LED Display Brightness Question

I’m working on a 7-segment, 0.56 inch tall, LED red display. The display must be readable in sunlight.

I’m looking for ideas to improve the segment brightness.

Should I consider another color of LED display?

What about putting some type of faceplate over the display to improve contrast? I’m thinking about the technique used on very early TI calculators. I don’t know what type of faceplate was used to make their displays visible under high brightness conditions.

What about a high current, low duty cycle drive to the display segments?

A PIC 16F676 multiplexes the three characters. Its limited to a source or sink current of 25 ma per pin. I’m using 220 ohm resisters to limit current from each display segment. Actual current draw is about 20ma on each segment.

Any other ideas?

Some HE red displays can be seen in daylight.

If you are multiplexing then you need to increase the current through your display because each segment is only on a fraction of the time. Check the datasheet for the maximum pulse current of the display and use transistors to increase it to what is allowable.

Mike.

Displays are Kingbright SA56-11SRWA

They run 4700 to 24000 ucandles so the displays are high intensity. The peak forward current is 155ma for a .1ms pulse width.

Ha anyone information on an optical cover. Should it be a polarized screen?

If the allowable peak current is 155mA and your running them at 20mA then they won't be very bright. Covering them will make them less bright. Just increase the current but make sure the average current is within spec.

Mike.

At 155mA you're not going to be able to drive them directly from the PIC. How many digits is your display?

Using a thin,tinted red plastic sheet should improve the display contrast.
I have used that solution in a few projects and it does make the display more readable.
Display filters are available from component suppliers but are very expensive.

Ideally you need it to be matched to the led colour but you could try getting an offcut sheet of tinted red acrylic sheet from places such as ebay.

Here is an article about the effect:
https://www.visionteksystems.co.uk/contrastenhancementfilters.htm

bobledoux said:

The display must be readable in sunlight.

Click to expand...

I put mine under a hood but it limits the viewing angle.

Three digits in the display.

What are you using for a column driver? Check out my Dragonfly schematic for NPN drivers.

Contact Texas Instruments and inquire about the opaque filter used in their calculators.

Use segment driver transistors (or IC) as well as column drivers. If 20 ma provides "full brightness" when at 100% duty cycle then you'll need approximately 60 ma "peak" current when running them at 33% duty cycle.

You're dropping too much voltage across the segment current limiting resistors (0.02 amps * 200 ohms = 4.0 volts)... 10 ohm resistors are probably more appropriate... or... eliminate them entirely... You'll still need segment drivers for full 60 ma "peak" current and full brightness...

Good luck... Mike

I realise you are really after using a LED display but have you considered an LCD 7-seg digit display? You can get transflective types that are easily readable in full sunlight, with a backlight that enables them to be read in total darkness.

Seven segment type LCD displays tend to be TN, rather than STN which now days gives them a great viewing angle as well as high contrast (only a few segments to drive rather than a dot-matrix type setup).

Not only that but with no backlight on they draw hardly any current <2mA.

I thought I better just mention it just incase it satisfies your application.

As for increasing the contrast of a LED display, the previous answers have covered alot, but I haven't seen anyone mention circularly polarising filters. These can block light that has gone through the filter, then reflected off of the display (when the segments are off) but allow any light created behind the filter to go through..sort of like a one way system..light can only pass through it once, reflections are strongly attentuated.

Even if you increase the current of the LED's, you are increasing the maximum brightness, but not minimising the minimum brightness when segments are off (sounds wrong but you know what I mean).

Also, using a display with a grey front rather than black would help. Even though it reflects sunlight tmore than black, it matches the colour of the segments when they are off, helping the eye to distinguish between off, and on segments.

Hope that helps.

Blueteeth

Thanks for the ideas.

The new display is supposed to replace an LCD display based on a 3 1/2 segment voltmeter. It needs to be read at a distance of about 8 feet. The LCD digit size and contrast is small making for reading difficulty.

The LED's are intended to improve contrast and readability. In normal room light there appeared to be little change in brightness as I reduced steady current flow from 15ma down to 5 ma. But at 15 ma the display washes out in bright sunlight.

The application calls for reading temperature from 0 to 255 degrees F. I have also considered using a 2x8 LCD display and creating custom characters so each 2 across and 2 down characters would be lighted for a single character. This would give me my three characters with vertical pair of LCD characters for each space.

The system is battery operated; currently wired on three AA cells. With the heavy pulsed current draw, when using segment switching transistors, I'd want to include a big cap in the supply line to smooth out current draw voltage spikes.

My solution sequence is as follows:

1. Wire the system up, as it stands, and see if the problem with visibility remains.

2. If there is a problem, add a red filter over the display and see if this solves the problem.

3. If problem remains, create new PCB with high current pulsing of characters and segments.

4. If all else fails, look for a big 2x8 LCD display and make characters that are two rows tall by two characters wide.

Thanks again for the feedback.

Hey again,

Sounds like you've got a good plan there. A 'solution sequence' is very sensible, I can't tell you how many bits of paper I have laying about my desk with bullet points, or numbered solutions for a problem

I assume you have already achieved the interfacing with the LED display part? It suddenly hit me whilst I was reading that, most 'LCD volt meter' modules drive the LCD display direct, which is very different from LED displays. If it has teh ability to drive an LED display, or you've wangled a way to rig up the LED display then thats all good.

One little thing I just thought of..coudl you use a 'green' LED display? The human eye has its highest dynamic range around the 532nm part of the spectrum, giving you the greatest sensitivity between light and dark...a lot more so than red or blue (blue is crap for display reading)

I don't want to make your life more difficult by adding *more* options, but brainstorming is what I do. Assuming you are 'fairly' well versed in microcontrollers, which I can only assume you are since you mentioned a 2x8 LCD display which at the least would require basic initializing with a micro...

4a. Use a 132x32 graphic LCD display. The microcontorller would have to store bitmaps for characters, but could easily 'draw' the characters. Plus you'll have no gaps in them, as the display is graphic. I've recently become quite fond of these little graphic LCD's. More hassle than character LCD's, but not as memory hungry or scary as their bigger 320x240 cousins. You can create very neat custom symbols/characters, without the restriction of 'character spaces' plus fill the entire display making the digits/characters as large as possible. Hassle I know, but its doable.

Should you decide to go down the 'LCD module' route, and assuming this is a one-off..I can highly recommend ebay. Theres usually lots of cheap 'large character' displays, as well as countless mini graphic ones. I've recently picked up some serial 7-segment LCD's (12 digits) which are great for apps like this.

However you decide to go, some interesting reading:

**broken link removed**

**broken link removed**

(3M will happily sent you free samples).

Thanks for the feedback. I'll try green LED's

The temperature gauge is for a hot air balloon. The gauge is a box 4" by 3" by 1.5" deep, mounted in the bottom of the fabric envelope. A wire pair goes to the sensor, which is a 3904 transistor wired as a diode, 50 feet up inside the envelope.

The system is operating. I'm using an a phono jack/ plug for the sensor connection.
The PIC 16F676 drives each segment with three pins for anode drivers. The individual numerals are defined as a table in the software. An LM358 op amp converts the PN junction into range for the 10 bit A/D converter in the PIC. Two 10 turn trimmers set a two point calibration for the sensor.

Three AA batteries drive the unit. So far, I've avoided using voltage regulation as experiments suggest little variation as batteries age. If voltage control is needed I may use the Spark Fun DC to DC converter that provides five volts out from two AA cells.