I would never use such a cannon for such a simple/cheap solution.
It’s like embedding a pentium to read the data of a temperature sensor…
I for one will stick to trying to figure out theese registers, and if it can’t work with a low mips mcu without a tonn of extra logic, then it isn’t worth it. There are plenty of other sensor chips that are better documented and I’ll rather use one of those.
For my design all I want is a sensor chip, a mcu and a memory device like a mini/microSD. No cannons to shoot this sparrow.
I may use the STM32 kit I have here to help facilitate the analysis and discovery, but that’s way too pricey for the final version.
As you can see, it’s humanoid. Camera is definately not using it’s auto functions or I’d not be purple. (Actually I was green untill I auto-balanced it.)
The image is still not full frame. Somehow it refuses to work with the correct area. It is however a good start in the general direction I want to go. This time I can reproduce it too, so atleast there’s a chance of learning what some more bits are doing.’
After the chineese hollidays I might get an ansver from china too. Hopefully a final version of the datasheet, not the draft… (Yes, I can say that now being finally deemed worthy!)
I am asking as I might do a few changes and etch a couple to you. Not just from the goodness of my heart. I also think the more ppl picking at the registers in this camera, the more a chance we have at getting some more usefull stuff out of it.
My current board has two regulators on it. A 1.6v for the core, and a 2.8v for the I/O. Both are 150mA regulators so they will handle the cameras requirements barely.
The thing is, many might want to use 3.3v or 5v on their microcontroller and that wouldn’t be too great with this setup. Even though 3.3v would eek by, it wouldn’t deliver much headroom for the 2.8v regulator. Frankly it’s not ideal for the I/O either on a 2.8v platform.
Maby some simple level conversion would be good. The complex stuff would be on the SDA pin as that is the only 2-way pin.
I just want your opinions and ideas. It might be a quick fix to make the breakout board something you would all find usefull.
I think just having all of the pins broken out to fit a standard dip socket would be what I would want. I could handle the voltage regulation and level shifting off board.
If I wanted more, it would only be to have regulation on board, the 1.6V and probably 2.5V as opposed to 2.8V, but that’s just me. With switching regulators, it should be a simple switch of a resistor to get the different voltage (and probably the same with the 2.8V). And depending on what type of regulator, you could probably get a buck/boost so the source voltage could be a fairly wide range.
I’ve been pretty bogged down with school work and work work, so I’ve not had much time to work on a board for this guy, but I do agree the more we can do the cheaper we should be able to get them.
I really wish SFE would let us know if they are already working on one though…
I already have 2 linear regs for the supply. Currently “MIC5205BMS TR” which has low drop and adjustable output with seperate ground reference. Changing output voltage is a simple equation. No ground-current compensation or such.
The cam is quite sensitive to switching noise so I opted to not use swmode regs. It doesn’t draw that much power so it seems fine as-is.
I’ll add 0-Ohms resistors to the signal pinns so one can add series resistors if needed. This might also help when running MCU at 5v or such to limit input-current.
Pullup resistors will be put on SDA and SCL as well as reset to ensure valid signals.
Spacing it out a bit more would also make it easier to solder.
The board will have two 2x5-pin headers carrying 8 signals pluss power and ground. One for data and one for control. I have matched these to STK500 so that the data port can be directly linked. Depending on the way you want to control the cam, the control port could possibly be directly linked too but for interrupt driving or such, a squid-cable could be used depending on MCU.
Just got me a pair of those cameras as well, and it seems someone is doing the PCB job for me - nice indeed! What about the clock - will there be an onboard oscillator or would I need to feed the clock signal through a pin myself?
There won’t be a clock as the mcu you use to talk to it should ideally be synched to it for ease of readout. Essentially, you should thus create the clock. (use HW pwm!)
reklipz There are two 8+2 pin headers. A total of 20 pinns.
First header is for data only + gnd/pwr.
Second has EXTCLK, VBLK, HBLK, DCLK, STROBE, RESET, SDA, SCL + gnd/pwr.
The headers are positioned on opposite side of the pcb so you won’t have any dangly wires on the side with the cam.
If this board was to be made professionally I’d ofcource do it 2-sided to tidy up the routing. Then I’d add silkscreen to connector-side so one would always have the pinout of the connectors available.
edit:
The large 1206 resistors are for all spots where user might want to change stuff. Only the power-system is set up with tiny stuff as that will be a 1-time affair. (0603 and 0805 for power system.)
Nice work! I for one would be interested of getting one of those (especially as I live in the neighboring country;) )!
A bit off topic perhaps, but when I did a PCB for the TCM8230 (the low-res equivalent cam) I used zener clamps to get 2.8V level on the inputs instead of resistors only. As I’m a software guy by learning, what are the pros and cons? For instance, could there be any problems using zeners with respect to high-frequency stuff like the clock?
Zener clamps is good but you need to combine with series resistors for them to work. It starts becoming a lot of components.
For a 1-sided layout like this I’d have a hard time getting both clamp to ground with zener and series resistors in there. The series resistors help as they can jump traces, but the clamping zeners would do the opposite and need to jump to ground all over.
Also, running avr at 5v might not allow it to detect the 2.5 or 2.8v signal from the cam as high. I was going to use pullup in avr for it to get higher levels from the 2.8 while dimensioning the series resistor to balance the effect. I think I’ll stay at 3.3v to make it easier.
Ah, I’m doing it doublesided so I’ve got it a bit easier of course. I’m also sticking to 3.3v - I haven’t had the time to test my board though so we’ll see if it works when the fab is finished I guess.
Haven’t added the small resistors that set the voltage yet. Other than that, it’s ready to go. (Not happy with the flux in the solder I am using, think I need to pick up seperate flux…)
If only the cam could stop delivering what looks like UV images I’d be happy. There is actuallty a bit for that. I’ll test it when I get home from work
What about the clock - will there be an onboard oscillator or would I need to feed the clock signal through a pin myself?
