What is your constant voltage source for powering the chip? If the power supply has stabilization of its output, then it can generate the second harmonic. Check the shape and amplitude of the supply voltage oscillations on the power bus with an oscilloscope when the amplifier is loaded at 4 Ohm 1 kHz.
Hey! Thanks again for all inputs. Happy that we are seeing the same. Sorry for my delayed reply.
I will take new measurements with dB activated instead. And try some other frequencies to see if anything differs.
When it comes to my voltage source, I do have a supply with stabilization that i have been using the most. However, I did make a 15v linear power supply circuit with some parts I have laying around, using a toroid-transformer. Will be interesting to see if I notice any difference when using that one.
Final design will most likely use something like this.
I will take new measurements with dB activated instead. And try some other frequencies to see if anything differs.
When it comes to my voltage source, I do have a supply with stabilization that i have been using the most. However, I did make a 15v linear power supply circuit with some parts I have laying around, using a toroid-transformer. Will be interesting to see if I notice any difference when using that one.
Final design will most likely use something like this.
Hi!
When measuring the DC power supply, set the scope to remove DC so you can increase scope sensibility to look only the AC component by magnifying it.
AC component will always be small compared to the DC component of a power supply.
When measuring the DC power supply, set the scope to remove DC so you can increase scope sensibility to look only the AC component by magnifying it.
AC component will always be small compared to the DC component of a power supply.
Hello!
So, finally got some time over to do some more measurements as per your ideas, thanks for that.
Still using the exact same circuit until all measurements is made.
This time using dBu instead. Now this is a measurement i am not familiar with at all, so bear with me and the settings 😀
Settings: 512 bins with window function "Hamming". The second harmonic is clearly visible in this mode also, so i think i got the settings right.
I did measurements both with my stabilized PSU and with a linear supply i pulled together out of scrap parts. consists of a toroidal transformer 12v 50VA and 4x 3300uF smoothing capacitors.
Surprisingly (or not?) the second harmonic is about exactly the same.
Now, the linear supply might need some more smoothing, I include a measurement of that also, if anyone wants to have a look at that if its acceptable cap charge ripple or not. However, don't think that might be the cause, since distortion is nearly identical between the different power sources
I did tag all the measurements in the pictures themself.
dBu with stabilized psu:
dBu with linear supply:
AC scope view of stabilized psu:
Smoothing caps analysis of the linear supply:
So, finally got some time over to do some more measurements as per your ideas, thanks for that.
Still using the exact same circuit until all measurements is made.
This time using dBu instead. Now this is a measurement i am not familiar with at all, so bear with me and the settings 😀
Settings: 512 bins with window function "Hamming". The second harmonic is clearly visible in this mode also, so i think i got the settings right.
I did measurements both with my stabilized PSU and with a linear supply i pulled together out of scrap parts. consists of a toroidal transformer 12v 50VA and 4x 3300uF smoothing capacitors.
Surprisingly (or not?) the second harmonic is about exactly the same.
Now, the linear supply might need some more smoothing, I include a measurement of that also, if anyone wants to have a look at that if its acceptable cap charge ripple or not. However, don't think that might be the cause, since distortion is nearly identical between the different power sources
I did tag all the measurements in the pictures themself.
dBu with stabilized psu:
dBu with linear supply:
AC scope view of stabilized psu:
Smoothing caps analysis of the linear supply:
Hi! You can see now that is much better to look in a dB scale when comparing very different values. The difference is around 40dB, which is 1% - it's the same circuit as you said. But now you know how to better observe harmonics. Just a hint: before going to this FFT mode, always make sure the scope input is not saturated and not so "small" - set the sinwave to be clearly within the scope screen and then go to FFT.
OK, so you concluded that the power supply is not the problem. And this is the normal troubleshoot procedure. We cannot disregard anything and we have to continue to narrow down until find the root cause.
The 120Hz ripple is not a problem, since most class AB amp has high common mode noise rejection.
We have to go back to the PCB design or even consider a possible IC problem (is it original? We never no).
Maybe now you could try the list of suggestions I gave you.
With dB scale is much easier to see if anything changes.
Hi! Thanks for checking it out.
I did check out the datasheet to see what the tolerances are.
Am I interpreting it correctly that it says typical 0.2% to maximum 1% distortion and I'm currently at 1%? Would this mean that the manufacturer would say this is OK even in their manufacturing tolerances? Or maybe that is another thing entirely that they specify there?
Is this 1% considered bad in my case?
It´s interesting thought about if the chip is original. It is clearly marked with the ST brand and its bought from a reputable component vendor local to me. However, no one can be trusted and I think they might have got this in a big batch as they had over 1k of them in stock. Might be fakes maybe? Fakes or out-of-specs even?
I think the next step, since now all measurements are done with this is to try all the suggestion in changes that you made. Then we will have a clear before and after measurements of the circuit. The reason for me not to have done the changes yet is that I really wanted all possible measures done in the current state, so we really learn and can pinpoint what the issue was, rather then changing stuff in the middle of diagnosing and then have no real positive clue what the error was.
If we suspect the PCB design in it self still, would you most likely put the blame on the grounding or is anything up for grabs?
Thanks!
I did check out the datasheet to see what the tolerances are.
Am I interpreting it correctly that it says typical 0.2% to maximum 1% distortion and I'm currently at 1%? Would this mean that the manufacturer would say this is OK even in their manufacturing tolerances? Or maybe that is another thing entirely that they specify there?
Is this 1% considered bad in my case?
It´s interesting thought about if the chip is original. It is clearly marked with the ST brand and its bought from a reputable component vendor local to me. However, no one can be trusted and I think they might have got this in a big batch as they had over 1k of them in stock. Might be fakes maybe? Fakes or out-of-specs even?
I think the next step, since now all measurements are done with this is to try all the suggestion in changes that you made. Then we will have a clear before and after measurements of the circuit. The reason for me not to have done the changes yet is that I really wanted all possible measures done in the current state, so we really learn and can pinpoint what the issue was, rather then changing stuff in the middle of diagnosing and then have no real positive clue what the error was.
If we suspect the PCB design in it self still, would you most likely put the blame on the grounding or is anything up for grabs?
Thanks!
Hi! Considering the datahseet, we cannot say that you are out of specs. I don't have much knowleadge about chip manufacturing and its tolerances. From my experience, most of time we get in the typical values, as I've got with the TDA2003 that I've tested here. Let's see if someone else comment on that.
I cannot hear THD below 1%. You can test your own ears buy using the waveform you've created with artificial distortion. Make the A/B listening (THD=0 and THD=1%) and check with a good speaker or headphone.
You live in a high developed country, where I believe you have less chance of getting bad components. But I really don't know your market.
I know mine, which is terrible, where a lot of effort needs to be taken to get some quality component. Even so, my cheap TDA2003 works fine.
For now, I'd consider you IC is good and let this suspecious to the last option.
If everything is good, PCB design could be the issue. But if PCB grounding is the problem, executing the tests I suggested will probably reveal that.
As you move GND power supply input to other points of PCB, make some thick wire shortcuts etc and you see a relevant THD changes, voilà!
If adding more 100nF capacitors or additional RC networks changes the THD, then it's a matter of changing the circuit.
2003 is long out of production.
You got NOS (next to impossible) or fake.
Bridged configuration will get 1% or higher THD.
Go for LM1875, genuine still available, single chip is enough for your needs.
LM3886, and other designs are also available.
Quite economical to buy a ready 2030 plate amp for $1.50, and use it.
The chips are Chinese clones, work okay on 12-0-12V supplies.
And 12V plate amps, with up to 40W RMS are sold for car audio, can think about that.
You got NOS (next to impossible) or fake.
Bridged configuration will get 1% or higher THD.
Go for LM1875, genuine still available, single chip is enough for your needs.
LM3886, and other designs are also available.
Quite economical to buy a ready 2030 plate amp for $1.50, and use it.
The chips are Chinese clones, work okay on 12-0-12V supplies.
And 12V plate amps, with up to 40W RMS are sold for car audio, can think about that.
Hello everyone, and thanks again for all replies.
NareshBrd, Just to note, this is a 2004, but maybe that is also out of prod?
Bridge configuration adding to THD is interesting for sure, did not think about that.
So, picked up the diagnostics today again, and just dabbed around a bit with grounding. Tried to ground on different places on the PCB, and also just straight to the GND-pin on the chip. What I noticed, a little bit to my surprise was the 2nd harmonic stayed virtually the same, no matter what I did or how i handled the PCB, it just stayed "stable".
Before pulling out some more caps from storage to try and add in the circuit as suggested, I did one more last sanity check. I used a different player, same.
Once again I measured the player output, the 2nd harmonic was still not there, so player good.
BUT, then I had the idea to measure the output on the player, but while in circuit and amp powered up, then the 2nd harmonic shows up exactly as it does on the output.
This to me points to a issue with player output/amp input while the source is "loaded"? Or can we get fooled in other ways while measuring player output while in circuit?
To add to this, if the circuit is powered down, with just player playing in to the circuit, the harmonic is not there. The amp need to be on, if that is of any indication.
I add a screenshot to show, this is measured at earliest input from player on PCB.
What do we think about this?
NareshBrd, Just to note, this is a 2004, but maybe that is also out of prod?
Bridge configuration adding to THD is interesting for sure, did not think about that.
So, picked up the diagnostics today again, and just dabbed around a bit with grounding. Tried to ground on different places on the PCB, and also just straight to the GND-pin on the chip. What I noticed, a little bit to my surprise was the 2nd harmonic stayed virtually the same, no matter what I did or how i handled the PCB, it just stayed "stable".
Before pulling out some more caps from storage to try and add in the circuit as suggested, I did one more last sanity check. I used a different player, same.
Once again I measured the player output, the 2nd harmonic was still not there, so player good.
BUT, then I had the idea to measure the output on the player, but while in circuit and amp powered up, then the 2nd harmonic shows up exactly as it does on the output.
This to me points to a issue with player output/amp input while the source is "loaded"? Or can we get fooled in other ways while measuring player output while in circuit?
To add to this, if the circuit is powered down, with just player playing in to the circuit, the harmonic is not there. The amp need to be on, if that is of any indication.
I add a screenshot to show, this is measured at earliest input from player on PCB.
What do we think about this?
Hi! Interesting.
So the amplifier itself is not creating this 2nd harmonic, it's just amplifying what is in the input.
I'd make 2 tests:
1) GROUND ISSUES: Check the ground point of the input source (player). Try to connect the ground to another place, such as close to the ground power supply or exact on the speaker ground. This way, we check if there is any spurious current flowing in the PCB track, generating voltages along the PCB and causing this distortion. Also when measuring, try to put the scope ground lead to different ground points such as the input ground, power supply ground and speaker ground.
Since you mentioned that with speaker connect the 2nd harmonic is there and when the speaker is not connect the harmonic is not present, it might still be a ground problem.
Ideally, all ground points on PCB should be equal, but we know the resistance or even the impedance of a track is not absolute zero.
2) INPUT IMPEDANCE: The minimum input impedance of this amp is quite high (70k ohm) and should not influence the source. But it might be not linear and might be loading your source of signal.
In order to check, you can lower the input impedance using some resistors (just for testing) or driving the TDA amp with an opamp.
Here's a suggestion, but you can play with different resistor values, The lower the resistor that goes to earth, the lesser will be the influence of the TDA2004 impedance, if this is the case.
So the amplifier itself is not creating this 2nd harmonic, it's just amplifying what is in the input.
I'd make 2 tests:
1) GROUND ISSUES: Check the ground point of the input source (player). Try to connect the ground to another place, such as close to the ground power supply or exact on the speaker ground. This way, we check if there is any spurious current flowing in the PCB track, generating voltages along the PCB and causing this distortion. Also when measuring, try to put the scope ground lead to different ground points such as the input ground, power supply ground and speaker ground.
Since you mentioned that with speaker connect the 2nd harmonic is there and when the speaker is not connect the harmonic is not present, it might still be a ground problem.
Ideally, all ground points on PCB should be equal, but we know the resistance or even the impedance of a track is not absolute zero.
2) INPUT IMPEDANCE: The minimum input impedance of this amp is quite high (70k ohm) and should not influence the source. But it might be not linear and might be loading your source of signal.
In order to check, you can lower the input impedance using some resistors (just for testing) or driving the TDA amp with an opamp.
Here's a suggestion, but you can play with different resistor values, The lower the resistor that goes to earth, the lesser will be the influence of the TDA2004 impedance, if this is the case.
The 2004 is out of production since about 1990 at ST.
UTC, no idea if they make them, but their production is marked UTC.
You can measure DC at input and output, also put a capacitor coupling on the output, for DC blocking and so on...for values, see datasheet.
UTC, no idea if they make them, but their production is marked UTC.
You can measure DC at input and output, also put a capacitor coupling on the output, for DC blocking and so on...for values, see datasheet.
NareshBrd,
Thanks for your input. Please note that this is TDA2004R, and ST has copyrighted the datasheet PDF as year 2013, so hard time believing that it has been out of production since 1990?
I did some more experiments, this time with the GND.
What i did was (Still using my PCB):
1. Keeping the normal GND feed, and then bridging another GND directly from power source to the individual GND points. No different, the 2nd harmonic stayed virtually the same all the time.
2. Skipping normal GND feed, and just feeding at each separate GND point, not even here was there any difference in the 2nd harmonic.
What is interesting here, is that there is virtually zero difference, which is weird, or good? I did expect something to happen at least, especially if its the circuit itself that create this distortion.
But then again, the PCB is constructed with GND-flood, so I think this makes sense.
As for the input impedance test, I don't have a good way to put a series resistor as I'm stuck with what i have before the POT.
However, for sake of science, I did add a 1k to GND at input, but no difference there either. That test though, might not be as effective as the one proposed!
Short of rebuilding the circuit on a breadboard, The only way to test the suggested input impedance change is to remove the pot and replace with the 10k ohm resistor. That might be the next step then I guess?
One thing i gathered so far with this, is that the 2nd harmonic is very "stable" no matter how i mess with it its always there and just changes amplitude together with the output.
Also, we do get this distortion on the input when amplifier is on. What could cause that?
Oh, and also the OP-AMP idea is interesting.. I think this might be my next way?
Thanks for your input. Please note that this is TDA2004R, and ST has copyrighted the datasheet PDF as year 2013, so hard time believing that it has been out of production since 1990?
I did some more experiments, this time with the GND.
What i did was (Still using my PCB):
1. Keeping the normal GND feed, and then bridging another GND directly from power source to the individual GND points. No different, the 2nd harmonic stayed virtually the same all the time.
2. Skipping normal GND feed, and just feeding at each separate GND point, not even here was there any difference in the 2nd harmonic.
What is interesting here, is that there is virtually zero difference, which is weird, or good? I did expect something to happen at least, especially if its the circuit itself that create this distortion.
But then again, the PCB is constructed with GND-flood, so I think this makes sense.
As for the input impedance test, I don't have a good way to put a series resistor as I'm stuck with what i have before the POT.
However, for sake of science, I did add a 1k to GND at input, but no difference there either. That test though, might not be as effective as the one proposed!
Short of rebuilding the circuit on a breadboard, The only way to test the suggested input impedance change is to remove the pot and replace with the 10k ohm resistor. That might be the next step then I guess?
One thing i gathered so far with this, is that the 2nd harmonic is very "stable" no matter how i mess with it its always there and just changes amplitude together with the output.
Also, we do get this distortion on the input when amplifier is on. What could cause that?
Oh, and also the OP-AMP idea is interesting.. I think this might be my next way?
Last edited:
Hi!, that's good news, it indicates a good ground (an it was expected due to GND-flood). So, let's search for another culprit.
1k ohm to GND is the key to the test I suggested - Series resistor was just to not load the source with 1k.
And 1k at the input is low enough to overcome input impedance issues. So, this is not the problem.
You can test with the opamp, if it's easy for you, cause opamp has a very low output impedance and should not be affected much by input impedance.
I can only think about a non linear amp input impedance affecting the source signal.
I don't have other ideas for now. This is a trick problem to solve.
Alright!
Since we get this distortion also on the input, while under load, it makes a lot sense that it has something to do with the input impedance.
Next course of action must be the OP-amp if this can rule it out.
I'm not familiar with exactly how to use that in circuit though, will have to do some research.
I see your diagram, but guessing I would need some further components, power and so on
Since we get this distortion also on the input, while under load, it makes a lot sense that it has something to do with the input impedance.
Next course of action must be the OP-amp if this can rule it out.
I'm not familiar with exactly how to use that in circuit though, will have to do some research.
I see your diagram, but guessing I would need some further components, power and so on
Hi,
The opamp in this case will just act as a buffer with unitary gain (G=1).
What comes in, goes out. Very simple.
Just power it up with, +V and -V if you have a symmetric power supply.
If you just have single power supply, it can work too, but you need some more components.
See below the options.
Symmetric Supply:
Single Supply (it can be the same you use for your TDA if it is up to +36V.
Resistors/capacitors are just one solution. You may vary them according to what you have on hand.
The opamp in this case will just act as a buffer with unitary gain (G=1).
What comes in, goes out. Very simple.
Just power it up with, +V and -V if you have a symmetric power supply.
If you just have single power supply, it can work too, but you need some more components.
See below the options.
Symmetric Supply:
Single Supply (it can be the same you use for your TDA if it is up to +36V.
Resistors/capacitors are just one solution. You may vary them according to what you have on hand.
The TDA2003/2004 was later succeeded by the 2030/2040/2050 at ST.
There are links that suggest Unisonic made them in Taiwan,
Another audio forum post says last made in 1996.
ST puts a watermark or otherwise marks the data sheet if the product is obsolete.
The copyright lasts for 75 years, they may have last published the data sheet in that year, it may have been left in place by accident, or with other data sheets published in that year.
There are TDA72xx series chip amps with stereo output even one with quadrophonic output, intended for later applications, developed after the 2004 by ST.
For a one off project, I would buy a locally available chip, and use it as is, without too many modifications.
For series production, I cannot compete with the many factories in Delhi, and China, and that is not my line of business in any case.
A ready plate amp, available for a small amount extra compared to parts cost, is worth my time and money.
There are links that suggest Unisonic made them in Taiwan,
Another audio forum post says last made in 1996.
ST puts a watermark or otherwise marks the data sheet if the product is obsolete.
The copyright lasts for 75 years, they may have last published the data sheet in that year, it may have been left in place by accident, or with other data sheets published in that year.
There are TDA72xx series chip amps with stereo output even one with quadrophonic output, intended for later applications, developed after the 2004 by ST.
For a one off project, I would buy a locally available chip, and use it as is, without too many modifications.
For series production, I cannot compete with the many factories in Delhi, and China, and that is not my line of business in any case.
A ready plate amp, available for a small amount extra compared to parts cost, is worth my time and money.
NareshBrd,
Thanks, however, according to STs own page, it is in production but in NRND-status. Thus still in production, but not recommended for new designs.
https://www.st.com/en/automotive-infotainment-and-telematics/tda2004r.html
As this does not really have anything to do with the topic otherwise, I think we can step away from it. I know there are other chips and even ready made boards. However, no fun in that 🙂
This is more of a diagnosis thing and we are trying to figure out where the error is. Thanks for your ideas and input though!
Now, back to the topic.
Ron68,
Perfect, thank you so much for your schematic and all help, really learnt a lot during this work.
I was able to find a UA741CP among my parts, that does seem to have the same pin-config as the TL071 and is a general op-amp.
Checking the datasheet, it looks to me that it might work as a substitution for TL071. What do you think? Maybe even your suggested values can be used?
I will be using single supply though, as that is currently the most convenient.
Thanks, however, according to STs own page, it is in production but in NRND-status. Thus still in production, but not recommended for new designs.
https://www.st.com/en/automotive-infotainment-and-telematics/tda2004r.html
As this does not really have anything to do with the topic otherwise, I think we can step away from it. I know there are other chips and even ready made boards. However, no fun in that 🙂
This is more of a diagnosis thing and we are trying to figure out where the error is. Thanks for your ideas and input though!
Now, back to the topic.
Ron68,
Perfect, thank you so much for your schematic and all help, really learnt a lot during this work.
I was able to find a UA741CP among my parts, that does seem to have the same pin-config as the TL071 and is a general op-amp.
Checking the datasheet, it looks to me that it might work as a substitution for TL071. What do you think? Maybe even your suggested values can be used?
I will be using single supply though, as that is currently the most convenient.
As far as I know:
TDA2002/2003 and double version 2004 was intended for single supply low voltage (max 18 V) automotive use, class B, optimized for low impedance down to 2 ohm.
TDA2030/40/50 is optimized for symmetric supply and higher voltages up to +/-25 V or 50 V single supply). Class AB and much lower distortion.
TDA2030-40-50 can be considered a high power op amp. TDA2002-3-4 not really.
I gave it a shot to hook up the 741 opamp i had with the diagram above, but no go.
It gives me some strange output, no matter on the input. I think i checked over everything at least 10 times, but must have missed something.
This is what the output from the opamp looks like, horrific to say the least 😀 :
It gives me some strange output, no matter on the input. I think i checked over everything at least 10 times, but must have missed something.
This is what the output from the opamp looks like, horrific to say the least 😀 :
Turns out OP-amp as not so easy as I was hoping for 😀
Loading the op-amp output with a dummy 10ohm resistor and still using the above schematics, i get something that might resemble a sinewave, if your imagination is good. Not really sure what is going on here?
Loading the op-amp output with a dummy 10ohm resistor and still using the above schematics, i get something that might resemble a sinewave, if your imagination is good. Not really sure what is going on here?