problem with the VU meter circuit

[Adhith]

Hello everyone...
I have posted a previous thread on seeking help  for the circuit for a VU meter. It has been made now and was working fine but now I encounted a new problem. The lighting of the LED was now very weird as if there is some noise in the input audio. Its lighting in more over like in the bar mode option(even when the pin 9 is floating) but with very low lighting response. What could be the  possible reason?? Does the peak detector at the input side is creating some sort of noise so that the response is bad. On the other hand it works perfectly some time and goes back to the bad condition again after. there is no loose joints or soldering since i checked it thoroughly.

just checked the pin 9 of the IC and its found that there is a voltage around 10.5 V with respect to ground. Removed the IC from the base and then probed the IC base leg for pin 9 and there is no voltage. So its clear that somehow the voltage is getting to the pin 9 through the IC. Is this a normal thing or it the cause of my problem??

The circuit diagram is attached below.

[Audioguru]

Instead of using pins 7 and 8 on the LM3915 as a voltage regulator for the pin 6 reference voltage, you use a pot on the 12V supply voltage that is probably noisy.
My LM3915 VU meter works fine in BAR or in DOT modes. You copied my peak detector circuit but I can't remember which opamp you used. My peak detector is fed from an opamp used as a mic preamp.
Are you using a shielded audio cable for the input source?

[Adhith]

Instead of using pins 7 and 8 on the LM3915 as a voltage regulator for the pin 6 reference voltage, you use a pot on the 12V supply voltage that is probably noisy.
My LM3915 VU meter works fine in BAR or in DOT modes. You copied my peak detector circuit but I can't remember which opamp you used. My peak detector is fed from an opamp used as a mic preamp.
Are you using a shielded audio cable for the input source?

Thank you very much for your reply. The circuit was working very fine in both bar and dot without any problem for the past few days. The problem arised just a day ago. Its seems very much like a loose contact problem since it will be working perfectly sometimes and sometimes it wont. But there is no loose contact issue since i checked it thoroughly and also re soldered every spot and still no use. So I guess the potentiometer wont be an issues.

Yes i coiped the peak detector that you have recommended for me in the previous thread. The IC used is LM358.
The LED display is lighting up and no issue in that but the problem is with the response of the lighting up. The LEDs are just blinking with poor response as if I have fed a really bad audio input and thus i was under an assumption that the peak detector is adding up any noise or something in the input audio.

No I'm not using a shielded cable but fed the audio input from another source and still the problem persist.

[Adhith]

founded out another problem which i think is the cause of the weird response of the VU meter. All the 10 emitter legs of the transistors are connected together which is then connected to the +Ve output terminal of the LED driver. The R, G, B output terminals from the driver are directly connected to the LED strips. But now I disconnected the +Ve supply and the R,G,B connection from the driver and instead connected a +12v terminal from a DC adapter to the transistor emitter legs and the negative terminal to the Red terminal of the strips. now it is found that all the strips are flickering continuously without any audio signal or even when the VU meter is turned off.
On further look it is found that while removing the LM3915 IC from the base there is no LED light up or flickering. So its clear that Somehow the IC is producing some noise at its output legs.. So I think this would be the problem but I don't know how to solve it.

[Audioguru]

Test that the LM3915 can turn off all the LED strips by shorting its pin5 input to ground. If the LEDs do not turn off then replace the LM34915 IC.
If the LEDs do turn off then short the input of the peak detector (disconnect the signal source from C4 and replace it with a short to ground). If the LED strips do not turn off then replace the LM358.
Did you properly disable the unused opamp in the LM358 so it does not oscillate and cause interference?

You MUST use a shielded audio cable to feed the signal source to the peak detector to avoid mains hum in the air causing an interference signal.

[Adhith]

the flickering of the LEDs are seen while powering the transistor with the 12V Dc adapter without powering th actual VU meter circuit. So the se of experiments that you mentioned should be done without powering the VU meter right??
Shorting pin 5 to ground doesnt turn off the flickering. nor replacing the LM3915 IC
Remove the LM358 from the base and still the leds flicker. So I guess removing the capacitor wont do anything good either.
Removed the audio input and pwered from a seperate amplifier externally(which used shielded wires) still no use. the problem presist.

[Adhith]

Test that the LM3915 can turn off all the LED strips by shorting its pin5 input to ground. If the LEDs do not turn off then replace the LM34915 IC.
If the LEDs do turn off then short the input of the peak detector (disconnect the signal source from C4 and replace it with a short to ground). If the LED strips do not turn off then replace the LM358.
Did you properly disable the unused opamp in the LM358 so it does not oscillate and cause interference?

You MUST use a shielded audio cable to feed the signal source to the peak detector to avoid mains hum in the air causing an interference signal.

the flickering of the LEDs are seen while powering the transistor with the 12V Dc adapter without powering actually powering the VU meter circuit(to pin 3 and 6). So the set of experiments that you mentioned should be done without powering the VU meter right??
Shorting pin 5 to ground doesnt turn off the flickering nor replacing the LM3915 IC
Removed the LM358 from the base and still the leds flicker. So I guess removing the capacitor wont do anything good either.
Removed the audio input and powered from a separate amplifier externally(which used shielded wires) still no use. the problem persist.
Does the transistor some how leakes the voltage of something because a voltage around 2.5 to 2.7V is seen at all the  collectors of the transistor with respect to ground.

[Audioguru]

Since the LEDs flicker with the LM3915 turned off then obviously the problem is caused by the WIFI driver or the darlington transistors.
Are the MPSA64 Darlington transistors real ones or are they ebay cheap Chinese fakes? The 10k resistors between the base and emitter of the darlingtons will turn them off when the LM3915 is turned off.

[Adhith]

Since the LEDs flicker with the LM3915 turned off then obviously the problem is caused by the WIFI driver or the darlington transistors.
Are the MPSA64 Darlington transistors real ones or are they ebay cheap Chinese fakes? The 10k resistors between the base and emitter of the darlingtons will turn them off when the LM3915 is turned off.

Thank you again sir. The MPSA64 are the real ones and not the cheap ones from ebay. also I have checked the transistor legs to see if its according to the standard pinout and it is a standard one. Since the board worked perfectly both in dot and bar mode before, i cant blame the transistor either. while the VU meter is turned off and the transistor is powered by the AC-DC adapter the flickering occurs and its absent when the LM3915 IC is removed from the base. So I guess the IC has some role in the flickering. replaced with another IC and still it persist. Im attaching a pic of my actual board here. In the board, all the transistor emitter is connected by a jumper wire above the board. The inverted C wire are those connections. So does it forms any sort of small inductor or something to create some disturbance or so since it is very much near to the transistor??

The reason that I'm guessing its some sort of disturbance is because it ones worked perfectly and now it doesnt' light up according to audio as before. and also it lights up more over like in a bar mode even when pin 9 is floating.