn, Genesis 1 (SMS, 32x, CD) w/ Everdrive, TurboDuo, 6 port jamma switch
NEC Y/C: Atari 800XL w/ MaxFlash, NES w/ PowerPak

I'm still acquiring certain VGA cables, switches, and flash carts, but I have majority of that list. Once the mux board is built, I'll be able to leave them all plugged in at the same time.

Who ever said I was going to f*** up the video?

At first I thought the mini only did CVBS & Y/C.

Also note that the RGB SCART signal from old consoles is not the same as VGA. Both use RGB, but the consoles' SCART uses composite sync, whereas VGA uses separate horizontal and vertical sync (RGBS vs. RGBHV). It also lacks the monitor ID, bidirectional data, and data clock lines of VGA.

I'm no expert on the difference between RGBHV and RGB, but lots of people have managed to get their old consoles to run on VGA monitors capable of 15khz. I assume this is what you use a sync splitter for. You can even hack SCART cables to RGBHV. Just in the last month I've acquired/made VGA cables for: Xbox 1, 360, PS3, Wii, Dreamcast, Gamecube/SNES/N64 (Nintendo A/V port 31khz style), Saturn, Genesis. I still need TG16, PC10 PPU, and the Atari RGB board.

Are you using SCART and HDMI A/V switches? Seems like you're going to be forced to use multiple switches, that can't be linked. I'm just saying that with RGBHV you can use the same switch box for everything.

The 360 supports VGA output, but the PS3 does not.

I take it you haven't seen Wii/PS3 VGA cables, yet.

Composite - Nintendo

Have you considered the PC10 PPU? Really not that much work. Especially considering you're making your own custom PCB for this project, that should be cake.

Don't you think this setup screams for an XP29?

A what?

The NEC MultiSync XP29 is the answer to upscaling and settling for an LCD. IMHO it is the absolute best monitor you can get for gaming. Capable of 15khz to 720p range resolutions (I think it even does 1080p). Since it's a MultiSync it doesn't step resolutions, so it will support all the resolutions that consoles and arcade pcbs demand.
It does RGBHV and I suspect RGBs (through VGA & BNC), YPbPr (combined with RGBHV BNC), and two S-Video ports. You can get composite by using S-Video->RCA plugs, and component from BNC->RCA. It probably has more input capabilities that I'm missing, too. Since it has its own I/O and RS232 remote port, practically anything can control it.

I've looked long and hard for the right monitor for a setup like this, and that's it. Unfortunately mine is the XM29 Plus model, which lacks component. Still a great monitor, though.

A 15kHz CRT is obviously best for 15kHz sources, but Street Fighter IV is crap in 480i.

15kHz on the XM29 is nice, so I imagine the XP29 would be the same. You can still play SF in HD and don't have to suffer when you drop the resolution for other games. But it's things like this that have driven me to go to a tri-monitor setup.