VU meter

[mik3ca]

I'm trying to make my own VU meter, and I only have partial success. What I would like is full LED from full volume. My LED doesn't dance to the music

The input is a resistor (which varies the strength) and a capacitor in series. This capacitor is connected to any audio source directly.

this circuit uses a 2N4403 PNP and a 2N2222 NPN. The last NPN is used to give the LED a boost.

I think I am OK with choosing resistor values, but the capacitor values are trouble. If I choose them too high, then I will have to wait while they charge.

How do I choose the correct capacitor values?

I am interested in converting normal audible sounds into LED signals.

View attachment 39864

 

[audioguru2]

Hi Mik3ca,
Your circuit doesn't have a current-limiting resistor between the transistors. So when the PNP has a high input level then its collector current is unlimited and one or both transistors will blow up if the battery can supply enough current.

I don't know why you have a filter capacitor. Usually a capacitor is used with a rectifier circuit to hold the highest level for a moment so you can clearly see it.

You are trying to make 1/10th of an LM3915 or LM3916 bargraph driver IC. Why not use one, with the Half-Wave Peak Detector circuit shown in their datasheet? One pin selects between many LEDs lighting as a bar, or just one lighting as a moving dot.

 

[mik3ca]

Hi Mik3ca,
Your circuit doesn't have a current-limiting resistor between the transistors. So when the PNP has a high input level then its collector current is unlimited and one or both transistors will blow up if the battery can supply enough current.

I haven't had a transistor blow up in this configuration.

I don't know why you have a filter capacitor. Usually a capacitor is used with a rectifier circuit to hold the highest level for a moment so you can clearly see it.

Talkingelectronics.com uses a diode in place of my capacitor. I tried that configuration without much success. When the diode is connected this way, doesn't it create a varactor? Which means there is guaranteed capacitance.

You are trying to make 1/10th of an LM3915 or LM3916 bargraph driver IC.

then I must be on the right track.

Why not use one

Because I don't want IC constraints to affect my design.

with the Half-Wave Peak Detector circuit...

I am going to search for that. Thank you.

 

[MP1]

mik3ca,
Here is a standard circuit of vu meter that is made with transistors. Hope it helps.

MP

View attachment 39867

 

[audioguru2]

In the VU meter project at Talking Electronics, the PNP transistor, diode, input capacitor and collector capacitor make a Half-Wave Peak Detector (rectifier circuit) with a fast rise time and a longer release time of its storage capacitor. Then each LED lights long enough that you can see its level clearly. Without the diode and storage capacitor the LEDs would be a dim blur.

The 47k resistors between the PNP transistor and the NPN transistors limit the current like I was saying so that the transistors don't blow up and that the battery isn't killed from excessive wasted current. The 47k resistors also allow the storage capacitor to discharge slowly.

The diode is a rectifier, not a varactor.

 

[mik3ca]

that was the circuit I was talking about.

That circuit barely works for me IF the resistor value is low enough.
Sometimes, when I use the diode, the LED just stays on bright, and never dims.

So the RC circuit in parallel determine the holding time?

What if I were to replace the diode with one that can handle more power like the 1N4007?

 

[audioguru2]

That circuit barely works for me IF the resistor value is low enough.

Which resistor, its input level pot? It needs a pretty high signal level, 0.65V peak which is 460mV RMS. You should amplify its input with a preamp like they show.

Sometimes, when I use the diode, the LED just stays on bright, and never dims.

The circuit is missing a resistor across the PNP's base-emitter to turn it off. Try 100k.
Maybe your input wires are not shielded and are picking up mains hum.

So the RC circuit in parallel determine the holding time?

Yes, about 22ms. A very quick blink.

What if I were to replace the diode with one that can handle more power like the 1N4007?

Power?  It is only micro-watts.
A 1N4148 has a clear case and light makes it behave like a photocell a little. Some 1N400x's have a black case and light won't affect it. Try it.

 

[mik3ca]

The circuit is missing a resistor across the PNP's base-emitter to turn it off. Try 100k.

I tried adding resistors, and making it more like talkingelectronics version, but that didn't help.

Maybe your input wires are not shielded and are picking up mains hum.

At the moment, my input wires are about 6 inches long, because this section is in the beta stage, and my radio portion works.

Yes, about 22ms. A very quick blink.

I want to minimize holding time, in case I pick up sudden changes in the signal. In fact, I would like to pick up some white noise too!

A 1N4148 has a clear case and light makes it behave like a photocell a little. Some 1N400x's have a black case and light won't affect it. Try it.

Why would I need a photocell to detect a signal?
Is there a way to avoid the need of a photocell? (or a diode that behaves like a photocell?)

 

[audioguru2]

I tried adding resistors, and making it more like talkingelectronics version, but that didn't help.

You said the LED stayed on. The PNP transistor is missing a resistor to turn it off.

At the moment, my input wires are about 6 inches long, because this section is in the beta stage, and my radio portion works.

6 inches of wire is a good antenna for interference pickup. The input wire must be a shielded audio cable.

I want to minimize holding time, in case I pick up sudden changes in the signal. In fact, I would like to pick up some white noise too!

22ms is a short blink. Try reducing the value of the capacitor for a shorter hold time then the LED won't be bright enough for you to see short duration signals.

Why would I need a photocell to detect a signal?
Is there a way to avoid the need of a photocell? (or a diode that behaves like a photocell?)

Maybe your problem is caused by the clear 1N4148 being affected by light. Try a black diode to eliminate the light problem.