Another music to LED project

After building Audioguru's LED flasher to music i want to build something more intense. That circuit was fine and did the job but I want something more tuned to bass. I did put a bypass filter guru's circuit but didn't lie the results.

So which brings me to my next project VU meter. Attached is a schematic of a VU meter i found on the internet. I wish i found it back when i was in my modding xbox phase . Now This circuit is designed for a "stereo output" RCA jack. I would like to have the flexibility to be able to use it for a mp3, computer speaker, or at the end of a stereo speaker (so i can put the LEDs in the speaker. is there anything i need to change? can i just put a potentiometer to limit the current(although there is already one in there)? I already see that this schematic has constant power and the creator just opens and closes the audio signal. Also, He combines the left and right channels is this going to pose a problem besides it being mono?

Also, In case i need to draw a schematic for you guys what would be a good program to draw it with?

and for 12 volts can i use a door bell transformer and use a bridge rectifier to convert to DC then add a capacitor and limiting resistors? seems to me the transformer might be to big for the job i rather just use a 9v although it couldn't do much for very long.

Thanks for any input,
Jeff

1) The LEDs are paired in parallel which is wrong. Two LEDs should be in series for each output of the LM3915.
2) The 680 ohm resistor sets the input voltage that lights the 10th LED and sets the current from each output at 17mA and the MODE pin is high so the IC is in the BAR mode. Since the supply to the LEDs is too high at +12V then if the LEDs are 2V red ones in parallel, when all LEDs are lighted the LM3915 must dissipate (12V - 10V) x 17mA x 10= 1.36W which is its absolute max allowed temperature (it is extremely hot). Even if the LEDs are paired in series and have a higher voltage, i would add a resistor in series with the supply to all the LEDs (and a bypass capacitor as shown in the datasheet) to share some of the heat.

If you use a "16V" door bell transformer then it makes a rectified and filtered 21V to 26V supply which is much too high. The circuit needs only 6V to 9VDC depending on the LED voltages.

The "peak detector circuit" made with the transistor's collector resistor charging the capacitor following the rectifiers has an attack time that is too slow. In my VU meter I use an opamp instead of the transistor and it drives an emitter-follower transistor that charges the capacitor very quickly. My opamp is a single-supply type so it automatically rectifies without the rectifier voltage loss of your circuit.
This is my preamp/peak detector circuit:

This circuit has a lot of problems with it. The bigest one as audioguru has pointed out, is that the LEDs are in parallel and not in series. The second problem is that it mixes the left and right audio signals together. It would be better if you could take an LM3915 and just make 2 of the same circuit for you purpose. You will need a peak detector if I'm not mistaken (I don't understand why though please elaborate if possible).

Music is full of short duration peaks.

A peak detector should charge its capacitor very quickly then discharge it slowly so that very short duration peaks of music are stretched in time to allow our vision to see the peak brightly. If a peak is less than about 20ms in duration then it will look dim or not be seen.

Thanks for spending time looking at the circuit. There is a reason for both sides being connected in parallel and that is that they are all doing the same thing. Putting them in series would create a voltage drop and both LEDs wouldn't be the same brightness. **broken link removed** this the project i am attempting minus the xbox. I do agree 12volts seem high though. I believe the ic needs to be in bar mode to perform its function, because only one LED would light if it wasn't.

I will give your OP-AMP a shot. If it doesn't make it for this circuit I'm sure i can find use for it somewhere ;-)

scarygood536 said:

There is a reason for both sides being connected in parallel and that is that they are all doing the same thing. Putting them in series would create a voltage drop and both LEDs wouldn't be the same brightness.

Click to expand...

Wrong!
The outputs of the LM3915 have the current regulated so voltage drops of the LEDs do not change the current.
Since each LED has a different forward voltage then in parallel one will be very bright and might even burn out and the other will be dim or not lighted.
Leds control the brightness with their current, not their voltage so when they are in series then their currents and brightnesses are exactly the same.

My VU meter project is almost the same as yours since one LM3915 drives 20 LEDs in two columns of 10 LEDs. My LEDs are paired in series.
My LEDs are much brighter than yours with a current of 27mA in each LED.
My supply is 7.2V from a rechargeable battery or 9V from the charger so I use a 10 ohm/1W resistor in series with the supply to the LEDs to share the heat.
Since my supply is less than yours then even though my current is higher then my total heat dissipation is almost the same as yours. My LM3915 and my 10 ohm resistor are warm, not extremely hot like yours.

My VU meter does not use a rectifier like yours so it does not have the error of the rectifier at low sound levels. So I made my project extremely sensitive so it can display a pin dropped on the floor in the next room and therefore I added an automatic gain control circuit that adds 20dB to its range and it also displays loud music, TV or my dog barking.

my project has a DOT/BAR mode switch but I like it in the BAR mode.

ok I think I understand what you are saying. Some fundamental electronics missing on my end. I think rather than share the heat i might put in a heat sink. I'll run it off of 9volts dc from a wall outlet charger. I will be ordering the parts soon.

And I think instead of a mic on the OP-AMP i am going to use regular inputs. (so i'll modify the gain and add in a pot.)

A switch on the mode doesn't seem like a bad idea either.

Thanks for your help btw!

It is very difficult or impossible to add a heatsink to a dual-inline plastic package. Instead share the heat with a power resistor like I did.
This is my schematic:

nice schematic. seems simple enough. just curious why didn't you include diodes in your circuit to limit the direction of current flow (esp having the battery attached to the same circuit as the charger)? What program do you use for your schematic? is it eagle or electronics workbench?

scarygood536 said:

why didn't you include diodes in your circuit to limit the direction of current flow (esp having the battery attached to the same circuit as the charger)?

Click to expand...

The charger charges the battery then trickle charges it most of the time. The charger also powers the circuit whenever it is not running on the battery.

What program do you use for your schematic?

Click to expand...

I use Microsoft Paint program to make all my schematics and to correct mistakes on schematics posted by others on these forums.
I copy and paste parts from other schematics and from datasheets.

I make most projects on stripboard (Veroboard). Here is the inside of my VU meter showing the DOT mode:

nice thanks for the tip! I look forward to attempting this project.... time to do some shopping at Jameco!!!

Thanks for all your help,
Jeff

https://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_23683_-1https://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_23683_-1 will this work for the op amp? Jameco only lets my buy the mc3317 in multiples of 2500 (that's a little out of my budget). Also, your 330nF (.33uF) capacitor of what type is it? ceramic?

An LM324 is a very old single supply quad (4) opamp. An LM358 is the same but has only two opamps and has the same 8-pins case as the much better MC33172 dual opamp that I recommend.
The LM324/LM358 have a fairly low bandwidth since they get into trouble above only 2kHz. The MC3317x opamps work well up to 35kHz. The LM324/LM358 have pretty bad crossover distortion. The MC3317x ICs have none. They have the same low voltage operation and low supply current.

Digikey have more than one thousand MC33171 single opamps in stock today for $1.02US each. They have nearly 2000 MC33172 dual opamps today for $1.47US.
They are in the DIP case. They also hve surface-mount cases. I didn't bother looking at MC33174 quad opamps nor look at Newark.

I use metallized plastic film capacitors up to 1uF for audio coupling becasuse they are fairly small and work perfectly. A ceramic capacitor causes low frequency distortion because its value changes with the modulation. It is microphonic. A ceramic capacitor makes a good supply bypass capacitor.

All he wants to do is flash a few LEDs, does it really matter if the op-amp has poor bandwidth and the ceramic capacitors distort the signal?

I bet you'll hardly notice the difference.

I selected opamps that work with all sounds. The lousy old LM358 and LM324 do not work with high frequency sounds that I can hear.
the MC33172 and MC33174 opamps fixed all the problems with the LMxxx opamps and have the same low supply current, operation with a supply as low as 3V, inputs that work at 0V when there is no negative supply and an output that goes close to 0V. A single opamp the MC33171 is available.

Ceramic capacitors are poor for coupling audio and are also microphonic which causes probems in a very high gain audio amplifier like in my VU meter.