The Portable Dog Killer / Quiet Canine Competition

[RiccardoGiuliani]

Well... I must admit my fault in giving details.

Stimulated by others contenders with their beautiful projects and reports I'm here posting the KiCAD scheme of my device, with the advice that you must substitute:
1) the LM555 with a NE555 (not important: but any case KiCAD has not NE555 model);
2) the 78L05 with the MC33164P (this last one not present in KiCAD).

Thank you (I saw three votes for me until now: I couldn't image this result for an inflated classical device made for playing).

[blankfield]

do you have a scope? or maybe at least a DMM set to %? I think most of the power is lost because you're not feeding a symmetrical signal to the transformer. When I played around with the trimpots, having the output "idle" high made the transformer heat up considerably. 30W of audio is a lot IMHO. Does the transistor heat up?

Hi dr_p,

Writing 100% was my mental shortcut, of course I'm feeding square wave 0 - 4.2V (4.2Vpp) with max 50% duty cycle @ 20kHz. Power losses are dissipated on series resistance of Mosfet Rds_on + resistance of wires + internal resistance of Li-ion cells. Some energy returns to GND via mosfet's diode while mosfet is shutting off.  Rest of power saturates ferrite core and transfers to secondary side. Voltage and current on secondary side is depended to transformer ratio, with a large simplification of course.

What you mean writing 'symmetrical' are you using symmetrical supply like +/-12V or you mean duty cycle?

blankfield

[dr_p]

I meant 50% duty-cycle but you already covered that. Maybe I'm wrong, but still 30W seems like a lot. I'm insisting because my experiment also used to eat a lot of power. I had a common mode choke with both coils wired in series, but generating opposing magnetic field. For no reason, it was just silly, and accidentally reversed one's polarity, so now they're not opposing any more. Current draw went down from 450mA to 20mA and it was almost as loud (at 3kHz). I'm not ace at inductors, but shouldn't the core NOT be saturated?

Anyway, how do the test subjects react to this?

[blankfield]

I'm using ferrite bead core capable up to 50W power applications so I'm in half a way to saturation. 30W is reasonable like for 4 piezo speakers, one can be driven at 10W. Power losses and heat in your project appeared because you're using wrong type of magnetic material. This small common mode choke have ferromagnetic iron core (iron dust), good for filter applications on 50/60Hz common lines. Useless for high power and frequencies like 20kHz. For this frequencies your choke 'have only  primary side' because 'power' (magnetic flux) at this frequency cannot pass this cheap material core. Tray to run typical 50Hz transformer at 20kHz and you will get same effect.  So you getting only heat on mosfet and primary side windings. Try ferrite core and fell (hear) difference  ;-). I'll made some update on weekend.

[RiccardoGiuliani]

Congratulations to blankfield and dr_p!

[blankfield]

Thanks Riccardo ,

I'm still working on my PBK, I've  some new clues according how to reduce audible frequencies on output.
For today I used 20kHz frequency but I still hear some noise, digital FFT in my scope showed why. On first picture we can see FFT for 20kHz output waveform captured at speaker connectors. Some part of spectrum covers audible range even down to 14kHz. Now I'm using 28kHz frequency and a lower margin of spectrum is above audible range.

I've also checked power balance,
ON state:
Power from battery cells RMS : 33.5W (8.0732A @ 4.15)
Power output RMS :  30.05W (1.7354 A @17.32 V)
Power dissipated mosfet + windings : 3.45W

Standby:
Power from battery cells RMS  :  3.94mW (950uA @ 4.15)

[ProBang]

Hello.

Is there some progress anywhere?
Is this competition still running?
Or is it closed?
And the winner is...?

[MJLorton]

Hello.

Is there some progress anywhere?
Is this competition still running?
Or is it closed?
And the winner is...?

Hi ProBang...it is "officially" over and a winner was announced...but if anyone wants to build one and demonstrate it working I will certainly cover it in a T4D video.

Cheers,
Martin.