TPA3255 - all about DIY, Discussion, Design etc

Hello,


Did someone identified suitable heatsinks for DIY TPA3251 and TPA3255 boards ? The reference used on the TI EVM boards does not seems to be a common off the shelf model.


And sometimes those heatsinks are a significant par of the final price...


Help on this side would be welcomed,


Best regards,


JMF

Hi

for more power use the original heatsink from the TI EVM boards
ATS-TI1OP-519-C1-R3 Advanced Thermal Solutions Inc. | Lufter, Warmemanagement | DigiKey


but at the chinese boards its maybe not enough space to mount it....

chris
 
Hi

for more power use the original heatsink from the TI EVM boards
ATS-TI1OP-519-C1-R3 Advanced Thermal Solutions Inc. | Lufter, Warmemanagement | DigiKey


but at the chinese boards its maybe not enough space to mount it....

chris


Thanks Chermann,


Your link provides a good starting point for me. I will have a design similar to the EVM boards.



There are SMD caps around the TAS3251 chit. I don't know if a flat bottopm heatsink can make the job, and not touch the surrounding components when pressed on the amplifier chip.


The TI heatsink seems to have some "central pad" to rest on the chip and have additional margins to avoid the SMD components around...


JMF
 
Moving slowwwwwllllllyyyyyyyyy... but forward :)
 

Attachments

  • JMFneatAmp2019.JPG
    JMFneatAmp2019.JPG
    159.1 KB · Views: 495
A lot of learning and trial and errors. And for sure I would need some advices of more experienced people:
- are there some ground loops ?
- GND copper plane on the top and bottom, or only on the bottom (currenltly there is one on both sides) ?
- other design flaws or checks to perform...


JMF
 
No, I understand. It was more about which caps you used. Which changes provided benefits in your experiments, and which ones had no effect :)


JMF


Ahhhh. ok....now i feel better ;)

at the TPA3255 chinese YJ board i was most happy with the 2x green UHE 10µF muse caps directly at the input (before the opamp) + the 4 x ukz 10µ (unluckly i just had the 50V version) at the TPAchip inputs.

as i wrote..there was a discussion at The TI EVM board about using electrolytic cpas instead of SMD...because the sound of the ceramic/tantal are not nice...

output filter i am happy with the 8R configuration= 10µH Würth 74439370100, TDK Epcos caps B32652A3684J or 0,47µF...
opamp 1602
Rail caps = by Doctor = Chemicon EGPD630ELL242MM40H
or similar (high ripple at 100khz low ESR/ESL)

Important is to seperate the voltage regulation (low noise) for the TPA chip and the opamps . Faux French did a +/- double voltage for the opamps to provide a better THD....at the TI EVM its not realiszed like this. i personally havn´t tried this.

As i do not have THD measrument equipment i can just say that my result was listening test.

chris
 
Ahhhh. ok....now i feel better ;)

at the TPA3255 chinese YJ board i was most happy with the 2x green UHE 10µF muse caps directly at the input (before the opamp) + the 4 x ukz 10µ (unluckly i just had the 50V version) at the TPAchip inputs.

as i wrote..there was a discussion at The TI EVM board about using electrolytic cpas instead of SMD...because the sound of the ceramic/tantal are not nice...

output filter i am happy with the 8R configuration= 10µH Würth 74439370100, TDK Epcos caps B32652A3684J or 0,47µF...
opamp 1602
Rail caps = by Doctor = Chemicon EGPD630ELL242MM40H
or similar (high ripple at 100khz low ESR/ESL)

Important is to seperate the voltage regulation (low noise) for the TPA chip and the opamps . Faux French did a +/- double voltage for the opamps to provide a better THD....at the TI EVM its not realiszed like this. i personally havn´t tried this.

As i do not have THD measrument equipment i can just say that my result was listening test.

chris


Thanks for sharing Chris. I have to have look at my circuit to se how it could be applied
 
Did anybody try to bypass that opamp and feed signal directly to TPA3255 input pins via 10µF caps?
Whats the purpose of that opamp? buffer?


hi zek


if i remember right. if you use a single ended signal -rca connector --> the opamps are therefore to "build" a inverted signal for the TPAchips input Input A....input D.
for differential input you can go directly to the tpachip over the 10µF.



there are some ideas to use bipolar caps here instead of the polarized eleytrolytics...but now idea what happend if the TPAchip get some DC voltage there....


chris
 
hi zek
if i remember right. if you use a single ended signal -rca connector --> the opamps are therefore to "build" a inverted signal for the TPAchips input Input A....input D.
for differential input you can go directly to the tpachip over the 10µF.

there are some ideas to use bipolar caps here instead of the polarized eleytrolytics...but now idea what happend if the TPAchip get some DC voltage there....

chris

Thanks Chris for an explanation.