• Content count

  • Joined

  • Last visited

Community Reputation

0 Neutral

About Keithgf

  • Rank

Profile Information

  • Gender Male
  1. Thanks Jack and Jaxartes for the replies. I've read them each a couple times. And read the section Jack linked to a few times, even before I posted originally. The FPGA does have configurable voltages, but I've been very careful to set them to +3.3v. It defaults to 8ma drive strength, but I tried adjusting it to "Maximum current" with no effect. I can double check the digital input pins for proper voltage. The 12" cable mentioned is on the digital side. Between the computing shield header and the FPGA GPIO header. I am using a regular proper VGA cable on the analog side between the VGA connector of the computing shield and the monitor. What bothers me isn't so much the dimness, but the lack of range. The advantage of a 4-bit per color DAC is so that you have 16 distinct brightness levels per color. My current setup has only 0xA through 0xF usable, anything less than 0xA is black. So I end up with about 5-6 distinct brightness options per color, which gives me sort of hundreds of colors instead of thousands of colors. Note I'm not really blaming the shield -- it may be the simplest part of the equation. There's lots of moving parts here, and something along the way is breaking down. I'm just trying to figure out what that is. I'll also connect the output of a color on the analog side to ground via a 75-ohm (or close) resistor, and then measure that voltage. I'll also try outputting blocks of color from 0 to 16 on each color. Thanks
  2. I'm using a Computing Shield with a BEMICRO MAX 10 FPGA board. It has +3.3v GPIO. I'm using (4) PMOD headers with a custom 12" cable. It was longer than I wanted, but it's what I've got. This is a cable made from Pololu Electronics -- basically DIY pre-crimped M/F header cables. I'm driving it with my own Verilog design. This is a custom computer. I've got the 12 digital color bits + hsync + vsync + GND hooked up. I've done VGA FPGA projects before and have some familiarity. I'm running it at 800 x 480, ~31.6khz/60hz, if it matters. Samsung 21.X inch IPS LED monitor. I've got the 12 least significant bits(MSB is on the left/top) of a 16-bit word mapped to RGB. The top bits are unused. So bits 11-8 go to Red, 7-4 go to Green, and 3-0 go to Blue. When I output 0x0FFF, I get a dim shade of gray, when I would expect to see a purer, cleaner, brighter white. Black at 0x0000 is fine. When I output 0x0F00, I get a pure red, but it seems awfully dim to me. The other colors like Green (0x00F0) or Blue (0x000F) are similar. Blue is actually the worst. This isn't a bright blue like I would expect. Only nibble values of 0xA to 0xF produce visible colors on the display. Anything under 0xA is so dark that the color turned on isn't discernible. I reverse engineered the pinout from looking at the PCB and using the Papilio Duo that it would normally connect to for the pin names. An image showing the BOTTOM of the computing shield with associated pin names would have really been helpful. The image is crisp and stable. I can use a scope and get output voltages. Will the computing shield output the full range of voltages? Meaning does a digital input of 0000 correspond to 0v, and 1111 to +0.7v? Any chance you have a chart the shows the corresponding voltages to expect given the input? Ie given your resistor values? I'm inclined to think I've got a wire/pin mapping problem. or reversing the order of the bits or something. Any thoughts or things to try? Thanks Keith