Thanks jkwilborn.
It's been a real adventure and education!
The buttons on the EL panel allow for 3 different levels of brightness rather than a continuous dimming so a finger arm arrangements might get to one of those levels but not those in between.
To give some background I'm an astrophotographer and one of the data calibration steps is to remove vignetting and other artefacts from the images caused by the optics. This is done by taking what we call 'flat' frames that are essentially exposures of uniform light that are subtracted from the actual image frames. These can be done by pointing the scope at the sky during twilight but most people now use the likes of EL screens for the purpose under programmatic control. Depending on the aperture size of the scope these can cost anything from about £150 upwards! So I figured I would build my own if I could get a cheap light panel and figure out a way to control it from an UNO.
Well since my last post I have played with various solutions. Forgetting about the problem of the switch on the panel I began to learn about basic electronics in general and how to control circuits that potentially require more power than than the UNO can chuck out. I played with the PWM pins connected directly to the panel via the supplied cable. I could turn the light on at 100% duty cycle but it would turn off at anything below that so I assumed it either didn't like PWM or the switch electronics was scuppering things!
Assuming it didn't like PWM I explored the possibility of supplying it with an analogue signal and put together a low pass RC filter. I also bought a USB oscilloscope so I could see what effect the filter was having on the PWM input. Amazed at how well that worked! I wrote a sketch to send a sin wave duty cycle to the PEM pin and observed that on one probe of the scope and the output of the filter on the other. After playing a bit with the R and C values I got the most perfect analogue sine wave! I know about Fourier analysis but seeing an application of it work so perfectly was awesome. I also bought a cheap DAC converter an MCP4725 to play with which gives 4096 levels as opposed to the 255 of the PWM output although 255 is more than I need for my purposes!
Plugging the analogue output of the filter into the panel did not solve the problem, the light would come on at max values but turn off at anything below that. So it clearly wasn't an issue with the panel not liking PWM.
I then figured I would explore powering the panel from a separate circuit that was turned on and off by the PWM signal from the UNO or the output of the low pass filter. This led me to sign up for a video course in basic electric theory so I explored Ohm law, Kirchoffs various laws, Equivalent circuits etc etc. Then I studied the transistor and played around with it acting as an amplifier and a switch. Then I figured I would try a cheap op amp so did a quick course in those and made a unity gain buffer, voltage follower using negative feedback to the op amp. Explored all sorts of things to do with op amps. Then I realised I had a number of Darlington chips lying around that where used to control a small motor and they had the necessary power and signal connections so I could use one of those to isolate and control the panel. At each stage I explored the various IV relations across the source and load circuits.
So I now had various ways I could control a load. Only problem was the 'load' still wouldn't play ball with that switch in the way. So I thought I would have a closer look at that and started to pull the back of the panel. This peeled away quite easily exposing the circuit of the dimmer which I'm guessing uses a capacitive touch pad alongside a PWM generator. Anyways the wires out to the LEDs were clear to see so I desoldered them from the dimmer circuit and connected them to my Darlington circuit. I played with both analogue and PWM signals from the UNO and both worked perfectly!!!! The brightness response was linear as well across the 0-255 range of first stage PWM. Awesome.
So i figured I would go back to stage one and see what happened if I connected the UNO PWM output (pin 6) direct to the panel circumventing the buffer circuit and that worked perfectly as well with apparently no overheating or any other side effects - I have the Uno connected to a 2A powered USB hub.
I have since done the same 'mod' to a second panel and have that running directly on pin 3 with no complaints from the UNO either. They are both controlled via ASCOM drivers - an observatory equipment control standard - that I have written.
So what a cracking journey I have had. I've learnt so much I'm glad I didn't just rip the back off at the beginning - although at that point I would have had no idea what I was looking at or why this might actually work.
John
It probably controls the pwm of the screen 'enable' by the internal micro who 'read' the switch activation's. That's a guess and that's where you going to stay until you confirm some of these assumptions.
