In The Ever lasting Chain Of Experiments The Next Project, ultrasonic anti-fouling for boats.
Part 12 Observations and conclusions
Made a small change in the circuit diagram, added a 6k8 resistor in the input pin 15 of the PIC where it sees the voltage from the curent transformer. This so that when that point goes negative, the resistor limits the current into the PIC (body diode).
Made some changes to the software to make the OLED display look better and more understandable:
Also changed the power off feature, it is no longer the way Microchip say you should do it, see subroutine pwm_init, but now it works.
All in all this all works great, now using 2 x 6 turns primary on an Al 1200 small core, and 120 turns secondary for about 600 Vpp on the transducer at 20 kH.
This last sentence is very Important.
Observation: First I have experienced my first RF burns in years,
40 kHz, 80 kHz, and even 20 kHz makes great RF burns at 600 Vpp. So that is why, especially in a wet environment, you need to protect the wiring and connections to the transducer, say use some PVC tubing as in the siliconchip article.Even if you do not die from burning, the reflex can make you hit your head really bad in a small space like a boat. I am sure some here have experienced that....
;-)
Second, I do notice that contrary to what you would expect, the transducer reacts like a SERIES LC circuit (but then mechanical). This means you can happily drive it at 20 kHz with 600 Vpp, and it only draws little current, that goes fine to about 35 kH or 37 kHz, but then it starts to load the transformer resulting in ZERO volts at resonance at 40 kHz, then coming back at 42 kHz and above that drawing ever less current again. The current limit is your friend here, keeping things in one piece, and keeping power consumption low enough so you do not need a special battery or generator on board, but 12 Watt I'v seen.
I notice this test transformer I made is just fine, need to get some other diameters transformer wire but really.
At high power levels (if you MUST drive at exactly 40 kHz) it may be a good idea to put the MOSFETs on a heatsink. as those go in my case into a transparent waterproof box, with waterproof connectors and waterproof power switch.. so forget about air circulation... and in the tropics... so I did order some large heastinks.
All in all I get nice waveforms on the transducer, programmed the range of frequencies to just leave out the range from 37 to 43 kHz for these transducers. later I will try to run the current versus frequency for a heavy mechanically loaded transducer and see if I can put the 37 kHz to 43 kHz range in again. Programmability in this software is great, you can set _any_ parameter via the serial port at 4800 Bd.
Next project: Building yer own nuculear reactor from dived up plutanium from old sunk Russian sub//
Na, no idea yet.