Oversized 7 Segment Display

oracacle

Senior Member
Despite the fact that I still haven't quite finished the last clock I was asked be build (just waiting on one last thing to arrive), I have been asked to build another.
This time its a digital one with oversized digits. The display should be white in colour with each number being about 3 inches tall. should ideally run from AA batteries. This clock will be for some one who is partially sighted.

At this point I am thinking 10 LEDs per segment, there is no need for decimal place, but there is for a colon separator between hours and minutes, second do not need displaying.
It would be ideal if it was able to pick up the MSF (or similar) clock signal as well as keep it own time if needs be.

The first hurdle I think is getting the seven segment displays to work as a normal seven segment. I am thinking that a transistor (maybe a mosfet will be better?) from each segment and a logic level mosfet to act as the common cathode would make the display work with most display drivers if I were to choose to go that route. It would be simplest I think to have a display driver, I am open to suggestion on components and circuit suggestions at this point.

I quick glance through some white LED specs show they are around the 3v vf, so wired in pairs with no resistor for 6 AA batteries should give fairly uniform results across the hundreds of LEDs.

I have no idea on how to deal with radio clock thing, it on the research list, but I suspect there are some here who have already dabbled in that.

any thoughts on this, am I going about it wrong?
 

Attachments

Buzby

Senior Member
Hi oracacle, ( or should that be ClockMaker ),

Do a few calculations.

How many segments lit at once ?

What current per segment ?.

What capacity does an AA battery have ?

I think you will find your customer will be spending a lot on batteries !.

Cheers,

Buzby

Regarding LEDs, Check out some thing like this : http://uk.rs-online.com/web/p/products/2359382/
 

oracacle

Senior Member
yeah I know, its an ideal. if its 23:59 20 segments will be lit, 10 LEDs per segment, that's the better part of 4 amps presuming 0.02A for each LED if they are all lit at once.
If the displays are multiplexed so only a single display is lit at once a max of 70 LEDs will be lit which drops things to 0.14A plus drive circuitry, which is a wee bit more manageable.
 

Buzby

Senior Member
Even at 0.14A ( 140mA ), an AA battery of say 2400mAH will last less than one day.

The only technologies that can provide continuous displays at low currents are LCD and e-ink, but neither are good for high contrast.

Maybe you could make the display triggered on demand, a bit like an LED watch, with a button or a sound detector.

Your challenge here is not how to drive the LEDs, but how to reduce current consumption to an acceptable level.
 

hippy

Technical Support
Staff member
AA and AAA batteries are rather impractical for powering higher current circuits for any continuous period of time. Assuming that 2400mAh rating, your battery life would be -

1 day @ 100mA
7 days @ 15mA
1 month @ 3mA

In terms of other aspects; is it really necessary to have 10 LED's per segment ? Fewer and a diffusing cover might actually work better.

For people with poor eyesight, higher brightness isn't necessarily the best solution for making things easier to read.

Two other alternatives are electromagnetic 7-segment displays like outside petrol stations displaying prices, or possibly a car odometer type display with digits on cylinders which can be rotated.

There are also mains alarm clocks which have large 7-segment digit displays which could perhaps be modified to be brighter.

Or, a 7" LCD display or monitor which can show a clock time. It might even be possible to use a digital picture frame and use a PICAXE to cycle through the images.
 

inglewoodpete

Senior Member
So much wasted energy! So much computing power not being used!

Perhaps you need a sensor to detect when someone is looking at the clock and only illuminate it when someone is :) :).
 

oracacle

Senior Member
He has wet macular degeneration, so big and bright and high contrast is exactly what he needs - its where the requirement for white LEDs came from

as an idea he cans struggle to watch sport a 50" TV from a 2 meters and be able to read the score.
I will speak with his daughter and see if I can wangle a wall wart in there instead batteries.

I may produce a couple of prototype displays with different LED numbers, so 5 per segment and 10, maybe a an 8.
I have some strip board that I could nock up some segments on for testing.
 

AllyCat

Senior Member
Hi,

Perhaps you need a sensor to detect when someone is looking at the clock and only illuminate it when someone is :) :).
A PIR might be a possibility. Or "face recognition" is quite advanced in modern digital cameras. ;)

Another way to improve power consumption (and clarity/contrast) is to flash the LEDs on and off, anything between 10 Hz and 0.1Hz (i.e. flashing on once every 10 seconds) might be appropriate. Also, I think you're being pessimistic about the current required for a "directly viewed" high efficiency white LED.

However, I wouldn't normally consider battery operation for a clock like this, except perhaps one using "dayglo" mechanical digits/segments as suggested by hippy, which of course still needs an adeqaute ambient light source.

If battery operation is essential, perhaps for safety (e.g in a bathroom or to avoid trailing cables), then I'd consider a "cassette/cartridge" system for the battery. For example two dedicated "battery packs" which are swapped between clock and mains charger every few days. The charger and clock would both be designed to NOT fully charge/discharge the cells, thus extending their lifetime to perhaps thousands of cycles.

Cheers, Alan.
 
Last edited:

BESQUEUT

Senior Member
He has wet macular degeneration, so big and bright and high contrast is exactly what he needs - its where the requirement for white LEDs came from

as an idea he cans struggle to watch sport a 50" TV from a 2 meters and be able to read the score.
I will speak with his daughter and see if I can wangle a wall wart in there instead batteries..
So you can use 144 LED/m strip
1m-5m-APA102
With 10 LED/segment, each segment will be 7cm long, and a full number 14 cm tall, and very bright !
Power need : 3W per segment ==> 84 W for 28 segments ==> 5V 17 A...


You also can use 5w-cob-900lm-3000k-warm-white 4.8 cm/segment

Dimming via PWM seems necessary...
 
Last edited:

hippy

Technical Support
Staff member
Graphical OLED displays may be worth considering. Though they probably won't be physically large, their excellent contrast may make up for that.

One can also get reasonably large digit displays on a text OLED and, when used in graphical mode, the blocks to make digits look nicer can be tailored and not limited to the 8 user-defined characters of text mode.
 

oracacle

Senior Member
Done a little looking around, and for the size requested OLED displays start becoming quite expensive.
I have drawn up a board to use 4 10mm LEDs per segment, giving and overall digit height of 10cm (4 inches if that's easier to understand). LEDs are connected together in pairs.
As it happens I ended up having segments connected on the cathode side, and the each digit controlled on the anode side. This could easily be changed just by putting the LEDs in the other way around on the board.
The next bit I were I would like to ask for some help in selecting components. the display is multiplexed so switching speed needs to be good. I am thinking P-channel logic level FET to the anode and maybe something like a ULN2803A on the cathodes.
 

Attachments

BESQUEUT

Senior Member
Done a little looking around, and for the size requested OLED displays start becoming quite expensive.
I have drawn up a board to use 4 10mm LEDs per segment, giving and overall digit height of 10cm (4 inches if that's easier to understand). LEDs are connected together in pairs.
As it happens I ended up having segments connected on the cathode side, and the each digit controlled on the anode side. This could easily be changed just by putting the LEDs in the other way around on the board.
The next bit I were I would like to ask for some help in selecting components. the display is multiplexed so switching speed needs to be good. I am thinking P-channel logic level FET to the anode and maybe something like a ULN2803A on the cathodes.
With apa102, no need for all that stuff...
 

oracacle

Senior Member
And to create segments' of the same size at lest twice as many would be needed which then causes the same power issues as before. Believe out not apa102 doesn't solve the problems with an old guys eyes
 

goom

Senior Member
Kingbright make a 7 segment common cathode red display with 4" digits (SA40-18SRWA). Each segment has 8 LED's configured as 2 banks of 4 in parallel. I have driven them through a ULN2803 from 12V and 240 Ohm resistors. They are quite bright.
 

BESQUEUT

Senior Member
And to create segments' of the same size at lest twice as many would be needed which then causes the same power issues as before. Believe out not apa102 doesn't solve the problems with an old guys eyes
Since #7, there is no power issues.
Plus : you can use 144 pixel/m, 120 pix/m, 60 Pix/m or 30 pix/m...
Plus : PA102 can be dimmed easily : so power need is exactly what you want or what you can assume.
Plus : APA102 can be white or any cool white, warm white or color more suited for readability.

Interface : forget FETs and others : direct connect to Picaxe...

But, why make it simple when we can make it complicated ?

He has wet macular degeneration, so big and bright and high contrast is exactly what he needs
 

oracacle

Senior Member
So dimmed LEDs are easy to see? I can get full brightness with 80mA and nice and easy to see, or half rightness and harder to see. Or full brightness and twice as much power consumption.

Complicated isn't an issue, things can be learned, we're ask her to learn. Its not actually complex per se, its essentially an and gate.

Out might be scary but not all led projects cash be solved with apa102
 

oracacle

Senior Member
Think I have the control Circuit sorted.

The MSF clock may have to wait, although I am adding the header for it. If I do get to it, or when as I will have some access to it after the new owner takes it I am going to have it check just once a day for the time, in the early hours of the morning. it shouldn't matter if the clock stops for a while to check the time. the rest of the time, time will be kept by a DS3231

suggestion for changes welcome, going to try and bread board some of it before committing to PCB
 

Attachments

Last edited by a moderator:

oracacle

Senior Member
By comparison to learning the correct programming for the LEDs, Surface mounting 200 Plus LEDs masking them out so the pcb can be painted and waiting for them to arrive from where ever they are coming from. That combined with the fact the this has to be some kind of running before Xmas and 6 night shifts to do at the same time, something where half the code already exists is a lot lit simpler.


Fyi all the bread boarding is done.
And believe it out boy you most likely don't have all the information about everything
 

techElder

Well-known member
Sorry for the pun in here, but this application is just 'SCREAMING' for an audio solution. You wouldn't hesitate for an audio solution if the fellow was totally blind.
 

oracacle

Senior Member
He's more deaf than blind. You can often hear the TV from the bottom of the drive duty all Windows and doors closed
 

oracacle

Senior Member
reasonably sure you have missed the point of the original LED density. and for that matter, the density of the 10mm LEDs, the gaps in the strips are not what one would call ideal, its why the LEDs are pushed hard up against one another, its why there would have to go onto a custom board and why there would have to be so many of them - its also the reason why they will cause the same power issues as the original design - something you evidently did not grasp

I will repeat this, your favourite programmable LEDs are not the answer for things that have more than one LED, its not always the easiest, nor is it always the simplest. For that matter it does into take into account design considerations, deadlines and a whole bunch of other things.

Edit: Oh dear, the cost would nearly double using APA102 too. cheapest I could find worked out at around £20 per meter for 72 LEDs per meter, 1.12 meters minimum length needed - double that to get rid of the caps in the segments
120 10mm LEDs, £10, drive components £4 - which gives a better output for the application.
APA102 doesn't really stack up unless you are doing a lot of colour changing - its probably the only place that they become a more practical way of doing things.
this display, like the last one only has to display one colour.
 
Last edited:
Top