Noise

Thread Starter

MikeD_72

Joined Nov 11, 2008
46
I've breadboarded this schematic and have a couple of questions about it.

Firstly, with power on and the audio input disconnected, the audio output becomes extremely noisy. I don't know what's causing this. I'm trying to improve this circuit, and getting rid of this noise is one of my top priorities. Any ideas what I can do about this problem?

Another improvement I'll be making is to switch over to bar graph drivers (LM3916) for my output.

The second question is about my power supply. I'm trying to squeeze +/-6Vdc out of a 12Vdc wall transformer. I don't think this idea ranks on the genius level :D At any rate, when I measure Vcc with my analog multimeter (I've got a digital on order!), the needle swings wildly below 6Vdc. Is this the result of the 7806 regulator dropping out repeatedly? Why doesn't the same thing happen to the Vee signal? It maintains -6Vdc quite well. I'm thinking of moving to 5V LDO regulators.

Thanks,
 

hgmjr

Joined Jan 28, 2005
9,027
One observation is that the +6 volt supply is the only supply that is powering the LEDs. The negative supply is never called upon to power the LEDs.

hgmjr
 
Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
Another item is that the 7806 and 7906 each have a 2v dropout. Since most of your load is on the +6 side, you'll wind up with around 2v on the +6 side, and a nicely regulated -6v.

You might use a power opamp like an LM675 to generate an artificial ground instead. See National Semiconductor's LM675 datasheet.

When you disconnect your audio input, the opamp begins oscillating because there is no load on the input. I'm curious why you started off with a 50 Ohm network, and then went into an inverting amplifier?
 

Thread Starter

MikeD_72

Joined Nov 11, 2008
46
hgmjr,
It was my suspicion that the current draw on the 7806 was causing it to drop out. Thanks for your feedback.

SgtWookie,
In actuality, when the LEDs are off, the circuit maintains a nice +/-6V reference. When the LEDs start turning on, then the +6V reference starts to drop - I suppose this is caused by the current dependence on the dropout voltage.

I'm not sure exactly how the LM675 is supposed to act as a split-supply generator. Perhaps you can clarify this for me? Looking on page 2 of the datasheet, I don't understand how V- is being generated using the resistor and capacitor. Also, this IC requires an input of >16Vdc. I only have 12V to work with.

The 50 Ohm resistance is part of the inverting amplifier. The fact that I stuck two in parallel is just a result of what components I have at my disposal.
 

SgtWookie

Joined Jul 17, 2007
22,230
I'm not sure exactly how the LM675 is supposed to act as a split-supply generator. Perhaps you can clarify this for me? Looking on page 2 of the datasheet, I don't understand how V- is being generated using the resistor and capacitor. Also, this IC requires an input of >16Vdc. I only have 12V to work with.
The 1 Ohm resistor and 0.22uF cap make up a Bucherot cell. This acts as a 'snubber' to eliminate high frequency transients.
http://en.wikipedia.org/wiki/Boucherot_cell
A smaller power opamp like the L272 would work fine in your application.

V- isn't being "generated"; only the "artificial ground" (GND) reference by the output of the opamp. It just takes what input voltage is present, and splits it in two.
 

Thread Starter

MikeD_72

Joined Nov 11, 2008
46
The 1 Ohm resistor and 0.22uF cap make up a Bucherot cell. This acts as a 'snubber' to eliminate high frequency transients.
http://en.wikipedia.org/wiki/Boucherot_cell
A smaller power opamp like the L272 would work fine in your application.

V- isn't being "generated"; only the "artificial ground" (GND) reference by the output of the opamp. It just takes what input voltage is present, and splits it in two.
I will look into this promptly. However, I was also just now considering a 10V regulator and splitting this with some 2% resistors. The accuracy would probably be decent enough for my application. Any thoughts on that?
 

SgtWookie

Joined Jul 17, 2007
22,230
I will look into this promptly. However, I was also just now considering a 10V regulator and splitting this with some 2% resistors. The accuracy would probably be decent enough for my application. Any thoughts on that?
It will result in very poor regulation.

Right now, most of your load is between ground and V+, and that load varies significantly depending upon what frequencies are passing through.

Seems like at the moment what's most critical is for you to have a stable V+ rail to ground voltage. But, if you used a single LDO adjustable regulator with a power opamp like an L272 generating a virtual ground, that should pretty much take care of your needs.
[eta]
See the attached. I used an LM317 regulator because the sim I used didn't have an LDO regulator available.
The LM317 has a 1.7v dropout. It's still much better than the 78xx/79xx series, and provides much better regulation overall. As drawn, the LM317 is dependent upon the L272 drawing at least 5mA current to provide guaranteed regulation. As the typical current consumption of an L272 is 8mA, this should be sufficient, unless very low LM317 output voltages are selected via R2.

R2 adjusts the voltage between V+ and V-.
R3 adjusts the voltages between V+/SimGND and SimGND/V-.

This configuration may be advantageous to you.

If you really wanted a fixed voltage between V+ and SimGND, you could replace R3 with a Zener from V+ to the wiper connection, and a constant current circuit from the wiper connection to GND - but that's getting into maximum overkill, methinks.
 

Attachments

Last edited:

Thread Starter

MikeD_72

Joined Nov 11, 2008
46
The circuit has problems:
(I edited the attachment)
Audioguru, thanks for your input. Your attachment erroneously says I'm using an LF357. I'm not, I'm using an LF347. Just pointing that out in case it's a point of confusion between us.

Can I take care of the electrical short you are referring to if I increase the resistances to 1k and a 5k pot? The pot would then typically sit at around 3k.

I was not aware some op amps had a minimum gain. Is there somewhere to find it on the datasheet?

I was suspicious that the capacitance in my breadboard may be causing the oscillations in the op amp when the audio input is disconnected. Do you suspect that the problem will alleviate itself when I go to protoboard/pcb?
 

Thread Starter

MikeD_72

Joined Nov 11, 2008
46
It will result in very poor regulation.

Right now, most of your load is between ground and V+, and that load varies significantly depending upon what frequencies are passing through.

Seems like at the moment what's most critical is for you to have a stable V+ rail to ground voltage. But, if you used a single LDO adjustable regulator with a power opamp like an L272 generating a virtual ground, that should pretty much take care of your needs.
Thank you for the advice, I will draft up a schematic and post it here.
 

Audioguru

Joined Dec 20, 2007
11,248
Sorry, I confused your old LF347 with the very old LF357 opamp.
The LF357 was a very wideband opamp in its days with a minimum allowed gain of 5.

Your LF347 is also not a low noise opamp.

The long jumper wires on a breadboard cause mains hum and buzz pickup and possibly oscillation.
 

Thread Starter

MikeD_72

Joined Nov 11, 2008
46
SgtWookie, can I analyze the network around the L272 using the golden rules of op amp design? The voltage drop over the 1k resistor seems to mess things up because V- will not equal V+.
 

SgtWookie

Joined Jul 17, 2007
22,230
SgtWookie, can I analyze the network around the L272 using the golden rules of op amp design? The voltage drop over the 1k resistor seems to mess things up because V- will not equal V+.
I don't see why you couldn't use the golden rules - unless you have your voltage references confused.

Keep in mind that the input should be considered "floating" - there is no real ground reference. This is a problem when you try to analyze it. To simplify things and allow analysis via the computer, I've used ground references liberally.

Keep in mind that the input impedance of the opamp will be very low, and that the Bucherot cell on the output will take care of any oscillations that may take place.

If the schematic is too confusing at the moment, replace the 1k resistor (R4)with a wire, and the 15k resistor (R5) with a wire. Then the power opamp will become a simple voltage follower. It won't respond to transients as well.
 
Top