Light Canvas - LEDs 6: Dim power

The previous post demo-ed all the LEDs roughly in place in the frame, running on the dimmer circuit - but not running bright enough. I've figured out (I think) the LEDs aren't getting enough power. Firstly I wanted to establish what is happening in the circuit:

• From it's specifications, the 2N222A transistor has a beta value of about 100
• The current applied to the transistor base (555 chip output current) measured at: 1.16mA
• Taking the above, 100*1.16mA = 0.116A
• This calculation is confirmed by the measured of current going through LEDs (via the transistor): 0.1A
• This 0.1A is a quarter of the current needed to fully light the LEDs - measuring the current when driving the LEDs directly from the power supply gave a reading of 0.39A (pretty close to the calculated value of 0.42A)
• The transistor's peak base current is 200mA, so the max emitter current (base current * beta value) should be 200ma*100 = 20A - considerably more than the 0.4A needed.
• It should be possible to get the 0.4A - I just need to figure out how to get more current into the base.

Answers on a postcard...

Light Canvas - LEDs 5: All the LEDs

Here are all the LEDs running - 12 panels, each with 5 LEDs. They're not fixed in place yet, just taped in the approximate location.

It turns out that the 555 dimmer circuit isn't quite right, as it seems that the LEDs aren't getting a large enough current - 0.16A, rather than 0.45A (when powering all the LEDs without the dimmer).

I guess it's something due to the transistor not putting/allowing enough current through...

Video:

Light Canvas - LEDs 4: Progress!

Having important things to do can be really motivating - for other projects. I've household repairs to organise, but can't figure out an elegant way for it to be done - so I'm ignoring that for the moment, and have made some progress on the Light Canvas :-)

Subsequent to my last post on this project, I satisfied myself that the circuit was ok, and ordered a bunch of components, and took the opportunity of a OxHack session to  breadboard the control circuit, which I did - except for the minor point that the circuit didn't actually work... Oh well, it was a nice event, and good to meet some more hackers and see what they were up to - I need to get along to more of that sort of thing!

Many months later, I've a working breadboard with the 555 astable circuit producing a pulse width modulation varied from 0-100% depending on a potentiometer. The 555 PWM doesn't have enough power to drive the LEDs directly, so it's used to control a transistor which effectively amplifies the power of the PWM enough for the LEDs. W00t!

This happened by:

1. Getting the 555 PWM working on a single LED directly,
2. then on a panel of 5 (on a breadboard) via the transistor, then (the next day),
3. making up 9 perfboard versions of the LED panel, powered via the LED.

Video evidence:

Next steps;

1. Convert the 555 circuit from breadboard to perfboard.
2. Modify the frame to house the circuit, controls, and wiring.
3. Modify the frame so the back can be attached, and hung on the wall.
4. Stretch the canvas onto the frame.
5. Install the electronics in the frame, and put it all together.
6. Put on wall and bask in its glow :-)

Light Canvas - LEDs 3: Circuit development

It's great what you can learn with some time and t'interweb - when finding time between band, family, and the rest of real life, that is...

This LED wizard (and generally informative site) helped me realise that "forward voltage" is the same as "voltage" as generally used, but relating to correctly wired in components, rather than what's coming out of the power source.

My previous tests were under-driving the LEDs not by current, but by voltage: A 12V source isn't enough to power four 3.3V LEDs - obviously 4x3.3 = 13.2V.

So, rather than four LEDs in series, I can have up to three. Three seems too few though, so after trying some different combinations, I've decided each column will have three 5mm in series, in parallel with two 3mm LEDs in series. This will add up to 420mA, which at 12V is 5.04W.

After a bit more reading I realised that just slinging a potentiometer between the power source and the LEDs wouldn't work (safely at least), as regular potentiometers won't handle power around 5 watts.

Some helpful people on the sci.electronics.basics Usenet group answered my query about what to use instead, and the consensus was to use pulse width modulation - a PWM circuit based on the 555 timer.

Although the intention with the light frame was to do something without that level of complication, it's just accelerating my schedule a bit, as PWM is something I know I'll need for other projects I have in mind.

Here's what I've got so far:

That schematic is missing quite a few important details - like values of the capacitors, diodes, and resistors surrounding the 555 chip. Oh, and there'll be 3 times the number of LEDs.

I've not yet learnt exactly why this circuit should work - specifically how the current flows around the diodes - but the basic principle is that the discrete components around the 555 tell it how long to turn pin 3 on for (the pulse width), with the variable resistor (R3) allowing manual modification of that duration.

Pin 3 goes to the base (B) of a high power resistor, telling it when to allow current to flow between its collector (C) and emitter (E) - acting as a switch between the power and the LEDs.

That gives us: a small (safe) current through the pot, making the 555 turn the transistor on & off, which varies the power going through the LEDs - controlling their brightness. Hurrah!

Now I need to satisfy myself that I understand all the circuit, and figure out the values for those discrete components to get the timing I want, then I can figure out a list of what I need to buy (AKA Bill Of Materials).

There's a possibility that I'll revisit the LED/resistor configuration to see if I can make it more efficient with a higher voltage source, but it's too late in the evening to think about that now :-)

Light Canvas - LEDs 2: Choosing and Testing

After shopping around a bit online, I bought 100 3mm and 100 5mm warm white diffuse LEDs - buying in bulk for this project and the next. As well as size, the two models differ in viewing/beam angle, and brightness. Two types to experiment with is manageable, and should let me play with brightness/diffusion.

One hundred pounds worth of LEDs don't come in a big box...





(I should've included something for scale in those pics. Nevermind.)

Once I had the frame put together, it occurred to me that three or four LEDs could be put serially in little columns on perfboard, each with their own resistor and connection to the power supply - something like the plan on the right.

Lets remove some unknowns: I want it bright, and nine columns of four LEDs should be pretty good, I reckon.

I've already got wires, resistors, and a breadboard so now I can do some testing.

For prototyping, I attached a couple of crocodile clips to a mains adaptor at it's lowest setting and ran that through a breadboard with one 5mm LED and the recommended resistor, to check that the basics were working, then bumped up the voltage to 12, as I plan to use in the final piece. That worked fine, so I put four LEDs in series and, using an online LED resistor calculator, found a (lower) value of resistor, accommodating the greater voltage drop caused by the additional LEDs (assuming a 2V drop per LED).

If I'm to learn anything here, I ought to figure the resistor stuff out myself, rather than using the online calculator.  For the moment though, I think I get the gist:

• Power source provides a electrical potential difference,
• generation of light by each LED is an electrical load, which uses some of that potential, hence
• voltage drop gives a lower effective voltage over the circuit, hence
• a lower value resistor is necessary.

I need to do some testing to see how this fits in with Ohms law so I understand it a bit better - but not right now, as there's more hands-on stuff first!

I was quite happy with the 5mm LEDs in a row of four, so did the same with the 3mm LEDs. Here's the breadboard:


...here it is lit up:


...with 80gsm paper as a diffuser:


...and with 160gsm card as a diffuser:


Of course, those pictures don't really convey the quality of the light - although they're HDR (taken with Bless N900) so they're not utterly terrible.

I'll need to play around a bit to see if I can/should use lower value resistors and get the LEDs brighter - they could take a higher current, but I need to verify that the voltages and currents used aren't going to be a fire risk, and look into whether higher current might give a significantly shorter LED lifespan - although this is probably a good point to find a canvas (or some similar looking material) to go on the front...

Light Canvas - LEDs 1: Known Unknowns

This is the more complicated part; or at least I have little idea what I'm doing when it comes to electronics. I know basic theory (e.g. Ohms law), and what various components do, but - so far - no practical experience beyond giving glowing red eyes to a toy raven.




The electronics for this shouldn't be too tricky though, as it's all simple analog circuits, with no digital components. The plan is to have:

1. A mains power source (I don't want the hassle and inefficiency of batteries), connected to an input socket in the frame, which itself connects to...
2. An on/off switch, which connects to...
3. A linear fader, to control brightness of...
4. A bunch of LEDs, set up to give an even glow through the canvas (some subtle pattern would be acceptable though).

So, onto the known unknowns:

• What voltage and amperage should the power source provide? It'll depend on the number and specification of LEDs  it'll be supplying - which I don't know yet. I'm thinking 12V is a fair place to start, and conveniently I've got a 3.5 - 12V adapter to prototype with.
• I've got a little toggle switch of the type I want to use, but don't know what power it's rated for.
• The fader could be a potentiometer - or might need to be something else, if the electrical load is too high for a pot. I bought a mixed bag of pots to play around with a while back, and it contained just one linear fader - hopefully it'll be suitable.
• LEDs - this is the biggest area of uncertainty. There are lots of LEDs with differering specifications available, and I don't know which type and how many I'm going to need for the canvas to glow like I want it to. Aspects to consider:
  
  • Brightness (enough to shine through the canvas)
  • Viewing angle (wider is better, as I want an even glow, not visible bright points)
  • Number of LEDs to use (more at lower brightness would give more even illumination)
  • Clear vs diffuse finish (at last, an easy one - diffuse)
    
  • Configuration - rows and columns, random distribution, or some other pattern?
  • How to choose and wire up the right combination of LEDs?
  • Cost - no need to start taking out loans for a hobby...

I think this a fair example of too many choices, not enough knowledge, so I decided that the only way to get started was to just buy some LEDs and fit everything else around that.

Time passing whilst I researched LEDs will be represented by this post ending, and a new one starting...

Light Canvas - Frame

I tend to think about all the different aspect of a project (in this case, frame, LEDs, canvas) at roughly the same time, gradually figuring out how each part will work, independently and together with the others. That would be an utter nightmare to write up though, so these posts will mainly focus on one aspect at a time. Let's start with the easy one: the frame

This is relatively simple - I know the dimensions of the canvas print (135x45cm), so I know I want my canvas to be 135x15cm.

I little research suggests that getting a pre-built frame this size will be difficult or expensive. To be honest, most of the research was watching YouTube videos on stretching canvases. Woodworking is kinda fun, so I elected to make this myself.

After a while trying to think of a way to add a bevel to regular planks (to make sure the canvas doesn't have lumps due to the frame, I realised I could simply use a simple beveled architrave.

The frame is simply two horizontal and two vertical pieces, glued and nailed together with mitre joints, plus vertical braces inside for strength.


Cutting a mitre joint with a plastic mitre box wasn't very accurate - I ended up sanding down the cuts, and hoping that glue would make up for where the surfaces didn't quite meet up.


Positioning the braces was one of those things where it was totally unclear how many were needed and how to space them, until the four frame pieces were laid out on the floor, at which point it was obvious that two braces at thirds of the width would be ideal. The braces are at the back of the frame, so the shouldn't stop light bouncing around the inside of the frame too much - I want the whole canvas to glow, rather than having several compartments. Here's all the pieces laid out (but not glued or nailed):


Also in that pic is a large piece of fibreboard from the Ikea clearance corner - thanks to H and Bic for letting me bring it back with us :-) This will be cut down to size, have the LEDs mounted on, and will be screwed onto the back of the frame.

In keeping with family tradition, my place is full of tools and materials in progress during all this:


Ideally I would've used some corner clamps to ensure that the frame was square when I glued the pieces together - but I don't have any, and the local DIY shop didn't have any. Instead, I lined up the pieces to a roofing square that I bought for the cupboard project, PVA glued them all, then nailed them all (with pre-drilled holes). The braces needed a bit of encouragement to keep them in the right place whilst waiting for the glue to cure.


Cut the fibreboard to size, and that's pretty much all the woodworking finished!



This is all about white light, so the frame gets painted with a matt white, to reflect as much light as possible, without specular highlights.


I'll need to make some modifications to fit electronics, but the frame is done for now.

Light Canvas

Another new project! This one started around October, and is a result of three different threads coming together:

1. I bought a canvas print for my bedroom, a rather nice false-coloured X-ray photo of flowers (probably based on Nick Veasly's work).

2. I need a decent bedside light to replace the rather inelegant broken anglepoise I use at the moment - and preferably something dimmable so I have the option to introduce myself to the day gently.
3. My guitar lights project (more on that later) has been stalled for around two years, and I could do with a simple LED project to learn and gain some experience with electronics.

The predominantly black canvas immediately suggested a short, wide white canvas emitting white light. I think it could look very cool, and give a nice diffuse light to the room through a rough canvas texture. Including a fader will allow me to control how bright it is.


The basic design is:

• A wooden frame the same width as the print, but a third of the height. I was vaguely thinking about the rule of thirds when I chose that height, but mainly it was a good fit for the space between the bedhead and print.
• White LEDs in the frame, with on/off and fader controls, and power supplied via an external switched adaptor (e.g. a typical wall-wart). The LEDs might need some sort of diffuser so the canvas emits light reasonably uniformly.
• A white canvas stretched over the frame. Obviously, this needs to be thin enough to let the LED light through, but be thick enough to be durable. It should have some texture, for the aesthetic.

From that design, there are a lot of details to work out, starting with:

• What LEDs do I need? How many? How should I arrange them? What rating power supply will they need? How do I wire it all up?
• Do I really want canvas - what other options are there?
• How do I get the frame? Buy pre-made or construct it myself?

Stay tuned for answers to all these questions and more...