LED circuit problem

[pizzachang]

Hi all,

 Since I'm a novice pcb user/builder, I seem to have made a mistake in my LED project. I did, to the best of my understanding, follow the schematic accurately for the project. Basically, it's a 4 separate array (48 each array) intended to be used as an extension signalling light, powered by another unit. What I'd like to do, is have someone with enough knowledge to test and explain what I'm doing wrong. I can add that all the pieces I built, do function, e.g. all the LEDs light individually and in their resistor strings, but when I connect the original signalling unit, nothing happens. Monetary compensation is not a problem.

Thanks in advance.

[Crossy]

Bang up a few photos and the schematics here, let's have a look what we are potentially getting into.

 

[pizzachang]

3 minutes ago, Crossy said:

Bang up a few photos and the schematics here, let's have a look what we are potentially getting into.

 

 

8 x 6 led array (5).pdf

[Crossy]

Can you post a couple of photos of what you've built?

 

Did you design and etch the boards or are they commercial boards which you've just populated?

 

[pizzachang]

The enclosed schematic here is using only 1 LED per signalling array. I wanted a larger, more visible unit, so,

I have changed that ULN from the 2003A to a 2803A .

[pizzachang]

9 minutes ago, Crossy said:

Can you post a couple of photos of what you've built?

 

Did you design and etch the boards or are they commercial boards which you've just populated?

 

The pcb's I used are double-sided [commercially available]and the 48 LED arrays should come on either 1 or 2 arrays at a time, according to the signal received. The original sending unit provides + or - 5 volts, when the LEDs are triggered and 0 volts at rest

[Crossy]

I was thinking more of photos of the actual works to see what standard it's built to and what the boards look like.

 

If you ground each output of the 2803 does the relevant string light?

 

Have you got 15V on the "common" and 0V on the "gnd" of the 2803?

 

If you put 3-5V on each input of the 2803 does anything happen?

 

 

[Crossy]

1 minute ago, pizzachang said:

The original sending unit provides + or - 5 volts

 

Your 2803 is not going to like having -5V on any inputs, are you sure that's what's coming from the controller? 

 

 

[pizzachang]

[pizzachang]

[pizzachang]

2 minutes ago, Crossy said:

 

Your 2803 is not going to like having -5V on any inputs, are you sure that's what's coming from the controller? 

 

 

I think that means up to 5 volts.   If you're interested, we could meet up somewhere in CM and you can examine what I've done. I have plenty of testing stuff.

[Crossy]

So, can you answer any of my earlier questions?

 

Sadly we're not in CM, but that's such a simple circuit you ought to be able to sort it with a little guidance.

[pizzachang]

3 minutes ago, Crossy said:

So, can you answer any of my earlier questions?

 

Sadly we're not in CM, but that's such a simple circuit you ought to be able to sort it with a little guidance.

I've only tested the strings of 6 and individual LEDs. they all light - I have not tested the using the ULN 2803.

[Crossy]

2 minutes ago, pizzachang said:

I've only tested the strings of 6 and individual LEDs. they all light - I have not tested the using the ULN 2803.

 

Time to go to the outputs of the 2803. You can quite safely short each output to gnd without fear of doing any damage. The associated string should light.

 

If they don't then check continuity from the 2803 to the relevant string and also check you are getting the 15V supply at the top of each string.

 

[dingdongrb]

It's difficult looking at the backside photo and understand exactly what is above. Looking at the backside photo I can say it's a mess. I mean you seem to be having stay leads laying across and potentially touching nodes that they shouldn't.

 

I didn't see a reply from Crossy's question, the one where it was asked if you have supply connections to the IC. If you take DMM what voltage do you measure at pin 10 of the IC when using pin 9 as the reference? (measured directly on the leads of the IC)

[pizzachang]

17 minutes ago, dingdongrb said:

It's difficult looking at the backside photo and understand exactly what is above. Looking at the backside photo I can say it's a mess. I mean you seem to be having stay leads laying across and potentially touching nodes that they shouldn't.

 

I didn't see a reply from Crossy's question, the one where it was asked if you have supply connections to the IC. If you take DMM what voltage do you measure at pin 10 of the IC when using pin 9 as the reference? (measured directly on the leads of the IC)

The "backside" is all cleaned up and neat.  The important part is all my leds light up as they are supposed to , when tested individually - so they are connected according to the schematic that uses the 2803A ULN.  I can also  get the entire strings of 6 leds to light from the resistor to the end of the string, using an adjustable power supply. I have not been able to light the entire 48 Leds in the array, all at once.

Are you located in Chiang Mai?

[thaibeachlovers]

15 hours ago, pizzachang said:

The enclosed schematic here is using only 1 LED per signalling array. I wanted a larger, more visible unit, so,

I have changed that ULN from the 2003A to a 2803A .

I'm not an expert by any means, so can someone explain why if the + side of battery is connected to common, and after passing through the lights it is wired back to + side they would work? I'd assume they should be wired to - side of battery to get a circuit, or is that just for normal bulbs and LEDs work differently?

[Crossy]

19 minutes ago, thaibeachlovers said:

I'm not an expert by any means, so can someone explain why if the + side of battery is connected to common, and after passing through the lights it is wired back to + side they would work? I'd assume they should be wired to - side of battery to get a circuit, or is that just for normal bulbs and LEDs work differently?

 

Yeah, "common" on the 2803 is confusing. It's not the common point of the switches (that's gnd), it's the common cathode of the internal protection diodes (for inductive loads).

[Crossy]

14 hours ago, pizzachang said:

The "backside" is all cleaned up and neat.  The important part is all my leds light up as they are supposed to , when tested individually - so they are connected according to the schematic that uses the 2803A ULN.  I can also  get the entire strings of 6 leds to light from the resistor to the end of the string, using an adjustable power supply. I have not been able to light the entire 48 Leds in the array, all at once.

Are you located in Chiang Mai?

 

Have you yet powered the whole lot up from your 15V supply and shorted the "bottom" of your strings to gnd/0V either at the string or at the 2803?? Result??

 

[thaibeachlovers]

9 minutes ago, Crossy said:

 

Yeah, "common" on the 2803 is confusing. It's not the common point of the switches (that's gnd), it's the common cathode of the internal protection diodes (for inductive loads).

Thanks for that. I was also confused why Gnd was connected to the battery, as I assumed it was GROUND( EARTH ) as in a normal 240 v circuit. Be easier if they used different terminology.

[pizzachang]

6 minutes ago, Crossy said:

 

Have you yet powered the whole lot up from your 15V supply and shorted the "bottom" of your strings to gnd/0V either at the string or at the 2803?? Result??

 

I'll do that today. I'm assuming by "bottom" , you mean the cathode of the last Led in the string?  I did get strings to light up by using the bench power supply, using the resistor connection and the last cathode leg in the string. I may have done something else wrong , as I couldn't get the same results on each of the strings for all the arrays.

[Crossy]

Just now, pizzachang said:

I'm assuming by "bottom" , you mean the cathode of the last Led in the string? 

 

Yeah.

If you can get each string to light up like that, then try at the 2803 outputs (it's quite safe to short the outputs to gnd/0V).

Take it step by step fixing each issue as you find it.

 

[dingdongrb]

2 hours ago, thaibeachlovers said:

I'm not an expert by any means, so can someone explain why if the + side of battery is connected to common, and after passing through the lights it is wired back to + side they would work? I'd assume they should be wired to - side of battery to get a circuit, or is that just for normal bulbs and LEDs work differently?

The ULN2803 is not a normal CMOS or TTL IC as one would think. It's a driver, a Darlington Pair transistor driver where the 'Common' is actually the circuit source (pull-up) voltage for the output.

[dingdongrb]

1 hour ago, thaibeachlovers said:

Thanks for that. I was also confused why Gnd was connected to the battery, as I assumed it was GROUND( EARTH ) as in a normal 240 v circuit. Be easier if they used different terminology.

'Grounds' can be confusing to say the least. There is 'circuit' (signal) ground, 'chassis' ground, and then 'earth' ground. In most cases, not all, the three are typically tied together but there are cases where they are kept separate. 

[pizzachang]

1 hour ago, Crossy said:

 

Yeah.

If you can get each string to light up like that, then try at the 2803 outputs (it's quite safe to short the outputs to gnd/0V).

Take it step by step fixing each issue as you find it.

 

Ok, have tested by using my bench supply connected at the power supply for one side (24 leds) and the cathode consolidation line (the line that runs to the ULN outputs) and the results:

Both green arrays fully light (as a unit of the 24, lighting at once)

One red led array (24 leds) fully lights and the other array of 24 has one string that does not light.

One white array fully lights and the other 3 each have the sma problem - with one white array having 3 strings not lighting.

Assuming I have damaged or failed to solder a quality connection on the non-functioning strings.

[Crossy]

6 minutes ago, pizzachang said:

Assuming I have damaged or failed to solder a quality connection on the non-functioning strings.

 

Good start.

 

Check you've not got a LED the wrong way round too.

 

If you dial back the supply voltage a bit you can start at the top of a non-working string and dab your ground on each junction until you find the device that's not doing its job.

 

Once you get all the strings lighting then you can look for the next fault.

[dingdongrb]

1 hour ago, Crossy said:

 

Good start.

 

Check you've not got a LED the wrong way round too.

 

If you dial back the supply voltage a bit you can start at the top of a non-working string and dab your ground on each junction until you find the device that's not doing its job.

 

Once you get all the strings lighting then you can look for the next fault.

To check each LED you can also measure the resistance across both twice. Once in forward polarity and once reverse polarity and see that all are good and orientated properly. If your DMM is a good one it might just have a diode check position. Also some DMMs produce a strong enough current that would allow an LED to light up just so slightly.

[canthai55]

1 hour ago, dingdongrb said:

 but there are cases where they are kept separate. 

My self converted motorhome, on an MCI chassis, was this way, for 120V AC

[dingdongrb]

I never did see any posts indicating exactly what fixed the problem....or was it not fixed?

[pizzachang]

On 5/11/2021 at 11:06 PM, dingdongrb said:

I never did see any posts indicating exactly what fixed the problem....or was it not fixed?

Not fixed. But I do have some well-made, 48 LED matrices, that work until I connect the machine that I want to supply the actual signal to operate the LEDS