Oversized 7-segment display

[sunray]

 

 

Here's my rather shabby test rig - built up with hot glue and then poured in.

 

 

With the uncovered tape, when poured on, the output is reduced, or at least the focus is knackered:

 

 

 

This was exactly my experience which I must admit, following Tims earlier comments, I was beginning to wonder if I'd remembered it wrong.

Interestingly the focus of the tape is better through the resin.

Edited March 10, 2020 by sunray

[the kid]

Bear with me -

 

We have some traxon http://www2.traxontechnologies.com/products/product_details/5705/ We made a massive clock in it 7 segment style , I dont have a photo but each "puck" is waterproof + RGB + so 3 channels a puck.

 

We do have tonnes of it as a possible option.

[musht]

Led neon won't be visible in direct sunlight.

 

The front firing profile in 96 LEDs/m has reasonable chance of being readable, usually mounted in a routed channel , perhaps a routed shader to sink the front face of the `neon` below the surface.

 

BTF also do empty silicone `neon` profile if you want the fun of pulling the strip through.

 

The cabochon style are definitely sunlight visible, they`re 5050s with polycarb diffuser.

 

5mm LEDs still have their uses, but its a load more wiring

 

https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20200310050752&origin=y&SearchText=led+gas+price

[TomHoward]

Ah, they look pretty much like what might be on the current signs already, or something already made up to 7-segment.

I was thinking two lines of 5050 60/m next to each other, something like this mockup in old school single (but to give the idea:

 

Cutting a laser / printed top to cover the strips apart from the holes might be a good idea. I was going to use black strip on a black backing already but I guess cover the white elements even more might work well in terms of visibility.

Going to make some test ones - currently making a couple of small ones for programming then I'll make some large ones for testing the visibility.

[TomHoward]

So far I've made a 3rd sized mockup.

The thing uses WS212b so has 21x pixels and draws about 5.25W or just over 1A at 5V.

The plan is to use WS2811 so it can be run at 12V and double up the strips as in the picture above, so the final one would have 126 LEDs, pull 35W so be 2.6A at 12V.

This is a bit of a problem as with nine in a display it could pull 23A max and actually need some fairy serious batteries if it were to be mobile.

We might have to scale the size down a bit for the mobile version if it's to be battery powered.

 

1/3rd sized character:

 

Wide shot on a messy desk, because it's hard to judge from the video just how bright it is:

 

YouTube link:

[timsabre]

Looks good. Limit the LED output in software so that if you have 2 colours lit (e.g. red+green for yellow) then you only run each at 50%. This gives you a consistent brightness across colours and reduces your LED drive current to 1/3 of what you have calculated. Unless you are really stuck for brightness and need the full power...

[TomHoward]

That’s what I’m going to try - I’ll still make a full sized one as prototype but make a single strip one as well and see whether one strip at 100% or two strips at 50% are more readable at around 200m.

 

I don’t know whether a colour might be more readable than white in daylight as well - eg red might actually be more visible than white maybe due to the contrast?

 

The only thing I’ve done since is print a cover with just the LEDs showing, which works well as a concept but is a bit of a shabby print as I tried just two layers thick and didn’t have the height set too well. Might try getting these laser cut instead which should be faster & neater.

 

 

[David Duffy]

LED tape will fail so having the whole thing encapsulated will render the digit useless and non repairable. Using WS28xx pixel may be easy for the data flow, but one dead pixel can kill everything downstream from it.

 

I strongly recommend you look at high brightness oval LEDs and make up digit circuit boards. Use shift register output drivers and ribbon cables to link the digits. With separate LEDs you can make each segment a string for higher voltage and less current.

 

There's a reason nobody uses LED strip for scoreboards. :)

[TomHoward]

To be honest, if I can make every digit modular and exchangeable, that'll halfway do as they are relatively low cost per digit.

I'm planning to cut tops and glue/silicone them on to make them semi-removable rather than pouring resin over them any more so they may be able to be cracked open to exchange lengths of tape.

It is true about the entire scoreboard running downstream of one pixel so I will explore running it in multiple circuits so hopefully only one line or one figure may be driven from one output if it's viable, so have several universes of output for each board to separate the data, so hopefully we might only lose one digit or one line.

We were thinking of driving static LED tape instead from shift registers but the amount of wiring and work really increased. I'm hoping to keep it really serviceable and then make a stock of spare digits that can be swapped out, as we are printing & cutting them on site hopefully we can keep them running.

[timsabre]

You can now get ws2815 tape which has a backup data line on each pixel from the one before to bypass faulty pixels. If the led sees no data on the main input it switches to the backup.

 

David might be correct about why scoreboards don't use tape, but I suspect the reason may just be that discrete led is cheaper to manufacture in bulk and/or tape is a relatively new thing and designs haven't changed yet.

[TomHoward]

It's a thought, the WS2815 is interesting, I'd seen the WS2813 but I didn't want to go 5V I'd rather go 12V as the current just goes through the roof at 5V and the batteries and power supplies aren't as easy either.

With the WS2815 tape, if there is a dead pixel, does the next pixel behave one address sooner (ie it all shifts one pixel to the right) or is the dead pixel bypassed but the address accounted for?

 

As if it all shifts one pixel on, it doesn't really work for 7 segment as it'll be lighting up one of the next segment and need repair / replacement much sooner anyway.

 

Edit:

It seems this rather basic question has been answered previously and it should drop the data from the missing pin when using the backup pin:

https://www.blue-room.org.uk/index.php?showtopic=73153

[timsabre]

It's a thought, the WS2815 is interesting, I'd seen the WS2813 but I didn't want to go 5V I'd rather go 12V as the current just goes through the roof at 5V and the batteries and power supplies aren't as easy either.

With the WS2815 tape, if there is a dead pixel, does the next pixel behave one address sooner (ie it all shifts one pixel to the right) or is the dead pixel bypassed but the address accounted for?

 

As if it all shifts one pixel on, it doesn't really work for 7 segment as it'll be lighting up one of the next segment and need repair / replacement much sooner anyway.

 

Edit:

It seems this rather basic question has been answered previously and it should drop the data from the missing pin when using the backup pin:

https://www.blue-roo...showtopic=73153

 

That is what happens, though if you lose 2 pixels in a row then it all stops.

In fact I have done a lot of pixel tape work and pixel failure is always at the beginning or end of the tape where the chip gets zapped by something external. I don't think I have ever had one spontaneously fail in the middle of a run.

 

[TomHoward]

Thanks, hopefully it'll be fairly sealed inside as it'll be like the one above with a lid on.

For the proper ones I'm thinking of either using transformer winding wire between then instead of bar (so it has some insulation) or printing divisions for the wires to keep them isolated as well.

I might run the 12V into the middle of the tapes separately away from the data if it's a bit tight.

[timsabre]

I would suggest you put a 100R resistor in series in the incoming and outgoing data lines from each digit. This will give some protection against static zap.

I don't see why you need insulated wire if you use solid core like you have done. It's not going to short like that.

 

[TomHoward]

Just reliability, in case of any stray bits inside etc, but it is a bit of a bugger as it may be harder to tin both ends. Originally when I was thinking of pouring resin in I was planning bare wire for ease because it would be cast in place.