Embedded Files
after browsing aliexpress searching for a decent 2.1 channel amplifier board, I finally settled on the ZK-AS21. this board is built around 2 (two) TPA3255 chip-amps, each capable of 300w x 2 stereo, or 600 x 1 mono, at approximately 50 volts DC.
the board is configured so 1 chip is for stereo L and R, and the other for the subwoofer.
the casing for the board/unit itself displays a power input value of up to 36 volts, and after some probing and reading datasheets, it became clear that the limiting factor was not only the heat it produced, but the 12v regulator having a maximum input of 38 volts.
that small IC on the left, mid center, near the capacitor is the 12v regulator in question, while the 5v regulator can be seen opposite, and is supplied by the 12v regulator, primarily because of limits secondary to excess power being dissipated as heat.
be aware, many of these cheap boards use inferior components. those two large green caps are/were fake nichicon, the potentiometers are scratchy, and we'd imagine the op amps behind each pot are cheaper clones.
If dave could have his way, he'd surely use wima capacitors for both input and output, bourne's audio potentiometers, and texas instruments op-amps -- though of course, you get what you pay for.
regardless, a quick order for a meanWell NSD05-48S12 DC-DC converter capable of 72v input/12v output and a bit of reverse engineering later, it is now being powered with 48v. it was important to keep the 12v and 5v sections for other circuitry, such as the op amps responsible for tone/volume control and the bluetooth IC (integrated circuit) that also controls switching between analog 3.5mm, bluetooth itself, and USB inputs.
for it to function correctly, we removed the 12v regulator and it's supporting components, which was the small 8-pin IC, two capacitors, and two inductors.
four wires were then soldered -- two to the GND pad, one to the V in, and then another to the V out of the old 12v regulator's pads.
then these were soldered directly to the MeanWell module.
pumping wattage into a subwoofer by feeding that small amplifier IC 48 volts will most definitely crank out some heat. as we document, we realize those red/black wires were added for powering the fan, yet as you can see it's receiving power from elsewhere.
the original heatsink is laughable, at best, when the full potential of these little power house chips is understood. from a product offering point of view, limiting input voltage to a maximum of 36v (minimum being 18v) simultaneously reduces heat, production costs, and increases power supply compatibility with chosen components.
it may have done well in prototype testing, but it became a question of how it would handle 48v at a high volume level
the quest was nearly " to make the ears bleed" --- which they literally did, but for other reasons entirely.
at this point, dave scavenged old parts and found a large solid copper heatsink from a PC we built long ago. miraculously, after some slight modifications, we made it fit and mounted a fan atop only to discover further insulation of components was needed. these were previously guarded mostly by the thermal pad seen on the original fan.
on to testing.... where the capacitors on the power rail were soon observed to swell and leak. they were, in fact, not rated for anything higher than 36v. (slight oversight!) off to Amazon to locate some 63v capacitors with a slightly larger capacity. they arrived within a few days and were promptly installed.
now we're cooking!
at least it was thought so.
this thing has so much low end grunt, it resets itself when a nasty bass note is played at half volume. we've yet to rectify that issue due in part to being mostly amazed with how much sound this little board can put out as it is, along with many other projects keeping us busy.
mind you, the subwoofer is just a 10" SKAR in a ported box -- it relentlessly devours music with intense bass, demanding all the electrons the amplifier can supply.
the difference in output between 36v and 48v probably cannot be fully realized at the moment, due in part to speakers being rated for lower power along with the power supply issues, but on paper it's significant.
perhaps two more capacitors in parallel, or maybe even a proper power supply utilizing a toroidal transformer and small bank of capacitors might be of use.
a bonus to not only power the amplifier, but also the mixer board as well.
ah well, we'll see..
in an attempt to increase audio quality and simultaneously reduce distortion, the operational amplifiers were replaced. the engineer who designed the board chose the popular NE5532. given this unit is manufactured en-mass among many other electronic items, it is almost guaranteed the components used to populate the board are of relatively low quality.
sourcing some genuine texas instruments NE5532 chips was of no hassle, and replacing them simply required plucking the old ones off with hot tweezers, and tacking the new ones in place.
the small 8-pin chips seen above each knob is an op-amp, with associated circuitry. left-most is essentially a variable low-pass filter with a volume control next to it, specifically for the subwoofer channel. the middle two are high pass and low pass filter, respectively, with the main volume control at the far right. the main volume control also contains a power switch, which is why it's slightly larger than the others.
dave also replaced the volume potentiometer from whatever it was to a Bourns PTR902-1015K-A103
there is no longer scratchiness when adjusting volume --
perhaps replacing the capacitors on the output lines to ones of higher voltage rating will eliminate the slight distortion that becomes apparent at moderately high volumes -- or, it's simply 'clipping,' as the TPA3255 datasheet states that the input should not exceed -/+ 4.5v.
so, becoming quite agitated with the amplifier's inability to deliver consistent sound [loudly], a proper power supply is being strongly considered. the 48v switching supply by MeanWell does not contain fairly large [thicc] capacitors on it's output, and is likely not designed for audio use. we'll begin with this 36/72 volt center tapped toroidal transformer dave found on eBay.
in this photo, red/black[115v] are the primaries, while green/blue are secondaries, at 32 volts. it has two primaries to retain compatibility with 240 volt countries. for us 120 volt american scrubs, simply connect both red, and both black, to parallel them.
we chose 36 volts due to the DC voltage after rectification, which turns out to be about 51 volts. then, subtracting voltage drops of the rectifier diodes, we get about 49.5 volts, and with an additional diode on the amplifier itself, it is further whittled down to 48.8
perfect.
so, being a center tapped transformer, there's 2 separate windings for the secondary, with which one end of each is at the 'center'. they may either be utilized together or separate -- together, in 'series, inner blue/green are connected and 64v is realized in the photo. the center connection itself may also be used, to form a "ground reference" that an amplified signal may travel along.
connecting both greens together and then both blues will effectively 'parallel' them, maintaining the 32 volts, but doubling the amperage capacity.
now, we need to smooh out the AC, and store some of that energy with a few capacitors that will hopefully keep the amplifier supplied with the instantaneous current it may demand.
a quick design contains a rectifier and smoothing capacitors that absorb the AC pulses... still need to add some decoupling capacitors, and potentially shrink the design/board.
that's the process so far -- i've yet to actually purchase the transformer -- however, the potential thrill of wall shaking bass encourages a handful of audio engineers of the fur kind to design the board and make it before the purchase of a transformer.
also, 7.7 cu ft ported enclosure with an 18" car audio woofer sounds exciting
Page updated
Google Sites
Report abuse