Was just tinkering with it some more (it’s a good excuse to slack off from my real work ;))
So I hooked up the scaler with T-SLG to my NEC. To answer your question: when the T-SLG is set to the ‘OFF’ position, I still experience the ‘image, blank, image, blank, image for 2 seconds, blank; repeat’ in the test.
HOWEVER (this applies to the NEC)!
With the T-SLG set to ‘WIDE’ mode with the scaler running at 640x480, I DO GET AN IMAGE (albeit one with wide scanlines). If I change to any other resolution, I get no scanlines, but **I STILL GET AN IMAGE! **
I’m not an electronics expert by any means, so I don’t know what that may mean, but I hope it helps diagnose what the problem may be. If I carve out some more time in the near future, I’ll run this test on the other two displays.
I’m testing using a PC, as I don’t have a VGA-1P or any other VGA upscaler currently.
I’m starting to suspect that HSYNC (and also maybe VSYNC) polarity is having some influence on whether scanlines are being shown or not (or even if image is being displayed at all).
By looking at the fourth table on this page, http://martin.hinner.info/vga/timing.html , it looks like only video modes with -vsync/-hsync polarities are supported? Well… just assumptions of a newbie who don’t have a clue about VGA signals
Anyways, I would also experiment to invert polarities if I had a hex inverter here… I will try to get some this Monday…
Yes. That’s the first thing I tested for each monitor: scaler and SLG only, powered on, and wait for No Input to display, followed by bringing up the menu for the scaler. Then, I hooked up the Sync Strike and tried 3 systems through it via RGB SCART: a Sega Nomad, SNES, and Dreamcast. I ran this entire test regardless of the monitor not displaying any image, so that I had some sembelence of a control.
Rosser: When you did your test with the prototype, you used Marvel vs Capcom on what I assume is the Dreamcast version via VGA. What make and model VGA box did you use? I’d like to try and track one down to see if it’s perhaps the Naki one I’m using causing my issue.
And we’re out of the T-SLGs. I’ll post up as soon as the production batch comes in.
Meanwhile, I’ll start digesting this awesome information from mistahsnart and soaking it in so I can understand and destroy it.
(mistahsnart, never ever worry about my feelings when it comes to facts. You’ve done me a great service and I’ll be doing my best to overcome this problem.)
No I’m just saying that I believe its a different chip that does the same thing but uses less voltage I’m just throwing things around in my head its prolly has nothing to do with it
Folks should be receiving their T-SLG and shields in the mail, so I should put up some assembly instructions for them.
Mainboard:
Assembly is easiest if you start with the shorter items first and work your ways up to the taller items. Even though the VGA ports are in the board when shipped, pop them out because you’ll be adding them last. All directions as top, bottom, left, right assume you have the board in front of you where the GodLikeControls.Com is properly readable from left to right. The Male VGA spot is on the left, and the female VGA spot is on the right, the slide switches on top, T-SLG text on the bottom. All pieces are installed on the side with the white silkscreen writing. All soldering is done on the other side that doesn’t have silkscreen writing.
Locate the two slide switches. These go into their spots on top so that the slides hang off the edge of the board. Solder them in place.
Locate the spot for the resistor on the board, marked ‘10K’ just to the right of the 74HC74 chip spot. Bend the legs of the resistor and put them through the holes and all of the way through so the resistor itself rests against the board. Flip over and solder the legs in place. Clip the legs short with a pair of wire cutters.
Locate the spots for the two 14 pin chips. The upper on has the label 74HC74 on it. Look at the labels on the top of the two chips included with the T-SLG, and take out the one with the matching 74HC74 label. You’ll notice a notch on the end of the chip, similar to the notch in the white silkscreen on the board. Put the chip in its spot so that the notches match. Repeat with the 74HC125 chip in the matching label spot. Make sure the notches match the white silkscreen; one is on the left side, the other is on the right. Triple check it again. Once you’re certain they are in and oriented correctly, flip the board over and solder the pins in place.
Next are the header pins. Locate the two pieces, each one a 2x3 of header pins. One goes to the right of the 74HC125, the other goes to the left of it. There is no orientation, so just put them in, flip over, and solder in place.
Last are the VGA connectors. Male one on the left, female on the right. Push them in until they are fully in; they should click and be perfectly flat against the pcb. Once in all of the way, flip the board over and solder the 15 pins of each on. The two larger posts on either side will NOT be soldered down.
Place the six jumpers onto the header pins. Each jumper goes left to right, so each header will have three jumpers all in the same direction.
Shield assembly.
There are three spots for three potentiometers. As long as the potentiometer goes on the side of the board with the white silkscreen writing, you can’t really get this on wrong Put in the three potentiometers and solder them in place.
The last piece is the single 2x3 female header connector. There are two spots, and only one connector; that’s normal, but means you need to be very certain which one to use. Even better, the connector piece should go on the other side of the board; the header connector will be on the side of the board that doesn’t have the white silkscreen. Wise man says ‘measure twice, cut once’. I say ‘check your placement twice, solder once’, so here is where the connector should be placed: http://img36.imageshack.us/img36/5820/shieldcon.jpg
Make sure that is where you have it, in the spot farthest from the potentiometer pins, on the side without the white silkscreen. Flip it over, and solder it in place.
Shield installation:
When you want to place the shield on the T-SLG, remove the three jumpers from the header connector ON THE RIGHT. These are the jumpers next to the female VGA jack. Slide the black header connector on the underside of the shield onto the now-exposed header pins. If the VGA connectors dont cleanly fit into the notches in the shield, then your shield wasn’t put on correctly; pop it off and place it back on evenly.
The jumpers on the left side connector, next to the male VGA port, should never be removed or else the device will not function properly (they connect the colors from the male jack to the female jack). The jumpers on the right connect the colors to the output of the 74HC125; they will not be dimmed at all, so leave the jumpers on or the shield on for proper functionality.
1024x768@60Hz.logicdata - Here we can see working scanlines: RGB_OUT is toggled on and off at every HSYNC rising edge. HSYNC pulses LOW. Same for VSYNC signal.
800x600@60Hz.logicdata - RGB_OUT gets pulsed only when VSYNC is HIGH (what happens only at the end of every frame). That’s because both HSYNC and VSYNC pulses HIGH!
800x600@60Hz_!VSYNC.logicdata - Here I tried to invert only VSYNC signal as an attempt to get RGB_OUT toggled on and off at every HSYNG rising edge. I’m still NOT sure why VSYNC is pulsing all the time, compared to the “800x600@60Hz.logicdata” sample where it gets LOW all the time until a frame is finished…
800x600@60Hz_!HSYNC_!VSYNC.logicdata - Here I tried to invert both VSYNC and HSYNC… Things get crazy! VSYNC pulses very rapidly, resetting the flip-flops.
I believe that if we find a way to make VSYNC to pulse LOW, just like 1024x768@60Hz (and also 640x480@60Hz), and maybe figuring out why inverting only VSYNC makes it pulse several times before a complete frame, we could be able to make this work for all video modes.
Any clues?
P.S: Toodles/guys, please let me know if this is not the appropriate place for posting this kind of stuff.
So, went through all of the resolutions supported by my cheapy Xbox360 cable. Only 640x480 and 1024x768 used a high at rest VSYNC; all of the others use a low at rest VSYNC line. That would leave the board unpowered when needed, keep the flip flops low when needed, and who knows what kind of effect on the lines with the chips trying to drain power from whatever they’re connected to.
So. Time to think. Alternative power is first, which I think I have and shouldn’t really add to the cost to implement. Next is deciding one of A) selecting whether to use VSYNC or !VSYNC for flip flop clearing, or B) forgoing flipflop clearing entirely. B is obviously cheaper, but runs into problems if there number of HSYNC pulses per frame not evenly divisible by 2 (normal) or 4 (wide).
My PC video card outputs +5V in the pin 9 of the VGA connector and I’m using it to power the circuit. Maybe you could use it to power the circuit instead of using VSYNC? The problem here is that not all VGA controllers output +V there, so it would be “incompatible” with scalers that don’t output voltage at pin 9. External power would be an alternative, but I don’t like it also.
I’d love to be able to, but it’s rarely implemented on non-PC video stuff. Xbox360 cable, DC VGA box, and the ebay scaler all dont. I even found a number of VGA cable that dont have it pinned at all (hence why wiki says ‘key/+5v dc’). If I cant count the cables to support it, no way I can rely on it as a source.
That’s good to know (that none of the eBay scalers support +5V on pin 9)!!
Anyways, I’m starting to believe that pin 9 is not a reliable/constant power source for some video modes. I’m saying that because yesterday I connected it directly to the RESET pins on the 74HC74 in order to simulate an “always on VSYNC” and my protocol analyzer showed several pulses on that line. So for some video modes it’s probably being used as the “key” thing we see in the VGA pinouts diagram. That could also explain the pulses I got when inverting VSYNC on 800x600.
So today I will hook up an external power source to the circuit in order to test scanlines for other video modes and see how it goes. If that works, so we’ll definitelly need an external power source in order to get scanlines for video modes different that 640x480 and 1024x768…
Put in the three potentiometers and solder them in place.
