I built a circuit that runs from my amp to my computer. It steps down the large speaker output voltage to line level. There is audio hum that is fairly loud, but whenever I touch the phone jack that connects the computer to the circuit, the noise goes away. How can I get rid of the noise without being there to touch the circuit?
sounds like a grounding problem,you could take a volt meter and put one probe on the computer and one on the amp and see if there is reading there ,if so you could run a ground between them or use a matching transformer ,and if you do use a transformer ,maybe a line transformer it has alot of taps you could use to step down the voltage ,Have you tried the headphone jack from the computer for signal to your amp it's already low level,,,,,good luck
EDIT:I forgot to draw the transformer from 120V AC to 12V. It immediately precedes the diodes.
Schematic
Here it is. I blackboxed some of the simple resistive networks. The circuit works perfectly (aside from the noise). The capacitor has about 12V DC across it and I change that to +/- 5 and ground using regulators. The voltage is 4 or 5 volts peak at the isolation transformer and about 100mV at the voltage buffer. I'm pretty sure tying the audio output ground to AC ground would be a pretty bad idea because of the 5V difference (and possibly because it is AC ground).
Schematic
Here it is. I blackboxed some of the simple resistive networks. The circuit works perfectly (aside from the noise). The capacitor has about 12V DC across it and I change that to +/- 5 and ground using regulators. The voltage is 4 or 5 volts peak at the isolation transformer and about 100mV at the voltage buffer. I'm pretty sure tying the audio output ground to AC ground would be a pretty bad idea because of the 5V difference (and possibly because it is AC ground).
Last edited:
If your schematic is accurate, you have mistakenly used a 7805 twice, for both +ve and -ve supplies.
This is impossible:
You must use a 7905 for the NEGATIVE rail!!!!
Good save...sorry, I wrote that up in a hurry.
Right now I have the power cord taped to the metal leg of my desk to keep it quiet. I would assume that neither my desk nor I are grounded properly. Why does the noise go away?
Right now I have the power cord taped to the metal leg of my desk to keep it quiet. I would assume that neither my desk nor I are grounded properly. Why does the noise go away?
LT1013CP Datasheet
I am reasonably sure that this is the right datasheet. The label on the IC is near impossible to read.
I believe it is unity gain stable. I haven't had trouble transmitting the signal. If I grab the power cord coming from the outlet, the noise goes away, so I don't think that is related to the op amp's stability. I'm almost positive it's a grounding issue. If I tie the signal ground to AC ground, I am worried that I will get a ground loop when I plug into my computer.
I am reasonably sure that this is the right datasheet. The label on the IC is near impossible to read.
I believe it is unity gain stable. I haven't had trouble transmitting the signal. If I grab the power cord coming from the outlet, the noise goes away, so I don't think that is related to the op amp's stability. I'm almost positive it's a grounding issue. If I tie the signal ground to AC ground, I am worried that I will get a ground loop when I plug into my computer.
It sounds to me like an oscillation problem , the hum is being carried by RF . If you havent got something like 100 ohms in series with the output of the IC i suggest you try it . Also is the IC fully decoupled with 0.1uf ceramic caps on each rail pin etc and have the voltage regs got all the required decoupling .
Last edited:
Try the decoupling as stated above, it may help, but a full schematic would help, with the IC supplies and decoupling. Look at the signal return in respect of the computer jack, it is a wire with an inductor on the end, I would try connecting it to the supply ground and see if the noise goes away.
If the upper and lower circuits are isolated, I would wonder what that ground between the voltage regulators does (besides making an ugly ground loop). It doesn't appear to be seen by the op amp, only the + and - 5V for the op amp power. Does this ground connect somewhere inside the volume control block?
The diode bridge is mirrored and has a resistor load, which I assume is just hasty error.
The diode bridge is mirrored and has a resistor load, which I assume is just hasty error.
Last edited:
Hum which goes away when you touch something is almost certainly due to RF oscillation getting modulated by a rectifier diode.
More complete schematic
Sorry about the earlier schematic. There are a few errors. Like I said, the circuit works as intended, it's just really noisy.
The zeners in this schematic are used instead of the voltage regulators (my circuit sim doesn't have voltage regulators)
Under what circumstances should the power be decoupled? I wouldn't consider decoupling the power to, say a 7400 chip.
What would be picking up the RF noise? Would it be the transformer?
Thank you so much to everyone who has been contributing to this thread. I greatly appreciate it.
Sorry about the earlier schematic. There are a few errors. Like I said, the circuit works as intended, it's just really noisy.
The zeners in this schematic are used instead of the voltage regulators (my circuit sim doesn't have voltage regulators)
Under what circumstances should the power be decoupled? I wouldn't consider decoupling the power to, say a 7400 chip.
What would be picking up the RF noise? Would it be the transformer?
Thank you so much to everyone who has been contributing to this thread. I greatly appreciate it.
That circuit contains an op-amp follower. All followers are prone to instability, especially if they see a whiff of capacitance at the output. Add a resistor at the output, as close as possible to the opamp output pin. The value will depend on your load impedance, but somewhere around 100 ohms might work.
Every circumstance,IMO.
7400 series stuff should have a decoupling cap next to each IC,preferably right on the power pins. Remember,digital stuff is all square waves,with very quickly changing voltages,and lots of harmonic content.
I seem to remember Texas TTL books recommend one ceramic cap per chip, plus one electrolytic per row of chips. All this with careful design of ground and supply rails to maintain low resistance and inductance. Modern opamps can be faster than 7400 logic, so may need better decoupling. Don't use a faster opamp than you really need.
Ok. I tried the decoupling capacitors and they had no effect on the noise. (I tried one from +5 to -5 and when that did nothing, I tried one from +5 to ground and from -5 to ground.) Then I tried the 100 ohm resistor in series with the output, and the noise is gone....WHY???? I can't think of a reason why this would work, but it does--very well.
As I said
A bit of cable is sufficient to give enough capacitive load for instability. The resistor stops the follower from seeing the capacitor.DF96 said:
Wow. I am only using about a 6 inch cable to connect from the op amp to my computer. Is there capacitance at the input jack of the laptop?
- Status
- Not open for further replies.