camera hardware documentation missing

[acornejo]

Can't find this on the webpage.

[robgithub]

this page is blank http://www.raspberrypi.org/documentation/hardware/camera.md is that what you mean ?

[acornejo]

Yup, that is exactly what I mean. I assume instead of blank page we should be greeted with a page that has info on the camera hardware (perhaps a schematic and list of components, maybe together with a brief description that explains to which port on the RPI it attaches to).

[bennuttall]

@JamesH65 wrote that page but it was all about the software so I moved it to another section. I left this page here (blank) in anticipation of it being filled in with information about the camera hardware. Not sure what will go there just yet. What information were you looking for?

[bennuttall]

There's also the camera section in the usage area.

[lurch]

I've just noticed that the bottom of http://www.raspberrypi.org/documentation/usage/camera/raspicam/raspistill.md links to this now-empty "camera hardware" page - maybe the link there should be updated to point somewhere different?

[bennuttall]

Or someone could write it?

[waveform80]

If anyone wants a head-start on this, I'd recommend just copying the first section of the camera hardware page out of the picamera docs. This covers the camera modes supported as of recent firmware updates (including the 90fps stuff). The latter section ("under the hood") might contain some interesting stuff, but is rather picamera specific and therefore probably doesn't belong in generic docs.

[acornejo]

The kind of info I was expecting under camera hardware was:

• schematic of the camera circuit with list of components
• datasheet of the IC's used (i.e. OVA56*)
• info on the CSI-2 MIPI protocol (i.e. pin function etc)
• list of supported modes for the camera

[ghollingworth]

Schematic is not to be released
Datasheet is not ours to release
CSI-2 MIPI is a closed protocol

List of supported modes... Tick...

[acornejo]

Regarding the schematic, if it is not to be released then i assume the rpi camera is not an open hardware design?

You can always link a datasheet (if you wanted to). Adafruit, SparkFun, seedstudio, and most other open hardware designers always link data sheets of the relevant components.

Csi-2 Mipi is not a closed protocol. It is justo i2c for control signals plus parallel lines for data.

The camera modes are almost more relevant for software than for hardware.

If no info on the hardware is to be posted I suggest saying that on the page up front to make sure people know it's not an open design.

That way people can look for other alternatives, since it is pretty hard to develop software and drivers for hardware without specifications (you are essentially restricted to reverse engineer any existing software and hardware).

[JamesH65]

I'll write up something this week, been very busy last couple of weeks so not had time. The datasheet for the sensor is confidential (not that there is anything of use in it anyway), so I cannot include that (there may be one on the net somewhere I can link to), but should be able to put in the basic specs, and how much of those specs we actually currently implement.

As for previous post, the camera board is not an open HW design. And it's not really possible or even sensible (cost more to make than buying one) to make alternatives (there is a Chinese knockoff which was reverse engineered) which is a good thing in my book, because the profits from the camera board go straight to the educational charity.

[bennuttall]

Thanks James

[acornejo]

Thank you James.

If the sensor is the ova5647, there is a document brief here: http://www.ovt.com/download_document.php?type=sensor&sensorid=66

I actually have a copy of the datasheet which I was going to link to here. However only after you mentioned that it was confidential I realized it has a watermark that says "confidential". What a bummer!

This is a short but useful description of the MIPI csi-2 protocol:
http://www.mipi.org/specifications/camera-interface#CSI2

Finally, regarding the camera board not being open HW, its a pity.

IMHO having an open hardware design encourages more developers to jump on board and collaborate/hack together more software and even create more documentation.

As you have already said, it is always cheaper and easier to buy hardware than it is to create your own. Honestly I will be surprised if you loose a single customer due to having an open hardware design (as an example of this, I am eagerly waiting for the RPI compute module, even though the schematics were released, I rather wait for the boards than to spend 10x time money and hours/days routing the circuit on eagle to get something that might not even work).

Also, as you have also pointed out, Chinese knockoffs are inevitable if there is a large enough market for them. Given enough time and a multimeter and/or bus pirate, you can reverse engineer any board (especially a board with so few components and traces as the camera board). Security through obscurity is rarely a good idea.

That being said, I understand there are always higher forces at play that may prevent RPI from being completely open, so no complaints here. I applaud the efforts of the foundation.

[RichShumaker]

Just posted the information I learned from the forum

http://www.richshumaker.com/raspi/raspicamera.html

[flakeshake]

Hello all ,
i've collected some information i found in the forum here
https://github.com/flakeshake/pidocs/blob/master/cameradoc.md

Would it make sense to use it as base for the official documentation ?
Please let me know if you can spot any mistakes.

[bennuttall]

Hi @flakeshake would you mind forking the repo and adding your contents to camera.md and creating a pull request?

Then @RichShumaker if you have any factual data to add, you could request for them to be added in the same way?

Thanks guys