Sparkfun Audio Detector Question

I searched but couldn't find an answer.

Building a silence detector for one of our remote radio stations.

Using the headphone output of a portable transitor radio as the input signal. Needed to build a peak detector/envelope follower circuit which would have complicated the power supply so I decided to try to use the Sparkfun Sound Detector instead since it appears to have all the outputs I would need. I've already got the code working to detect and send an email upon detected silence. I just couldn't get the signal clean enough without a peak detector or envelope follower.

My Idea is to remove the Electret mic and replace it with the headphone output from the radio. Is this feasable? I looked at the data sheet for the mic but it doesn't appear to list anything for output. My concern is over-driving the Sound Detector's circuit with the headphone signal.

I realize I can decrease the gain if needed by adding a resistor to R17 but I don't want to blow it up right off the bat and do not understand the circuit (and these mics) enough to know if this is even possible.

Thanks in advance for any help!

kgoods, welcome to the new forum. Yes it is definitely doable. You will have to use either a transformer or build an attenuator. For the attenuator you take the output of the radio through a ~50K resistor You connect the other end of the resistor to a second resistor that is connected to ground and connect them and also the input of your sound detector. Ground connections are critical, if they are not solid you will get hum which may not cause problems. The resistor values are just a guess you may have to adjust them. The The Electret mic has a DC bias, which you do not need so you may have to input via a capacitor unless you can remove the bias on the Sound Detector. Posting a Schematic, not a frizzy thing will help us help you along with links to technical information on the hardware devices.

Thanks for the quick response!

Here's the original schem:
schem1

New users can only post one picture per post.... cont...

And if I understood you correctly I need something like this:
testschem1

Am I on the right track?

Thanks so much again!

You forgot to remove the DC bias, via R1. Delete R1, change the "lower" 50k to 2.2k.

You're right on R1, it was my plan to ask if I could just take it off the board, forgot. :slight_smile:

Thanks for the clarification on the L-Pad value, I hadn't got around to reading up on them to understand how to choose values. That's next on my list.

Thanks for your input, helps a bunch!

How are you ‘raising the alarm’ back at base ?
(Only curious, because I started a company 20 years ago for remote monitoring tv and radio services…!)

Since this station is too far away for us to monitor directly, I'm building this self-contained monitor with a radio, arduino, and sound detector. It will be plugged into a network at a friends business in an area that has reception and will email us if the station has dead air. We transmit the signal to the tower over the internet but we have no way of knowing if the transmitter is actually transmitting without being able to listen via radio.

You could place a receiver in the vicinity of the transmitter and sends its audio back via internet. If it disappears for a period of time it could send a text message as well.

Just disconnect C4 from IC1G1 and connect it - C4 - to the transistor radio output directly.

More sensible than attenuating the signal for it to be amplified again! :sunglasses:

Possible but problem is radios don't work very well in the transmitter shack (always at the least accessible location on top of a mountain somewhere). Laptops don't work at all with two 100KW and one 25KW transmitter inside a metal shack. :slight_smile:

Plus I'd have to run a computer up there to accept the analog, convert to digital and resend the signal over the net. By the time you shield the computer to get it to work reliably we'd have much more cost than the route I chose.

Don't think I hadn't thought about it. :wink:

I did think about that as well but wasn't sure how it would affect the next circuit block. I'm not an electronics guy, more network admin/programmer.

My thought was if I could get the headphone output to nearly match the mic then I'd still have the gain adjustment via R17 and R3 if needed. And as far as I can tell if would most likely be a simple matter of adding a resistor to R17 to reduce the gain.

I'm planning on using the gate output from the sound detector as I think it will simplify the programming side of things.

Here's the next circuit block:
schem2

And the final circuit block showing the gate output:
schem3

Finally, here's the traces of the various outputs:

This is what got me looking at the sound detector. That gate output looks ideal for what I need. My initial problem was the headphone output of the radio was only ~1V and by the time I added a diode to block the negative going voltage it only gave me ~250mV to monitor which wasn't enough to do so reliably.

@Paul_B
Thanks for your input. If I had a scope I may have looked harder at bypassing the first amp stage but with my limited understanding and lack of a way to actually measure what's going on I'll stick with the L-Pad and see what happens.

I'll be sure and post back once complete with my findings.

For the stated purpose, you don't need an op amp. A radio has sufficient output level to drive an analog input directly (but through a resistive attenuator/limiter circuit of course). All the peak detection etc. can be fairly easily done in software.

From the headphone jack? That hasn't been my experience.

I figured the speaker output would have enough but didn't want to break the radio apart if I didn't have to. It appears not to be the type you can open and close back up without breaking tabs and glue joints. I'd like to keep it looking semi professional. LOL!

I'd still need to remove the negative going voltages to protect the arduino input and then have some kind of envelope follower/peak detection to keep from fighting sampling issues in software, no?

From my experiments so far it appears that this little 10 buck detector will save me a bunch of headaches given my limited experience. :slight_smile:

There is plenty of power at a headphone jack to power an Arduino input. More than you need. 100 times more than you need. Perhaps your experience is not based on the correct designs. Actually you would not need to remove the negative voltages, why are they important? Of course you would need input protection, but that's easy with resistors. If you're really obsessed with that, a simple voltage divider bias source does it. So far you are expending a lot of effort to adapt a module when a handful of passive parts will do just fine.

Since the software only has to indicate signal or absence of signal, simple averaging of the input is all you need in software.