Unfortunately I didn't listen the amp too much before mods. Currently (for me) the sound quality totally acceptable for me (with Allo Boss V1.2 DAC, chinese relay attenuator (modded to 10 kohm input impedance and Canton CT1000 v3 speakers and cheap DIY cables). I replaced the electrolitic capacitors (to Panasonic FC, Os-Con, NIC Components and Silmic) and opamps for reliable working condition from trusted source not ebay and aliexpress. (banzaimusic and some hungarian webshop)
Some like OPA627/637 could change in character from OK to excellent and after hours to again not that excellent, then being very good the next day again. It really needs some more time especially the board itself.
You could replace the 6 input/coupling caps for fine Wima MKS2 10 µF in 5 mm pitch too. Better than any electrolytic cap.
https://www.reichelt.de/mks2-pet-kondensator-10-f-5-50-vdc-rm-5-mks2-50-10--p172683.html
Last edited:
Jean-Paul, I wonder If its as easy lowering the gain on this board as it is easy to lower the gain on Hypex ncore 125 ? In Hypex, all one has to do is to desolder 2 smd resistors , lowering the gain about 13dB leaving the buffer at unity gain. The soundquality with this mod made the sound slightly better.
Can one do the same with this board ? I guess one half of the ne5532 is amplifying the signal maybe 15 dB, ( unessessary ) and the other half is doing conversion to balanced signal to feed the 3255 chip ?
Picture show the gain mod on my Hypex amplifier - those two smd resistors are simply desoldered to lower the gain about 13dB. Makes it perfect to drive with my Rega dac.
Could be but I never had any TPA32x device, only a few TPA3116/8 where adjusting gain was done in a breeze. Since I only play at 1W maximum I am more into Merus chips based stuff.
BTW if you think the Wima are too expensive and you can live with good quality 4.7 µF 50V 5 mm film caps (usually electrolytic caps are oversized but calculate/check!) I can send these 6 for shipping costs within Europe. Just a stamp really.
BTW if you think the Wima are too expensive and you can live with good quality 4.7 µF 50V 5 mm film caps (usually electrolytic caps are oversized but calculate/check!) I can send these 6 for shipping costs within Europe. Just a stamp really.
Last edited:
Thanks Jean-paul, but the WIMA caps are already bought at Audiophonics and are on their way to snowy Sweden
About the output coils on this board , according to Audiogue they are really good . They are paralelled 22uH per channel= 11uH with very low resistance. They are better than the original coils inside the Aiyima a07 , and about as good as the expensive Wurth.
”The dc resistance of the dual core paralleled to 11uH is only 11 mOhm and they seem to be flat copper wire type”
Its the middle picture that has this board coils.
”The dc resistance of the dual core paralleled to 11uH is only 11 mOhm and they seem to be flat copper wire type”
Its the middle picture that has this board coils.
Last edited:
This tpa board has some surprises. First, we have the power capacitors, some of them coupled in series making them have higher resistance than paralelled. Then we have the output coils that are paralelled making them having very low resistance and apparently low distortion with saturation at 10A.
Those 4 output caps has to be replaced with 4 WIMA 1uF 63V, lowering the distortion above 2 KHz ten times.
Those 4 output caps has to be replaced with 4 WIMA 1uF 63V, lowering the distortion above 2 KHz ten times.
Last edited:
Usual practice with chinese boards. X7R already being too costly.
BTW it seems that gain is always 21.5 dB with TPA3255.
BTW it seems that gain is always 21.5 dB with TPA3255.
Last edited:
Yes, but only If the IC before the 3255 is a unity gain buffer . Those Chinese boards often has some gain in the preamp, for unknown reasons . I can take my modified Aiyima a04 tpa3251 as an example. Almost as they expect that the source will give less than 1V out…
After 1 day burn in the sound has clearly changed for the better .
Continuing modification - bypassing the two Elna 10uF caps at the input. My rega dac R RCA:s output is free of DC , being digitally driven by my WiiM ultra, so this can be done.
The sound went clearly better, so now I must replace the 4 Elna:s before the tpa3255 chip with some WIMA capacitors. This will be easily done cutting the elna:s and solder the new wima:s below the board.
The sound is now, in some areas, better than my Hypex ncore. Im talking about bigger soundstage and more open above 6 KHz . Maybe Im fooled by the outputcap distortion ( they are still original )
Continuing modification - bypassing the two Elna 10uF caps at the input. My rega dac R RCA:s output is free of DC , being digitally driven by my WiiM ultra, so this can be done.
The sound went clearly better, so now I must replace the 4 Elna:s before the tpa3255 chip with some WIMA capacitors. This will be easily done cutting the elna:s and solder the new wima:s below the board.
The sound is now, in some areas, better than my Hypex ncore. Im talking about bigger soundstage and more open above 6 KHz . Maybe Im fooled by the outputcap distortion ( they are still original )
Last edited:
Don’t the Wima’s fit on the right side of the PCB? Question of a few mm I think. In such cases it may help to align them and glue them together with tiny dots of mounting putty. Then pretin the steel lead wires, clean flux and bend the lead wires in the pattern. Then cut and solder (in that sequence).
Yes the Rega is DC free but maybe not at power on/off. Or a new DAC has power on/off phenomena. Maybe a source has an opamp that one day dreams of being a battery. Maybe the SMPS of a source one day decides to do what eventually most switchers do.
Fail safe is a well working principle.
Yes the Rega is DC free but maybe not at power on/off. Or a new DAC has power on/off phenomena. Maybe a source has an opamp that one day dreams of being a battery. Maybe the SMPS of a source one day decides to do what eventually most switchers do.
Fail safe is a well working principle.
Last edited:
I've also thought about replacing the capacitors on the output, but the problem is that a good quality polypropylene WIMA capacitor has a leg spacing of 22.5 mm, and I'm afraid that longer legs (because I could only install the legs bent and longer on the bottom side of the panel) would cause some problems (due to the high frequency), and I don't have an oscilloscope to check if everything is ok.
Only use similar sized compact parts that fit and definitely without lead wire extensions.
Otherwise the medicine likely is worse than the disease.
Otherwise the medicine likely is worse than the disease.
Last edited:
You can use 4 WIMA MKS2 1uF 63V polyester capacitors and solder it under the board , just cutting the replaced caps , price below 1 dollar each. Those capacitors was what Audiogue used, and the ones I have bought on audiophonics .
The gods of the semis will teach it one day 🙂
BTW it seems some of the electrolytic caps are relabelled ones. Or at least undesired ones. I might be wrong but I have a hunch.
BTW it seems some of the electrolytic caps are relabelled ones. Or at least undesired ones. I might be wrong but I have a hunch.
Last edited:
Jean-Paul, you’re right . But it might be a risk if it’s hard to unsolder the replaced component from below , possible destructing the board if the soldering used on the board is only melting at very high temperatures .
I have no clue if that’s the case here - but my modified Aiyima a04 was a little tricky .
One can try unsolder the original capacitor with a careful hand using desoldering braid and not using very high temperatures. If it goes just fine - do it and mount the wima capacitor through the hole on the upper side of the board, soldering below the board .
I have no clue if that’s the case here - but my modified Aiyima a04 was a little tricky .
One can try unsolder the original capacitor with a careful hand using desoldering braid and not using very high temperatures. If it goes just fine - do it and mount the wima capacitor through the hole on the upper side of the board, soldering below the board .
Last edited:
IF they are hard to remove (I doubt that) then cut the lead wires from above and destroy the cap, just leave 2 mm of both lead wires sticking out. Heat the pad and then either desolder and pull the lead wires out with tweezers or push them out with a stainless steel needle while heating up. The solder does not stick to the needle. Clean the PCB with isopropyl alcohol and a cotton swab.
The OPA1612 modules have debris and should also be cleaned with isopropyl alcohol and a cotton swab. ESD safe.
The OPA1612 modules have debris and should also be cleaned with isopropyl alcohol and a cotton swab. ESD safe.
Last edited:
Great work documenting these mods. Enjoying watching your progress. One question for you... Do you think this board could be modded for PFFB?
Thanks again.
Thanks again.
I would think so 🙂
My RCA sockets arrived today .
I took Jean-Pauls advice to use coupling caps at the input. I had two 3.3 uF polypropylene capacitors and used a common ground for both channels The sound is much better than with the Elna caps, but slightly worse than without caps, but its a very small difference - I can live with this. Soon, my 4 Wima 4,7 uF will arrive and I will desolder the 4 coupling caps ( same elna as input ) for the tpa3255 , expecting even better sound .
I have painted the front of the box with black metallic colour , so I listen without lid and front right now.
Strange thing is, when i put back the lid on the amplifier the sound turns a little less bright and more clean ,- the sound of less RFi.
My RCA sockets arrived today .
I took Jean-Pauls advice to use coupling caps at the input. I had two 3.3 uF polypropylene capacitors and used a common ground for both channels The sound is much better than with the Elna caps, but slightly worse than without caps, but its a very small difference - I can live with this. Soon, my 4 Wima 4,7 uF will arrive and I will desolder the 4 coupling caps ( same elna as input ) for the tpa3255 , expecting even better sound .
I have painted the front of the box with black metallic colour , so I listen without lid and front right now.
Strange thing is, when i put back the lid on the amplifier the sound turns a little less bright and more clean ,- the sound of less RFi.