Yes. I solder small ceramic caps directly to the power pins of chips, on the underside of the board. In fact it's better sometimes to use just one cap between the two power supply pins. If you use two caps they should be grounded to the noise ground for that channel.
I didn't look at your layout very closely, but haphazard layout is bad with high gain circuits. Specifically, there needs to be a power ground and a noise ground, and they should only be connected together at a star ground. Read this
Grounding Principles - The Signal - Archives - TI E2E Community . There are some other articles of interest too.
Finally, there are chips designed for car radios that have four bridge amps in one package, operate with a bare minimum of external components, and are so foolproof that it is almost impossible to screw up the layout.
So, does that mean I can put a cap across -12V and +12V?
This concept is a bit over my head, but I'll think about how can I make my circuit layout in that fashion, since there are multiple components, like caps and resistors which also connect to GND, so I need to see where exactly they would connect in the grounding configuration.
Erm, I think I'll rather try to make this circuit I am working on perform properly, but that's an option as well.
Thanks for helping me so far! 🙂
EDIT: Almost forgot. There is something that I need to ask about; since I use a PC PSU for the amp, when it's fan is spinning, the same spinning noise is audible from the speaker I use. When I force the fan to stop, the noise is gone. The same thing happened when I attached a CD drive to the PSU as a load for the 5V line: any clicks the drive made also were audible at the output of the amp.
Could this mean I have inadequate PSU caps? They are 100uF electrolytic caps; one for the +12V and one for the -12V line. They are not too far from the IC, but could their value be too low?
Last edited:
You need to understand grounding and bypass if you're going to build chip amps.
Do you at least understand "dirty ground" vs "clean ground"? You must understand this. It looks like you just put all your grounds in a row. You can't do that! Active devices can go nuts with that kind of layout.
It sounds like you just piggybacked onto your computer power supply. That's asking for all kinds of trouble. And no, 100 uF is way too small for power supply ripple caps. I find it ironic that you're trying to build a "bass amp" and you employed such tiny capacitors! I'm using 1000 uF in a headphone amp.
Maybe you could try a dedicated power supply and see how it works. Keep it simple.
Do you at least understand "dirty ground" vs "clean ground"? You must understand this. It looks like you just put all your grounds in a row. You can't do that! Active devices can go nuts with that kind of layout.
It sounds like you just piggybacked onto your computer power supply. That's asking for all kinds of trouble. And no, 100 uF is way too small for power supply ripple caps. I find it ironic that you're trying to build a "bass amp" and you employed such tiny capacitors! I'm using 1000 uF in a headphone amp.
Maybe you could try a dedicated power supply and see how it works. Keep it simple.
Last edited:
Just to mention it here at the end, distorted bass immediately raises suspicions about the power supply being too small. Because those big bass notes will drain the power supply capacitors and clamor for more.
At the same time, computer power supplies usually don't have any reserve power to speak of, and their 12-volt line often borders on the barely adequate. If I'm not mistaken, the negative 12 volt line is not currently used by computer manufacturers, so its status is "if any."
All of which is just to say, basics first, basics first, basics first.
At the same time, computer power supplies usually don't have any reserve power to speak of, and their 12-volt line often borders on the barely adequate. If I'm not mistaken, the negative 12 volt line is not currently used by computer manufacturers, so its status is "if any."
All of which is just to say, basics first, basics first, basics first.
Last edited:
Your *first* problem is inadequate supply, period.
Your *second* problem is that you should cut some bass to make life easier for your amp and speaker, don´t disagree with that at all, but first you must feed it proper power.
Forget your PC SMPS, get a 12+12VAC transformer, rated for 3 or 4A , a nice >6A bridge rectifier and 4 4700uFx25V capacitors (2 per rail) and Life will smile to you. 🙂
PS: the transformer might be labelled 12+12V , 50 to 75VA , means the same.
PS2: you´ll need a relatively large heatsink and, besides, I´d use TDA2050 instead of TDA2030 . Only a few cents difference.
Your *second* problem is that you should cut some bass to make life easier for your amp and speaker, don´t disagree with that at all, but first you must feed it proper power.
Forget your PC SMPS, get a 12+12VAC transformer, rated for 3 or 4A , a nice >6A bridge rectifier and 4 4700uFx25V capacitors (2 per rail) and Life will smile to you. 🙂
PS: the transformer might be labelled 12+12V , 50 to 75VA , means the same.
PS2: you´ll need a relatively large heatsink and, besides, I´d use TDA2050 instead of TDA2030 . Only a few cents difference.
.
I got curious and did a little research. It's odds-on the negative 12 volt line on your computer power supply is rated for less than one amp, which is to say less than one watt.
Obviously this is not going to do the job for a bridged amp of around 30 watts. For a 30 watt amp my rule of thumb says a 50 watt power supply.
Just to mention it in a sidebar, is this amp for general listening around the computer? If so 30 watts is overkill by about 25 watts. For just general small-room listening something like a 5 watt amp is probably more than enough. Note that unless you go to rock concerts, your ears have never in your life been subjected to more than about one watt of power.
Not that anything is wrong with using a larger amp, except that you have to keep the volume turned way down. Amps usually like to operate with the volume knob more or less in the middle third of its range.
.
I got curious and did a little research. It's odds-on the negative 12 volt line on your computer power supply is rated for less than one amp, which is to say less than one watt.
Obviously this is not going to do the job for a bridged amp of around 30 watts. For a 30 watt amp my rule of thumb says a 50 watt power supply.
Just to mention it in a sidebar, is this amp for general listening around the computer? If so 30 watts is overkill by about 25 watts. For just general small-room listening something like a 5 watt amp is probably more than enough. Note that unless you go to rock concerts, your ears have never in your life been subjected to more than about one watt of power.
Not that anything is wrong with using a larger amp, except that you have to keep the volume turned way down. Amps usually like to operate with the volume knob more or less in the middle third of its range.
.
I might try to do a new revision of the PCB. It's good that you have pointed out some very important stuff that I have never thought of.
To the others: I have an update. I modded two PC PSUs so I used the +12V lines for both +12V and -12V. Although there IS some improvement, the distortion is still there.. I'll try larger caps and see what happens.
Anyway, the transformer I am thinking to buy is a 30W one (I am unsure if that is enough, but the next available power is 60W and that is double the cost of this one). Also, it has ±12V and ±18V outputs. However, as far as I know the voltage is multiplied by a certain factor after using a diode bridge to convert it to DC, so I'll probably get ±14-15V from the 12V output. Please let me know if the 30W one will be okay.
Thanks!
.
Sorry to tell you that you'd need a room full of capacitors to make that computer power supply work, and it still wouldn't. Moving along to your much better idea of a real power supply:
AC voltage rectifies to its peak value, which is 1.414 times the meter reading. So a +/-12 volt supply (24 volt transformer center tapped) rectifies to +/-17 volts
You then you lose 1.4 volts across the bridge rectifier (2 diodes at .7 volt each), which puts you around plus and minus 15 volts, which is dandy for your purpose.
I'm speaking in round numbers, obviously. There's really no point in worrying about every teeny fraction of a volt with power supplies, because the performance of the amplifier chip is only minimally affected. Much better to run the chip a couple of volts under than a couple of volts over. Nothing audible happens with a volt or two.
About that transformer...
Are you, by chance, talking about a toroid? You don't need one, you really don't. No kidding, you don't. Think it over: you're going to put a bridge rectifier, and some capacitors, downstream of the transformer, so what's the advantage of the mysterious and godlike toroid? Hint: none.
But your judgment in your affairs is better than mine, of course. So aside from all that, the 30 watt transformer will work fine until you turn up the volume, and then the power supply will be inadequate. This is largely because transformers don't run at 100% efficiency, more like 60-80%, with the emphasis on 60. So don't hook an 18 watt horse to a 30 watt carriage and expect full speed.
BUT a 30 watt amp actually puts out 30 watts practically none of the time, so yes you can squeak by. You'll be doing what lots of manufacturers do, and they get away with it (mostly), so why not you? But again sorry, more capacitors won't help.
Ummm...on the other hand...is this a gig amp? You're in a group, or want to be? Go for 60, less won't do.
Hope this might be of some help.
.
Sorry to tell you that you'd need a room full of capacitors to make that computer power supply work, and it still wouldn't. Moving along to your much better idea of a real power supply:
AC voltage rectifies to its peak value, which is 1.414 times the meter reading. So a +/-12 volt supply (24 volt transformer center tapped) rectifies to +/-17 volts
You then you lose 1.4 volts across the bridge rectifier (2 diodes at .7 volt each), which puts you around plus and minus 15 volts, which is dandy for your purpose.
I'm speaking in round numbers, obviously. There's really no point in worrying about every teeny fraction of a volt with power supplies, because the performance of the amplifier chip is only minimally affected. Much better to run the chip a couple of volts under than a couple of volts over. Nothing audible happens with a volt or two.
About that transformer...
Are you, by chance, talking about a toroid? You don't need one, you really don't. No kidding, you don't. Think it over: you're going to put a bridge rectifier, and some capacitors, downstream of the transformer, so what's the advantage of the mysterious and godlike toroid? Hint: none.
But your judgment in your affairs is better than mine, of course. So aside from all that, the 30 watt transformer will work fine until you turn up the volume, and then the power supply will be inadequate. This is largely because transformers don't run at 100% efficiency, more like 60-80%, with the emphasis on 60. So don't hook an 18 watt horse to a 30 watt carriage and expect full speed.
BUT a 30 watt amp actually puts out 30 watts practically none of the time, so yes you can squeak by. You'll be doing what lots of manufacturers do, and they get away with it (mostly), so why not you? But again sorry, more capacitors won't help.
Ummm...on the other hand...is this a gig amp? You're in a group, or want to be? Go for 60, less won't do.
Hope this might be of some help.
.
Last edited:
So true. Switching power supplies are tricky for audio circuits. I prototyped a chip amp to run off a switching power supply. I used inductors in all three legs of the power supply local filter and it worked very well. I scavenged the inductors from computers.
Large filter capacitors just heat up a switching power supply and negate any efficiency gains. Power amps that employ switching power supplies have been used commercially, but I don't think that they used a glorified computer supply.
I don't think I would ever build a permanent version of an audio circuit that employed a switching power supply.
So true. I have tested commercial amplifiers and they rarely deliver the steady state power that they claim. There are notable exceptions; but typically an amplifier that is rated "50 watts RMS" 🙄 will barely deliver 12 watts a channel continuous without clipping. Even with a tone burst test, some barely deliver 20 watts a channel.
.
It makes little sense to tell a guy to build a power supply, then not tell him how to build it. So herewith a schematic.
The cost of this circuit is in the "a few dollars" category, even including the transformer. Depending on whether you're locked into toroid thinking.
Jameco's 24VCT (24 volts center tapped), 48 watt (2 amp), #112513 transformer goes for $13 plus probably 5 bucks shipping. And by the way Radio Shack's power transformers are something of a best kept secret, they're on eBay these days, free shipping. I'm not talking toroids, of course.
Electrolytic capacitor notes. These are my own prejudices.
Multiple small capacitors are better than single big capacitors. This is largely in the hope that errors will average out.
The same theory applies to using 105-degree capacitors. A more robust electrolytic = fewer errors, one vaguely hopes.
Same thinking with the voltage rating of the capacitors, and higher is better. The schematic calls for 50-volt capacitors because of what's readily available, but I use 100-volt capacitors when feasible. "Feasible" means prices can quickly elevate from pennies to dollars, and these hoped-for benefits are admittedly questionable.
Layout of power supplies is not terribly critical. If you don't use a toroid, be aware that the magnetic field of the transformer (possible source of hum) is weak where the coils are surrounded by the steel laminations, strong otherwise. Try to put the mounting tabs of the transformer at 90-degrees to the amp circuit. Distance helps.
The costs with these things goes: Components: cheap. Transformer: not so cheap. Housing: can be a freakin' fortune. Stop throwing soup cans away.
Hope this might be of some help.
.
It makes little sense to tell a guy to build a power supply, then not tell him how to build it. So herewith a schematic.
The cost of this circuit is in the "a few dollars" category, even including the transformer. Depending on whether you're locked into toroid thinking.
Jameco's 24VCT (24 volts center tapped), 48 watt (2 amp), #112513 transformer goes for $13 plus probably 5 bucks shipping. And by the way Radio Shack's power transformers are something of a best kept secret, they're on eBay these days, free shipping. I'm not talking toroids, of course.
Electrolytic capacitor notes. These are my own prejudices.
Multiple small capacitors are better than single big capacitors. This is largely in the hope that errors will average out.
The same theory applies to using 105-degree capacitors. A more robust electrolytic = fewer errors, one vaguely hopes.
Same thinking with the voltage rating of the capacitors, and higher is better. The schematic calls for 50-volt capacitors because of what's readily available, but I use 100-volt capacitors when feasible. "Feasible" means prices can quickly elevate from pennies to dollars, and these hoped-for benefits are admittedly questionable.
Layout of power supplies is not terribly critical. If you don't use a toroid, be aware that the magnetic field of the transformer (possible source of hum) is weak where the coils are surrounded by the steel laminations, strong otherwise. Try to put the mounting tabs of the transformer at 90-degrees to the amp circuit. Distance helps.
The costs with these things goes: Components: cheap. Transformer: not so cheap. Housing: can be a freakin' fortune. Stop throwing soup cans away.
Hope this might be of some help.
.
Attachments
.
Thinking about it, I don't think I achieved my dual goals of information and clarity with the original schematic. After all, any schematic is confusing to the point of terrifying if you're not accustomed to reading them.
So here's the same schematic, but with the notes removed. One transformer, one rectifier bridge, and as many capacitors as you feel like messing with--could be only two. Power supplies don't have to be complicated.
.
Thinking about it, I don't think I achieved my dual goals of information and clarity with the original schematic. After all, any schematic is confusing to the point of terrifying if you're not accustomed to reading them.
So here's the same schematic, but with the notes removed. One transformer, one rectifier bridge, and as many capacitors as you feel like messing with--could be only two. Power supplies don't have to be complicated.
.
Attachments
Agree and add: minimum total capacitance per side would be 2200uFx25V , 4700uFx25V much better and what you should *really* use.
As of transformer power rating, if a 48 "W" (really VA or voltampere) costs just $13 , I *hope* you can get a 60 to 70 VA one for a few bucks more.
I´m considering that it won´t be a "PC speaker amp" but a **Bass** amp, and those tend to need up to the last watt.
30W is fine for home practice and probably a garage band ... if the drummer is not a heavy hitter.
But I DO encourage you to build it, it´s the smallest size which is really useful (as in: it can be used outside the bedroom) and will be an excellent stepping stone towards your next build: a LM3886 based 50/60 W one 🙂
As of transformer power rating, if a 48 "W" (really VA or voltampere) costs just $13 , I *hope* you can get a 60 to 70 VA one for a few bucks more.
I´m considering that it won´t be a "PC speaker amp" but a **Bass** amp, and those tend to need up to the last watt.
30W is fine for home practice and probably a garage band ... if the drummer is not a heavy hitter.
But I DO encourage you to build it, it´s the smallest size which is really useful (as in: it can be used outside the bedroom) and will be an excellent stepping stone towards your next build: a LM3886 based 50/60 W one 🙂
Alright, I managed to put my hands on a 30W (at least it says so) transformer (not toroidal one!), built the power supply according to the schematic given by bentshake and I can say there is a great improvement now! The distortion is pretty much gone (well, the most of it) and I can get more power from the amp now, with supply voltages of ~±16V.
However, I have two issues:
-there is a certain buzz even if nothing is connected or there is a device which is turned off. The transformer is about 60cm (two feet) far from the circuit and the speaker. I used 2x2200uF caps for both +12V and -12V lines. The caps are somewhere in the middle between the transformer/rectifier and the circuit. Also, I have connected the marked points together here: Screenshot by Lightshot
Please let me know if I did something wrong.
Also, when I plug in my bass, and don't touch the strings, that buzz is amplified even more. However, when I touch either the strings or the ground point of the jack, the buzz lowers. This is a problem that some friend of mine has with a commercial amp too. Can you give me any help to fixing this?
Thanks!
However, I have two issues:
-there is a certain buzz even if nothing is connected or there is a device which is turned off. The transformer is about 60cm (two feet) far from the circuit and the speaker. I used 2x2200uF caps for both +12V and -12V lines. The caps are somewhere in the middle between the transformer/rectifier and the circuit. Also, I have connected the marked points together here: Screenshot by Lightshot
Please let me know if I did something wrong.
Also, when I plug in my bass, and don't touch the strings, that buzz is amplified even more. However, when I touch either the strings or the ground point of the jack, the buzz lowers. This is a problem that some friend of mine has with a commercial amp too. Can you give me any help to fixing this?
Thanks!
You have grounding problems; including, but not confined to, a ground loop.
I think you have demonstrated that your circuit is fine, and suffers only from a poor, haphazard layout. Read and understand the article I referenced, make a new board with correct layout, and claim victory.
I think you have demonstrated that your circuit is fine, and suffers only from a poor, haphazard layout. Read and understand the article I referenced, make a new board with correct layout, and claim victory.
.
I agree with Fast Eddie. A buzz that changes when you touch something is petty much certain to be a bad or missing ground. Loose wire, no wire, bad solder joint, something. Are the connecting cables you're using shielded, and are the shields doing their job (try another cable)? This could also cause distortion.
However, with your bass there's another issue. You're just plugging the bass directly into the amp? Less than ideal. Instrument pickups are very high impedance, and amplifiers are low-to-lowish impedance. Strictly speaking you should use an impedance matching device. Musicians have a name for these that I forget, but essentially what you need is a buffer made up from an NE5532, the cost would be pretty close to zero. Making matters worse, the output from an instrument pickup might or might not be enough to drive an amp, which is why there's usually a preamp or mixer.
Ummm...I don't mean to be smarty mouth...but is there any critical reason for using this particular circuit, which in my opinion is so horribly complicated that mistakes are practically mandatory?
How about a nice LM1875 ($2 on eBay)? ( http://www.ti.com/lit/gpn/lm1875 )
The LM1875 would only put out around 10 watts with the power supply you now have (thanks for your confidence), but those are 10 real watts that will make you jump, guaranteed. I'm posting a schematic I happen to have around. Notice that the amp circuit itself only has 5 resistors and 3 capacitors, the other capacitors are associated with the power supply.
You can get more out of the LM1875, or go to an LM3875, which is a bunker buster. But you'd need to redo your power supply, so why not make something that works first? Just a thought.
.
I agree with Fast Eddie. A buzz that changes when you touch something is petty much certain to be a bad or missing ground. Loose wire, no wire, bad solder joint, something. Are the connecting cables you're using shielded, and are the shields doing their job (try another cable)? This could also cause distortion.
However, with your bass there's another issue. You're just plugging the bass directly into the amp? Less than ideal. Instrument pickups are very high impedance, and amplifiers are low-to-lowish impedance. Strictly speaking you should use an impedance matching device. Musicians have a name for these that I forget, but essentially what you need is a buffer made up from an NE5532, the cost would be pretty close to zero. Making matters worse, the output from an instrument pickup might or might not be enough to drive an amp, which is why there's usually a preamp or mixer.
Ummm...I don't mean to be smarty mouth...but is there any critical reason for using this particular circuit, which in my opinion is so horribly complicated that mistakes are practically mandatory?
How about a nice LM1875 ($2 on eBay)? ( http://www.ti.com/lit/gpn/lm1875 )
The LM1875 would only put out around 10 watts with the power supply you now have (thanks for your confidence), but those are 10 real watts that will make you jump, guaranteed. I'm posting a schematic I happen to have around. Notice that the amp circuit itself only has 5 resistors and 3 capacitors, the other capacitors are associated with the power supply.
You can get more out of the LM1875, or go to an LM3875, which is a bunker buster. But you'd need to redo your power supply, so why not make something that works first? Just a thought.
.
Attachments
Last edited:
I agree; the circuit is atrocious. I would never look twice at it.
I prefer to design my own, instead of wondering if somebody else's circuit will do what I want. Plus building a circuit blindly is a crap shoot.
Perfect beginner circuit, with nice performance
Also heed the advice about signal level and impedance matching.
Electric guitars are notorious for picking up and injecting noise.
.
<< built the power supply according to the schematic given by bentsnake >>
Maybe I should add that this is by no means my power supply circuit. This circuit is pretty much "the way you do it," and has been since Tesla joined with Westinghouse to make AC (not Edison's DC) the power standard of the world.
Moving along, second verse same as the first. You can't tell a guy to build something, then not tell him how to build it. So posted is an NE5532 buffer circuit with a following gain stage.
Circuit notes:
This circuit will work happily off your present +/- 16v supply. Current draw is a few milliamps.
R1 is overcurrent protection for the NE5532.
C1 shunts radio frequency energy to ground.
R2 is a load resistor, put there because you usually don't know what the input signal is going to come from. By providing a current path to ground R2 ensures that there will be voltage at the input side of C1. R2 in parallel with R3 sets the input impedance, which here is about 20k.
(If R2 is not present, then the value of R3 is the input impedance.)
C2 and R3: C2 blocks DC, and with R3 forms a high-pass filter, also called a bass rolloff filter. Here the rolloff starts at just over 20 Hz.
U1a (1/2 of an NE5532) is a buffer. It's characteristic of op amps that connected this way they have unity gain, extremely high input impedance, and extremely low output impedance. The buffer isolates the input source from the following circuit, effectively "looking at" the input source without affecting it.
Rvol is the volume control. The value can actually be anything from 2k (no lower) on up.
U2b (the other half of an NE5532) is the gain stage, which here is arbitrarily set at a gain of 11 (21 dB). Gain can be anything theoretically, but keep it to 20 or so in the real world. The formula is simply R4 divided by R5, and add 1. That is:
Gain = R4/R5 + 1
For preference, leave R5 at about 2k and adjust R4. Higher resistance for R4 = more gain. R5 no lower than 1k, if you must.
The output is low impedance, and should be suitable for the line-in input of a mixer, preamp, or amp. If you build the LM1875 amp above, the output of this circuit connects to Vin, the top of R1, without any additional components.
NOTE: The 20k input impedance shown here is much too low for instrument pickups, such as on a guitar or bass. It will work, but not really well. To be strictly correct with an instrument pickup, change both R2 and R3 to 1meg, for a 0.5meg input impedance. Change C2 to 0.068uF
Caution: 1meg is a very high value in audio terms, and noise problems can multiply. Keep leads short, use shielded wire, and ground everything, including the body of pots (Rvol).
.
<< built the power supply according to the schematic given by bentsnake >>
Maybe I should add that this is by no means my power supply circuit. This circuit is pretty much "the way you do it," and has been since Tesla joined with Westinghouse to make AC (not Edison's DC) the power standard of the world.
Moving along, second verse same as the first. You can't tell a guy to build something, then not tell him how to build it. So posted is an NE5532 buffer circuit with a following gain stage.
Circuit notes:
This circuit will work happily off your present +/- 16v supply. Current draw is a few milliamps.
R1 is overcurrent protection for the NE5532.
C1 shunts radio frequency energy to ground.
R2 is a load resistor, put there because you usually don't know what the input signal is going to come from. By providing a current path to ground R2 ensures that there will be voltage at the input side of C1. R2 in parallel with R3 sets the input impedance, which here is about 20k.
(If R2 is not present, then the value of R3 is the input impedance.)
C2 and R3: C2 blocks DC, and with R3 forms a high-pass filter, also called a bass rolloff filter. Here the rolloff starts at just over 20 Hz.
U1a (1/2 of an NE5532) is a buffer. It's characteristic of op amps that connected this way they have unity gain, extremely high input impedance, and extremely low output impedance. The buffer isolates the input source from the following circuit, effectively "looking at" the input source without affecting it.
Rvol is the volume control. The value can actually be anything from 2k (no lower) on up.
U2b (the other half of an NE5532) is the gain stage, which here is arbitrarily set at a gain of 11 (21 dB). Gain can be anything theoretically, but keep it to 20 or so in the real world. The formula is simply R4 divided by R5, and add 1. That is:
Gain = R4/R5 + 1
For preference, leave R5 at about 2k and adjust R4. Higher resistance for R4 = more gain. R5 no lower than 1k, if you must.
The output is low impedance, and should be suitable for the line-in input of a mixer, preamp, or amp. If you build the LM1875 amp above, the output of this circuit connects to Vin, the top of R1, without any additional components.
NOTE: The 20k input impedance shown here is much too low for instrument pickups, such as on a guitar or bass. It will work, but not really well. To be strictly correct with an instrument pickup, change both R2 and R3 to 1meg, for a 0.5meg input impedance. Change C2 to 0.068uF
Caution: 1meg is a very high value in audio terms, and noise problems can multiply. Keep leads short, use shielded wire, and ground everything, including the body of pots (Rvol).
.
Attachments
Last edited:
- Status
- Not open for further replies.