Hardware Removal and conflicting functionality

[keel]

TLDR;
Will the ML code function without the Eos/Ef flash or mirrorbox connected to the camera?

Hey Guys,

Before i get blasted for even asking (before doing it myself) i thought i would bite the bullet and ask for some patience with lack of coding knowledge...

Has anyone stripped down -any- camera to basic functionality without connecting all parts. Such as, running a camera without a flash, or a mirrorbox, or any other (non essential) hardware?

I would like to rebuild the housing of my cameras to suit my needs, and in the process, potentially remove unnecessary parts. While code will stay inside the firmware for these functions, are there obvious issues popping up immediately with this (train of thought)?

For example, i need to remove the mount, mirrorbox, and flash. I really only theoretically (patience please) need the sensor, pcb, digic, card reader, battery, lcd (or hdmi out) and a few buttons to navigate the menu. In an ideal world,running ML only on a Canon product but with a custom body.

I apologize if this answer is obvious, such as "don't remove any hardware"... But ultimately this is what a eosm is vs a t2i.
The reason i ask here, is.because im a user of ML and respect the knowledge of those who made it. I cannot find any info online and ive built a camera on an arduino before, so i understand the most basic concept of code via "processing". As far as i got was assuming it might work but could i null functions of missing hardware, or will the firmware not boot of parts are missing, etc.

Thanks for the great work. I appreciate all nonsarcastic responses immensely.

[a1ex]

While ML gives you the tools to do this, the difficulty level is a lot higher than you think.

The best exercise IMO is to look at how various firmwares are emulated in QEMU. In some cases (see e.g. 700D and EOS M), some hardware devices that are not emulated properly are disabled in order to allow the simulation to advance.

Once you understand how that works, and can write similar patches for other models, you've got about 5% of the software knowledge necessary for your mod.

When you'll manage to emulate the MPU and understand how various hardware devices are connected to it, the percentage will jump to somewhere around 50%. Tip: the TX19A documentation is public, and QEMU already emulates regular MIPS (without TX19A extensions), so it's doable.

That's just my limited understanding about this firmware.

[keel]

Awsome response, thank you. I am willing to pursue it but i hate not having guidance. Months vs minutes in learning curves. This is a definitely appreciated. Google has been a slight nightmare. Headed to the local college to start asking around for some advice. Sounds ridiculous but they got a great technology department. Between 3d printing a new body to reassembly of the parts (even as a whole without the code changing) it is complicated enough. When changes /removal comes either false voltage signals or code adjustment. The electrical approach seems more suitable for me to do but for others i think the code would more suitable. Incase anyone is interested in hardware hacking. Adding cooling comes to mind on 50d (issue i am also considering).

[dmilligan]

I think you missed the point.

Once you understand how that works, and can write similar patches for other models, you've got about 5% of the software knowledge necessary for your mod.

This task is beyond my current knowledge. I've been contributing to ML for over 4 years now. So that means for me 4 years is less than 5% of the knowledge required, therefore, best case scenario it would take me something like 76 years to achieve what you are talking about. Good luck with that.


Look at it from another perspective: I'd say the amount of work required for something like this would be at least on the order of thousands of hours, if not tens or even hundreds of thousands. Is the opportunity cost of that much work really worth it for your mod? (In other words, would you really pay $50,000 for a 5D3 with no mirror?)

[keel]

I definitely didn't.

Every coder, I have every met ... seems to talk the struggle. I don't have any intention of becoming a professional coder, or to monetize a 5d, nor spend much on the project myself.

Just wanted to see if anyone had disconnected a hardware piece, or modified the code, and ultimately, would the code/camera function. I have enough hardware to break one open, but before I do, I would like to involve the code community and see the vocabulary to understand your responses. When someone who works with code says 5% of the knowledge, I can assume it is 500% more difficult than anyone thinks. Still, somebody learned it once. No reason to not try and communicate through a better vocabulary.

My only goal right now is to rebuild my 50d's into a better shape with better airflow. Ideally, i have already jammed the mirrorbox and I only use the raw video. I would love to use C mount lenses but box and mount would need to be removed. Frankly, the shape and design of the camera isn't suitable at all for 70% of what I need. A power switch, record button, and basic menu functions for ISO, and zoom, plus the LCD or just HDMI out... that is really all that is needed. The WB seems to sit at 6400K on the Raw videos, and I am sure I am forgetting some functions but this is essentially all I use.

[dmilligan]

By "digic processor" I assume you are talking about the main general purpose ARM CPU core (the CPU that ML code runs on)? If so then, yes, the raw image data from the sensor is copied directly to and from main memory and other hardware modules using specialized hardware (EDMAC).

[keel]

Yes, that is exactly what I mean. Thank you !