LM3914 Cct Just wont work

Thread Starter

transducer

Joined Jul 28, 2007
25
Here you go,
Photos of my breadboard, and point to point soldered version.
My apologies in regards to the 'tidyness' of the layout. The first time I did it I did take my time but some 30 times later asthetics didn't come into it

The layout is as follows

Supply Voltage 8V from 317 Regulator

Pin 1 to 1st LED
Pin 2 to 0V
Pin 3 to 8V
Pin 4 to 0V REFLo
Pin 5 Open at the moment
Pin 6 through 10k pot to 8V giving 4V I/P to pin 6 FSD REFHi
Pin 7 through 1K2 to 0V
Pin 8 0v
Pin 9 8V

Does anyone looking in on this thread have the bits laying around to knock this cct up, to see if the problems are the same, maybe even send it to me. I will gladly pay for your time, components, postage, daughters wedding, anything to try to get to the bottom of this

Many Thanks again for all your help
 

Attachments

hgmjr

Joined Jan 28, 2005
9,027
What strikes me about the photo of the layout is the conspicuous absense of any power supply filter capacitors.

I see the LM317 linear regulator in one of the pictures but no large filter capacitor on its input or output. I would imagine a value of 100 uFd should be adequate for the circuit. I would also sprinkle a few 0.1 uFd filter caps about the board between the 8V power bus and ground. I would recommend shortening the ground connections but your point to point soldered breadboard appears to be nice and compact.

It may be that your solution lies in taming your 8V power supply.

hgmjr
 

Thread Starter

transducer

Joined Jul 28, 2007
25
Thnx for the quick reply.

The only reason I left off the caps was for ease of construction. When I 1st built the cct I included the caps on both the I/P & O/P but it made no difference. I have also tried a different supply I/P. Straight from my regulated desktop PSU, again made no difference
 

nomurphy

Joined Aug 8, 2005
567
Try making the modifications per the redlined schematic I've attached, I know this works, so if you have problems it's in the connections.

1) attach pin-6 to pin-7
2) attach 1.2K between pins6-7 and pin-8
3) attach 2.4k from pin-8 to GND.

The formulas are:

Vref = 1.25[1 + (2.4K / 1.2K)] + (2.4K * 80uA) = 3.942V

I_led = (12.5 / 1.2k) + (Vref / 2.2k) = ~12mA

If you use 1% resistors, they would be 1.21k and 2.49k -- which would raise the reference voltage to ~ 4.02V.
 

Attachments

Thread Starter

transducer

Joined Jul 28, 2007
25
Just tried the cct you amended for me (cheers)
but it is still the same, all LED's light up as soon as power is applied
The voltages on the pins are:

pin 1 6.09
pin 2 0.00
pin 3 8.08
pin 4 0.00
pin 5 0.9 ish, Led's dim during measurment ??
pin 6 3.78
pin 7 3.78
pin 8 1.90
pin 9 8.03

This is really really annoying me now
 

nomurphy

Joined Aug 8, 2005
567
When you say the LEDs dim when you measure pin-5 causes me think you may have oscillations (possibly hi-freq) on the input.

Disconnect pin-5 from the signal source, and short pin-5 to pin-4 (GND) at the chip and see if the LEDs go off. If they do, then the above is most likely the problem.

BTW -- you do have a common ground between the input signal source and the LED board ...right? See attached:
 

Attachments

Thread Starter

transducer

Joined Jul 28, 2007
25
Yup, V high oscillations when I measure pin 5
Yup the LED go out when I short pin 5 to 4
Yup I do have a common ground

Please tell me you have a theory :)
 

nomurphy

Joined Aug 8, 2005
567
What is your signal source? Is it only DC, such as an adjustable power supply, or some type of AC from a signal generator? If you have a scope, take a look at it. What is the amplitude and frequency of the oscillations?

If the source is a DC voltage only (supposedly), and it doesn't change/adjust very quickly, try adding a 1K series resistor to pin-5, and add a 10uF capacitor from pin-5 to GND. This will be a low-pass filter that dumps signals beginning at 16Hz and above, Fc = 1 / (2pi*R*C).
 

Thread Starter

transducer

Joined Jul 28, 2007
25
Sorry I am a bit lost now, I understand your formula for FR but if I ground pin 5 then how can i i/p a signal into the circuit to measure on the led's

The source I am using is a DC power supply with presets 3,4.5,6,7.5,9 & 12 Volts, but this I am using at 12v to feed my regulator, this then gives me 8v into my circuit.

No scope, sorry

As soon as I power on no matter what I try the LED's all light up, the only time thry dont is when I ground pin 5
 

hgmjr

Joined Jan 28, 2005
9,027
Try taking a couple of 1K resistors and use them as a voltage divider off of the 8V power supply to create a 4V dc signal and connect that to pin 5. If all is working normally that should light up half of the LEDs in the array.

On second thought you may need to take the input to the voltage divider from the same voltage that you are supplying to the R(HI) pin.

hgmjr
 

Thread Starter

transducer

Joined Jul 28, 2007
25
I have just set up a voltage divider onto pin 5 with 2 x 1k resistors with:

Pin 4 to 0V RefLo
Pin 6 to 8V RefHi
Pin 7 to Pin 8 via 1K2
Pin 8 to 0V

Result, I had 4 bars lit on my array, so I changed the 2 x 1k resistor to a 100K pot and as I wound up I got FSD at 4V all led's lit and they went out 1 by 1 as I lowered the voltage

:DHAhaHAhaHAha:D
It is official, hgmjr is a GOD

I cannot begin to tell you how greatful I am for the help received

The conclusion is this:

Pin 5 has to have a reference voltage applied for the circuit to work, leaving Pin 5 open is the same as FSD

I thankyou all again, I can now move on and get to the next problem

Best Regards

Transducer
 

hgmjr

Joined Jan 28, 2005
9,027
In reality, others offered similar suggestions regarding the treatment of the input signal to pin-5 so we all had a share in the final successful outcome. I was just one of the throng that responded to your request for assistance.

Let us know if we can be of further help.

hgmjr
 

Thread Starter

transducer

Joined Jul 28, 2007
25
Let us know if we can be of further help.
hgmjr
Red flag to a bull me thinks

My next problem is this:

I now have a working circuit of 20 leds on two LM3914
When the last LED lights ie FSD it flashes the LED array on and off. This all works good apart from it only flashes on & off on the last ten LED's, not all twenty.

Can anyone assist in telling me where I need to modify my circuit to enable all twenty LED's to flash on FSD (cct attached)

Thanks again
 

Attachments

hgmjr

Joined Jan 28, 2005
9,027
It is possible that all you need to do is connect the two "OUT" pins together so that both devices receive the same pulse when the last LED fires.

hgmjr

PS. I am not familiar with the abbreviation FSD. What does it stand for? Is it "Full Scale Deflection"?
 

Thread Starter

transducer

Joined Jul 28, 2007
25
That was my theory too, but it just loaded the last set of ten led's.
Where, looking at the cct diagram would you make the connection?
It may be I am connecting in the wrong place.

FSD= Full Scale Deflection, good guess

Cheers
 

hgmjr

Joined Jan 28, 2005
9,027
OK. Then how about taking another 1000 uF and 470 ohm resistor and connect it to the same signal source at the 1000 ohm resistor and feed it to "OUT" pin of IC2. This will provide a dedicated flash control for IC2 that is the same as the one now feeding IC1.

hgmjr
 

hgmjr

Joined Jan 28, 2005
9,027
Hey. Don 't knock it. That could be sold as a feature.

It might be helpful if you could take another stab at describing what you would like to see the lights do.

hgmjr
 

Thread Starter

transducer

Joined Jul 28, 2007
25
The original circuit does what it is supposed to do. As the voltage increases the led's light up until they are all lit. When the 10th LED lights up the whole array flashes on and off to indicate FSD. What I have done is extened my circuit from 10 to 20 led's. But only the 10 leds connected to the cap and resistor flash, the other 10 just stay lit.

So basicallly what i want is for all 20 led's to flash on & off when the array hits ref hi
 

hgmjr

Joined Jan 28, 2005
9,027
.....So basicallly what i want is for all 20 led's to flash on & off when the array hits ref hi
The requirement to flash all of the LEDs in the array sounds like an ideal place to employ a 555-timer chip. Is that something you would be interested in trying?

hgmjr
 
Top