Testing a TPA3255 PCB Audio Amp with the QuantAsylum QA403

[@8Junio76]

Good video. Cheers.

[@boatanchoraudio6453]

Glad you enjoyed it

[@imiafamily]

Единственный вменяемый инженер во всём мировом интернете! И цифры, и данные - всё на месте, как полагается! Респект и уважуха, как говорится! Впечатлён!

[@Douglas_Blake_579]

8:50 -- 20 volts? I've tested a few of these, myself. 20 volts is too low for reliable testing. If that is a true TPA3255 chip (not a TPA3251) 20volts is very low in it's stable operating range. It will go into under-voltage protection at 16 volts. It's output will be closer to the noise level than with higher voltages and because it is barely into it's stable range the THD will be somewhat higher. It is going to perform far better in the range of 36 to 48 volts, with a 10 amp supply. At 48 volts with a 10 amp supply you should be able to pull almost 220 watts/ch off that board with 4 ohm loads at less than a percent distortion. (110w on 8 ohms) The TPA3251 maxes out at 38 volts and seems to like 32 volts the best. Here you will get about 100w/ch on 4 ohms again with less than a percent distortion. (50w on 8 ohms) Unlike many TPA325x boards, that one appears to be laid out properly. The output coils are kept well away from the heat sink which definitely improves the sound quality at the outputs. Many designers will put the coils right up against the heat sink, in an attempt to keep them as close to the chip as possible. This is a mistake, even with toroidal coils as the minor magnetic fields on the outside of the coil's windings will interact with the mass of metal causing eddy currents, inductive heating and some non-linearities in the PWM pulse chain... effectively de-tuning the output filters. It will be interesting to see what you get from your commercially produced TPA3255 amplifier.

[@boatanchoraudio6453]

Thank for your input, Douglas. I certainly agree that 20 volts is too low for reliable testing. However, we've also tested with a 32 volt brick and the results were not that much different; I'll have to post the results on my web site, but they're pretty much the same. As for the test results, I have to say I'm surprised at how well it did for a $30.00 "eBay Special". We also have a popular commercial Class D amp with the TPA3255 we tested at 32 volts and the results, for the most part, we not much better than the PCB amp. I bought the unit as advertised with a TPA3255, but it's anybody's guess who made the chip. As for the PCB itself, it seems fairly well laid out except that they have four electrolytic caps parked right next to the heat sink; I wonder how long they'll last there. Regards, Marty

[@Douglas_Blake_579]

@@boatanchoraudio6453 True enough, once the chip is in it's stable range (roughly 24 to 48 volts) the frequency response and distortion tests will be pretty close. What changes most is the power at clipping and signal to noise ratio. Most of those online specials are actually commercially made boards that either ended up as excess stock or have minor manufacturing defects. They "dump" them into the DIY market, rather than taking a total loss, knowing that most of us have the wherewithal to fix at least minor problems before putting them into any kind of permanent use. The list of problems goes on for a while, but the main ones are lack of vertical clearance between the heat sink and the parts under it, the 12v buck converter being somewhat unstable and capacitor voltage ratings being too low. All of which are fairly easy to fix. I've seen a lot of fake NE5532s on those boards (Which I replace with RC4558s or TL072s) but I've not seen a fake TPA3255 chip. I have, however, seen TPA3251s substituted and those boards will then go into overvoltage protection at about 38 volts. To be honest... the caps against the heat sink is far less of a problem than putting the coils there. If that's a concern we can always replace them with 105 degree equivalents. Or you can use my favourite cheap hack and saw one fin off the heatsink to create some clearance. Power supplies are a problem. The brick supplies you get with the finished ones most often will allow 100% overhead for brief moments before going into an over-current panic, which is a good thing. But the DC cords are too thin and too long to provide good transient response. Once the on-board reservoirs are drained the resistance of the DC cord slows the recharge a bit which leaves the amplifier unprepared for the next transient. I have mine connected to a 36 volt 10 amp supply using 12 inches of 16ga wire... works a treat.

[@dennissinger7880]

​@@Douglas_Blake_579i looking many time the wondom tpa3255 whith dsp . The Qualität looks verry good and one board for 45$

[@user-pp6jd6qj4x]

Response is terrible, do you know anything of design?