Hi experts 
Is it possible to sent us pic or schematics from your mods.
i am asking because pages before i try to get more information
https://www.diyaudio.com/forums/class-d/309813-wrong-tpa3255-30.html#post5498623
i really appriciate your help, thanks
chris
Is it possible to sent us pic or schematics from your mods.
i am asking because pages before i try to get more information
https://www.diyaudio.com/forums/class-d/309813-wrong-tpa3255-30.html#post5498623
i really appriciate your help, thanks
chris
Hi
my not fried board-opamsp changed to ad8599
status: left channel loud noise -R channel not soo loud and more lower frequency noise.
he r-channel opamp has still the "oldisch" ceramic caps 22pf and if i send a signal i get it but with noise behind.?
as doc wrote i try the 1n F at the noisy channel - left channel has the original caps smd 22pf and the changed 1nF -if i send a signal its very low level and the pink noise is still there - so practically not really a difference
ideas?
thanks, chris
hi
sorry but actually i have not enough time for audio...but i try to proceed with my
"AMP light"
i did the cabeling. i still keep the idea that i want to change the tpa boards so i kee the big connectors. to go ahead with my "favorite" board i need more time for comparing some ideas...no time now
until the 28th of September i have to !.....= introduction of my amp to friends
any other suggestion if experts look at the pics?
thanks
chris
my working board is setup like this:
TL02 opamp, changed input caps and the regulator setup, decoupling doctor´s magic caps
output filter 7µH epcos --> real 7,6µH + 0,680MF TDk caps --> real 0,698µF + CBTL real 0,93µF
this setup is for 4ohms speakers.
on the a very good and neutral speaker = DISCO M (4ohm, 5 ohm) =
the red one:
Die DISCO M - ein hochstwertiger Kompaktlautsprecher | felbi.at | Website von Thorsten Felbinger
i found my amp very hard whiz, fizzle sounding
i changed the CBTL to 0,33 each and its much better
anybody the same experience?
as expected the TPA3255 is damaged.
pic 1 shortcut as doc mentioned - hopefully correct done
massiv noise at left channel - right channel generates a switch off the amp while i am but a speaker in parallel to the 8ohms load.
pic 2 power consumption is starting with 14Watt and stop at 7,7Watt
...normal is about 31V - 4 Watt....so a new chip cost about 11 euros..and i do not know what is additionaly damaged.....complete board costs 44euro..
Attachments
Last edited:
Thank you very much for this very very good hint
Following your writings I just placed an external 12V power supply beside my TPA3251evm board and fed the 12V at the pin after removing the jumper J32. Very nice sound improvement. Soundstage, resolution and highs are profiting from this simple modification. I used a spare power supply I build for an other project : R-Core transformer + Reference STUDER900 power supply board with a cooling plate | eBay adjusted to 12V + 1000µF FM + 3xFKP1 0.1µF.
I was no completely convinced by this amp, but now I'm very satisfied by the great improvement.
Head to head in stock mode, the 3e sound is better than the EVM.
The EVM has the extra 2 x 4700uFD which 3e has left out. Opting
to pass the buck to the power supply.
The 3E has a better sounding "dynamic" NJM4580 over the dated
"soft" sounding NE5532 on the EVM. Both are under $1!
I too plan to bypass the 12V supplies on both the 3e and the EVM.
Simple on the EVM (J32 as mentionned) and almost as easy with the 3e (U6).
The 7812 can be removed replaced by a wide band discrete regulator.
Additionally, the better sounding SMPS500 and 600 from Connex
that I have been using have an additional +/-15VDC output which
can be regulated to replace the 12VDC on the 3255.
The EVM has the extra 2 x 4700uFD which 3e has left out. Opting
to pass the buck to the power supply.
The 3E has a better sounding "dynamic" NJM4580 over the dated
"soft" sounding NE5532 on the EVM. Both are under $1!
I too plan to bypass the 12V supplies on both the 3e and the EVM.
Simple on the EVM (J32 as mentionned) and almost as easy with the 3e (U6).
The 7812 can be removed replaced by a wide band discrete regulator.
Additionally, the better sounding SMPS500 and 600 from Connex
that I have been using have an additional +/-15VDC output which
can be regulated to replace the 12VDC on the 3255.
Guys, I can't get rid of noise. I can hear it from distance of 1m played on 90dB/2.83V speakers. I did my own design based on EVM. Power Supply is Meanwell HRS-150-36. Front end opamps are LME49720 which have their own LM2940 voltage regulator. TPA3251 got another LM2940. They are all supplied by 15V from LM5010 (just like EVM). Input coupling capacitors for TPA are Wima 4.7uF/50V. Without opamps, input shorted to ground trough capacitors the noise is almost the same, I would say very little less. It is capable of delivering more than 50W per chanel to 8ohm load without problems. Music plays, but its not as clean as my old LM3886 amp. Noise is much lower if I use PFFB, but still high (heard from 10-15cm). I currently didn't populate the board with PFFB components. 2 prototypes were made, both behave the same.
Any suggestion is very appreciated.
Schematic Schematic_TPA32xx-v2_20181001163724.pdf - Google Drive
Any suggestion is very appreciated.
Schematic Schematic_TPA32xx-v2_20181001163724.pdf - Google Drive
I will try this as soon as possible. the best match I have here is polymer panasonic 16sepf1000 12mOhm or electrolytic panasonic FM 1000 50V 16mOhm. Although electrolytic is giant compared to polymer, I have tendency to prefer those. Or may be I place 2 sepf in parallel = 6mOhm.
Difficult for me to imagine another improvement in clarity after external 12V PSU upgrade. So good sound already.
TPA3251 normally outputs 60uVrms noise. Which is about 95dB below 2.83Vrms. With 90dB/W at 1m distance you should expect -5dBA acoustic noise level - definitely far below the hearing threshold.
So I assume your noise level is some 20~30dB above the expectation - maybe there is some pcb-layout problem.
Well, thats quite hard to resolve. Just to check if onboard power supply is OK, I made one on the other PCB, with LM317, but it was the same. Tried capacitance multiplier, but without success.
Also AVDD blocking capacitor (suggesterd up to 1000uF) did not influence the noise.
If its a PCB layout problem, I must have made a HUGE mistake to get results that bad. I followed EVM layout for the output and power part.
Yes, BTL. Lost somewhere noise measurements from sound card, all I have is from scope. Sounds like regular white noise.
TPA3255 - all about DIY, Discussion, Design etc.
Now I realized how many threads there are for this chip.
Ah, so upgrading the 12V rail does make a difference!
I will be interested to hear from greinedo if adding a big cap to AVDD achieves even more improvement.
But I must say that's quite a serious power supply, and with its own transformer, will require extra space in your chosen enclosure.
Yes, that's an interesting approach, too.
In a similar vein, would anyone care to comment on the suitability of this product -
LT30451 + LT8608 Low Noise RF Audio DAC ADC Linear Power Module-in Replacement Parts & Accessories from Consumer Electronics on Aliexpress.com | Alibaba Group
It's a combination buck pre-regulator (LT8608) + LDO final regulator (LT3045-1). Only the LT3045-1 needs to be adjusted for voltage, because it controls the upstream LT8608 to provide optimal pre-regulation voltage value, via a feedback connection.
This seems an elegant solution to me, since it can be fed from the primary DC supply, and thus appears to be a more sophisticated version of what the TPA325x EVM already provides, conveniently avoiding the need for a completely separate 12V power supply.(?)
hi
i recheked my ordering
mybe the wrong SMD resistors?
0606 10k 0,1% tolerance
but 0,1Watt power disap. - is this eventuelly a problem if i solder ?
I will try something this weekend. I already identified that it will challenge my limited soldering skills. On TPA3251evm, tiny C41 is on top, under the heatsink, so difficult. Placing a cap on bottom side, between AVDD via and ground seems more possible.
But I realized in the meantime that a big cap there will stress the internal AVDD regulator during startup phase (inrush current). Is there any risk damaging something with time ? There is no indication in documentation except 1µF on AVDD and "is important to note that the linear regulators integrated in the device have only been designed to support the current requirements of the internal circuitry, and should not be used to power any additional external circuitry". My poor knowledge here cannot help me to evaluate if there is a risk. Thus every opinion is welcome.