LM3915 with diffused 3mm LEDs

[dragosmze]

Hi,
I have build a simple vu-meter with LM3915, I have used diffused 3mm LEDs and I have a little problem when one LED lights up the other ones that are ON are flashing a bit. I do not like the 'effect'. Is something wrong in the scheme? Maybe a too low current for the LEDs
It's there a simple change I can make to let the LEDs keep the same brightness 100% of the time (once they are ON)?

[Jan Dupont]

Make sure your supply can deliver at least 100 mA. Can you post a schematic of the power supply ?

[dragosmze]

I don't think that should be a problem, it's made with a LM117 down form a large supply (I also power up other things from it but it's not even near the max current it can give; but I'll check to be sure

[Mooly]

If the LED's are next to each other make sure you are not seeing "light bleed" from one to another... probaly not but just though to mention.

I've never used the LM series of drivers so can't comment on how uniform the led brightness is. Best way to tell may be to apply a DC input and make sure they light correctly and evenly as each in the chain is illuminated.

Is there no time constant for attack and delay ? selectable by caps etc. Maybe not.

I'm thinking as I go along.... you might need a sample and hold type circuit on the input so that it see "slow" changing signals... think that's your issue

[Mooly]

I mean peak hold... not sample and hold. That's a simple circuit with fast attack and slow decay so that transients are registered and then they are slow to decay so the meter shows the LED's as illuminated for a few milliseconds or longer.

[Mooly]

Something like figure 4 here,
LED Audio VU Meter

[dragosmze]

Actually I have added a small R to drop the voltage a bit on the LEDs, but without it I can still see the flashing (the LED's are also brighter).

If I apply 12V to the input they all light with the same intensity .. and by eye checking I can say it's the same like when I take the signal from the output of my power amp (less the flashing thing).

I also used a piece of paper to separate the first LEDs, just to be sure my eyes does not play a trick on my

I think they go OFF just for a small amount of time .. and that's why I see them flashing, not sure why.

[dragosmze]

Because it was hard for me to get a variable voltage from somewhere just now
found a AAA battery and a pot and from what I can see .. they all will dim a bit with each new LED that will go ON

[Mooly]

A series resistor will cause the brightness to drop as more LED's illuminate and is to be expected.

The LM3915 is an LED bargraph driver and if you feed it audio then it will illuminate irregularly. That's why you need the peak hold type circuit to firstly rectify the audio into DC so that it can be presented to the meter and secondly to "hold onto" the peak DC levels for a reasonable time so that the meter illuminates cleanly and you have chance to see it.

If you feed audio into an ordinary analogue meter or a DVM it won't register correctly. The audio has to be rectified first

[jean-paul]

You could use a decoupling cap directly from the power supply pin to GND just as with all other electronics. Just like the data sheet says to avoid oscillation and/or flickering of the LEDs.....