SparkFun Forums

[SOI_Sentinel]

I'm extracting this from the camera-working thread.

Since we have a working camera spec now, I'm working on a modular camera system. My first gen system that I'm working on right now interfaces the camera through an Averlogic video FIFO to a dsPIC to a dual CAN connection on the backside.

Now, what I'd LIKE to do is build a stackable system. A "video" breakout board has a standard multipin header (say, a 16 bit parallel interface, Vsync line, Hsync line, clock in, clock out, I2C, and a 4 wire SPI interface) to support the E700 camera and many possible future cameras (like an NTSC board).

A second board would be between this board, and would be populated with an Averlogic board. This allows you to chose to use the FIFO if you so chose.

The mainboard would contain the main CPU and probably the comm units. This would let you use a PIC, dsPIC, ARM, or whatever processor you chose with whatever comm buss you want too.

[nova]

I'm extracting this from the camera-working thread.

Since we have a working camera spec now, I'm working on a modular camera system. My first gen system that I'm working on right now interfaces the camera through an Averlogic video FIFO to a dsPIC to a dual CAN connection on the backside.

Now, what I'd LIKE to do is build a stackable system. A "video" breakout board has a standard multipin header (say, a 16 bit parallel interface, Vsync line, Hsync line, clock in, clock out, I2C, and a 4 wire SPI interface) to support the E700 camera and many possible future cameras (like an NTSC board).

A second board would be between this board, and would be populated with an Averlogic board. This allows you to chose to use the FIFO if you so chose.

The mainboard would contain the main CPU and probably the comm units. This would let you use a PIC, dsPIC, ARM, or whatever processor you chose with whatever comm buss you want too.

Sounds cool! The only problem I can see is that many people are using this camera because it is small, and putting 3 huge boards below it doesn't exactly keep that benefit, shall we say. I personally really like the idea - I would like to have a USB and gumstix (2 projects!) connection, so all I have to do it build an interface between my stuff and the averlogic with nice headers, instead of dealing with impossibly small connections (as an example. I'm personally also building an FPGA based cam board, which is just crazy). Just a couple of questions:
1) Where do you get the Averlogic modules? I couldn't find any good sources on their website.
2) Would this stack have all the nasty power regs, shifters, etc on it so that people would just have to worry about a single 5V or 3.3V signal (even on the i2c lines)?

thanks!

[SOI_Sentinel]

Crud, lost a long reply....

Anyway, the stack would be taller than standard, but possibly covering less area (I'm less volume limited in my needs than area limited).

The plan IS to have each board have it's own LDO/switching power system for the populated components, unless it's specifically meant or needed to be separated. Target voltage is 3.3V. The camera may not need level shifters if running at 2.8V as the Averlogic reads down to 2.3V as a 1 when running at 3.3V. I'm considering building the board with programmable pots on the power system so I can digitally experiment and fine tune the voltages feeding all but the CPU. I'd require either a semi-regulated 5V source feeding the system (although a LION or 3 cell battery system may work fine), or perhaps have a power board on the bottom.

I probably would have to put a level shifter on each board (for the bottom "incoming" signals) for the I2C lines. Only for those that need it, though.

The common bus notion and control systems is fairly related to the PC/104 standard. The idea is that this is a fairly decent data flow system. While I'd have to figure out a way to chain Averlogic FIFOs (they only really support width expansion), this would make a great way to allow someone to, say, build an NTSC board and do frame capture or build an analog board and do one or two storage oscilloscope capture, or maybe a small logic analyzer. Most people would opt for a USB PIC or related for a project like this. I would too if I was doing a DSO, but if I want to, say, do my own JPEG compression or analyze a video frame for objects and then output the suspected vectors onto a CAN control bus or ethernet, it'd be a matter of picking your favorite processor card instead of a full board redesign.