I have a project in mind that involves lights from a small LED (SMD would be nice) and i've come across this big problem but apparently there's no solution for it on Internet. I need that LED to be omnidirectional without using a reflector.

Ideally it should be only one (Maybe 2 as suggested) because its power source will be very tiny (A capacitor constantly being charged by electrostatic induction), it should light an led for a brief moment.

My main idea is to remove the mold from one (Leaving only the chip and connections for it), but that seems very complicated and i don't know if is even possible with chemicals.

LED structure

Is there a simple way of doing this?, maybe i'm thinking in the wrong way and there are options to buy on the market (Although i already searched). Thanks.


closed as unclear what you're asking by Chris Stratton, pipe, Nick Alexeev Jul 31 at 16:18

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    \$\begingroup\$ Why not just use multiple LEDs pointed in different directions? \$\endgroup\$ – Ron Beyer Jul 30 at 3:29
  • 1
    \$\begingroup\$ Do you mean omnidirectional in 360 degrees? The PCB will creeate a shadow. But some surface mount LED's are little more than an LED junction. Their radiation pattern is very wide. Nearly omni-directional in 180 degrees. If you put one of those on each side of the PCB, then it will be pretty close to omni. Don't answer in the comment section. Hit "edit" and update your question. \$\endgroup\$ – mkeith Jul 30 at 3:51
  • 4
    \$\begingroup\$ Many white LEDs uses a phosphor to convert blue/UV into white light. Make a phosphor ball with pinhole in it and use a Blue LED to activate it. \$\endgroup\$ – Jack Creasey Jul 30 at 4:55
  • 2
    \$\begingroup\$ sand-blast the domed-surface; that will cause radiation in all directions. \$\endgroup\$ – analogsystemsrf Jul 30 at 6:13
  • 3
    \$\begingroup\$ Where did you get that drawing? It suggests that LEDs have an internal reflector, which you are interpreting to mean that LEDs are inherently a spherical light source (like every other kind). They're not. \$\endgroup\$ – Harper Jul 30 at 15:20

Telling us what you wish to do in much more detail would be helpful.

I need that LED to be omnidirectional without using a reflector. Ideally it should be only one (Maybe 2 as suggested) because its power source will be very tiny (A capacitor constantly being charged by electrostatic induction), it should light an led for a brief moment.

You need to specify what you want more clearly.

  • By "omnidirectional, do you mean "half a sphere (2 Pi Steradians) or a full sphere (4 Pi steradians.)

    • A half sphere source mounted on a flat plane would be visible from any direction and height above the plane.
  • You say "to be omnidirectional.
    That is VERY imprecise

    • Do you wish to be able to discern the occasional pulse by eye? If so, from what range.

    • Or have it illuminate a volume?
      If so, how large. What surface.
      What brightness?

  • Does the illumination have to be evenly distributed across the while half-sphere or sphere?
    If not, what level of 'dropoff' at extreme angles (close to the plane) is acceptable?

LEDs of the type you have shown have an epoxy resin lens. Their restricted illumination angles are mainly a result of this internal lens.

Most surface mount LEDs have a silicone rubber encapsulant (especially higher power ones) and in most cases the radiation angle is extremely wide - often approaching the 180 degree half-sphere mentioned above. Light drops off at extreme angles away from vertical but the LED is still often visible at close to 90 degrees from the main axis.

You say "without using a reflector".
A simple 'lens' would do what you want.
A wide angle SMD LED (see below) capped by a dome of a suitable plastic (several are suitable) would give you a true half sphere emitter. Two LEDs and a clip together spherical lens would allow a light viewable from any direction.

LEDs are (or were) available that come close to your half sphere need.

Here is an example with a 180 degree viewing angle in X & Y directions.
They are obsolete, low power, inefficint and not available - apart from that one of these is visible across a half sphere and two would illumionate (dimly) a full sphere.

enter image description here

This 170 x 130 degree Osram LED comes close to what you want and is available. A few in a cluster would give half sphere of full sphere output. Output is down to 20% of max at 80 degrees off vertical in one direction and to 60% of max in the other.

In stock at Digikey $1.46 in 1's.

enter image description here

More candidates here

  • \$\begingroup\$ I've gone with the standard definition of omnidirectional. \$\endgroup\$ – uhoh Jul 30 at 23:31
  • 1
    \$\begingroup\$ @uhoh Yes. An interesting answer. Possibly not overly useful to the OP, but interesting :-). It would be good to know what he REALLY wants. \$\endgroup\$ – Russell McMahon Jul 31 at 1:56
  • \$\begingroup\$ pseudo-omni, para-omni or hemi-omni perhaps? ;-) \$\endgroup\$ – uhoh Jul 31 at 2:12
  • 1
    \$\begingroup\$ @uhoh All that. Maybe. But/and I've learned from too long experience that taking terms at precise face value when a question is vague often leads to the death of 100 clarifications. Hence my ATTEMPT to address the two most likely (imho) variants and my asking questions aimed at a decent degree of refinement / clarification. Does it work? -> sometimes :-). [2600 answers - still can't read people's minds with complete precision :-) ]. \$\endgroup\$ – Russell McMahon Jul 31 at 3:49

You can get LEDs as surface mount in very small sizes. Why not mount an array of them on a globe and use a diffuser?

Here’s a crazy-small one (0201 package): https://www.digikey.com/en/product-highlight/s/sunled/nanopoint-0201-series-leds

  • 1
    \$\begingroup\$ 0.65 mm x 0.35 mm x 0.2 mm - That's smaller than just "crazy small" to me... \$\endgroup\$ – Volker Siegel Jul 30 at 22:22
  • \$\begingroup\$ If I were confronted with this as a task for a product, I'd do something like lay out a flexible substrate that would be machine-assembled first, then 'fold' to a sphere-like shape. \$\endgroup\$ – hacktastical Jul 31 at 1:36

I put a dob of hot-melt glue on the lens end. I have a sort of cloudy / transparent version.

It destroys the focussing action of the lens, and acts as a sort of light-pipe and diffuser, and increases the effective light dispersion to about 300 degrees, until shadowed by the leads.

Purpose, Christmas tree decoration.


Omnidirectional LED, is it possible?


  1. Revive the business plan of Ball Semiconductor.
  2. Move the technology from silicon to some direct bandgap III-V material.

Each of steps will be extremely challenging and expensive and probably fail, but it might be "possible".

enter image description here enter image description here

above: from here and here from the (US) National Science Foundation

below: from Let's Fabricate ICs in Space

enter image description here

Or you could dice your direct bandgap material into tiny little cubes and build junctions on all six faces. The Lambertian distributions from all six faces added together would give you a fairly uniform omnidirectional distribution.

Python script for plot: https://pastebin.com/eQBMVbPz

The indices are the die normals, not crystallographic directions.

Lambertian distribution from a cube

  • 1
    \$\begingroup\$ Cool post. Need to power it somehow. \$\endgroup\$ – mkeith Jul 31 at 3:59
  • 1
    \$\begingroup\$ @mkeith BallSemi demonstrated function circuits, connected through bonding pads (those small balls in the first image). The OP says in the question that they will use electrostatic induction to charge a capacitor and blink infrequently, and that can all be intergrated i̶n̶-̶d̶i̶e̶ er... on-the-ball. \$\endgroup\$ – uhoh Jul 31 at 4:03

Every illumination lamp in history (except for lasers) emits light in a sphere. If you stick the lamp on a tall pole, it will throw light in every direction (except in the direction of its wire terminals/socket). We are talking about

  • Incandescent
  • Arc
  • Neon
  • Fluorescent
  • Low Pressure Sodium
  • High Pressure Sodium
  • Mercury Vapor
  • Metal Halide

All of them are spherical.

So it is understandable that you presume LEDs are also spherical. They are not. The spec sheets make clear what they are. Your drawing in OP shows a reflector; it may help that particular indicator LED reach its narrowbeam specification, but most illumination LEDs do not have a reflector.

One big thing in favor of LED is that 99% of the time, people do not want a sphere. They want a wedge. LED is more efficient there. In the rare case you do want a semi-sphere or sphere, you simply use an array of LEDs aimed in different directions.

  • 4
    \$\begingroup\$ Lasers have been around since the 1960's, so I would consider them part of history at this point. \$\endgroup\$ – The Photon Jul 30 at 16:14
  • \$\begingroup\$ @ThePhoton Sorry I wasn't clear enough; I meant ones used for illumination. \$\endgroup\$ – Harper Jul 30 at 18:24

Not the answer you're looking for? Browse other questions tagged or ask your own question.