LED drivers

@foxy My initial thought is series wins as it eliminates the current variability then just pick a voltage that provides the light intensity you want, it could even be a pot as long as it was limited to the acceptable safe range for the specific color LED.

@zander, @foxy - Actually, I think you design for the current for the light intensity.  For small strings, I use this http://ledcalc.com/ .  You pick the "current limiting" resistor based on how much current (and brightness) you need.  Although a certain colored LED might take 20 mA, it will be fine for just an indicator light on a panel at even 5 mA current and will last longer than at its limits.  

Now, I've never tried this calculator with even 20+ LED's... but then with those numbers I'd just use the cheap strings of LED light you get for $12 that includes the power supply, computer controlled dimmer/color picker and has current limiting resistors for every couple LED's.  No fuss.

@inq I use a different calculator but same idea.

After looking over the data sheets for these things it almost looks like they were designed to be impossible to connect them in series so I'll have to give up that idea. So back to the drawing board.

Hi @foxy,

 I haven't tried it, but 1 current source with 50 LEDs in parallel sounds like a design disaster. 'Single colour, Identical' LEDs are very unlikely to be 'identical', so that some will 'hog' more current than the others.

A serial chain of single colour Leds powered from a constant current source might work ... and is obviously trivial to repeat for three colours.

But if you wish to use a 'basic' RGB LED, with either common anode or common cathode, then as you have already realised, it is not suitable for building into serial chains.

This seems to leave possibilities in which each LED has some form of 'personal' current control, such as:

• The addressable LEDs (WS2812 and similar based), where each RGB LED is individually controllable
• A variation on your 3 power supplies and LEDs in parallel suggestion, in which LED has its own series resistor .. on the basis resistors are very cheap and the surface mount ones can be small. Hence one of each pair of parallel wires would go to the LED, the other to the LED's resistor.

Our excellent host, Bill AKA @dronebot-workshop has produced a video and blog note on the WS2812 which is well worth a watch and read.

I think there maybe other discussions on the subject within this forum that may also have some valuable info.

They seem to be available in a number of formats ... including tape-like strings and individual RGB LEDs, wired like Christmas tree lights.

https://dronebotworkshop.com/rgb-leds/#Addressable_RGB_LEDs

https://dronebotworkshop.com/?s=led

and the first example I found of Christmas tree lights... (NOT a recommendation, just shown as an example)

www.amazon.co.uk/WS2811-Pixels-digital-Addressable-String/dp/B00MXW054Y/ref=asc_df_B00MXW054Y/

The video is about 4 years old, so what is on the market may have changed a bit ... I recommend a serious 'Google session' looking for the best product, if you are thinking about following this route.

NB, using addressable LEDs means you only need a single voltage (5V?) power supply, that can supply enough current, plus an Arduino or similar to control it.

NB2... I recall another forum member had problems directly soldering to the LED strip ... you may need to find matching connectors, etc. if you use them.

Good luck .. maybe someone else has a smart idea! Best wishes, Dave.

@davee 

Thanks for your thoughts on this.  Actually this relates to an on-going project with my daughter who is a physicist working in the crystalography / gemmology field.  The current effort is to get an improved dark field illumination for a microscope.  The old traditional method, probably invented right after Edison invented the light bulb, uses an incandescent light bulb and reflectors  which gets hot as a firecracker (and no colour control).  We have already put together a few versions using the neo pixel ring and strips and machined plastic mountings which give low heat and controllable colour, both very desirable features but rather more stray light than we like which makes the dark field illumination sort of grey field.  The neo pixel strips could probably be made to work well if they would bend edgeways as well as flat but this is impossible.  The current effort is to mount multiple individual RGBW LEDs with the light aimed to a point just above the stage right on the subject.  This will involve careful machining of a block of transparent plastic to mount the LEDs  plus the LED drivers which will use one of the Arduino family controllers. 

I"m just hoping I can come up with a good control using a current regulator and power amplifier to drive 40 or 50 LEDs of each colour in parallel. I appreciate any thoughts on this.

Hi @foxy,

Interesting and worthwhile project, and I can appreciate at least some of your issues.

--------------

LEDs tend to towards omni-directional emission, whilst you might be looking for a more directional light source.

I don't know about gems, but in other microscope based studies, the ability to move the light source, relative to the sample, can be useful. e.g. being able to observe light reflected from the top of the sample as well as light transmitted through the sample. I think such light sources tend to providee collimated beams, to minimise the interference from scattered light.

If you are looking for an extra challenge 🤩 , you might consider moving the LEDs a little further away from the stage and using mirrors and/or lenses to create a collimated beam, possibly with the ability to aim it at the specimen from different angles.

------------------

As for driving the LEDS themselves. If you are thinking about 'bare' single colour LEDs, consisting of just the diodes itself, then I think each diode needs its own current control in series. The 'current control' can be just a resistor in series. Then the combined diode and resistor circuit forms a 'new' component, which can be replicated as many times as you like, and these 'new' components can be paralled, which can be driven by a single controllable supply, which could be either voltage or current controlled.

This same principle, could be extended to the 'bare' RGB LEDs which are just 3 LEDs with either a common anode or common cathode.

If (say) they are common cathode, then create a 'new' component, by connecting a resistor to each of the anodes. This 'new' component can now be replicated as many times as required. The resulting parts can then be paralleled by:

• connect all cathodes together
• connect all free red LED resistor ends together
• connect all free green LED resistor ends together
• connect all free blue LED resistor ends together

Now have three circuits, each needing a controllable supply, as before.

Note using common cathodes, implies the three power supplies have their negative terminals connected together, and the three positive terminals will go to their respective block of LEDs.

If the LEDs are common anode, then swop cathode with anode, and positive with negative in the preceding description.

........

I presume from your reply you have already discovered that the 'smart' LEDs with WS 2812 or similar controllers can also be purchased as individual components, albeit in pack quantities of say 50 or 100. If not, I saw some in AliExpress. Obviously these will need some careful soldering, but might be useful. I presume they are electrically identical to those in the strips.

Good luck to both you and your daughter in your respective projects. Dave

Thanks for your thoughts to DaveE and all others who have kicked in on this.  I think I've found out what I was looking for---Your (and other's) thoughts are about the same as mine and there seems to be no magic solution; just the brute force type.

foxy