That would be excellent.
Please also count some € for a participation in the work.
I would be very interested!
But we’d have to find out the specs and chip if not decided yet.
My use would be professional and needed information would be: 1. Color temperature, 2. Green/Magenta correction and 3. exposure in diaphragm/speed.
But of course, the device must comply with the OP needs first.
Just thinking… Perhaps generic interface boards with TOH pads tracks-linked to soldering pads would even be of interest?
This way, anyone could solder wires that would easily connect to the phone.
Well, it’s not as simple as that. You basically need a small controller on the TOH side since the requirements (ie. addressing and signalling to the phone) won’t be covered by generic sensors, even if they have I2S support. So, a generic TOH with a small controller might be a good idea… there is some work in this direction (with an esp board) at https://forum.sailfishos.org/t/cots-lora-3d-printed-toh-shell though I don’t think I’d go with a module, per se. One could piggyback on that project using the footprints it provides for the modules…
Ah thanks, I was not aware of that.
Yes yes yes
for those who have no idea on how to make/design/order boards.
This really needs some kind of brainstorming session. Choosing an mcu and making it extensible is non-trivial, though going the module route, ie. using a common size like the waveshare or seed mini boards might to the trick …
Are you really sure that a microcontroller with firmware is needed? From what I know:
- You can directly connect I3C and I2C target ICs to the same physical bus and use all of them at their respective maximum speed.
- A buffer/translator IC between I3C and I2C is only needed for connecting I2C target devices that use features for which I3C is not backwards-compatible, for example: clock stretching. The datasheet of the I2C chip in question should contain this information. I guess, such buffers/translators can be bought off the shelf.
- From the bus-side of things, the Jolla Phone only requires a small I2C memory chip at target address 0x50 inside the TOH shell which contains identification and initialisation information as specified here. It does not require complex stuff like a microcontroller or whatever.
- On that very same bus, you can then connect any I3C/I2C chips or ready-made modules of your preference.
EDIT: Here is how this basic identification circuit looks like in the case of the Snow White TOH: Jolla Phone 2026 The Other Half - How does it work? - #7 by moripeluka
THIS! ![]()
(20 characters)
Well, let’s see, so I need to have a circuit with a specialized chip to manage the bus. That’s one issue. Second issue, I probably need to manage communication on the bus, given that the defaults stuff like ‘wake on’ are not going to be implemented by generic modules. That’s issue number 2. And the list grows, the more I think about it. But, I’d be happy to be proven otherwise by a concrete example. For instance, a simple TOF sensor.
EDIT: For my part, I use a number of I2C devices and generally (from my driver end of it) use the default bus IDs which, if I’m using 2 TOF sensors, a display and a DAC would not be possible (I believe) with the chip in the linked document.
SOOOOOOOOO NEWS!
Thanks to @mad_dev amazing work on his RAWfish and Camera2 I can now finally reach the correct reflective readings from the camera on the Jolla Phone 2, which means the fiat lux app works, just like that, without a TOH (I’m still interested in making one, but this is a milestone, for me at least).
So, Thanks to @mad_dev Jolla Phone 2 owners can have a reflective light meter for analogue photography ![]()
I hope to be able to release a bug free version this weekend. Got some small stuff to iron out first ![]()
Not necessarily: https://www.ti.com/lit/pdf/scpa067
I’m not sure what you mean, but if you simply want to communicate with an I2C chip, just start talking to it whenever you like. No startup sequence whatsoever needed AFAIK.
Every I2C chip has a non-unique target address. For example: The EEPROMs containing the TOH identification information all have target address 0x50. Some I2C chips only have one target address, others let you statically configure it within a specific range by hardwiring defined pins on the PCB. Just study their datasheets to find out. Needless to say, that every chip on one I2C bus must have a different target address to avoid conflicts.
If you use 2 TOF sensors of the same type and they do not have dedicated configuration pins, then they would have the same target address and could indeed not be used on the same I2C bus. You would need an additional I2C multiplexer (that’s how they’re called AFAIK) to logically separate them from each other. If however the TOF sensors you chose let you configure their individual target address and your display and DAC use yet other target addresses, you can safely connect them to the same physical I2C bus without conflicts or complicated management stuff on the MCU side.
Here it is!
I’m sure there must be a bug or so still, but I can’t find them atm.
Hope you enjoy, I’ve been needing this since I got the phone, maybe you have too? ![]()
Very nice! ![]()
This also applies to your sophisticated descriptions on OpenRepos. ![]()
And if you read that application note you will find that the bus management solutions from TI are listed in a table at the end. Those are the specialized chips I was refering to. Add to the eeprom and the actual device, management and you see where I’m going, I hope?
Most I2C devices have low power modes. Those have to be poked on and off. This is kind of vital on a phone.
Yes, but by chance it’s often the case that different ‘classes’ of device have different ranges. The OLED devices I use 0x3C (0x3… range) and the Time of Flight sensors have 0x2… range. Since I have a design with two TOFs I need to drive the XSHUT pins to set the adresses which is fairly trivial. The point is that the design from Jollas’ side is expecting a single address to manage communicating with the device. This is unlike working directly with a MUC which has one or more I2C busses. In a sense I’m being lazy, but the design of the TOH is complicated by having to take on the address management in hardware instead of doing it from the software/driver side.
EDIT: What it really boils down to in the end is that prototyping something that is interesting to me will become too expensive without doing like a kickstarter. Even with JLC with the part count I’m looking at, we’re talking about > 40Euros a pop with a minum of 5. And that’s before having an enclosure.
Heh, I was curious to see if it runs on 5.1 experias and vollas. Seems to work correctly on the Xperia 10III but is missing bits on a 5.0 volla22.
Have you compared with a light meter?
It’s just, I had to re-calibrate the jolla phone readings by 4 stops to get close to my real meter, this might be off for any other phone.
Good to know, anyways, that it works for more phones than just the new Jolla! ![]()
No I have not. All of my analog cameras are Polaroids that I use for film transfers
Metering was never of much import.
Someone will have a meter to hand
@ric9k ?
Yes, here.
What…? oh ok, test the app against meter values?
(so monotasked on mad_dev’s RAWfish atm, that I forgot other threads…)
I’ve got a load of light meters, but don’t see where there is any lux value in the app?
There isn’t, any lux, you can find the EV on the measure button once you measure. But lux in the app name stands for light, as it fiat lux - let there be light. So it just suggests setting for your camera. It is no incident light meter ![]()
EV is lux on a base-2 log scale with some constant for calibration depending on the meter. EV is used since it accounts for shutter times vis-a-vis film speed. With a standard (for manual meters) constant it’s 2.5 × 2^x to get the approximate LUX from EV.