Sorry about the wait, ive been busy with uni, so, so, so, soooooo busy, but i have been researching.
let me recap the final step of the backlight monitor driver issue, taking the signal from the monitor and having the microcontroller read that;
the cintiq generates a negative voltage analog signal, i have no clue why, but it does, possibly for safety reasons since this signal influences an oscillator circuit that generates high voltages, so little blips of high voltage will trend towards zero and short rather than overload the oscillator. maybe. anyway, reading a negative voltage isnt a simple thing, the first and most obvious solution was to find a negative voltage supply to run the microcontroller off of, if the input and output of a boost converter are totally isolated then ground is whatever pin you decide is ground, but finding negative regulators is hard and they come with some drama to boot.
Next solution which i had to learn about before trying to apply in uni, was voltage dividers with a pull down. the negative voltage signal would be used as the ground refference, so, id use the 0 – -5v signal, once divided, to pull the signal up and down by 2.5v. but there was issued with stability, and i had to protect the microcontroller from the negative voltage, there were too many unknowns and a lot of instability, and in the end i found out too the ADC of an ESP is not totally suited to this task to begin with, without some modifications.
But finally i found a solution which is just totally perfect. I will be using an external ADC, the ADS1015. its a VERY high resolution analog voltage sensor essentially, costs about $3, but it has 4 channels which is the important part. ADCs cant read negative, thats a given, but what you CAN do, is rather than comparing ADC-ground, you can compare the difference between ADC1 and ADC3. this can be used to safely read the signal as-is, with high stability to the extent no software “smoothing”. just for refference, even averaging out 3 seconds of signal input on an ESP8266 resulted in oscillations which would cause the backlight to flicker. But using an external ADC chip none is required, aside from a capacitor on the actual signal itself. This solution is one that either did not exist when the project begin, or took some time for the pricing dropped before it was viable, or perhaps i was simply not educated enough to know to look for it, im not sure, but i found it now, and its very much plug and play, and essentially infallible. Im currently away on holiday until around the 13th of jan, when i get back i will be resuming and finishing the project before my semester break ends, to finish this once and for all.
After the 24HD backlight driver is done, i have a 22HD i obtained cheaply which will then receive my attention (inverter is fine but its LEDs are blown, so will be doing an attempted fix of the LED strip, or, replacing entirely), as well as some really cool other projects which i will post about in the next week or so.
Some have asked if i will be making and selling the backlight drivers premade, plug n play, the answer is yes, i have actually been prepairing for that among other things, thinking about and learning circuit fab. In the end an ESP is the ideal microcontroller for this just for the singular reason that, i can make 2 sets of code, one for a full featured board, either my kit, or a self-build of it, and then the other will be for an “open loop” version, which does not attempt to sense the backlight signal and instead relies on a web interface if you want to change how bright the screen is, so you just log into the ESPs own little wifi, go to its home page, enter the % brightness and thats that. taking the “on” signal of the monitor is a no brainer so this, unlike how my monitor is right now, would still turn off when turned off, it will just have less convenient brightness control, but it doesnt require anything more than the ESP microcontroller, stepdown and LEDs, plus a little wire, very simple and graspable for anyone.
I was considering initially to use an STM based microcontroller since for voltage sensing and, basically everything this project needed it was perfect, however, the silicon shortage hit hard and caused even the $2 bluepill pricing to skyrocket, and consequently, the STlinks also went up, meaning the buy in price for just one STM microcontroller will be something like $20-30 depending on where in the world you are, for something which last year cost $2-5.
My goal is to finalize this project once and for all before the end of Jan. finalizing the project consists of the following milestones:
- build (or rather, technically just assemble at this point) and install a fully functional backlight driver into my 24HD cintiq
- document the construction of the final driver which includes some more detailed dissasembly footage with VERY thorough photography.
- produce a final build instruction set for remaking it yourself, a parts list, and code
- publish code for a “lite” version where you change brightness through wifi instead of reading the signal generated by the monitor.
before the end of feburary, ideally, i want to do the following, i dont know the timeline so, im uncertain if i could, since a lot of things will be done for the first time like;
- design a PCB schematic file for the replacement driver
- self-etch or have the the PCB manufacturered (timeline for either is uncertain, iv never sent schematics to be manufactured before)
- find the name of the cable connector the wacom uses internaly so nobody needs to do DIY cable splicing and can literally just go plug n play in the most literal sense . finding these is very difficult.
- I will begin reselling self-made cables with labels attached if it turns out i can only bulk buy or only get un-wired connectors, or at very least just providing them at-cost to anyone who asks me for them, besides of course providing its name/part number for anyone to attempt to self-source.
- finally, i will attempt to list the first batch of complete drivers for sale, bearing in mind though they are first batch/for testing. revision 0. I dont forsee how they could fail since the PCB will essentially be a sort of shield, or just mounting board rather than wiring all the bits together, but who knows. all it is is just a circuitboard to replace a wiring loom, but still, the first batch will be sold conditionally like a sort of beta testing.
In march ill see what improvements, if any, can be made, then if all looks good, ill make a bulk-order of the driver PCBs and assemble, as well as just start selling the circuit board with self-assembly kits (including the female connector so no wire-cutting is needed).
this timeline i think is fairly reasonable, as of now, besides the fact im not at home for the next 2 weeks, i have everything that will be needed already sitting and waiting. but the 2nd stage, things to be done in February, will be dependent upon transit and processing times of entities other than myself, ive been told it can be quick, and done locally, but, its something im unfamiliar with. If time ends up being an issue i do have the capacity to self-etch a few of my own boards too, but thats all extra stuff for consumer products. If this doesnt go as fast as id like, i do have the ability to make semi-assembled plug n play connectors for the kits that will be a little messy but, will be essentially impossible to muck up yourself, because for this project, not to downplay the competency of the end users, we are talking about artists, people who for a living have mastered art, which is arguably the polar opposite of electrical engineering, working on their, usually irreplaceable $2000 art monitor which is sometimes the center of their income. even i myself felt very unnerved when i first cracked open the monitor, so imagine how it must feel for someone who probably doesnt even know what they are doing, but NEEDS to do it. with those people in mind i plan to make the repair as simple as possible with the smallest margins for error that i can manage.
So thats that.
Later ill also be posting some updates on other projects i have going or planned. Im currently visiting the fam up north, and half the time im bored as heck, storming up SOOOO many ideas i want to do once i get home again. Among them will include beginning the official Ronox engineering youtube channel. the final documentation/ dissasembly and backlight construction and installation/reassembly video will be posted on the new channel, to kind of christen it, and while your subscription would be appreciated, for now theres no need. In case anyone is wondering why im changing channels/accounts, its because my account is, in a way, corrupted, or broken, stuck with certain settings i cant ever fix, and honestly, monetization is my goal, thats something you need to maintain, so having any manner of head start wont really make a difference, its just as hard to go from zero, as it is to start with 20% of the subscribers and view time necessary. Also i may have bungled the new account too calling it ronox engineering, instead of just ronox, which i might not be able to change, so theres that too, youll be seeing more of me soon and ill soon be getting VERY active, i have too many plans now to remain idle.