Interfering circuits

[Dr. Val Rousseau]

Hello,

I have two circuits that work perfectly well when connected to independent power supplies (see figure below). The first one is just a standard square wave generator using the NE555. The measured frequency is 960 Hz. The second circuit is an audio mixer that allows me to select via a switch between an internal electret condenser microphone (ECM), and stereo RCA input. Since the ECM produces a very low signal, it is preamplified in order to bring the signal to a level of the same order of magnitude as that present on RCA input. Then the selected input, microphone or RCA, is amplified or attenuated by adjusting the potentiometer in the feedback of the second op-amp.

The problem arises when the two circuits are plugged to the same power supply, and that the microphone input is selected, AND that the gain of the second op-amp is set too high. Then the symptoms are that the output of the audio mixer becomes chaotic. I tried to plug some 100n and 470uF capacitors in parallel to each of the ICs, but without any improvement.

Any idea on how to solve the problem?

 

[73's de Edd]

If the 555 circuitry is not even powered up is there being that instability problem?
Is there a good shielded pairing involved in the RCA connectors if being remoted off the board?
I hope that the pot is not being remoted off the board.

73's de Edd

 

[Dr. Val Rousseau]

There is no instability at all if the 555 is not powered up. The problem does not occur when the RCA input is selected (RCA connectors are soldered on the board). It occurs only with the internal microphone and high gain.

Now, the potentiometer is indeed off the board, connected by 4-inch long wires. Is that a problem?

 

[73's de Edd]

Sir Val . . . .

Since the 555 circuitry is pulling but a mere 11 and 9/17ths ergs, why not drop in a 220 / 330 / or 470 ohm isolation resistor and filter with a 100ufd at the 555 pin 8.
Then see what the difference is .

Is there any introduced capacitance from your hand at that pot or is a non metallic shaft and a isolative plastic knob involved.

Disconnect your wiring feeding out to the pot and then evaluate at two different gain levels by tacking in a 10K fixed resistor right AT pins 6 and 7 and evaluate, then sub in a 100 K to evaluate.

if it fails in that testing, you still have the option of backing off on stage 1 gain.

73's de Edd

 

[AnalogKid]

There is no decoupling of anything anywhere.
There is no indication of how grounding is handled among the circuit elements.
There is no coupling capacitor between the first and second opamp. Left amp offset voltage could drive right amp near or into saturation at high gain, giving chaotic results.

ak

 

[Audioguru]

I agree that many opamps go crazy when their power supply is not bypassed (decoupled) with some suitable capacitance.
The TL082 and many other opamps oscillate at a high frequency when driving the capacitance of a shielded output cable a meter long or more. Then a 100 ohm isolation resistor should be in series between the output of the opamp circuit and the cable. A TL072 dual opamp is a low noise audio version of a general purpose TL082.
Why change the gain of the circuit? Why with a remote pot? Use a volume control instead.

When you get your NE555 running then its 400mA (!) switching current spikes will cause power supply interference and will be heard loudly. Use a Cmos 555 instead. The current spikes are described and shown in the datasheet of the lCM7555 Cmos timer.

 

[Dr. Val Rousseau]

There is no decoupling of anything anywhere.
There is no indication of how grounding is handled among the circuit elements.
There is no coupling capacitor between the first and second opamp. Left amp offset voltage could drive right amp near or into saturation at high gain, giving chaotic results.

ak

How could it be an offset voltage, since it works perfectly well (even at maximum gain) when the 555 is not connected?

 

[Alec_t]

I agree that decoupling of the power supply to the opamps and microphone is needed, and that the 555 should be replaced by a CMOS version.

 

[Chris Hill]

I agree with Audioguru and Alec_t. The bipolar version of the 555 (which I think is the one you are using) is notorious for "crowbaring" the power supply when its output changes state. In other words, when the output changes from high to low or vice versa, the device can instantaneously draw a very large current. This current spike can cause all kinds of noise and instability problems in nearby circuitry. The CMOS version of the 555, the 7555, is pin-compatible and doesn't have the bipolar version's crowbar problem. Could you substitute in a 7555 and see if the situation improves?

I have also had this problem with the dual 555, the 556 (also bipolar) where the two halves of the device were interfering with each other! Solution? 7556!

 

[Audioguru]

An LMC555 and a TLC555 are also Cmos versions of the 555.
Hee, hee. Texas Instruments make both of them now.

 

[Dr. Val Rousseau]

Thank you for all your answers. I will try this when I get back to my lab, and let your know if this solves the problem.

 

[Dr. Val Rousseau]

Alright, I replaced the NE555 by a TS555CN which is also a CMOS version, and it didn't solve the problem at all. So now I'm considering inserting a decoupling capacitor between the two op-amps, even though I don't see why an offset would appear only when the 555 is connected. But I have to I don't want to ruin my printed circuit board for no reason, so I will try it on a bread board first.

 

[73's de Edd]

Sir Val . . . .

When I saw that you were merely wanting a meek audio signal level and not a totally different situation.
Like in the order of a HIGH current POWER driver for a Transistor / FET with the then implied "400mA (!)"

Initially try the choices of the 220-330-470 or even 1K isolation resistors with the mentioned electrolytic capscitor at the 555 pwr pin.
Therein the max available currents even possible, would then be in the comparative 22 /15/10/5 ma range, but keep an eye on the resultant available Vcc voltage.
They spec at 4.5VCC low end, but I have had them work below that level and work well.
Throw in your .1 ufd HF ceramic shunts if you like.
If the noise persists, the search for inter circuit proximity or ground loops might be involved next .

73's de Edd

 

[Dr. Val Rousseau]

Ok, so far I have:

1. Replaced the NE555 with a CMOS TS555CM.
2. Inserted a 1kΩ resistor between 5V and power pin of the 555.
3. Added 470μF and 100nF capacitors in parallel to power pins of the 555, and also to the TL082.
4. Added 2000μF and 100nF capacitors in parallel to +5V and GND, and same for -5V and GND.
   
5. Inserted a 2.2μF between the two op-amps in order to remove any residual offset from the output of the first one.
6. Removed remote potentiometer and replaced it by a resistor of same value as the value at which the behavior becomes chaotic.
7. Connections are made as short as possible in order to limit interferences.

Absolutely NO CHANGE!

 

[Alec_t]

So you still haven't decoupled the supply to the mic? You could try ~1k between the 2k2 resistor and the +5V rail, and add 100uF or so from the junction of the two resistors to ground.
Do you have a star ground arrangement so that the oscillator ground currents aren't affecting the mic amp?

 

[73's de Edd]

Hmmmm . . . . Can you come up with (3) 1.5 dc cells and solder tack them in series , ( or use a holder IF one is even available ) .
Then examine the foil path circuitry involved with the 555 +5 vdc supply ; to see if a mini( later . . .solder bridgeable) gap cut across the foil path will totally isolate the 555 + circuitry from the rest of the board.
That then let's you test again with a then totally isolated PURE DC supply for the 555 . ( Less the 555 negative return to ground and possibly experiencing a shared ground loop there, with the op amp.)
If no definitive answers after that, can we then see a close-up photo of the foil path side of the PCB.
( And its component side . . . . so that we may inter relate.)
BTW did you do the close in test of a 100k right at the i.c. terminals, with the previous wire remoted 100k AND its wiring temporarily removal.

Having missed Sir Alex_t's comment during my prep time, I also heartily concur , since it's the most VERY sensitive input circuitry path.
You might then transfer the battery supply over to the supply of the electret mike, using the same isolative supply aspects.

73's de Edd

.

 

[Dr. Val Rousseau]

So you still haven't decoupled the supply to the mic? You could try ~1k between the 2k2 resistor and the +5V rail, and add 100uF or so from the junction of the two resistors to ground.
Do you have a star ground arrangement so that the oscillator ground currents aren't affecting the mic amp?

That's it! It works!
For the record, I also had to add a 100nF in parallel to the 100μF in order to completely get rid of the chaotic signal!

Thanks a lot for all your answers!

 

[Alec_t]

Good to know it now works.