Cannot share power source between audio module and 555 circuit

[pmolsen]

In post https://www.electro-tech-online.com/threads/stretching-audio-output-pulse.155785/ I outlined a circuit I was building to activate a solenoid with voice using a small Arduino sound detection module - **broken link removed**.

The output pulses from that module are too short to drive the solenoid directly so I had to build a monostable 555 circuit to stretch the pulses to around 50ms.

The problem is if I try to run both boards from a single 5v plugpack the sound module stops working, even with the solenoid disconnected.

The sound module output pin is active low so I connect that directly to the 555 trigger pin 2. I connect the grounds together and run power to the sound module only. It works perfectly and the output LED turns on and off rapidly in time with normal voice input.

The moment I connect the +5V power from the sound module across to the 555 circuit the sound module stops working. The output LED immediately goes hard on. I adjust the trimpot on the sound module so the LED just goes off, but the module is then too insensitive to respond to voice. I have to blow hard on the electret mic to trigger it.

The plugpack I am using is a 1.8A Google Home Mini power supply but I have tried various others with the same result.

If I run each module from a separate 5v plugpack, with the two circuit grounds connected together and the sound module output running directly to the 555 pin 2 it works perfectly and drives the solenoid properly.

I tried installing a 1000uF capacitor across the inputs to smooth out any ripples that may have been occurring in the power supply but no change. The current draw of the 555 circuit is around 1.5mA and even less for the sound module.

The 555 timing capacitor is 10uF and the timing trimpot is set to around 4.5K ohms. 10nF cap between 555 ctrl pin 5 and gnd. Mosfet connected to output pin 3 with flywheel diode across mosfet drain and +5V. 10k (or 47k) pull-up resistor on 555 input pin 2 (so I can test it without the sound module connected). 10k pull-down resistor on mosfet gate (not needed and works without it as 555 output pulls it low when off.

Can anyone suggest why sharing a common power supply should cause problems and what I can do to try to fix it please?

 

[pmolsen]

I stumbled upon part of the solution purely by chance. I was previously running the power to the audio module first then via wires about 70mm (3 1/2") long to the 555 module. I connected the power the other way, to the 555 first (via the USB connector shown) then across to the audio module and the audio module now works perfectly and triggers the 555 as required. 3 boards that I built had the same problem and all now work.

When I connect the solenoid to the mosfet output however the original problem returns. The solenoid triggers immediately when power is applied. When I adjust the trimpot until the output turns off the sound module is no longer sensitive enough to operate.

I put a scope on the power supply. The voltage sags by around half a volt when the solenoid is on (it draws 1 amp), which would no doubt upset the comparator in the audio module. What I do not understand however is why in the steady state with the solenoid off the function of the audio module changes completely when the single power supply is used.

 

[Nigel Goodwin]

You appear to have a complete lack of power supply decoupling, which is essential in any circuit, and particularly so with 555 circuits, as they cause large current spikes (the CMOS ones are better in this regard).

 

[audioguru]

Intersil, Texas Instruments and a few other companies make the Cmos 555. Intersil's datasheet shows that the old fashioned ordinary 555 draws 400mA(!) from its power supply each time the output switches. You should use the Cmos 555 because it is much better without the problems of the old one.

 

[pmolsen]

I am using a CMOS 555 (7555) as per the photo. I did also have a large 1000uF decoupling capacitor as per the holes in the board in the photo, which did not help (but not a small ceramic one for higher frequencies), however that is likely immaterial.

I suspect that the solenoid drawing 1 amp is always going to cause problems in the comparator circuit, when using a small, cheap 5v charger. Even the 1000uF capacitor made no difference to that.

 

[audioguru]

I am sorry that I did not see the faint i755x numbers in the photo.

The transistor in the sound module is biased the worse way that I was taught never to use. I suspect that the transistor is almost cutoff or it is almost saturated that changes when the temperature changes. Measure and report its collector voltage when there is no sound.

The 100nF capacitor at the collector of the transistor cuts frequencies above 160Hz and the 100nF capacitor to ground cuts frequencies above 360Hz.

The output of the LM393 is overloaded (7.4mA) if it has minimum spec's (4mA). Its datasheet says that it should have some hysteresis added.

 

[AnalogKid]

The transistor is intentionally "dangle-biased" to act as a comparator, not a linear amplifier. That output square wave-ish is then coarsely integrated by the collector 10K and 100 nF into a waveform that very roughly varies with the energy content of the audio. This is then "energy-level detected" by the LM393.

ak

 

[Nigel Goodwin]

The transistor is intentionally "dangle-biased" to act as a comparator, not a linear amplifier.

I'm with AG, it's just really poorly 'designed', and highly critical of the exact (not just the type, but the specific example) transistor fitted in the circuit .

 

[pmolsen]

This was the ultimately successful objective -

 

[AnalogKid]

I'm with AG, it's just really poorly 'designed', and highly critical of the exact (not just the type, but the specific example) transistor fitted in the circuit .

Still don't think so.

If you replace the transistor with a series diode (and move the 10 K resistor in parallel with the 100 nF cap), the circuit looks more like a traditional envelope detector driving the comparator. The transistor is acting as a "rectifier with gain". It is much more sensitive to small amplitude signals than a simple diode because it does not have a 0.6 V detection threshold of its own.

Note - I'm *not* saying this is a piece of wonderful or magical design, only that it is based on deliberation, compromise, and choices by an experienced designer. Granted, its operating point will wander around with temperature and part-to-part variations; Life is choice.

ak

 

[audioguru]

How can I make my Google Home Mini talk to me back and forth like that? I must say, "Hey Google" before each question I ask it.