Im planning to build a home security video surveillance to record video and stream it. This is a personal project but I want to approach it with a production ready mindset, as a challenge for myself.

The RPI general board P1/2/3 support for camera module out of the box. https://www.raspberrypi.org/products/pi-noir-camera-v2/

3 questions:

1) My goal is to integrate a camera module onboard what would be the approach to this? Is the compute model 3 hardware open sourced so I could modify it If I want?

2) What would be the approach for building a custom board and integrate the compute model with a camera module from Raspberry or other HD camera ? My goal is create a compact print that is less in size than a regular RPI

3) I don't have a background in EDA design so this is a danting challenge. Could anyone point out some resources on how to integrate the compute model 3 in a opensource project and how to approach this best? Since all RPI run Debian I think I could start with a regular RPI with camera module for prototyping and port my code lateron?

Preferrable im looking for an opensource RPI Compute 3 solution where there is already a camera onboard or references in this regard.


1 Answer 1


first I just realized you are over thinking the heck out of it. There are chinese hardware .264 usb cameras that are just board and lense mount like the pi cam for 30 bucks. grab 2 of them and a pi cam. they have 1080p versions for like 20usd more. I've ran 3 cams on a pi 3 before. 2 of them were usb with out compression one was the pi cam. I then bought 2 of the cams i described and framerate was way better. I could stream like 15fps over wifi/lan when all cams were .h264.

unless you want a flying robot this should be sufficent with out using the compute module. One thing you need to do is set the program you use to do the recording (likely "motion") to record to a ram drive. If you do not do this you will burn up your flash cards quick

if you have your heart set on compute module(its a mistake for what you have described):

imho you are going to pay through the nose for pre made boards to use with the compute module.

a board consists of 6 voltage supplies, which costs $18 bucks of IC at most, if you want nice parts and are just ordering one of each. it can probably be done for $10 if not less

say each usb port costs you a $1 and the hdmi costs $them 3. There are some passives and rfi and static/esd chips but places like ada fruit charging what they do for the mobo is nuts. The ddr port it plugs into is maybe $3 bucks.

the little connectors you need for the cams maybe cost $0.75 flex connectors is what you should search.

. eagle cad has a free version, they also have versions one my make a choice to acquire with a vpn. Kicad is completely free cad software one or the other will likely fit your current mind set better. There are part libraries for both and sites like Digikey have a checkbox to search parts with included CAD models. honestly you are best off learning to input a part from the data sheet/scratch.

My first pcb took more time finding pre made parts than if i had just suffered that learning curve

if you google something like "pi compute 3 minimum host eagle cad "(or kicad). you may find someone has provided a file with the psu components and the ddr2 socket already laid out. from there you will need a flex connector for the cammera.

Datasheets are your friend. They tell you about the part and its connectors. most connectors have a name and pin count. this enables you to search for and buy them.

cad programs have an option called "generate BOM" or something similar. Its Bill of materials.

If you find a file with the basics laid out, outputing the BOM tells you what you need to order for their design.

when it comes to getting a pcb made you have several options in us and china. osh park has like 5 day turnaround and priced 5 bucks per square in of pcb. This is for a 2 layer pcb and you get 3 copies of your pcb for that price..

if you dont mind waiting a month some pcb houses found on ebay will do it for less. keep in mind, working with a company in the developing world whos employees may or may not " the english so good" can add a layer of difficulty

path less traveled is to etch @ home. I had plexi laying around when I got into etching. One can assemble a thin vertical tank with plexi and paint thinner (or solvent glue). tanks can be had for about a hundo premade.

At the bottom you place an airline and use an aquarium pump to bubble air through your acid. Go with cupric chloride and your etchant(acid) will be reusable indefinitely because air regens it. almost every etchant becomes illegal to dump down drain after it eats copper off the board. the nice thing with cupric is its reusable so you never have to pay to get rid of it(assuming you may etch again)

Cupric Chloride can be made from Pool Acid(30%hcl) bits of copper wire weighted out and hyrdogen peroxide Hair bleach (volume 40 is the product name). Alternatively the 3% peroxide from drugstore will work but it will take about a week of bubbling air to get the chemistry right. the "volume 40" is like 15% peroxide and you will be ready to go that day.

There are 2 main choices for how to make the mask/template which will protect the copper we want to remain on pcb etched @home. What isnt covered will get eaten/etched away by acid.

the blank pcb is called a copper clad. I will refer to it as such for the rest of this.

choice 1) toner transfer. One needs a laser printer, an iron or cheap ($25 ish) laminator a witch doctor and said witchdoctors recommendation on what special paper to try to get the best results. He or she may later tell you that the printer brand you have has angered the spirits. I recommend skipping this method.

choice 2) uv photolithography. this is what the industry uses and by far produces better results if done with care. It is also repeatable with the same printed transparency choice 1 is not

you will need a printer for either choice.

UVlith process is compatible with inkjet or laser. If you are in the market there is a 3rd option which requires a lot of DIY to your printer, but may allow direct printing of pattern to Copper clad.

With that in mind, when i needed a printer for the lith process I bought an epson c88+. Printers that can use pigment inks are key but im not going to cover the 3rd method/direct print here. I just mentioned that printer because its cheap, it works and it may have that advantage later if you look into it.

List of stuff needed for UV - Lithography

  • printer,

  • laminator (amazon basics(cheap)),

  • photosensitive acid resist film. UV photoresist. (dont get the stuff from amazon. it sucks). The brand to have is Dupont riston. sadly Its hard to find and dupont didnt even answer my .biz domain asking for a free sample. Here is a link to a supplier I finally found. It apears to be real, Im testing it later this week. I already like how it went on the board better than the amazon/china crap


  • developer: This is just washingsoda...Sodium carbonate (Na2CO3)(NOT baking soda). you can actually find this at walmart (arm & hammer brand and its in laundry isle. probably by borax).

  • UV exposure light source: lucky for you that big yellow thing in the sky tends to work. I bought a 365nm 4.1w led from digikey.

Glass to sandwitch your transparancy and the copper blank when exposing. Do not use plexi it will block uv. Just take the glass from 2 picture frames and you are set.

transparency film for your printer type. A lot of people again go to witch doctor land here. I highly recommend you go inkjet for printer type then pick up Fixxions brand screen printing transparencies. I bought like 5 kinds, then found that the screen printing comunity had the answer


they have a coating that is designed to be clear and make the printer ink stick just like the paper it was made for. Beautifully dense/dark and never runs. Those are the two qualities you want because one equals minimum spacing resolution and the other would equal holes in your copper where you dont want it as the printed transparency determines where the etch resist stays on the board during exposure.

  • "copper clads" (for either toner or litho method): these are the blank pcbs. they are a copper film that is applied to fiber glass on either one or both sides. you can not make these at home. Sure the film is just glued on but unless your home has an autoclav your attempt at adhesive will not be sufficient. amazon and ebay carry them.

Misc things you need are: ajax/commit cleaner acetone from hardware store green dish pad(not used) rubber gloves(disposable) dremel, hand drill or drill press soldering iron for assembling board after(probably figured that )

process: So, if you put that kit together with an order or two and getting what you had in one pile... the process is: PRINT: 1. design in CAD software
2. export/print 2 copies of negative on
transparency film

  1. supper glue them together (blocks more light) 3.5) if you have a back layer repeat. sometimes a piece of pcb glued in between top and bottom can make a jig to slide your copper clad into to ensure alignment between layers. this is crucial if you have actual traces on the back. if the back is all ground it doesnt matter so much. if you miss and drill out the side of a
    pretty much flat pcb..i dont know what to tell you


  1. dawn gloves -scrub your copper clad (green dish pad and ajax/commit) rinse

  2. grab another pad and acetone can. Scrub with acetone. Its clean enough when a paper towel/white cloth and acetone do not turn grey. set it somewhere clean and dry when finished, you will need it again soon

  3. Get yourself under LED or very dim incandesnt light. Remove Riston film from its roll tube. LED is the way to go as they emit only a narrow band of color in lights and no UV. LED lights are usually safe full brightness for handling the UV sensitive film. Once under suitable light, cut yourself a square of film the size of your copper clad. cut another if you are doing double sided. If your house is a grow op and you have grow leds everywhere...stop this process and move out before you get busted. also... you just exposed all your film.

  4. cut a piece of clear tape.

  5. stick clear tape in corner of uv film and rip away quickly. The uv/riston film has 2 protective films (each side) we need to remove one and this is the quickest way.

  6. Lay film on copper clad, (side you just took protective film off down.) you will know what is film and what is protective layer because film is blue, or maybe green depending on brand, protective is always clear.

  7. laminate film to copper clad. I use the amazon basics cheap
    laminator. note: To avoid air bubbles I like to take the protective layer off one side of film, Place Copper Clad in a tray of distilled
    water, Float the film on top of water and pull board up to it (gloved hands, oils will mess things up). After that i send it through laminator. Some people find ways w/o water. that just works best for me. Bubbles are bad.

you can do multiple boards and store them for a few months if you want a ready to go supply. just keep them in a light proof container.

expose 1)put transparency over/over and under(2 sided) your riston coated Copper clad 2)sandwitch between glass sheets 3) bring outside to expose 4)set in dark for 15 min to let chemical/plastic stuff happen
5)drop in developer and brush gently w/soft brush or gloved finger tip for 1min. You will start to see the areas that were exposed get washed away/ come off and the rest stays. This is your "acid resist" pattern. 6) rinse under tap water or distilled to stop the developer

Etch 7) drop in acid tank for 15ish min 8) remove and rinse with h2o to stop acid 9) drop in acetone or scrub acetone and green pad to remove the resist and reveal the shiny copper wires/traces/pads you just created on your board 10) scrub with like 5% hcl and water to remove copper salt deposits that dont come off with just water.
11) wash your hands, secure chemicals or decide life isnt worth it and drink them etc etc.

12) drill holes through layers. solder on parts...etc.

obviously some googling on board etching will be required but I highly recommend you try it. Unlike sending out and waiting a week, a mistake can be corrected and new boards made in 20min to an hr for just a few dollars of waste board. If you make a mistake on a board you had professionally done; you might be out a few hundred.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.