Hi. I have a measurement device that has an in-built speaker, which I'd like to record the audio from directly. I'd like to do this by splitting the wires driving the speaker, and converting this to a level that I could input to a PC microphone input (So the speaker still works). Can someone help with this? I think I need a little circuit like the one shown here: http://arrakis-systems.com/pdfs/Microphone Input to Line Level Adapter.pdf
Could this cause any danger to the device?

(Terrible schematic to make sure you know what I mean!)

 

You could try putting a microphone up to the speaker.

 

Could this cause any danger to the device?

No.
But you may need to increase the value of the 100Ω resistor to get sufficient signal level.

 

Could this cause any danger to the device?

Possibly.

Mic inputs on a PC usually have one of the wires (the braided shield) at GND level.
If one wire to the speaker is also at GND level you could do some damage if you connect the two incorrectly.

 

Possibly.

Mic inputs on a PC usually have one of the wires (the braided shield) at GND level.
If one wire to the speaker is also at GND level you could do some damage if you connect the two incorrectly.

I'm thinking 600R telephony transformer from an old PC modem board, they're usually 1:1 and isolating.

probably needs to feed into the mic input via a resistive voltage divider. Whatever drives the speaker shouldn't even notice 600R strapped in parallel, but if the speaker is to be removed, I'd put in a dummy load resistor.

 

Possibly.

My answer was based upon the circuit he referenced (below) which had a resistor in both leads from the speaker.
The resistors should prevent any damage, no matter how it's connected.

 

Many amplified speakers have both wires to the speaker with signal (Bridge Tied Load) and no ground wire. If your resistor attenuator has its ground connected to the speaker amplifier's ground then the part of the speaker amplifier that is driving that speaker wire will be shorted and will probably be damaged and the mic input on the PC also might be damaged by the DC on each wire of a BTL speaker connection.

 

Hi, thanks for the responses! If I was able to use the linked circuit this would be the easiest option. Can you advice on calculating a better value for the 100 ohm resistor? I've managed to find the schematics for the speaker (the speaker wire is the only accessible part):



I really appreciate your inputs, my electronics know-how is poor, and I don't want to damage anything! - cheers

 

Hi, thanks for the responses! If I was able to use the linked circuit this would be the easiest option. Can you advice on calculating a better value for the 100 ohm resistor? I've managed to find the schematics for the speaker (the speaker wire is the only accessible part):

View attachment 138402

I really appreciate your inputs, my electronics know-how is poor, and I don't want to damage anything! - cheers

Start with 100. If your signal is too low, try 330, then 1k and 3300...

If none of those work, start over with lower value resistors feeding the 100 ohm resistor. It will take some iterative work.

 

Your amplifier is BTL and drives both wires of the speaker with no ground at the speaker. Each wire has DC on it which might mess up the microphone preamp. If the host audio has its ground connected to your amplifier's ground and you use the simple two resistors attenuator that joins the host and amplifier grounds together then one wire output from the amplifier is shorted and might destroy the amplifier.

 

If I used the 1:1 isolated transformer like mentioned above would this be safer?

 

Use two capacitors to AC couple the signal from the speaker to the MIC input and GND

 

If I used the 1:1 isolated transformer like mentioned above would this be safer?

A transformer might cut low and high audio frequencies.

 

Use two capacitors to AC couple the signal from the speaker to the MIC input and GND

The capacitors would feed the input and ground of the attenuator and the attenuator would feed the mic input and its ground. There might be mains hum.

 

Thanks, so I think I will end up with a circuit like this -

 

10uF into 47k ohms will pass earthquake frequencies down to 0.34Hz. Use 0.15uF film capacitors for a cutoff at 23Hz.

 

Thanks!