Yes, PCB ground is provided via the screws that hold the thermal block to the PCB as I have done on my design, but not connected to chassis “dirty ground”. Chassis ground will pick up noise from environment. Usually you want a 10ohm ground loop breaker or NTC and a small film cap between chassis ground and PCB ground.
Any opinions on any soundwise advantage bypassing the op amps with a 0.1 uF polycap at the power pins/legs ? Im talking about ne5532, lm4562 and opa1612 , 1656.
Does it sound better / more musical / more flowing with bypass capacitor at +V and -V or will the sound only be brighter and more sibilant ?
Does it sound better / more musical / more flowing with bypass capacitor at +V and -V or will the sound only be brighter and more sibilant ?
The important feature for power supply bypass is low ESR and closeness to the chip pins. X7R caps located 2mm away are going to be hard to beat and a film cap on long legs won’t be as good. There are specialty film SMT caps although they are quite large. Film caps make more of a sonic diff when in sound signal path. For instance, the 10uF electrolytic caps between the op amps and the amp audio in. On my amp I have choice of 10uF Silmic or Wima 10uF MPT.
Hi guys, I have read some of your posts in this thread. I'd like to buy a cheap chinese TPA3255 to "feel it" for a while, but I am afraid of fakes. With 3116D2's I was lucky until now, after carefully seeing reviews and choosing I bought some (2 years ago) and no-one was fake.
Still the TPA3255 is more expensive, even the cheapest is around $28-30. I have seen several films on youtube complaining about fake XL7015, fake capacitors, fake preamps, but none about fake 3255. And I am wondering if really in 4-5 years since they reached the market no fake has been made?
I'd like to supply mine up to 53-55V directly from batteries (ultimate voltage in datasheet is 69, of course nobody will push it until there) and I really dont want to fail it.... No need to say that I will not go with the output into clipping when supply will be at 54V 🙂
If the main chip is genuine, it should have no problem with 54V. Abot the 12V auxiliary, i can replace it if is fake.
After some time, if the unit is performing how I expect and is not fake, I will eventually buy another one and make my own board.
Please tell me if someone have found fake 3255 on these very cheap boards.
Still the TPA3255 is more expensive, even the cheapest is around $28-30. I have seen several films on youtube complaining about fake XL7015, fake capacitors, fake preamps, but none about fake 3255. And I am wondering if really in 4-5 years since they reached the market no fake has been made?
I'd like to supply mine up to 53-55V directly from batteries (ultimate voltage in datasheet is 69, of course nobody will push it until there) and I really dont want to fail it.... No need to say that I will not go with the output into clipping when supply will be at 54V 🙂
If the main chip is genuine, it should have no problem with 54V. Abot the 12V auxiliary, i can replace it if is fake.
After some time, if the unit is performing how I expect and is not fake, I will eventually buy another one and make my own board.
Please tell me if someone have found fake 3255 on these very cheap boards.
There are no fakes. The usual Chinese practice of rewriting the product code to sell weaker versions as the more expensive, stronger one, is not possible with the 3255, because of it's different pin layout. TI knows it's mates.
If you want a well made, cheap 3255, I would recommend the BDM8 from Sure/ Wondom. This one for example:
https://de.aliexpress.com/item/1005...9687666.1005004883731361&gatewayAdapt=glo2deu
If you want a well made, cheap 3255, I would recommend the BDM8 from Sure/ Wondom. This one for example:
https://de.aliexpress.com/item/1005...9687666.1005004883731361&gatewayAdapt=glo2deu
The xl7015 can be a rebranded chip with only 50V spec . It’s hard to tell if it’s fake . There are some reported examples that blows up with 48 V and the reason was the xl7015.
The Fosi v3 mono seems to blow up in some cases , using 48V.
I have to ask - The tpa3255 has its lowest distortion at 35 V , so why use 54 V ?
Using 48V and above you need at least 63V spec capacitors and a cooling of the chip against the box shassie, or maybe a use of a fan .
If you are using 48 Volt a 50 Volt capacitor is perfectly OK. There is no serious reason to get 63 Volt caps. This is an often mentioned misconception of people that never learned anything about electronics, but are in the laypeople's mindset of "more is always better". If a 50 V cap would only be good for 47V it would be named as such.
If you have a fridgerator 6 feet tall it will go through a door 6' 1" without any problem. You do not need to enlarge the door to 7 feet, just because " it is better".
If you don't need the full headroom of the 3255, you are free to run it on 35 volt or less, if you want the full dynamics, you should go for 48 Volt or even closer to the limit. Distortion is not audible different. You will not be able to tell 0.010% from 0.016% distortion. Things get ugly in the full percentage region, depending on harmonic order.
If you have a fridgerator 6 feet tall it will go through a door 6' 1" without any problem. You do not need to enlarge the door to 7 feet, just because " it is better".
If you don't need the full headroom of the 3255, you are free to run it on 35 volt or less, if you want the full dynamics, you should go for 48 Volt or even closer to the limit. Distortion is not audible different. You will not be able to tell 0.010% from 0.016% distortion. Things get ugly in the full percentage region, depending on harmonic order.
My house is completely offgrid and have 2 lines of supply (actually 3, one being 19.6V, but is not relevant to our point here).
One line is 230V (from a low frequency inverter with a huge idle current, thats the main problem for the moment and maybe for another 2-3 years) and one at 48V directly from a bank of 16 cells of LiFePO4. The 48V is suppling all my house lights, but also is running around all walls and supply all the electronics in the house. Through mostly 4 converters, the 48V line is made to 19.6V to achieve this.
The 230V is used sparsely, especially in winter, I'm starting the inverter mostly for kitchen appliances, power tools and so on. and stopping it after that.
The 48V is actually around 53V most of the time. LiFePO4 are very tight with their voltages, most of their juice is between 3.25V and 3,35V per cell. My BMS (made by myself) is doing egalization at 3.400V per cell and the charging stops at 3.410V per cell, thus around 54.55V full battery.
Of course I could do eventually a "small" linear LDO circuit, at around 10A average and 20-30A peak, to drop the excess voltage. (saying "small" because probably you know that dissipating an eventual 2-3V drop at 5-10A needs a quite big (so, not very small at least) heatsink and maybe a fan).
Or a smps eventually, but that really needs a great effort for the extremely low drop-out needed most of the times.. and so, don't worth the effort (prototypes, testing, and so on..)
I really prefer to mod a bit the chinese circuit, instead of doing that. at least for this year (2025). House is not yet ready and i still need to put most of my time next year and maybe another one in the interior (and a little to the exterior) of the building process. So, unfortunately not really too much time for electronics.
One line is 230V (from a low frequency inverter with a huge idle current, thats the main problem for the moment and maybe for another 2-3 years) and one at 48V directly from a bank of 16 cells of LiFePO4. The 48V is suppling all my house lights, but also is running around all walls and supply all the electronics in the house. Through mostly 4 converters, the 48V line is made to 19.6V to achieve this.
The 230V is used sparsely, especially in winter, I'm starting the inverter mostly for kitchen appliances, power tools and so on. and stopping it after that.
The 48V is actually around 53V most of the time. LiFePO4 are very tight with their voltages, most of their juice is between 3.25V and 3,35V per cell. My BMS (made by myself) is doing egalization at 3.400V per cell and the charging stops at 3.410V per cell, thus around 54.55V full battery.
Of course I could do eventually a "small" linear LDO circuit, at around 10A average and 20-30A peak, to drop the excess voltage. (saying "small" because probably you know that dissipating an eventual 2-3V drop at 5-10A needs a quite big (so, not very small at least) heatsink and maybe a fan).
Or a smps eventually, but that really needs a great effort for the extremely low drop-out needed most of the times.. and so, don't worth the effort (prototypes, testing, and so on..)
I really prefer to mod a bit the chinese circuit, instead of doing that. at least for this year (2025). House is not yet ready and i still need to put most of my time next year and maybe another one in the interior (and a little to the exterior) of the building process. So, unfortunately not really too much time for electronics.
53vdc is perfect as the max voltage for TPA3255. That’s what I spec my custom SMPS for my amps. Don’t exceed 57v though or it will blow the chip.
For those needing to convert from 220/110vac - The SMPS 650SO PSU for the TPA3255 is available from MicroAudio.
For those needing to convert from 220/110vac - The SMPS 650SO PSU for the TPA3255 is available from MicroAudio.
Very nice amplifier indeed 🙂. And 63V capacitors to. Cooling of the tpa3255 chip against the amplifier chassie is the right way to do it.
No fake regulators.
—————
However, I wonder about the reability with any random Chinese made board and using 48V or even more. Will the amplifier work after 5 years ?
Its not only the tpa 3255 chip that must stand the voltage , its all the other components on the board that has this voltage at its input ( xl7015 regulators and such ).
Topping pa5 v.1 and Fosi audio v3 mono is two examples that dont do so well with 48V.
No fake regulators.
—————
However, I wonder about the reability with any random Chinese made board and using 48V or even more. Will the amplifier work after 5 years ?
Its not only the tpa 3255 chip that must stand the voltage , its all the other components on the board that has this voltage at its input ( xl7015 regulators and such ).
Topping pa5 v.1 and Fosi audio v3 mono is two examples that dont do so well with 48V.
Last edited:
I think my DCDC buck converter that makes 15v from the 53v was the part that blew up when I exposed it to higher voltages. This happened when I had a “simulated sine wave” UPS kick in on my main power strip rack. It uses a 3 step stair case function to simulate a sine wave at 60Hz from batteries. The peak voltages on that were then amplified by the SMPS (which was not a PFC type) and net result was higher voltage blew up some parts. So be careful if you are using an inverter to power your amps. Make sure it is true sine wave.
Check cap ratings for sure. Most of the time it’s the regulators that can’t handle it if not the caps.
Check cap ratings for sure. Most of the time it’s the regulators that can’t handle it if not the caps.
Yes, I'll be careful. Thanks.
Also I just read some phrase on another 3255-related topic (around year 2021) who put me on guard over an unexpected (by me), but understandable (from the manufacturer POW) behaviour.
"Make sure the voltage is below 48V. Above ca 53V the chip itself shuts down and since many transformer psu deliver much higher voltage than spec'd on low loads that can often be the culprit."
Is that the case?
Also I just read some phrase on another 3255-related topic (around year 2021) who put me on guard over an unexpected (by me), but understandable (from the manufacturer POW) behaviour.
"Make sure the voltage is below 48V. Above ca 53V the chip itself shuts down and since many transformer psu deliver much higher voltage than spec'd on low loads that can often be the culprit."
Is that the case?
You could look into the data sheet to get reliable information. Anyway, 53.5 Volt is the maximum working voltage for the TPA3255. If you insist to run it at 53.6 Volt, it will shut down. In this state theoreticaly it can stand 69 Volt, but will not work.
And, no, there is no more margain you can "overdrive" it.
This doesn't mean that other components on a board "made somewhere in China" will not blow at more than 48 Volt or even less, if they saved half a Won on even cheaper parts.
https://www.ti.com/lit/ds/symlink/t...A%2F%2Fwww.ti.com%2Fproduct%2Fde-de%2FTPA3255
And, no, there is no more margain you can "overdrive" it.
This doesn't mean that other components on a board "made somewhere in China" will not blow at more than 48 Volt or even less, if they saved half a Won on even cheaper parts.
https://www.ti.com/lit/ds/symlink/t...A%2F%2Fwww.ti.com%2Fproduct%2Fde-de%2FTPA3255
I have seen the datasheet first time I heard of this amp (always doing that when buying some new stuff), but overlooked about the shutdown when pass 53V barrier. Thanks
In the meantime am looked through the pdf again and did not seen any reference about a shutdown function on overvoltage (yes, it seems could be useful IMO), but instead I saw that the chip have actually as a workable parameter 56.5V at 6 ohms load. Page 6, row 2 in the table, Recommended Operating Conditions,.
So far so good...
So far so good...
One other aspect is the loudspeaker that will be used with the amplifier.
A single power supply with 36V 10A for two channels ( 360W ) will probably give more power than a 48V 5A supply (240W ) , if using 3 or 4 ohm stereo speakers.
I would go for 36V 10A if I had two 4 ohms speakers to drive.
Last edited:
I'd like to build XRK's 3255pffb if/when a fully populated unit becomes available.
For the moment I don't have other than 2 x 4 ohms speakers, but i do not plan to go with the amplifier nowhere near clipping for a while. Sincerely the only points I was worried were about fake 3255 (my initial question) and the 53-54V barrier who someone indicated after. In the meantime I continue to read at this topic and most probably I will order a board in the last day of 24...