Hello all, this is my second post, the first one led me to choose popular tpa3116d2 amplifiers for my new speaker build.
I know there is a monster thread on tpa3116d2 amplifiers, but i wasn't able to find information regarding this particular board design through the forum search function.
Here is the amplifier i'm talking about, it is dirt cheap($2.30 is almost what the chip itself is), has minimum number of components, but lacks heatsink, output lc filter and output snubber.
There is also an image of the board in question in attachments, just in case you dont want to click external links.
Regarding the absent heatsink, i can easily install one.
For the absent output LC filter, datasheet says it is not necessary when the speaker cables are less than 125cm in length. I can satisfy this condition by directly placing the amplifier inside the speaker box, turning my loudspeakers into active ones. I was already planning to bi-amp my cabinets so putting one stereo amplifier board inside each cabinet is sensible.
But i cant find a subtle information about the snubbers necessity, datasheet just says that "Additional EMC improvements may be obtained by adding snubber networks from each of the class-D outputs to ground.". My web search also didnt yield anything.
My questions are:
When exactly the snubber is required on the output of a tpa3116d2 amplifier circuit?
Can i just get away without having a snubber on my amplifier board, for the sake of cost savings?
Does any of you have experience with this particular design? Are they suitable for active speaker design, or even better, for driving passive speakers with not so long speaker cables?
I also noticed that it lacks big capacitors in the power input, just some "biggish" smd capacitor chips. Can i also get away with such small power supply filters if i power this amplifier with a laptop power brick (the 19V ~5A ones)? Do those power bricks supply a clear enough power for this application in general, or should i measure the ones i have?
Thanks in advance!
I know there is a monster thread on tpa3116d2 amplifiers, but i wasn't able to find information regarding this particular board design through the forum search function.
Here is the amplifier i'm talking about, it is dirt cheap($2.30 is almost what the chip itself is), has minimum number of components, but lacks heatsink, output lc filter and output snubber.
There is also an image of the board in question in attachments, just in case you dont want to click external links.
Regarding the absent heatsink, i can easily install one.
For the absent output LC filter, datasheet says it is not necessary when the speaker cables are less than 125cm in length. I can satisfy this condition by directly placing the amplifier inside the speaker box, turning my loudspeakers into active ones. I was already planning to bi-amp my cabinets so putting one stereo amplifier board inside each cabinet is sensible.
But i cant find a subtle information about the snubbers necessity, datasheet just says that "Additional EMC improvements may be obtained by adding snubber networks from each of the class-D outputs to ground.". My web search also didnt yield anything.
My questions are:
When exactly the snubber is required on the output of a tpa3116d2 amplifier circuit?
Can i just get away without having a snubber on my amplifier board, for the sake of cost savings?
Does any of you have experience with this particular design? Are they suitable for active speaker design, or even better, for driving passive speakers with not so long speaker cables?
I also noticed that it lacks big capacitors in the power input, just some "biggish" smd capacitor chips. Can i also get away with such small power supply filters if i power this amplifier with a laptop power brick (the 19V ~5A ones)? Do those power bricks supply a clear enough power for this application in general, or should i measure the ones i have?
Thanks in advance!
Attachments
Last edited:
Yes, they are useful even without a snubber. These are minimum configuration boards intended to be completed by the user. My boards are red by the way.
The TPA3116 chip requires some peripheral components as a minimum. These peripheral components are arranged on the minimum board. The rest you have to add according to your application.
Snubber circuit: A snubber circuit is not shown as mandatory in the TPA3116 datasheet. Anyway, the snubber circuit has to be designed to match the output filter and no output filter is implemented. A snubber circuit is further a matter of EMC-regulations that differ in areas around the world. For DIY use, I would not bother about a snubber circuit, at least before I had decided about the output filter.
An output filter is not stated to be mandatory in the datasheet. In practice, it more or less is mandatory. A TPA3116 chip can operate as low as from a 5V supply in which case an output filter may not be mandatory. For normal operation with 12-24V supply, the TPA3116 is so powerful that it will generate too much noise without a filter. A forum member has described that his TPA3116 board, without an output filter, prevents him from receiving FM radio. I would consider an output filter mandatory and add one with a 4-5A capacity as a minimum. And, the zobel-element shown in the datasheet schematics. A snubber circuit only if you find it necessary.
The power line decoupling capacitors (perhaps 1uF each) are there on the board to avoid chip oscillation. These small decoupling capacitors are clearly insufficient for decoupling in the audio-band. Add some 2200uF to 10000uF as close to the small board as possible.
Cooling: The TPA3116 can often operate without much heatsink. Low quiescent current and low-impedance output switches are the reasons. If needed, add as you find appropriate.
The TPA3116 chip requires some peripheral components as a minimum. These peripheral components are arranged on the minimum board. The rest you have to add according to your application.
Snubber circuit: A snubber circuit is not shown as mandatory in the TPA3116 datasheet. Anyway, the snubber circuit has to be designed to match the output filter and no output filter is implemented. A snubber circuit is further a matter of EMC-regulations that differ in areas around the world. For DIY use, I would not bother about a snubber circuit, at least before I had decided about the output filter.
An output filter is not stated to be mandatory in the datasheet. In practice, it more or less is mandatory. A TPA3116 chip can operate as low as from a 5V supply in which case an output filter may not be mandatory. For normal operation with 12-24V supply, the TPA3116 is so powerful that it will generate too much noise without a filter. A forum member has described that his TPA3116 board, without an output filter, prevents him from receiving FM radio. I would consider an output filter mandatory and add one with a 4-5A capacity as a minimum. And, the zobel-element shown in the datasheet schematics. A snubber circuit only if you find it necessary.
The power line decoupling capacitors (perhaps 1uF each) are there on the board to avoid chip oscillation. These small decoupling capacitors are clearly insufficient for decoupling in the audio-band. Add some 2200uF to 10000uF as close to the small board as possible.
Cooling: The TPA3116 can often operate without much heatsink. Low quiescent current and low-impedance output switches are the reasons. If needed, add as you find appropriate.
Last edited:
Unless you're really, REALLY strapped for space, i'd be more inclined to go with some of those blue PBTL TPA3118 boards (or, if you're willing to pay 1-2$ more, there are stereo TPA3118 boards too).
Yes, they're a bit bigger than this "bare-bones" design, but you DO get proper output filters (if a bit undersized), and the 3118 has a belly-pad, using the PCB itself as a heatsink (so no need to mess about with mounting a heatsink). I got mine a couple months ago for well under 2$ a piece from AliExpress, btw 🙄 Or, well, technically, $30.33 for 15pcs including shipping.
I'm using a pair of those PBTL TPA3118 boards to drive my active "recycled" surround speakers. They're not the most sensitive ones out there, but even so, they can go to (relatively) stupid loud SPL without even breaking a sweat, and i'm powering them with 19V laptop "bricks" 😀
Yes, they're a bit bigger than this "bare-bones" design, but you DO get proper output filters (if a bit undersized), and the 3118 has a belly-pad, using the PCB itself as a heatsink (so no need to mess about with mounting a heatsink). I got mine a couple months ago for well under 2$ a piece from AliExpress, btw 🙄 Or, well, technically, $30.33 for 15pcs including shipping.
I'm using a pair of those PBTL TPA3118 boards to drive my active "recycled" surround speakers. They're not the most sensitive ones out there, but even so, they can go to (relatively) stupid loud SPL without even breaking a sweat, and i'm powering them with 19V laptop "bricks" 😀
@Khron thanks i was not aware of those boards, they are much cheaper and has output filters. I'm not really strapped for space, indeed i have plenty of space. I'm just worried about the cost and i don't want to spend too much on a pair of loudspeakers that might not turn out that good. The cheapes pbtl tpa3118 mono boards i found are $1.70 including shipping while the cheapest stereo ones are $6 so i think i'll go with 4 of those blue pbtl boards.
@FauxFrench i didnt know that it might cause that much radio interference so i guess i wont go without those filters. I'm not worried about the fm/am interference but other radio interference might cause trouble.
@FauxFrench i didnt know that it might cause that much radio interference so i guess i wont go without those filters. I'm not worried about the fm/am interference but other radio interference might cause trouble.
If your speakers connection cables are <10-15cm then an output filter mostly isn't necessary. You may use a simple ferrite bead filter if AM/FM is a problem and your connection wires are >15cm and/or you place them near sensitive devices.
I made some extensive tests with the TPA3128 (Which is a 3118 in a different package) in the past. It can clearly be seen, that a ferrite+C filter is enough even with 150cm between amp and speaker.
TPA3132D2 vs. MAX9709 Messungen EMI/EMV – #360customs
I made some extensive tests with the TPA3128 (Which is a 3118 in a different package) in the past. It can clearly be seen, that a ferrite+C filter is enough even with 150cm between amp and speaker.
TPA3132D2 vs. MAX9709 Messungen EMI/EMV – #360customs
I have a friend who is a very skilled radio-amateur. Perhaps you are also a radio-amateur. A TPA3116 amplifier supplied by 24V and operating without a filter, directly connected to speaker wires, is two low-impedance, power, 24V square-wave signals at 300-500KHz, on long cables acting as antennas.
For radio-amateurs, this is worse than any movie made by Stephen King.
Krohn's suggestion is, as usual, very good.
For radio-amateurs, this is worse than any movie made by Stephen King.
Krohn's suggestion is, as usual, very good.
As said, it depends on the connection wires length. I'd tested with 150cm on the link above and also verified with an FM receiver.
Using long wires without filters does exactly what you describe. Have you checked the plots?
Using long wires without filters does exactly what you describe. Have you checked the plots?
It's my personal project blog/page for everything I find worth sharing with others. Some projects are available for sale, mostly bare PCBs for DIY but not on the commercial site. 🙂
So if i understand correctly output filters are not a must if you would build for example a small bluetooth speaker? Or is the bluetooth signal strong enough to distort the audio without an output filter+snubber circuit when it's close together?
If you use a low supply voltage, the TPA3116 may function well without an output filter. Due to the low supply voltage the currents are less. The TPA3116 should function down below 5V.
Your Bluetooth concern I do not understand. The Bluetooth radio signals are much higher in frequency than the TPA3116 carrier and the power level very small. You may have a valid concern the other way around such that the fast carrier switching may jam the Bluetooth reception with high harmonics.
- Status
- Not open for further replies.