Hey.
So I have built a TDA2030 bridge amp while completely following the schematic in the datasheet. I works pretty well when I play some music from my cellphone or when I plug in a guitar, but when I plug in my bass, the sound is pretty distorted when playing low notes. It improves a bit when I turn the treble knob all way up, but otherwise it distorts. Just for the record - when playing some music from the phone, if there is some more prominent bass in the song, there is distortion again.
This is when plugging in any device directly, with no preamp or anything between the device and the amp.
I tried using a simple booster circuit made with a TL072 (I think) IC. Now the distortion gets even worse on the bass side. Additionally, when the booster circuit is turned near to the max, there is some nasty sound on the output, even when there is nothing on the input of the booster (the guitar is unplugged or just nothing being played). In attachments there is a short track showing the actual noise generated. Also, the same thing happened when I use a distortion pedal and turn it up even a little bit.
The speaker is definitely not the culprit. I use the ±12V rails from a PC PSU to power the circuit. The max current on the -12V side is 0.8A (as from what's written on the PSU at least).
The chips are properly cooled.
I am unsure if this could be some grounding problem (I am using shielded cable on the input BTW), or just a bad power supply, or maybe a faulty board (I double-checked the connections though), so any help is appreciated.
Thanks!
So I have built a TDA2030 bridge amp while completely following the schematic in the datasheet. I works pretty well when I play some music from my cellphone or when I plug in a guitar, but when I plug in my bass, the sound is pretty distorted when playing low notes. It improves a bit when I turn the treble knob all way up, but otherwise it distorts. Just for the record - when playing some music from the phone, if there is some more prominent bass in the song, there is distortion again.
This is when plugging in any device directly, with no preamp or anything between the device and the amp.
I tried using a simple booster circuit made with a TL072 (I think) IC. Now the distortion gets even worse on the bass side. Additionally, when the booster circuit is turned near to the max, there is some nasty sound on the output, even when there is nothing on the input of the booster (the guitar is unplugged or just nothing being played). In attachments there is a short track showing the actual noise generated. Also, the same thing happened when I use a distortion pedal and turn it up even a little bit.
The speaker is definitely not the culprit. I use the ±12V rails from a PC PSU to power the circuit. The max current on the -12V side is 0.8A (as from what's written on the PSU at least).
The chips are properly cooled.
I am unsure if this could be some grounding problem (I am using shielded cable on the input BTW), or just a bad power supply, or maybe a faulty board (I double-checked the connections though), so any help is appreciated.
Thanks!
Thanks for the answers!
The schematic I used: Screenshot by Lightshot
It's an 8 ohm speaker.I measured the current on the -12V line while the amp was at peak volume - it exceeded 1A. The exact voltage while zero load on the -12V rail is ~ -11.2V and it dropped by ~0.5V when loaded.
The schematic I used: Screenshot by Lightshot
It's an 8 ohm speaker.I measured the current on the -12V line while the amp was at peak volume - it exceeded 1A. The exact voltage while zero load on the -12V rail is ~ -11.2V and it dropped by ~0.5V when loaded.
You used that exact circuit? You didn't change some of the capacitor values, and now you wonder why there's distortion when you use it as a bass amplifier?
Datasheet circuits are illustrative and often not suitable for practical use. This circuit is the perfect example of that. All the TDAxxxx datasheet circuits that I've seen suffer from the same deficiencies. The fix is easy.
Calculate the poles for C2, C6, and C1. Then you will know why your circuit isn't performing the way you want.
Datasheet circuits are illustrative and often not suitable for practical use. This circuit is the perfect example of that. All the TDAxxxx datasheet circuits that I've seen suffer from the same deficiencies. The fix is easy.
Calculate the poles for C2, C6, and C1. Then you will know why your circuit isn't performing the way you want.
The pole for C2-R2 and C6-R7 is 10.6 Hz. This assures sonic intrusion (distortion) all the way up to 106 Hz. And if C2 and C6, and R2 and R7 aren't matched perfectly (which I can assure you they're not), even greater amounts of complex distortion will be introduced. 
The pole for C1-R1 is 7.2 Hz, which assures sonic intrusion to 72 Hz. This is a very poor aspect of this design too.
To mitigate distortion from C2-R2 and C6-R7, you want the pole for C1-R1 to be significantly higher. You want that pole to be the dominant pole. So you see, this schematic wasn't very well thought out at all; the feedback poles dominate.
I didn't learn all this in engineering school. They didn't teach these concepts, at all. I had to figure all this out on my own (with a little help from Walt Jung ). Maybe that's why there's so much mediocre equipment out there. As far as the TDAxxxx devices, they're very popular in industry and I always see the datasheet circuits employed; even though it's super easy and super cheap to greatly improve them. There's a guy here named Daniel that has fiddled with these devices extensively. Hopefully he'll chime in.
The pole for C1-R1 is 7.2 Hz, which assures sonic intrusion to 72 Hz. This is a very poor aspect of this design too.
To mitigate distortion from C2-R2 and C6-R7, you want the pole for C1-R1 to be significantly higher. You want that pole to be the dominant pole. So you see, this schematic wasn't very well thought out at all; the feedback poles dominate.
I didn't learn all this in engineering school. They didn't teach these concepts, at all. I had to figure all this out on my own (with a little help from Walt Jung
). Maybe that's why there's so much mediocre equipment out there. As far as the TDAxxxx devices, they're very popular in industry and I always see the datasheet circuits employed; even though it's super easy and super cheap to greatly improve them. There's a guy here named Daniel that has fiddled with these devices extensively. Hopefully he'll chime in.Oh come on. I already explicitly stated which caps were causing the distortion, and why. Go back and read what I wrote, and then tell me which caps are the offenders and how they are introducing distortion into the circuit. Show me that I'm not wasting my time, and I'll come back and tell you what values to use, a few more critical components to change to reduce distortion even further, and the analysis of the newly configured circuit.
The components will only cost a few dollars, but the engineering fee is going to cost you dearly (just kidding).
I don't like spoon feeding people. I do like teaching people, though.
The fact that you actually have the circuit functioning means that you're halfway there.
The components will only cost a few dollars, but the engineering fee is going to cost you dearly (just kidding).
I don't like spoon feeding people. I do like teaching people, though.
The fact that you actually have the circuit functioning means that you're halfway there.
You are not wasting your time! I am following you completely from the start. What I was trying to say is that I am not expert enough to be able to calculate the needed replacement values, so I need you to tell me them. My apologies if that means that I am asking you to spoon feed me.
Thanks!
Thanks.
So having in mind what you said, that the C1-R2 pole freq. should be significantly higher than the other two, feedback poles, I suppose these values would work fine:
C1 - 47uF - ~49Hz freq.
C6 & C2 - 100uF - ~23Hz. freq.
The resistor values stay intact (or should they?)
This might be completely wrong on multiple levels, so please correct me as needed. By my logic, I would go even higher on the cap values to further reduce the frequency, but something tells me it might cause other issues.
Again, I don't really have an idea of what I am doing, so please let me know if I am wrong.
Thanks again!
So having in mind what you said, that the C1-R2 pole freq. should be significantly higher than the other two, feedback poles, I suppose these values would work fine:
C1 - 47uF - ~49Hz freq.
C6 & C2 - 100uF - ~23Hz. freq.
The resistor values stay intact (or should they?)
This might be completely wrong on multiple levels, so please correct me as needed. By my logic, I would go even higher on the cap values to further reduce the frequency, but something tells me it might cause other issues.
Again, I don't really have an idea of what I am doing, so please let me know if I am wrong.
Thanks again!
Ideally it should be significantly higher that the other two poles, yes. In practice you can juggle the ratio (it becomes a practical necessity usually), but you must always make the input pole (C1-R2 in this circuit) higher than the feedback pole.
The reason for this is because the distortion introduced by the input pole is significantly less (and less complex) than the distortion introduced by the feedback pole. This concept is simple parsing and provides an easy and inexpensive way to reduce distortion.
The effects of the distortion introduced by the feedback pole are greatly exaggerated in a bridge amp application. The reason for this is due to any mismatch of components; and by far the worse offender in this circuit would be the feedback capacitors. So "burying" their poles by making them way lower than the input pole will greatly mitigate the distortion introduced by them.
Your math is off, but you're moving in the right direction.
C6 and C2 = 100 uF yields 2.3 Hz, which is much better.
C1 = 47 uF yields 0.5 Hz, which is way too low for a 2.3 Hz feedback pole. You want the input pole to be higher.
Try C1 = 2.2 uF for a pole of 3.2 Hz. This will yield flat response and minimal phase shift down to 32 Hz.
Try C6 and C2 = 220 uF which would yield a pole of 1 Hz. This pole spread will help to mitigate distortion from the feedback capacitors and is a tremendous improvement over the original circuit.
Using such large capacitors for C6 and C2 might cause objectionable thump when turning the amp on and off, or worst case scenario might cause latch up problems when turning it on and off. This can be mitigated or entirely eliminated by the addition of an anti-latch up circuit; which for this circuit would consist of four small signal diodes.
To be continued.
Listening to the OP's audio sample is hard to judge but it kind of sounds like an oscillation which can be many problems but is likely poor circuit layout. Many new amp builders fall into this trap. They follow the schematic to a T, but the actual board layout is terrible.
Lowering the pole frequency will certainly improve the bass performance but it won't fix the farting sound his amp makes if he's got layout issues.
Lowering the pole frequency will certainly improve the bass performance but it won't fix the farting sound his amp makes if he's got layout issues.
I didn't even realize there was an audio clip. They aren't usually very helpful anyway.
A bridge amp can introduce distortions that a single ended amp can't. If the amplifier is being fed with frequencies where the amplifier has different phase shift and different beta gain in both legs, very nasty things can happen. Remember, in a bridge amp any mess at the output will be injected right back into the feedback network of both legs.
That is why it's imperative to make the input pole the dominant pole. You do not want to feed the amplifier frequencies where the imbalance from the feedback capacitors becomes significant; you want to avoid that completely.
The better balanced a bridge amp is, the better it will perform. Remember that a bridge amp will theoretically cancel some distortions; but the flip side is that if the beta gain and phase parameters of each leg are not closely matched, it will introduce more distortion than a single ended amplifier will.
I agree with you about layout too. In fact poor layout will exacerbate the phenomena I explained in my previous post.
It can be difficult to come up with a good layout if you're trying to make things compact. Long circuit board traces can be troublesome but the way to mitigate them is to place all passive components (resistors, capacitors, etc) as close as possible to the leads of the active devices (chips, transistors, etc) and connect the other leg of the resistor to the longer circuit board trace. In fact I'm building a headphone amp I designed using a 49600 buffer and I was stuck with a long circuit board trace to the input of the buffer, which has a very high input impedance. Of course I want to put a resistor on the output lead of the driver chip, so I split the resistance; I put a 47 ohm resistor right on the output of the driver chip and another 47 ohm resistor right on the input lead of the buffer. This way the long circuit board trace is connected to a resistor lead on both ends, and not directly to an active device.
It can be difficult to come up with a good layout if you're trying to make things compact. Long circuit board traces can be troublesome but the way to mitigate them is to place all passive components (resistors, capacitors, etc) as close as possible to the leads of the active devices (chips, transistors, etc) and connect the other leg of the resistor to the longer circuit board trace. In fact I'm building a headphone amp I designed using a 49600 buffer and I was stuck with a long circuit board trace to the input of the buffer, which has a very high input impedance. Of course I want to put a resistor on the output lead of the driver chip, so I split the resistance; I put a 47 ohm resistor right on the output of the driver chip and another 47 ohm resistor right on the input lead of the buffer. This way the long circuit board trace is connected to a resistor lead on both ends, and not directly to an active device.
Last edited:
Nope. No joy.
I first tried 220uf C2 & C6, 2.2uF C1, and the amount of distortion remained pretty much the same. It is the same even after placing 470uF C2 & C6 and 4.7uF C1. Just to note, no noticeable thump when turning on/off...
Could it be something with the caps' quality? I am using ordinary electrolytic caps, 0,10$ each.
Also, I tried with the only one IC output connected to the speaker with the other speaker connection to GND, and the same thing happens, on both ICs. BTW, that sound that I got when the booster was on max setting, occurs only on the left IC, while on the right one it's much less exaggerated.
So I am a bit confused now. The pole changes were supposed to at least improve the circuit to some point, yet the exactly same problem persists.
If it helps somehow, here is the PCB I used: Screenshot by Lightshot .. I drew it by myself with no considerations for the quality of the layout. If you think there might be anything that might be related to my current issue, please let me know. The trace width is probably too small, but I am unsure if that might cause any issues (about 10 mils).
Thanks!
I first tried 220uf C2 & C6, 2.2uF C1, and the amount of distortion remained pretty much the same. It is the same even after placing 470uF C2 & C6 and 4.7uF C1. Just to note, no noticeable thump when turning on/off...
Could it be something with the caps' quality? I am using ordinary electrolytic caps, 0,10$ each.
Also, I tried with the only one IC output connected to the speaker with the other speaker connection to GND, and the same thing happens, on both ICs. BTW, that sound that I got when the booster was on max setting, occurs only on the left IC, while on the right one it's much less exaggerated.
So I am a bit confused now. The pole changes were supposed to at least improve the circuit to some point, yet the exactly same problem persists.
If it helps somehow, here is the PCB I used: Screenshot by Lightshot .. I drew it by myself with no considerations for the quality of the layout. If you think there might be anything that might be related to my current issue, please let me know. The trace width is probably too small, but I am unsure if that might cause any issues (about 10 mils).
Thanks!
Well I didn't really scrutinize your layout in detail, but it does have some long circuit board traces going to the chips. Input pins should be connected to short traces.
I wouldn't worry about the caps right now. Furthermore, cheap generic electrolytic caps can work very well in some audio signal applications, if the circuit design is correct. But there are better caps for almost every application if you just read the datasheets and manufacturer literature.
I would look for other things before worrying about layout. You need a high frequency bypass capacitor for the chips. You need to solder a 0.1 uF ceramic capacitor as close as possible to the power supply pins; I often solder them directly to the pins on the bottom of the board.
I don't know how carefully you assembled it. I solder in one component at a time, starting with passive (resistor and capacitor) components. Then I measure and make sure that there is good contact to the traces. I do not leave anything to chance because it is so easy to slip up or screw up. Wiggle the components and see if you can see anything loose on the copper side of the board.
You must also do a DC test. Since you say that one chip distorts and one doesn't, see if the DC voltages are the same for each circuit. Any discrepancies will lead you to the problem.
Cliff notes -
1) High frequency bypass capacitor right on the power supply pins of the chip - this isa a must if you haven't already done it.
2) Wiggle test / inspection - you just might turn up your problem. Make sure there are no cold solder joints. Reflow any joints that you find suspect.
3) DC test - Just do it. Be extra careful not to short anything out. I have tiny little hook clips that work great on component leads. but they still aren't foolproof.
I wouldn't worry about the caps right now. Furthermore, cheap generic electrolytic caps can work very well in some audio signal applications, if the circuit design is correct. But there are better caps for almost every application if you just read the datasheets and manufacturer literature.
I would look for other things before worrying about layout. You need a high frequency bypass capacitor for the chips. You need to solder a 0.1 uF ceramic capacitor as close as possible to the power supply pins; I often solder them directly to the pins on the bottom of the board.
I don't know how carefully you assembled it. I solder in one component at a time, starting with passive (resistor and capacitor) components. Then I measure and make sure that there is good contact to the traces. I do not leave anything to chance because it is so easy to slip up or screw up. Wiggle the components and see if you can see anything loose on the copper side of the board.
You must also do a DC test. Since you say that one chip distorts and one doesn't, see if the DC voltages are the same for each circuit. Any discrepancies will lead you to the problem.
Cliff notes -
1) High frequency bypass capacitor right on the power supply pins of the chip - this isa a must if you haven't already done it.
2) Wiggle test / inspection - you just might turn up your problem. Make sure there are no cold solder joints. Reflow any joints that you find suspect.
3) DC test - Just do it. Be extra careful not to short anything out. I have tiny little hook clips that work great on component leads. but they still aren't foolproof.
Last edited:
I inspected ALL circuit traces/joints, and I found nothing. I have actually done it before too, once while I was soldering the elements, and once the soldering of the board was finished.
DC test went well too. ~10mV on the left chip and ~5mV on the right.
There are two .1uF polyester caps which should serve as bypass caps, but they might be a bit too far from the chips, I would say. (Their positions: Screenshot by Lightshot )
As these chips don't have a GND leg, should I solder the caps +12V to GND and -12V to GND, with the one connection of the cap to the GND somewhere on the board, right?
Also, should I put two caps (for +12V and -12V) to each IC (that being 2x2)?
Thanks!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.