how to isloate audio ground from the amplifier

[Adhith]

hello friends,
I'm working with a speaker project which also has a vu meter along with it. For the VU meter the audio input is fed from the amplifer output (from the left channel positive and ground). but while connecting the ground wire from amplifier output to the VU meter input, the speaker produces a distortion sound and I believe its because of lack of isolation between the audio signal from the amp and the VU meter circuit. So what should I do to isolate the ground?? I have already completed the project so im looking for smaller solutions to fit inside the enclosure. does a diode at the audio ground solve the problem or should I use an audio transformer??but I don't know how to use it. Could anyone help me out??. Thank you in advance.

[Audioguru]

The output of your amplifier might be bridged. A bridged amplifier output has two amplifiers. Each amplifier feeds one speaker wire but the signals are with opposite phase so that the output voltage across the speaker is doubled which produces almost 4 times the output power. Then your connection of the VU meter ground is shorting one output of the amplifiers to ground.
Connect the input ground wire of the VU meter to the amplifier's ground instead then it will show half the speaker's voltage.

[Adhith]

Thank you very much for your reply. I'm using a 25W+25W amplifier module for my project which is based on the class D chip TDA7492P. I have attached a pic of the board below. Could you explain a bit about where to find the amplifier ground??

[Zero999]

The amplifier is definitely bridged.

The schematic is probably similar to the one found on the data sheet and the ground connecting can be found, using the continuity test function on a multi-meter.
http://www.st.com/content/ccc/resource/technical/document/datasheet/35/01/97/6b/20/ce/40/6f/CD00211320.pdf/files/CD00211320.pdf/jcr:content/translations/en.CD00211320.pdf

[Audioguru]

Are you the guy who is feeding only one channel of the Bluetooth stereo to only one speaker?
The TDA7492 in your photo is missing the heatsink shown on the product from AliExpress.

[Adhith]

The amplifier is definitely bridged.

The schematic is probably similar to the one found on the data sheet and the ground connecting can be found, using the continuity test function on a multi-meter.
http://www.st.com/content/ccc/resource/technical/document/datasheet/35/01/97/6b/20/ce/40/6f/CD00211320.pdf/files/CD00211320.pdf/jcr:content/translations/en.CD00211320.pdf

Thank you for your valuable reply. Actually im not that good in electronics..I learn it by doing these kind of experiments and thus my knowledge is limited. And while seeing an SMD PCB thing are too confusing for me. If you could give me some details about where to locate the ground, i mean the position or area on the board it would be a great help.

[Adhith]

Are you the guy who is feeding only one channel of the Bluetooth stereo to only one speaker?
The TDA7492 in your photo is missing the heatsink shown on the product from AliExpress.

No sir I'm not that guy. I'm using both the speakers for both the channels. since the vu meter circuit which i have build have only one channel as input, I'm using an additional taping from the left channel for it.
The board that you have shown is for 50w+50w i guess. Mine is a 25w+25w one and brought it from banggood. I think this 25w one dont comes with any heat sink in any online sites, anyhow i'm planing to fit a small heat sink for it in my project to be on the safe side

[Buriedcode]

The modules ground is the bottom side of that DC barrel socket (from your picture).  I cannot see a specific pad for ground on that board so it looks like you'll have to solder a wire to the DC socket.

You mentioned an audio transformer which is one option, but I feel I am not qualified to recommend a particular part number.

Another option, possibly overkill would be a differential amplifier using an opamp: https://www.electronics-tutorials.ws/opamp/opamp_5.html 
You would still need to connect the amplifier boards ground (again that DC socket) to the ground of the VU meter.  But it would mean the VU meter shows the full voltage swing.

There is actually two differential amps on that PCB because the output from the bluetooth module is differential, which is then converted to single ended by the on board dual opamp (NE5532, one opamp per channel).  I'm unsure why they have done this since the input to the TDA7492 is differential, so that part of the circuit only adds noise.

[Audioguru]

Each speaker wire of the bridged amplifier has a DC voltage that is half the supply voltage. If your VU meter measures  DC then a coupling capacitor must be used in series to block the DC but pass the AC to the VU meter.

[Adhith]

The modules ground is the bottom side of that DC barrel socket (from your picture).  I cannot see a specific pad for ground on that board so it looks like you'll have to solder a wire to the DC socket.

You mentioned an audio transformer which is one option, but I feel I am not qualified to recommend a particular part number.

Another option, possibly overkill would be a differential amplifier using an opamp: https://www.electronics-tutorials.ws/opamp/opamp_5.html 
You would still need to connect the amplifier boards ground (again that DC socket) to the ground of the VU meter.  But it would mean the VU meter shows the full voltage swing.

There is actually two differential amps on that PCB because the output from the bluetooth module is differential, which is then converted to single ended by the on board dual opamp (NE5532, one opamp per channel).  I'm unsure why they have done this since the input to the TDA7492 is differential, so that part of the circuit only adds noise.

Thank you for your valuable suggestions. By the bottom side of the DC barrel socket did you mean the ground power supply to the board?? I'm also attaching a pic of the bottom side of the board with this post so that you can have a look.
You mentioned about another option of using opams but I have a doubt. does the small opams could work at the output voltage and current levels of this amp??

[Adhith]

Each speaker wire of the bridged amplifier has a DC voltage that is half the supply voltage. If your VU meter measures  DC then a coupling capacitor must be used in series to block the DC but pass the AC to the VU meter.

My Vu meter is based on LM3915N IC. The circuit which I'm using has already a diode at its positive audio input side to consider only the positive cycle of the audio. The circuit works fine since I have already used for an another project (with a another amplifier based on LA4440 IC ), but I don't know how to make it compatible to this class D amplifier IC board i.e the noise by connection the ground wire

[Adhith]

The modules ground is the bottom side of that DC barrel socket (from your picture).  I cannot see a specific pad for ground on that board so it looks like you'll have to solder a wire to the DC socket.

You mentioned an audio transformer which is one option, but I feel I am not qualified to recommend a particular part number.

Another option, possibly overkill would be a differential amplifier using an opamp: https://www.electronics-tutorials.ws/opamp/opamp_5.html 
You would still need to connect the amplifier boards ground (again that DC socket) to the ground of the VU meter.  But it would mean the VU meter shows the full voltage swing.

There is actually two differential amps on that PCB because the output from the bluetooth module is differential, which is then converted to single ended by the on board dual opamp (NE5532, one opamp per channel).  I'm unsure why they have done this since the input to the TDA7492 is differential, so that part of the circuit only adds noise.

I have removed the DC socket for the project and I have attached the pic of the board. You said that the ground will be below the DC socket and I have marked it on the board. Is this the amplifier ground??

[Audioguru]

The LM3915 does not need a diode in series with or to ground at its positive input since it measures only positive voltages. It does nothing with negative input voltages. Adding a series diode causes its measurements to be 0.4V to 0.8V too low depending on the diode's current and its temperature. The datasheet shows peak detector circuits and active rectifier circuits to feed the input of the LM3915 that do not cause its measurements to be too low. The peak detector holds the peak voltage long enough for our slow vision to see it.

The LA4440 amplifier is an ordinary stereo amplifier with one wire of each speaker grounded (its outputs are not bridged) so your LM3915 is measuring only the live output signal. Its datasheet shows that both of its amplifiers can be connected to make a single bridged amplifier, that is like both amplifiers in your TDA7492 amplifier.

Your TDA7492 amplifier has two bridged amplifiers. You connect the input of your VU meter circuit through a series coupling capacitor to only one of its speaker wires and the ground wire of your VU meter to the ground of the amplifier. Shorting one of its speaker wires to ground will destroy the TDA7294 amplifier. Then your VU meter measure only half of the speaker voltage.

Use a multimeter set for OHMS to measure the ground on the pcb with the ground on the DC power jack.

[Audioguru]

The instructions that came with the TDA7492 amplifier should show the positive and negative (ground) connections of its DC power jack. We haven't seen the instructions and we have not measured the jack so we don't know.

[Adhith]

The LM3915 does not need a diode in series with or to ground at its positive input since it measures only positive voltages. It does nothing with negative input voltages. Adding a series diode causes its measurements to be 0.4V to 0.8V too low depending on the diode's current and its temperature. The datasheet shows peak detector circuits and active rectifier circuits to feed the input of the LM3915 that do not cause its measurements to be too low. The peak detector holds the peak voltage long enough for our slow vision to see it.

The LA4440 amplifier is an ordinary stereo amplifier with one wire of each speaker grounded (its outputs are not bridged) so your LM3915 is measuring only the live output signal. Its datasheet shows that both of its amplifiers can be connected to make a single bridged amplifier, that is like both amplifiers in your TDA7492 amplifier.

Your TDA7492 amplifier has two bridged amplifiers. You connect the input of your VU meter circuit through a series coupling capacitor to only one of its speaker wires and the ground wire of your VU meter to the ground of the amplifier. Shorting one of its speaker wires to ground will destroy the TDA7294 amplifier. Then your VU meter measure only half of the speaker voltage.

Use a multimeter set for OHMS to measure the ground on the pcb with the ground on the DC power jack.

Thank you very much sir for you help. I'm still having some doubts regarding it. I know its very annoying to ask these sort of simple things but for me its the only hope sir. I still dont get it on how to measure the ground on the pcb with the ground on the DC power jack. All the copper pads are coated with insulation right. I have an exposed power ground on the dc socket pin but dont see the ground of the amplifier. (by ground you mean the ground of the bridge amplifier right not the left and right ground right??)

[Audioguru]

On a bridged amplifier no speaker wire is grounded, each speaker wire is powered from its own amplifier, then the speaker voltage is doubled and the speaker power is almost 4 times higher. If you ground a speaker wire then you cause distortion and probably will blow up the amplifier.
The amplifier has a ground which is the 0V or negative of its DC power supply. Therefore one or more terminals on its DC power jack is a ground connection but you must measure it to see which terminal.

[Buriedcode]

You should read up on the terms "bridged" and "differential". I used the latter term, but in the context of speaker amps its the same thing as bridged (albeit probably less specific).

What has been said a few times here is - each connection to the speaker is driven by an amplifier, neither of these connections are at ground.  You are still thinking in terms of "single ended" amplifiers, where one side of the speaker is connected to ground, whilst the other side is driven by an amplifier.  In that case, your VU meter would have no problems, but your amplifier board - and many modern amplifiers - drive speakers differentially. 

So, if you use a scope and measure each speaker terminal with reference to ground (the DC jack) you will see a signal.  One will be the opposite of the other - as one side of the speaker is driven high (with respect to ground) the other side will be driven low.  So if an amplifier is powered by 12V, it can create ~12V across the speaker in one direction, but also in the other direction.  Like a H-Bridge! Check out the datasheet on page 6: http://www.st.com/resource/en/datasheet/cd00205863.pdf     

So what this means is... you have two signals for each speaker, neither of which are ground.  You can either just look at one of these signals, which means your VU meter will see a signal sitting at half the power supply voltage, and only give half the indicated voltage (you can use a series capacitor to block DC so it only shows the signal and not the DC offset) ... OR... combine the two differential signals into one, which is why I mentioned a differential amplifier.  This creates a single signal, referenced to ground, that will work with your VU meter.  It doesn't have to be accurate, so even an LM358 would do.

I'll post an edited picture of your boar din a minute to confirm the ground connection.

[Buriedcode]

SO yes, the area you indicated is "ground" (0V reference).  I have also attempted (badly) to show the signals on the speaker outputs.  Neither of them goes negative, but the signals are centered on half the supply voltage, and are opposites - the speakers on care about the difference between these signals.  Connecting your VU meter to one of these terminals, and our actual ground (as indicated on the diagram) would mean your VU meter would show half voltage at all times.  Using a DC blocking capacitor would mean it only see's AC, but as each output is only half the signal it would show half the voltage.

Try it yourself - connect your VU meters ground to the amplifier boards ground, and the "signal" to one of the speaker outputs - any of the 4 outputs will do.  And see what it shows.

[Adhith]

On a bridged amplifier no speaker wire is grounded, each speaker wire is powered from its own amplifier, then the speaker voltage is doubled and the speaker power is almost 4 times higher. If you ground a speaker wire then you cause distortion and probably will blow up the amplifier.
The amplifier has a ground which is the 0V or negative of its DC power supply. Therefore one or more terminals on its DC power jack is a ground connection but you must measure it to see which terminal.

Thank you again for your reply. Ok sir now things are getting cleared about the ground connections.

[Adhith]

You should read up on the terms "bridged" and "differential". I used the latter term, but in the context of speaker amps its the same thing as bridged (albeit probably less specific).

What has been said a few times here is - each connection to the speaker is driven by an amplifier, neither of these connections are at ground.  You are still thinking in terms of "single ended" amplifiers, where one side of the speaker is connected to ground, whilst the other side is driven by an amplifier.  In that case, your VU meter would have no problems, but your amplifier board - and many modern amplifiers - drive speakers differentially. 

So, if you use a scope and measure each speaker terminal with reference to ground (the DC jack) you will see a signal.  One will be the opposite of the other - as one side of the speaker is driven high (with respect to ground) the other side will be driven low.  So if an amplifier is powered by 12V, it can create ~12V across the speaker in one direction, but also in the other direction.  Like a H-Bridge! Check out the datasheet on page 6: http://www.st.com/resource/en/datasheet/cd00205863.pdf     

So what this means is... you have two signals for each speaker, neither of which are ground.  You can either just look at one of these signals, which means your VU meter will see a signal sitting at half the power supply voltage, and only give half the indicated voltage (you can use a series capacitor to block DC so it only shows the signal and not the DC offset) ... OR... combine the two differential signals into one, which is why I mentioned a differential amplifier.  This creates a single signal, referenced to ground, that will work with your VU meter.  It doesn't have to be accurate, so even an LM358 would do.

I'll post an edited picture of your boar din a minute to confirm the ground connection.

Thank you again for your kind reply. could you help me with the type and value of the capacitor that can be used for the dc blocking. I'm planing to first use this amplifier ground and if its not that good then i'll move on to the differential amplifier option.

[Adhith]

SO yes, the area you indicated is "ground" (0V reference).  I have also attempted (badly) to show the signals on the speaker outputs.  Neither of them goes negative, but the signals are centered on half the supply voltage, and are opposites - the speakers on care about the difference between these signals.  Connecting your VU meter to one of these terminals, and our actual ground (as indicated on the diagram) would mean your VU meter would show half voltage at all times.  Using a DC blocking capacitor would mean it only see's AC, but as each output is only half the signal it would show half the voltage.

Try it yourself - connect your VU meters ground to the amplifier boards ground, and the "signal" to one of the speaker outputs - any of the 4 outputs will do.  And see what it shows.

I just used a multi meter ohm function and found that the dc jack ground and the left channel ground is having a resistance of 350 ohm.

[Adhith]

Ill connect  the ground to the vu meter let you guys know about it.

[Buriedcode]

I just used a multi meter ohm function and found that the dc jack ground and the left channel ground is having a resistance of 350 ohm.

The left channel has no ground. Nor does the right channel.  No speaker is connected to ground.  Both terminals of the speaker are driven by push-pull outputs. In this way the reason they label them "+" and "-" is for phasing.  If you connect speakers backwards you will still hear music it will just be out of phase with the other channel, it has nothing to do with ground reference.

[Zero999]

Ill connect  the ground to the vu meter let you guys know about it.

Yes, that will work.

As this is a class D amplifier, you might find the existing filter is inadequate at removing all the high frequency content, so the VU meter might indicate there's a signal, even though there isn't one. Another RC low pass filter might be required, before the VU meter.

[Audioguru]

As this is a class D amplifier, you might find the existing filter is inadequate at removing all the high frequency content, so the VU meter might indicate there's a signal, even though there isn't one. Another RC low pass filter might be required, before the VU meter.

Good point. We cannot hear ultrasonic switching of the class-D but the VU meter might show it.