Count down timer for my soccer club

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
A compromise would be to make an identical PCB for each segment with mounting holes, then mount them to a cheap board. This would save time and money and allow you to pick the LEDs you want.
As far as constructing a board that's a great idea - one small board per segment. A number of individual segments could be made on a 12" x 12", then chopped up ready for the leds.

Last night I googled scoreboards again and had a look at some "manuals". One was listing replacement parts (for large digits) and they had clip in "segments". It didn't click (excuse the pun), but now that you mention it - yes, that would be easier. A PVC mounting board with segments that connect in. This would make for easy replacement if needed, and easy construction. This is definately in my "to do list" if I go down the way of making them in lieu of purchasing.
 

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
Hi Bill / elec_mech.

I've been doing my homework the past few nights - google crazy. This is what I've found:

Green - clearest colour to see (next to white) and usually brighter than most other leds colours.

I've seen a score board close up. LEDs are 5mm oval, green, 3200mcd, 140/40 viewing angle (140 on the horizontal plane and 40 on the vertical plane).

I drew the soccer pitch (105metres x 68 metres + area around) and sure enough it's about 138 degrees if standing at the furtherest point on the left and right looking at the scoreboard! I check the horizontal line of sight, and assuming the board set up say 2 metres off the ground, someone standing infront of it (at say 1500mm high) compared to someone on the other side of the 68 meter width pitch and up a metre or 2 (on say a stand), is infact 23 degrees. So 40 degrees has some safety margin.

At night the leds dim right down - that's the secret to seeing it clearly at night. Mine were originally way too bright so I couldn't distinguish the numbers. They blurred into a bright blinding blob. Logical now that I sit back and think of it. During the day, digits on a black background is sufficient. No need for colored filters.

A 16" digit needs around 500 leds. This give a good intensity when in direct sunlight.

As I've done on my latest sample, resistors should not be installed on the driver circuit board, but on the digit board so as to avoid a bunch of leds shorting out if grounded accidentally - via moisture on a connection pin. These leds also need to be coated (with?) so they don't rust over time.

To dim the leds, the voltage needs to be dropped. I assume this can be done with maybe 1/2 dozen light sensors set at different settings to do a gradual change (triggering a change via transistors?) - your expertise needed here.

I've just found a place in Hong Kong that supplies this spec led. I was originally told they cost in the order of 20 cents each (AU), however I found them for 3 cents (in quantity of 8000). Bargain! The company makes these with mcd and viewing angle to order!
http://www.qxyled.com

Given there are now 500 leds to solder per digit (1000 connections), the digits are best mounted directly on a "printed" circuit board. The board needs 1000 drill holes for the anode/cathodes of the leds. With 15 boards, 15,000 holes, and 15,000 solder joints, I think I need made up circuit boards.

Again I googled. I found a site in China that does an online instant quote based on size, holes, board type, etc, etc:http://www.pcbcart.com

I filled in what I think the board should be made like (could you please check it):
View attachment Printed board quote.pdf
The price for 15 boards ended up being around $20 each. That makes my digits around $35 a piece plus resistors and connectors. I'm hoping I haven't made any major mistake, but if not, this is a definite goer. I'm going to order 500 sample leds, build a single digit, work out the led layout, build it, vary the voltage to test at night and also day, and fingers crossed it all works well. Finally some light (excuse the pun) at the end of the tunnel.
 

Wendy

Joined Mar 24, 2008
23,408
Like I said earlier, I would mount into a substrate material such as wood or metal, but I can understand why you want to do it your way. Still, the cost of under $1 per panel vs $20 (per PCB) is significant. I am a natural born cheapskate.

You can spray parts with a conformal coat to protect against the elements. Acrylic Krylon has been recommended to me, though I have not tried it yet. It is going to be used in my project to lightly coat a home brew PCB. If the coating is light, so I am told, you can solder right through it without a problem. I'll find out eventually.

A suggestion, build a small mock up score board to test your electronics. This will allow you to test and troubleshoot your electronics indoors. When I built a light tree and all the trimmings I did something similar, it save some time and hassle.

The Completed Projects forum has several new projects that would have been nice for this, and from what the author/creator has said, he is not done. They would have applied nicely for your whole effort.

Project: PIC 4-Digit 99 Minute Timer
Project: PIC 4-Digit DS18B20 Temperature Monitor

The author, MMcLaren, has hinted that he has another in the works that would have made a simple score keeping counter, if I read it right. Life is like that sometimes.
 

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
I know I'm going a bit for the rolls royce here. I figure that with my weekends at soccer (playing and helping in the Canteen Saturdays and Sundays) I'll be a bit strapped for time. The job will also look alot more professional.

I've now also had the president show some interest (his line of work is "lighting"). He says that if I can get the board up and running, and also find the easiest/most economic way of manufacturing (well the digits), he may take it on and offer production for other soccer clubs in the area.

I'll ask my local electronics store for the spray you suggested.

For the "mock up", good idea. I'll do that. Could save alot of work in the long run.
 

Wendy

Joined Mar 24, 2008
23,408
Non conformal spay is electronic, but the acrylic is off the shelf hardware store.

If you can sweet talk the other gentleman to modify his code you could get the scoreboard down to 4 chips and LED drivers. I haven't priced these µC chips, but I suspect they are really cheap.

Anyhow, like I said, whatever I can do to help.
 

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
Hi Bill.
I've built my 16" digit - I think brightness with the LEDs I bought looks good in daylight (against the black mounting board).
I built the prototype PCB board and spent the last 4 nights trying to get it to work. Can't find the problem.

Breadboard (works):
6v DC supply voltage for circuitry. 24v DC supply for led display. I connected the 24v Vcc directly to the led common. I have darlingtons to run the 7 segments (run by the gnd of the 24v supply). I linked the gnd's of both supplies. Via remote, digit counts as it should.

Prototype PCB (doesn't work):
Same setup (and I "think" wiring as far can see). However, the counting floats. As I lay the board down on the kitchen table, the digits start to automatically count. If I lift the board, it works. If I wave my hand over the transistor that triggers pin 1 of the 4518, it starts to float again. I do have the base of the transistors pulled down to gnd via a 10k resistor and 0.1uF. I've measured the voltage across gnd and pin 1 of the 4518. On the breadboard it's around 0.4v. On the PCB it's around 1.4v.

View attachment Count down timer PCB_V7.pdf, View attachment Count down timer wiring_V7.pdf

So I'm really confused. Can you make any suggestions.
 

Wendy

Joined Mar 24, 2008
23,408
I am about to go to bed, but I will be a second set of eyes for you. You have build this?

Don't worry too much about it, even big business goes through this. I have seen enough rewiring to last a lifetime, I suspect yours is simpler.

As a clue, do you have any symptoms? When you say floats, what are you trying to describe?
 

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
I built a prototype 16" led digit.
So in order to test it easily outside without carting my breadboards with me, I thought I'd use my prototype pcb board that I built earlier (score keeping one). This way not only could I test the digit independantly, but at the same time test my prototype PCB.

So no, it's not the final one, therefore I don't mind making alterations at all. It already highlighted some problems that I've got to change anyway.

When I connect the digit to the pcb, the number "races" as if on steroids! It doesn't stop counting until I lift the board up off the table. But waving my hand over it, starts it counting again.

I don't know if it's because I've only got 1 or the 2 digits connected to it. Grounding? It definately seems like a grounding problem.
 
Last edited:

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
As a clue, do you have any symptoms? When you say floats, what are you trying to describe?
Bill, got the PCB working. Advancing digits due to an LED it fed off the "count" inputs. A a "future fault find", I placed an LED on the input that fed directly to the base of a BC547 that gave a pulse to pin 1 of the 4511 to count. What I didn't realize is how sensative the transistor was (showing my lack of knowledge here). I therefore connected the LED pin 1 directly (which is how my breadboard was connected).

I then found that the count ended up being by one's, two's and sometimes more. I changed the 0.4uF on the base of the transistors to 0.1uF. Perfect.

What I do need now is to work out how to auto dim the leds. I bought a phototrasistor to test dimming an LED. I want to somehow dim these LEDs at night, otherwise they turn into a bright blur.

I'm running them on 24V. I found that if I run 2 of the 16" digits via a 470Ω resistor, the voltage drops to around 16.8V and the intensity looks about right. Note, each digit at full brightness will draw around 830mA.

Can you suggest any way I can dim these?
 

Wendy

Joined Mar 24, 2008
23,408
I would use an LDR myself, though a transistor works.

You will need to make a basic circuit, possibly with some optical filters to cut down the light level reaching the transistor.



Then measure the voltage levels at the various light levels you want to change at. This will let you make a circuit that will interpret the light levels. It will be very analog.

With me so far?

With an LDR (AKA electric eye, photocell, etc) it is much easier, but either will work. That is to say, I have a lot more experience with a LDR.

You feed the light level you want to change at into either circuit, and adjust the voltage output with the pot to be ½ Vcc.

For multiple set points you make more than one circuit, adjust each for the light level trip point it is sensing. There are much more elegant solutions, I'm aiming for quick and dirty. A LDR will assume a specific resistance for a specific light level. If you were to measure the resistance you could "look up" the PWM duty cycle for the LEDs.

A off the shelf solution is always better, but if you want to pursue this I'll go through how to digitize this and use if for the project. It will involve some PWM circuits. PWM is more consistent across the board, and once you have the sensor network done, not too hard to implement (or expensive).

Pulse Width Modulation

The idea would be to turn on/off the ground path for all the LED displays. With PWM the change in brightness would be the same.

Are you using a mock up display to test it out?
 

Attachments

Last edited:

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
Are you using a mock up display to test it out?
Yes.

An LDR is what I bought and have been experimenting with, so yes I'm all for it. I'm guessing that I need 3 light levels:

1. Night time (dim)
2. Day time cloudy (normal)
3. Direct sun (brightest).

Is it possible to make 3 circuits to cater for the above, then join them so that when sunny only "1", "2" and "3" kicks in, then when cloudy only "2" and "3", then at night only "1"?

You feed the light level you want to change at into either circuit, and adjust the voltage output with the pot to be ½ Vcc.
So do I use a circuit similar to:
http://www.kpsec.freeuk.com/images/tranldr2.gif

If so, how do I connect 3 together and how do I tackle the amps (around 10A in total)? Or does something like this circuit do what I need:

http://electroschematics.com/wp-content/uploads/2010/07/light-sensitive-switch-with-ldr.jpg

Excuse my ignorance, but does the above mean that the LDR draws little current and the transistors are what handle the high current?

When you're talking about PWM is the theory behind this that you pulse the LEDs at a high frequency so you don't see flicker. Then to get them dim, you only turn them on for say 10% duration of each pulse? If so, that's ingenious - and it probably lengthens the life of the leds too!

If I'm on the right track, does the circuit (see link below) with a 555 timer do this trick?

http://www.dprg.org/tutorials/2005-11a/index.html

Could I build 3 of these circuits, triggered by LDRs which pick up different light levels and adjusted to how bright I want the digits displaying?

Sorry for all the questions.

PS. Last night at the soccer pitch, I tried my test PCB (count only), 2 digits (16"), and remote control. I dropped the 24V to around 16.8V and the numbers look perfect. Everyone was happy with the brightness and size. Also, one of the test digits is made from "oval" LEDs and the other from "flat top" LEDs. The flat top seem to have a more even light distribution from any angle (as wide as about 160 degrees) - with no "bright spot" head on.
 

Wendy

Joined Mar 24, 2008
23,408
I'm been thinking about it, I have some ideas you might be able to used, but I will get back with you. I haven't had a chance to review your posts properly, but I will.

DPRG, I know it well, I go to a meeting now and again. :D
 
Last edited:

Wendy

Joined Mar 24, 2008
23,408
OK, here is a though, take your LDR, attach it to a ohm meter. Take measurements for the light levels you want. It is possible (not guaranteed) that we can make a circuit that smoothly varies the light level vs light intensity. If we know the resistance range I can draw something up.

A LDR acts like a variable resistance, this can be used to an advantage.

Does this seem doable and reasonable?

I am a good draftsman, and I also am a fair designer. I suspect you can isolate the grounds to all your displays, and have them connect to one common point (not quite a ground, but very close). I'll draw what I am talking about in a bit.

Do you know the max current all the LED displays will add up to? Exact numbers aren't extremely important, and it is possible not all your LEDs go to a common point. I will have to look at your schematics to see.
 

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
Suggestion sounds reasonable.

Unfortunately due to winter, it's cloudy today. I've measured the LDR infront of a 50W halogen indoor spot light (brightest thing I can get hold of). I assume outside in direct sunlight would be brighter??? Reading is 50Ω infront of the light. In darkness, it is close to 10MΩ.

Specs for my LDR:
http://www.jaycar.com.au/productView.asp?ID=RD3480

Each of my digits runs 7 segments. 4 of the segments are 8 banks of 7 leds (with 330Ω resistors and powered by 24V). 3 of the segments are 7 banks of 7 leds (330Ω at 24V). I'm running each LED at max 15mA, hence (4 x 8)x15 + (3 x 7)x15 = 795mA. I have a total of 15 digits to run (hence 12A total).

The minimum I need however I'm not sure yet. I did last night run each digit through a 150Ω resistor (795mA draw on 24V supply) on the common and the intensity was just right. Don't know how to convert that for my LDR.
 

MMcLaren

Joined Feb 14, 2010
859
That's close enough. Thanks Chris.

If you decide to take it into production, you should really consider a microcontroller based design which would reduce parts count to something like; (A) one microcontroller, four 8-bit constant current sinking driver ICs, and eight MOSFET driver circuits for a multiplexed solution, or, (B) one microcontroller and eight 16-bit constant current sinking driver ICs for a non-multiplexed solution. While the non-multiplexed solution uses more ICs, it reduces the peak current demands of a multiplexed solution. Both multiplexed and non-multiplexed solutions use the same amount of current but the multiplexed solution drives the LEDs at a higher 'peak' current in order to maintain the same 'average' current and brightness that you would have in the non-multiplexed solution. BTW, the constant current sinking driver ICs eliminate the need for segment current limiting resistors and they cost less than $2 a piece.

Good luck with your project.

Cheerful regards, Mike
 
Last edited:

Wendy

Joined Mar 24, 2008
23,408
I was thinking of using your several completed projects myself. By mixing and matching modules you could make whatever type of scoreboard you want, there are certain fundamental types of displays on any scoreboard. I'm working on parts, and there will need to be some programing changes, but over all it is very doable, and would be much smaller and manageable.

As for the LDR, it is going to be very subjective. I'll start drawing some ideas, and to some extent we will just have to play with them to see if they can be made to work. Mentally I've broken it down into several pieces that are independent, but in the end will interact.


  • Light levels needed - The PWM circuit is pretty simple, but not what is show for DRG. I can make a simple percentage of intensity circuit to help determine what the levels (in %) are needed. Once you have numbers designing is easy.
  • Optical levels needed - This one is a bit trickier, but no big deal. Measure ohmage in full sunlight (it does not have to be perfect), measure full darkness. This will give us the range needed to measure. If the curve of light level to optical level is not linear there are ways to compensate the curves. The other way to do the same thing is stepping functions, have 3 or more discrete light levels that will match specific optical levels, it will be a lot more digital in nature, and a bit more (not extremely) complex.

I have not reviewed your latest schematics yet, there could be problems waiting. Basically PWM will turn the LEDs on/off several thousand times a second, varying the percentage on on time (this is duty cycle). Since it is a percentage of on time the brightness will vary identically for all the LEDs, they will stay matched.

First we'll breadboard this circuit, using a linear pot (this is critical). We mark the pot between 0% and 100%, and record the percentage of light levels needed for your applications. We will try to interface the variable PWM to the various LED circuits, once this is done about 80% of your problems are solved, maybe more. A simple MOSFET works very well as a switch, and can be used to interrupt the current to the LEDs, perhaps one can do all of them at once. If they are "wired in" to a multiplex circuit then you may have to use more. The components are cheap, the work is hard.

OK, using my Figure 5.3 out of my LEDs, 555s, Flashers, and Light Chasers I have come up with this. It is a fundamental concept circuit, details will need tweaked.

 

Attachments

Last edited:

Thread Starter

chrischrischris

Joined Feb 18, 2012
313
I've just measured the reading in full sunlight to be 300Ω.
I also measured it in indirect light to be 1500Ω.
What I would need to also do I suppose is give you a reading at night (with the flood lights on).

I just tested the digits in full sunlight. Flat tops perfect, oval ones again a bit bright head on - so flat tops definately the go. Pics:

digits.JPG digits sunlight_2.JPG 16 inch rear.jpg

Regarding reviewing my schematics, hold off for a while. I'm just checking my other 3 wiring diagrams against my breadboard against my new PCB layouts - it's a hell of a job!

So I'll just finish this off if ok, then start breadboarding.
 
Top