Small Prototype Engineering
6,5" active subwoofer. 2022-12-12
Adding a small active subwoofer to the previous project to handle the lower frequencies. Using the same same digital amp but configured as bridged mode to drive the subwoofer with 100Wrms.
Did not have all the correct components in stock so I used what I had. The audio result is at the moment not as good as expected, would almost state it's bad. It feels lik the amp is to weak to controll the subwoofer. the base sound just weird and blurry. It does'nt really produce any high soundpressure either, but that could be caused of it's small size. I have never built a ported enclosure before so it might be the case that i have tuned the box badly.
Tried to play around with the DSP and if i narrow the frequency span i can filter out the blurry sound.
I will connect the whole system later on and see if i can tweak it to sound good enough later on. Othervice i'll try to experiment with the enclosure and the port length.
Digital 2Ch amplifier. 2022-10-11
Have done a few amplifier projects but never did a class D amplifier which would be a fun thing to design. I have a room in my garden where it would be nice to play music, so this amp did fit this application. To really optimize the sound i will use a MiniDSP 2x4. The speaker are also home made and work excellent but will need a small subwoofer to sound as they should
The amplifier features are
-Class D
-DCDC powersupply
-2Ch
-Autostart at detected incoming signal
-433MHz remote
-Digital potentiometers
-Temp controlled
DCDC power supply.
As a part of the design i wanted to learn a bit more about high power buck design. So i designed a 60W 20V switched power supply.
I did use TI's webbench to design the power supply and reused their layout example. The design is based on LM2696. It calculates allt the component values and it can optimze for BOM, price or efficincy.
Measurements below show the output loaded with 100mA, 500mA and 3A which is maximum aoutput current of the buck controller. Ch1 is connected to the switch node and Ch2 to the output. 24V input voltage!
Class D amp
TPA3116 is a small class D amp that will work with amost no heat sink. It has extremaly high efficiency and small footprint. Its pretty straightforward to use it and there is alot of information on how to concect it as with all switched electronics care must be taken when drawingt the layout. One thing to have in mind is that the inputs of the TPA3116 are balanced (differential) but most home aduio equipment are unbalanced (common mode).
This was solved by a specific TI audio circuit OPA1632. Input gain can also be adjusted here to fit the power stage!
The layout was done very carefully and i used the layout of an evaluation board for the TP3116 as reference. On the bottom there is a ground plane and a power plane connecting power to the IC and big capacitors.
Output is measured with input signal of 1Vpp/1KHz. No switching ripple can really be found at the outputs.
Notice that the output swing is 3,64Vpp+ and 3,7Vpp- for the left channels and 3,7Vpp for both +- on the other channel. The difference of 60mV is quite small and cannot be picked up by the ear.
Also a conducted FFT was meausred on the L channel. I'm not quite sure how well the scope actually performs this it was mostly done of interrest.
Temperature controll, autostart, digital potentimeters and remote controll are all managed by PIC16F18346. It measures the temperature locally att the buck controller IC and the audio amp IC. if any of the temperature limits are exceeded the MCU will shut down the amplifier by controlling mute and shutdown pins. There is also a small fan that can cool the heatsink for the audio amp i temperature is increasing. For the remote controll the HC11 433Mhz module is used and will controll volume, mute and power ON.
For volume controll im using 2pcs 64 step digital potentimeters MCP4011 wich are both controlled by the mcu thru a simple interface with. Everytime the amplifer turns on the volume is set to a default level.
Finally the autostart is made of a single OP that connects to a ADC input of the MCU. If ther is a input signal it's detected and enables the whole board in proper order.
Below shows the protocoll for one volume down step. And also the command for defalt star volume, first it lowers the sound to minmum by sending 70 times the lower command. Then it will raise the volume by 15 steps or so.
Setting and calibrating the temperature controll and limits
Finally the speakers and amplifier casing was built. During development i tried a couple of different speakers and the best result was with my hifi speakers KEF-LS50, they sounded great and hade excellent base response and no dsp or external sub was needed. But the speakers i intended was the 2 way kit from Morel Ultimo2. They had hard time with base response even when i tried the DSP. There will be an added subwoofer in the future.
Nano33IoT released. 2022-07-30
Nano33IoT has now been ofificially published and it has been recieved well. Now some more updates and improvements will be implemented to be produced in a bigger batch
Nano33IoT_adapterboard. 2022-06-09
My 2nd design for Episcope.
This adapterboard is intended to house both a Raspberry Pi 4 and a Arduino nano IOT. It will act as a inteface to the outside world with high protection level. The board is desigend to fit into a industrial casing that will be sealed.
One thing i used in the design was a small DCDC power module from Traco which has almost the footprint as the standard TO220 linear regulators. They are expensive but works very well with low heat dissipation at even medium to high load.
Heat tests were performed at 40degC ambient. All power nodes were loaded externally with electronic loads.
There was a requirement to have some kind of external and independent watchdog that controlled the reset pin of the Arduino. It recieved a signal each minute from the Arduino and if this was not sent the WD issued a reset signal to the Arduino.
I tried to find some reset circuit but they are all made for shorter timeout delays. I therefore used a Pic12LF1572 and created a WD. To test the function i used my arbitary signal generator and created an output signal that sends out 14 pulses. One/ minute and each 1s long. After 14 pulses it will than wait some time so that the watchdog can send out 3 reset signals. The signal was verified both with oscilloscope and with a longer test using a multimeter with logging function.
Test signal created with Siglent waveform generator