LM3914 LED supply Voltage

[chrisspartan]

Hi, my first post so please be gentle.

I am building a fuel gauge with a LM 3914 and have read the data sheet a number of times and read quite a few of the posts on here.

I understand basically how the LM 3914 works - what I dont understand is what voltage is applied across the LED's ie where LED V+ comes from? is it just the supply voltage V+ or is it an independant voltage source just for the LED's?

I understand the current through the LED's is controlled by 10 times current out of pin 7 but as I am planning to use standard 2.5V 20 mA 5mm LED's so I cannot work out how you apply the required 2.5V across the LED's.

I know I may be missing something fundamental but I don't get it?

Thanks

Chris

 

[duffy]

Yes, VLED is almost always connected to V+. They show it separate in the spec sheet mainly to confuse noobs. They are being dicks. Look lower down the page and you will see they are connected together.

https://www.electro-tech-online.com/custompdfs/2009/01/LM3914.pdf

 

[audioguru]

LEDs are driven with current, not voltage.
The LEDs set their own voltage. The LM3914 sets their current.
If the LM3914 is set for a bar-graph then when all the LEDs are lighted the LM3914 might get too hot because it must dissipate 10 times the current of one LED times the voltage across the output transistors of the LM3914.
Then a power resistor can be used to feed the LEDs and the resistor will share the heat. The datasheet shows this and shows a 2.2uF capacitor to ground.

 

[duffy]

That too, but mainly because they are being dicks.

 

[k7elp60]

Just because the LED is rated at 20mA doesn't mean you have to pass that current thru it. The current affects the intensity, they may be perfectly bright will 10 or even 5mA.

 

[chrisspartan]

OK thanks for that fellows, I was thinking that if the supply voltage was say 5V then it would blow a 2.5V LED but obviously not.

I was planning to lower the supply voltage down to 5V or so using an LM317 to try and avoid the heat problem (I assume the LM317 will get hot instead) but will also bear in mind I can disipate some heat/voltage through a LED power reisitor.

Thanks - probably not for the last time!

 

[duffy]

Yes, lowering the voltage would be the thing to do.

>>I assume the LM317 will get hot instead

NOW you're thinking like an engineer. Yes, you are moving the heat from the chip to the series pass regulator. At 20ma and 5V and 10 red led's (1.7V drop), the 3914 is dissipating over half a watt - 660mw. Lower that voltage to 3.5V and now it's only dissipating 360mw, nearly cuts it in half.

 

[audioguru]

The LM3914 regulates the current for the LEDs so you don't need a voltage regulator for the LEDs. Just use one simple cheap power resistor to share the heat instead of the LM317 regulator.

 

[duffy]

What, that current programming resistor? It doesn't handle significant wattage, it just sets the internal regulation. The chip dissipates the heat.

 

[audioguru]

What, that current programming resistor? It doesn't handle significant wattage, it just sets the internal regulation. The chip dissipates the heat.

I was talking about using a resistor in series with the supply to the LEDs.

Your example of a 5V supply (which is too low), 1.7V LEDs and all 10 are lighted at 20mA each results in dissipation of only 660mW in the LM3914. Its max allowed dissipation is more than double that.

If the supply is 12V, the LEDs are 1.7V and all 10 are lighted at 20mA each then the LM3914 will try to dissipate 2.06W. A 39 ohm resistor in series with the 12V supply to the LEDs will dissipate 1.56W then the LM3914 will dissipate only 0.5W.

 

[duffy]

My apologies, I see you posted that in an earlier response. Why are you saying 5V is too low? The spec sheet says it can work all the way down around 3V, and I see no constraints on the project's operating voltage so far.

 

[audioguru]

The input offset voltage affects voltages that are low.
The voltage reference is regulated so you should use a supply as high as possible (its max supply voltage is 25V).

 

[duffy]

The input offset voltage affects voltages that are low.

Do you mean the input voltage range? The input offset is only 3mv and unaffected by supply voltage. At 5V he still has 0-3.5V, plenty of dynamic range. I've used this chip many times at 5V.

The voltage reference is regulated so you should use a supply as high as possible (its max supply voltage is 25V).

I disagree. Don't stress the chip, limiting resistor or not, run it at a low voltage.

 

[audioguru]

Hi Duffy,
Where can you buy a "typical" LM3914? Nowhere.
The max input offset voltage is 10mV, not 3mV.
With a 5V supply then the max setting of the Vref and the max input voltage is 3.5V.
The turn-on voltage for the 1st LED is only 350mV so the 10mV input offset voltage represents an max error of almost 3%.

If you use a 12V supply then the max error is less than 1%.

I have used the LM3915 many times. It is more than 2 times more sensitive than the LM3914.

 

[duffy]

I've used the LM3915, too - it's logarithmic. Peas and apples.

With a 5V supply then the max setting of the Vref and the max input voltage is 3.5V. The turn-on voltage for the 1st LED is only 350mV so the 10mV input offset voltage represents an max error of almost 3%.

Another way to look at it is that each discreet step represents over a quarter of a volt of input change with a maximum uncertainty of .01V at each step. Quite a bit of precision for a bargraph.