Your VU meter circuit is completely wrong:
1) The LEDs are backwards with the anodes at ground and the cathodes driven to +12V.
2) The input diode passes DC to the input of the LM3915 and ruins the accuracy of the lower LED turn on voltages. The diode and the 2.2uF capacitor it charges form a very poor peak detector that shorts the amplifier output badly each time the input goes 0.7V or more positive that could blow up the amplifier and/or the diode.
The datasheet shows a simple opamp peak detector circuit that I have used and it works perfectly. It is accurate and has an input capacitor to block the 6.4VDC from the amplifier output. The opamp needs a dual polarity supply or an opamp that has inputs that work all the way down to ground like one of the two opamps in an LM358.
You cannot add a series capacitor to your horrible diode circuit.
Since the amplifier output is +6.4V then it must have a +12.8V supply. Then if the horrible unregulated pot in your circuit is set so that the 10th LED lights when the input peak is 10V, the series diode's voltage drop prevents the 1st and 2nd LEDs from lighting when they should. A peak detector with an opamp will allow all LEDs to light when they should.
In your circuit, the +6.4VDC will cause the 8th LED to light in the DOT mode or cause LEDs 1 to 8 to light in the BAR mode with no signal.
The horrible pot in your circuit is its unregulated reference voltage but the LM3915 alredy has a regulated reference voltage at pin 7 that is set with two resistors on pin8.
You show 2N3906 transistors with a maximum allowed current of only 200mA driving LED strips that might draw 500mA each?
You have each output of the LM3915 driving 1k resistors in series with the base of the transistors. Then when all LM3915 output are driving these resistor the LM3915 will be too hot and will fail. The datasheet recommends reducing the output currents so the heating is not too high. How much current does each LED strip need?
Here is my peak detector circuit:
Thank you very much Sir for your detailed suggestions. I now understand that I should have taken much more seriously in understanding the faults in the circuit. the circuit was found from the internet and I had build another project based on it and it works, so i didn't go in depth in understanding the concept. I now totally know that I shouldn't have gone blind over the circuit.
1) yes the leds are connected backwards but I have connected connected correctly in my project, sorry about that mistake
2) I'll remove the diode and capacitor from the circuit
3) In my setup i"m using two voltage input for the VU meter board. the voltage fed to the LEDs during the transistor switching is given by the +ve output of a WiFi led controller. The voltages to the pin 3 and and the input to the 47K potentiometer is fed by a 12V buck/boost converter.
I'm already using these boards in the same project for other functions,so just used these to use on VU meter board also for a steady supply.
4)I'm using 12 V RGB led strips for the 10 outputs of the VU meter . the strips are have three individual leds per section. I'm using a continuous 4 sections for each of the outputs of the VU meter board, so that makes 4x3=12 individual leds in each row of the 10 output sections of the VU meter output line.
5) So is it over 200mA current limit?? but I wont be using it in too brightness.
6) The LM358 is a single supply op amp right?? is it capable of working in the output ranges of the of the amplifier?? in the datasheet its written that it could withstand upto 32V
7) Yes i remember some heating issues with the IC in my previous circuit based on the same IC and also after few months 3 strips were not working, but back then I thought it was normal for these ICs to heats a little on continuous use and was under the assumption that it might be the loose wiring to the leds that makes it not to light up. Now I totally understand the case. So what should be done to reduce the heating??