2 LEDs is tricky, because it means the illumination isn’t uniform any more; arranging several LEDs to cover a wider area (or even using a matrix of LEDs like Adafruit sell) is an option, and indeed people have published some cool results making synthetic images by illuminating one LED at a time and combining all the resulting images.
The lens in the illumination does two important things:
- it concentrates the light from the LED into an area just a little bit bigger than the area you’re imaging, which means
a. you improve contrast because there’s less light hitting the objective from outside the area you’re imaging
b. you improve brightness because you’re not wasting much light
- It increases the numerical aperture of the illumination, i.e. the light converges on the imaging area from a larger range of angles.
The lens means you’re illuminating an area around 1mm across - so it’s concentrating the light by a factor of hundreds or more. So just adding a few more LEDs will improve brightness but not by anywhere near as much. However, if you have a fairly bright LED, brightness is often not what limits you, it’s stray light (which you really can’t fix without a lens) and numerical aperture. You can fix the numerical aperture problem by using a diffuser instead of a lens - try sticking some white electrical tape over the aperture where the lens should go. It will hurt brightness even more, but should improve image quality if you have a bright enough LED to start with. You’ll want the condenser fairly close to the sample.