I am connecting it to the output on the pc.

my problem is in dot mode more then 1 led is on at a time (like 3-4).

ive tired a cap and that didnt help, checked all the wires on the bread board for nada.
input voltage is 12v.

using this sch
http://www.national.com/mpf/LM/LM3914.html
except that r2 is a 10k pot.

any ideas?

 

You have noise somewhere; either on the supply to the LM3914 or on your input signal.

Quiet the noise down with caps.

You may need to add some resistance between the signal source and the LM3914 signal in, and add a cap from the LM3914 signal in to ground.

Let us know the values of the components you try, and what the results are.

 

.1u & 2200pF & 10u
470 ohm and 100k
(had these right at hand)
tried all the combos of those.

at best, i got the more to light up and shut off quicker.

but had number 1 on and the middle one on. middle being a constant on with either side flashing(and i could change the middle by moving the pot).

also, my power supply is a pc power supply.

 

PC power supplies are switchers. A certain amount of signal is necessary. You need to regulate this voltage with a linear regulator.

The reason is a switcher turns on and off between two voltages, creating a steady triangle wave on the voltage output.

 

Thanks, had a slight feeling it might be because of the switching power supply.

But just tried it was some batteries and it wasnt any better

 

Just occured to me what you've been saying. This is a bargraph display, it lights up a column up to where the voltage is. This is what you want, right?

 

Thanks, had a slight feeling it might be because of the switching power supply.

But just tried it was some batteries and it wasnt any better

According to the datasheet, the LM3914 has a 1mV overlap on changing to the next LED in dot mode. So at anytime either one or two LEDs is ON.

Use a battery as input signal to test the operation of your LM3914. If only one or two LEDs turns ON, then the problem must be your input signal if more LEDs are ON.

 

No overlap, bargraph. It doesn't light one LEDs, but all the LEDs under the voltage.

▐
▐
▐
▐ ▐
▐ ☻
▐ ▬

 

No overlap, bargraph. It doesn't light one LEDs, but all the LEDs under the voltage.

▐
▐
▐
▐ ▐
▐ ☻
▐ ▬

LM3914 can be set for bar mode or dot mode.

 

Yes is it in dot mode.

Tried the battery and yes that works except that led 1 is always barely on.

Should i just go for a mic circuit instead of straight to pc?

 

LM3914 can be set for bar mode or dot mode.

I stand corrected. Sorry about that. The block diagram doesn't really show how they do that.

Have you tried giving it a pristine voltage source, say a 12V battery and a potentiometer?

 

yea have done 12volts in batteries.

have also used a battery and 2 on the signal side and it worked properly, 1 led on at a time dependent on the voltage, except that led number 1 was always faintly on.

so i think we narrowed it down to signal side.

now the question is can i fix it or is it just my pc, and if not should i just make a mic circuit to feed the signal instead of straight out of the pc.

 

have also used a battery and 2 on the signal side and it worked properly, 1 led on at a time dependent on the voltage, except that led number 1 was always faintly on.

This is by design. If you read the LM3914 datasheet, you will notice the description quoted below.

The solution is to parallel a 10K resistor across LED#1.

An auxiliary current source at pin 1 keeps at least 100μA flowing through LED No. 1 even if the input voltage rises high enough to extinguish the LED.

 

The LSB is toggling because of resolution. I WOULD NOT WORRY ABAOUT THAT . With zero in it may toggle it cannot make up its mind if on or off.

 

Thanks, hadnt read that far into it yet since that wasnt my main problem yet (with most of them being on i couldnt tell).
haha

 

The LSB is toggling because of resolution. I WOULD NOT WORRY ABAOUT THAT

Your comment is misleading.

There is no LSB, only low voltage level and high voltage level for comparators on a resistive chain inside the LM3914.

If a higher input voltage level is reached, the comparator for the lower voltage(s) will NEVER toggle.

This is not a SAR.

 

This is by design. If you read the LM3914 datasheet, you will notice the description quoted below.

The solution is to parallel a 10K resistor across LED#1.

I couldn't find that quote in the National datasheet.
There is a discussion in the datasheet about putting 10k across LED #11, but I don't think our OP is using two LM3914s. Did I miss something?

 

I couldn't find that quote in the National datasheet.
There is a discussion in the datasheet about putting 10k across LED #11, but I don't think our OP is using two LM3914s. Did I miss something?

It's on page 9 and you are correct it is mentioned for the case of two LM3914s.

However, I am reading this also to mean that aux. current source being there and is active all the time if LM3914 is in DOT mode.

The datasheet does not mention how one can turn off this current source in DOT mode so it is logical(for me at least) to assume that the same current is actively flowing on Pin1, even with a single LM3914, giving rise to a faint LED as reported by the OP.

 

The datasheet for the LM3915 has some circuits for "peak detectors" that sample and hold the level long enough for your slow vision to see.

Without a peak detector circuit then audio is a dim blur.
With a peak detector then audio levels are very bright and clearly only one LED lighted at a time.

My Sound Level Indicator project has been working perfectly for more than 3 years.
It has an electret mic feeding a preamp and a peak detector circuit feeding an LM3915. It is always set to the BAR mode.
It detects a pin dropped on the floor of the next room and loud music in its room.

 

would the half peak work(one with the PNP and diode) or are you using the on with the op amps (Precision Full-Wave Peak Detector)?

though i do have some LM324Ns laying around if they would work...

thanks for all of your time