Hi all,
I guess I am gonna build an amplifier with the TDA2050 too.
I have used a TDA9822 (Class D) amp in one of my projects but I am gonna replace it (TDA2050?). The TDA8922 interferes with the FM receiver, very bad!! I tried everything: shielding, seperate power supplies but it sucks
@ashok: is your toroid really 2x18V? This means rectified 18*2^0,5 = 25,5 Vdc minus 1,4V (rectifier) = 24,1 Vdc. This is when the toroid nominally loaded. But without (significant) load, the voltage can rise up to 27 Vdc (no load) >> does your TDA survive this?
@casiomax: try to change your supply to dual (+/- 18V), the negative power supply ripple compensates the positive
Regards, Rolf
I guess I am gonna build an amplifier with the TDA2050 too.
I have used a TDA9822 (Class D) amp in one of my projects but I am gonna replace it (TDA2050?). The TDA8922 interferes with the FM receiver, very bad!! I tried everything: shielding, seperate power supplies but it sucks
@ashok: is your toroid really 2x18V? This means rectified 18*2^0,5 = 25,5 Vdc minus 1,4V (rectifier) = 24,1 Vdc. This is when the toroid nominally loaded. But without (significant) load, the voltage can rise up to 27 Vdc (no load) >> does your TDA survive this?
@casiomax: try to change your supply to dual (+/- 18V), the negative power supply ripple compensates the positive
Regards, Rolf
@casiomax:
You might try to make the connection from input to power ground through a 2,7 Ohm resistor, and isolate the input connector from ground, e. g. metal parts of the case that are grounded.
Other possible solutions are small capacitors (~47pF) across the rectifier. Or small capacitors (100nF) next to the TDA in parallel to the power supply capacitors.
Separate wires as far from each other as possible. Where audio wires come close to wires that carry AC (or rectified but not yet filtered DC), they pick up hum.
Shielding might also help. I. e. put something in between the power section and the audio section, e. g. aluminium sheets, aluminium foil or a copper-plated PCB connected to ground. Needs careful fixing, so that no short circuits are produced.
Did you use a potentiometer with mains switch attached? I have a small amp like that, and it is picking up hum from there. To reduce that you would have to disconnect the switch and use a separate switch separated from the audio lines.
@IceCube:
Maybe this can help you in your future power supply designs.
If the power supply capacitors are of the right size (~1.000µF/1A load) you get 1,27 times the nominal transformer voltage at nominal load (before rectifier). 1,41 times is what you get under no load conditions.
Here are the most common reasons for higher voltages than above stated
You might try to make the connection from input to power ground through a 2,7 Ohm resistor, and isolate the input connector from ground, e. g. metal parts of the case that are grounded.
Other possible solutions are small capacitors (~47pF) across the rectifier. Or small capacitors (100nF) next to the TDA in parallel to the power supply capacitors.
Separate wires as far from each other as possible. Where audio wires come close to wires that carry AC (or rectified but not yet filtered DC), they pick up hum.
Shielding might also help. I. e. put something in between the power section and the audio section, e. g. aluminium sheets, aluminium foil or a copper-plated PCB connected to ground. Needs careful fixing, so that no short circuits are produced.
Did you use a potentiometer with mains switch attached? I have a small amp like that, and it is picking up hum from there. To reduce that you would have to disconnect the switch and use a separate switch separated from the audio lines.
@IceCube:
Maybe this can help you in your future power supply designs.
If the power supply capacitors are of the right size (~1.000µF/1A load) you get 1,27 times the nominal transformer voltage at nominal load (before rectifier). 1,41 times is what you get under no load conditions.
Here are the most common reasons for higher voltages than above stated
- - Overdimensioned capacitors may lead to something about 1,5..1,6 times transformer output voltage. This is the reason, why capacitor overkill should be considered carefully. More is not always better.
- Cheap transformers with high no-load voltages (e. g. print types with easily 25..30% above nominal)
- Overdimensioned transformers, which will never reach their nominal load, thus always supplying more than nominal voltage.
- Mains voltage above nominal, (allowed range -15..+10%) which should be taken into consideration, when designing a power supply.
Hi pacificblue,
You say 1,27 times with nominal load, before rectifier. Maybe you mean after the rectifier. IMO before = AC, after = DC.
Smaller caps will NOT change the voltage at NO LOAD!! No load does not mean no speaker, there is always a small (quiescent) current through the amp.
You can use 1,27 times the nominal voltage, but this depends on the capacitance, the current, the voltage and voltage drop across the diodes. For example, if you use a 3V ac transformer at nominal load, the rectified voltage will not be higher than 2,83 Volt (due the voltage drop of the diodes).
If you use smaller caps, I agree the average DC voltage (at load) will be lower, and also the ac ripple is higher. The peak current through the rectifier diodes will also be lower.
Regarding the secondary voltage at no load, in general: the lower the power, the higher the no-load voltage (print types are usually low power).
The bypass capacitors (across the rectifier) will not reduce hum (50/60Hz), but shorts HF currents, and so reducing interference and noise.
Oscillations may sound as hum, but is totally different. In casiomax's case it is important to see what causes the 'hum'. So connect the input to ground (with 100 ohm) and connect an oscilloscope to the output and see what the hum is (50/60 Hz) or other (maybe oscillations).
Regard, Rolf
You say 1,27 times with nominal load, before rectifier. Maybe you mean after the rectifier. IMO before = AC, after = DC.
Smaller caps will NOT change the voltage at NO LOAD!! No load does not mean no speaker, there is always a small (quiescent) current through the amp.
You can use 1,27 times the nominal voltage, but this depends on the capacitance, the current, the voltage and voltage drop across the diodes. For example, if you use a 3V ac transformer at nominal load, the rectified voltage will not be higher than 2,83 Volt (due the voltage drop of the diodes).
If you use smaller caps, I agree the average DC voltage (at load) will be lower, and also the ac ripple is higher. The peak current through the rectifier diodes will also be lower.
Regarding the secondary voltage at no load, in general: the lower the power, the higher the no-load voltage (print types are usually low power).
The bypass capacitors (across the rectifier) will not reduce hum (50/60Hz), but shorts HF currents, and so reducing interference and noise.
Oscillations may sound as hum, but is totally different. In casiomax's case it is important to see what causes the 'hum'. So connect the input to ground (with 100 ohm) and connect an oscilloscope to the output and see what the hum is (50/60 Hz) or other (maybe oscillations).
Regard, Rolf
@AndrewT:
Is that so? So all my teachers must be wrong and most of the literature I've read. Maybe you can enlighten me about the wrongnesses you encounter.
But you might just as well try to help solving the hum problem instead, since you seem to have the answer to everything.
@IceCube:
You are of course right. It is after rectifier and filter caps.
I am not aware that I wrote anywhere smaller caps were changing the voltage. I was referring to your previous post, where you said it was 1,41 times the nominal voltage at nominal load. However I learned that it is 1,41 times the nominal voltage at no (or next to no) load, and that the voltage drops to about 1,27 times nominal at nominal load with filter caps of 1000µF/1A load.
The bypass caps do exactly what you say. And you obviously understood what I was aiming at, because you wrote
Well, maybe casiomax knows the difference. And maybe casiomax also has access to an oscilloscope. Until I know that for sure, I try to give easy-to-try solutions that might help, and that are no big waste of time and money, if they don't.
Is that so? So all my teachers must be wrong and most of the literature I've read. Maybe you can enlighten me about the wrongnesses you encounter.
But you might just as well try to help solving the hum problem instead, since you seem to have the answer to everything.
@IceCube:
You are of course right. It is after rectifier and filter caps.
I am not aware that I wrote anywhere smaller caps were changing the voltage. I was referring to your previous post, where you said it was 1,41 times the nominal voltage at nominal load. However I learned that it is 1,41 times the nominal voltage at no (or next to no) load, and that the voltage drops to about 1,27 times nominal at nominal load with filter caps of 1000µF/1A load.
The bypass caps do exactly what you say. And you obviously understood what I was aiming at, because you wrote
Well, maybe casiomax knows the difference. And maybe casiomax also has access to an oscilloscope. Until I know that for sure, I try to give easy-to-try solutions that might help, and that are no big waste of time and money, if they don't.
You are right, I was not mentioning that I meant peak voltage (instead of average voltage).
To be clear, just an example
:
Transformer: 18Vac nominally loaded
DC current: 1A, Capacitor: 1000uF
After rectifying: peak dc voltage = Uac * (sqrt)2 - Udiodes = 18 * 1,41 - 1,4 = 24Vdc (this voltage will also be present at no load).
The average dc voltage (full wave rectifier @ 50 Hz: dt= 10ms):
First calculate the ripple across the cap (top-top):
dU = I * dt / C = 1 * 0,01 / 0,001 = 10V!!
Then calculate average voltage:
Uavg = Utop - (1/2 Uripple) = 24 - 5 = 19 Volt
So in this example the conversion factor is: 19 / 18 = 1,1
If you take 2000uF the ripple voltage will be 5V and the average voltage will be 21,5V so the conversion factor is 1,19
Regards, Rolf
To be clear, just an example
:Transformer: 18Vac nominally loaded
DC current: 1A, Capacitor: 1000uF
After rectifying: peak dc voltage = Uac * (sqrt)2 - Udiodes = 18 * 1,41 - 1,4 = 24Vdc (this voltage will also be present at no load).
The average dc voltage (full wave rectifier @ 50 Hz: dt= 10ms):
First calculate the ripple across the cap (top-top):
dU = I * dt / C = 1 * 0,01 / 0,001 = 10V!!
Then calculate average voltage:
Uavg = Utop - (1/2 Uripple) = 24 - 5 = 19 Volt
So in this example the conversion factor is: 19 / 18 = 1,1
If you take 2000uF the ripple voltage will be 5V and the average voltage will be 21,5V so the conversion factor is 1,19
Regards, Rolf
Ashok: Just visited Ritchie St. today and got generic Taiwanese brown dipped polypropylene 1 uF, 250V at Swastik for Rs.5 each. He also has 2.2uF, 250V for Rs.8. If these are indeed polypropylene and are sonically clean, I'd say that this is a very reasonable price.
Are yellow box-type ECQ-series polypropylene? Swastik has those also, but I skipped getting them.
Are yellow box-type ECQ-series polypropylene? Swastik has those also, but I skipped getting them.
Linuxguru: The problem with the local suppliers is that they cannot give you the same brand everytime. I'm told they mostly get surplus material and so prices are relatively low. I tried some of those boxed yellow caps. I didn't like them. I tried a few grey boxed types and they were OK but the next time I tried to get them they were still grey but sounded 'very' different ...poorer ! A lot of their material has tinned steel leadout !They don't flex easily either.
Heat up the cap with a soldering iron and check if it's a polyprop with a capacitance meter. I'd be surprised if it is really polyprop.
You'd have to do it carefully to avoid burning the cap.
Cheers.
Heat up the cap with a soldering iron and check if it's a polyprop with a capacitance meter. I'd be surprised if it is really polyprop.
You'd have to do it carefully to avoid burning the cap.
Cheers.
New tda 2050
Just finished my first chip amp, a 2050, running of a 2 switching power supplies. All I can say is amazing! The sound is so clean, and sharp.
The bass is tight, and that is without any more capacitance than what the PC power supplies had. I have some 4700 that I plan on hooking in next. should be fun.
Thanks for the info in this thread!
Just finished my first chip amp, a 2050, running of a 2 switching power supplies. All I can say is amazing! The sound is so clean, and sharp.
The bass is tight, and that is without any more capacitance than what the PC power supplies had. I have some 4700 that I plan on hooking in next. should be fun.
Thanks for the info in this thread!
Try some experiments with this board.
I just got back from a short much needed vacation by the sea side !
For all those who tried the TDA2050 I would recommend that they try switching the chip to LM1875 . No changes required to the board or components. You should hear a difference . Both chips are good but you might develop a preference.
Don't look back at other old posts that talk about this . It might bias your opinion before you try it . Do this with a clean (!) mind .
I haven't been able to listen to these boards for some time due to work pressure. I should pull them out ( I have both chips on different boards ) and do a second round of comparison and see if they still (?) sound different !
I just got back from a short much needed vacation by the sea side !
For all those who tried the TDA2050 I would recommend that they try switching the chip to LM1875 . No changes required to the board or components. You should hear a difference . Both chips are good but you might develop a preference.
Don't look back at other old posts that talk about this . It might bias your opinion before you try it . Do this with a clean (!) mind .
I haven't been able to listen to these boards for some time due to work pressure. I should pull them out ( I have both chips on different boards ) and do a second round of comparison and see if they still (?) sound different !
the bass boom in a different listening area is almost certainly due to the alignment chosen by the manufacturer to give the impression of strong bass in a very cheap to make and package and transport cabinet.
High Qbox leads to boomy bass in many environments. A lower Q, requiring a lower Q driver or a bigger box is more expensive. Changing amplifiers is unlikely to cure the original design choice to maximise their profit.
High Qbox leads to boomy bass in many environments. A lower Q, requiring a lower Q driver or a bigger box is more expensive. Changing amplifiers is unlikely to cure the original design choice to maximise their profit.
Yes as AndrewT says, the boom is most likely from the speaker . Try changing the location of the speaker. Keep them away from a corner and / or the walls.
If the box has a duct ,try stuffing it with a sock or some other material. This will convert it to a sealed box and roll off the bass earlier. But corner placement will bring it up enough to sound OK. Try it and see which option works.
The chip amp is capable of producing very good bass if the speakers can.
If the box has a duct ,try stuffing it with a sock or some other material. This will convert it to a sealed box and roll off the bass earlier. But corner placement will bring it up enough to sound OK. Try it and see which option works.
The chip amp is capable of producing very good bass if the speakers can.
Thanks Guys,
I presumed that the speaker was the culprit, as they are a cheapy set of bookshelves...
Maybe should start working on a new diy set, I will ponder it while soldering up a new amp
Speaking of new amps, has anyone tried any of the other TDA chips? I was thinking about trying an active crossover, with maybe 6-8 channels, using a bunch of cheap TDA chips as a test run. I would like something that will have small power requirements as the supplies cost a bit... Any ideas?
I presumed that the speaker was the culprit, as they are a cheapy set of bookshelves...
Maybe should start working on a new diy set, I will ponder it while soldering up a new amp
Speaking of new amps, has anyone tried any of the other TDA chips? I was thinking about trying an active crossover, with maybe 6-8 channels, using a bunch of cheap TDA chips as a test run. I would like something that will have small power requirements as the supplies cost a bit... Any ideas?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.