I'm afraid I've done it again. I have a COMPLETELY new (and SO MUCH BETTER) design...

[Steve Gibson]

Gang...

Here are the details of a completely

new and MUCH superior design...

This new approach is 3 to 4 times more efficient, in that it's able to put a MUCH more pure sine wave -- nearly perfect -- of 70-80 volts p-p (27 Vrms) across our piezo tweeters while consuming less than 100mA. This moves its power consumption down into the range that a single 9v "transistor radio" battery can easily handle -- so that's the operating voltage it is designed to use.  This new design employs only one inductor, plus one active and three passive components. It's AMAZING.  :)

It is NOT dynamically tunable (I will be working on that next), but I have chosen and documented a very nice and inexpensive line of inductors, with a large variety of available values surrounding 1000uH. So you can purchase a range (at 42 cents each) to set the operating frequency wherever you choose, given your component tolerances, individual tweeter capacitance, and personal frequency preference.

Here's what happened...

As I last wrote, I was working on the problem of a continuously tunable system. (That problem remains outstanding, and I will return to it now that this most recent breakthrough is detailed and documented.)  I have always been annoyed by my original use of a "buck-style" voltage boost converter, since it inherently wastes so much power while the transistor is on.  In looking at the wave shapes and frequency responses of the resonating filter being driven by the buck-style booster, I was wishing that I could provide not only a positive impulse to the resonator, but also a equal negative impulse.  My feeling was that if I could drive the resonator with something even a bit "less of a spike" -- perhaps more of a square -- then the wave shapes I was seeing might have at least some better chance of being less unruly.

Then I remembered the beautiful little Microchip MOSFET driver IC I had used in the first "Big Mama Mega Blaster" I built for Mark.  Because I was building only one, and I didn't want to spend my life optimizing it, I used a MONSTER overkill power MOSFET.  But a MOS transistor necessarily has a large enhancement mode channel to suppor large current flow. And that requires a large gate and accompanying gate capacitance.  Since that first device was driven by a simple 555 timer, the timer's output drive was limited.  So I added a "MOSFET Driver IC" between the 555 timer and the overkill power MOSFET's gate. Such a driver has a beefy complementary MOS (CMOS) output stage... essentially two large MOS transistors, one N-channel and one P-channel which are used to rapidly charge and discharge an external power MOSFET's large gate capacitance. That's all it does. The front end of a MOSFET driver is typically a high-impedance input, and this one also had 300mV of hysteresis to eliminate output jitter and uncertainty if the input is noisy or wandering around a bit. These MOSFET drivers are available in both inverting and non-inverting versions. I always over-order parts from Digikey on the theory that (a) if I order 10 they're cheaper each, and (b) something I used once is more likely to be used again... that was certainly the case this time. So I already had some of those $2 MOSFET driver ICs on hand.

I assembled the bits the pieces of a now-familiar architecture: an inverting driver, driving a high-Q resonator consisting of an inductor and capacitor in series (the capacitor being both the circuit's load and our tweeter), taking feedback from the inductor/capacitor junction, through an RC phase-shifting filter, and back to the input of the inverting driver...

It took off like a bat out of hell and began to sing!

I then spent the next 24 hours or so studying it closely and optimizing its operation. As you can see, it could not possibly be any simpler.

I believe we have our final design for a non-tunable, high-frequency, high-power audio oscillator.  When I saw that the circuit was drawing so little power, I experimented with raising its operating voltage from 6 to 9 volts... and it only made it louder.  I am getting NO distortion, NO harmonics, and NO misbehavior of any kind.  The R1 and C2 values are non-critical, though I think they are optimal for the frequency range we want.

One thing we learned during the past few weeks, is that the tweeter capacitances vary widely, but so too do the inductors (10%). And we've also seen that people have varying preferences for the operating frequency of their devices. Some people don't mind if it's audible to humans, others prefer to be completely stealthy.  So rather than specifying a single value for the device's frequency-setting inductor, I selected a very nice, Digikey-stocked, family of inexpensive (42 cents each) inductors. The family offers a full range of values clustered around 1000uH (1mH). My advice would be for people to purchase as many different values in the region as they may need and want. Lower inductance values will yield higher frequencies, whereas higher values will bring the frequency down. The mid-point of the range I have provided is 1000uH, which nominally oscillates at 12 to 16 kHz depending upon the individual tweeter's capacitance.

The rule of thumb should be to use the lowest frequency (highest inductance value) that you're comfortable with, since we don't yet have sufficient experience with the response of these piezo tweeters to know how much power they emit at super-sonic frequencies -- nor for that matter with canine responses.

Inductors could either be swapped in, one at a time, to select the one you prefer, OR a simple selector switch could be used to switch between a collection of them at will:  http://www.digikey.com/product-detail/en/A20605RNCQ/CKN9488-ND/416324

In conclusion...

I am EXTREMELY pleased with the surprising evolution this aspect of the project has taken. I believe that armed with the schematic and bill-of-materials (attached to this post), anyone with basic electrical/electronics assembly skills can now quickly, easily, and inexpensively assemble a state-of-the-art fixed (or incrementally variable), high-performance high-power acoustic sound generator.

It is my hope that the people here will be interested in assembling this much-superior, version 2.2.0 of the fixed-frequency oscillator system, so that we might all share in everything that can be learned about it. It is absolutely true that all of the feedback provided about the performance of the previous design was instrumental in getting me to the place where this one would be born.

Thanks everyone!!  :)

/Steve.

[bhal...@gmail.com]

I'm sitting here reading this as our neighbors' dog is barking loudly. I am ready to build one of these, but I have only very beginner electronics assembly skills and understanding. The last electronics project I did was a simple, Radio Shack kit to build a tunable radio to listen to local police and fire radio transmissions. It had very easy-to-follow instructions and I was able to make something useful using the parts included. I'm hoping that this bill of materials and the schematic will be easy enough to figure out so that I can put this together as well. I have a soldering iron, and basic electronics assembly tools. What I'm missing is the bigger picture how-to instructions. For example, what sort of circuit board do I use to assemble this? What about the overall project enclosure, etc? I've spent hours today researching and learning about basic electronics theory in order to prepare myself for a deeper understanding of how this thing actually works (although I'm sure that understanding may not come very soon). I apologize for the n00b questions, and maybe I'm just not being patient enough (I understand that you will soon be publishing a how-to guide on GRC.com), but where can I go to find out more about basic electronics assembly and how to actually put all of the components in the schematic together?

Thanks so much for all you and others in this community do!

Brian Hall

[Steve Gibson]

Hi Brian,

I'm sitting here reading this as our neighbors' dog is barking loudly. I am ready to build one of these, but I have only very beginner electronics assembly skills and understanding. The last electronics project I did was a simple, Radio Shack kit to build a tunable radio to listen to local police and fire radio transmissions. It had very easy-to-follow instructions and I was able to make something useful using the parts included. I'm hoping that this bill of materials and the schematic will be easy enough to figure out so that I can put this together as well. I have a soldering iron, and basic electronics assembly tools. What I'm missing is the bigger picture how-to instructions. For example, what sort of circuit board do I use to assemble this? What about the overall project enclosure, etc? I've spent hours today researching and learning about basic electronics theory in order to prepare myself for a deeper understanding of how this thing actually works (although I'm sure that understanding may not come very soon). I apologize for the n00b questions, and maybe I'm just not being patient enough (I understand that you will soon be publishing a how-to guide on GRC.com), but where can I go to find out more about basic electronics assembly and how to actually put all of the components in the schematic together?

I'll bet that many people here can point you at some terrific sites for filling in your background.

And you're right that the GRC pages will eventually become the final resting place -- or static documentation -- for this entire project.  The TEMPLATE for the pages can be seen in the "link block" at the bottom of the main page... though NOTHING ELSE is there yet...

http://www.grc.com/tqc/thequietcanine.htm

/Steve.

[Brian Hall]

Cool, thanks! If anyone has any recommendations for good learning material for how to read, understand and assemble a project from a schematic (including the board, soldering tips, and anything else helpful for a beginner), please share.

Thanks!

Brian Hall

[Thomas Trostel]

So …. if you haven't seen Fritzing before, its a hyper cool very easy design tool capable of creating pretty nice quality PCB design patterns based on breadboard layouts!

Anyhow …. using your design, here's a start for the last design as a PCB.  Please don't laugh too hard ;-)  What do you think?  If anyone wants to help clean it up that would be great too!

[Thomas Fors]

One great way to learn about building electronic circuits might be to find a Maker Space a.k.a. Hacker Space in your area. Many of them offer soldering classes that are open to the public and have members with a wide range of expertise that are usually very willing to share their knowledge.

Here's a list of hacker spaces. (I have no idea how complete it is.)

http://hackerspaces.org/wiki/List_of_Hacker_Spaces

--Tom

On Dec 22, 2012, at 6:18 PM, Brian Hall <bhal...@gmail.com> wrote:

[Steve Gibson]

Gang...

My continuing research into designing a digital tunable high-power harmonic oscillator has borne a new and immediately applicable bit of fruit:

This is a Single Inductor ZPL (Zero Power Loss) approach to tuning the v2.2.0 circuit.

Since, as has been observed here several times, capacitors do not dissipate power, when an additional capacitor is placed ACROSS the (nominally 0.13uf) tweeter, the voltage (and thus the loudness) remains as high as ever across the tweeter... while the FREQUENCY of the entire system will be reduced... retaining a beautiful near-perfect sine wave.

Since many experimenters have collections of capacitors lying around (they'll need to be non-polar and at least 50 volts capable), and since caps are both a LOT less expensive and a LOT lighter and smaller... this "capacitor gang" approach is superior to the "inductor gang" approach I suggested in the root posting of this topic.

You'll note that I specified the SINGLE inductor's value at 820uH.  I chose that, since the highest frequency I've seen with a 1000uH inductor has been 16 kHz.  So 820uH that moves that operating point up a bit further up, allowing parallel capacitors to then bring it down to different frequencies for experimentation.

I'm still working on a micro-controller-based solution... but since I stumbled upon this improved approach for creating a mechanically switchable device, I wanted to share it sooner rather than later.

/Steve.

[Steve Gibson]

Holy Crap!

While showering for a Christmas Eve dinner with friends, another tidbit fell into place:

Assuming that the capacitor tunability described in my previous posting holds as we scale inductances down and capacitances up...

We can LARGELY REMOVE the this solution's frequency dependence upon the tweeter's capacitance!

If we used a 100uH inductor (1/10th the size) with a 1.3uf parallel capacitor (10x the size)... then the system would oscillate at approximately the same frequency... but then the ADDITION of the tweeter, at 1/10th of the capacitor's value would have a MUCH smaller overall effect.  So that the effects of tweeter capacitance variation would be much diminished.

I haven't verified this experimentally yet... but if I'm sober enough after dinner tonight... it's the first thing I'll do!  :)

/Steve.

[Kindanyume]

Hey Steve 2 things

> While showering for a Christmas Eve dinner with friends, another tidbit fell
> into place:

that just sounds very wrong :P

> I haven't verified this experimentally yet... but if I'm sober enough after
> dinner tonight... it's the first thing I'll do! :)

It's Xmas eve.. STOP WORKING!!!!

:)

[Matthew Yakel]

Awesome, I haven't followed this group in a while, but I have really enjoyed the mentions of this in the podcast lately. Personally I too suffer from the dog problem (neighbors) so count me in when I can purchase a kit. I'll probably buy two. I don't know much about the deep core of these components but will gladly accept the challenge of assembly and setup. Thank You again for the great content and contributions to the tech community. 

Matt

On Saturday, December 22, 2012 4:55:36 PM UTC-6, Steve Gibson wrote:

[Steve Gibson]

Whoops.

The concept was correct, but something else got us:  Since the energy stored within an inductor is one half its inductance times the square of its current, if we reduce the inductor's size by a factor of 10, we also reduce the amount of energy it's able to store for a given current by a factor of ten.  And in this case it no longer had sufficient power to adequately drive the tweeter at a good level.

So...  Maximum Loudness will always be to use the LARGEST inductance that's practical -- remembering that the larger the inductance the lower the frequency.

In my test rig, one of my more expensive ($18) "MADE IN PHILIPPINES" test tweeters measures at 0.135uf.

The oscillation frequency of a purely harmonic LC oscillator is 1 / ( 2 * PI * SQR( L * C )), where L is the inductance in Henrys and C is the capacitance in Farads.

So with that 0.135uf tweeter, my favored 1mH inductor yields a calculated frequency of 13.7 kHz.  But what I'm seeing is a 16 kHz sine wave.

I don't have a means for directly measuring the inductor's true inductance, but even if we allowed for a -10% tolerance decrease to 900uH, that would still only account for a rise in frequency to 14.44 kHz.

But... we DON'T actually have a "purely harmonic LC oscillator".  In a purely harmonic LC oscillator, the driving feedback to keep it going is also a sine wave.  But in our system we're driving the oscillator with a square wave... and relying upon the LC system to do the best job it can of smoothing things out.  This "square drive" has the effect of "steepening the slopes" of the sine wave thus both deforming it a bit AND increasing its oscillating frequency by about 14%.

Since this is probably a constant factor, a better formula for computing the frequency from L & C when under a square wave drive, is:

Freq(hz)  = 1 / ( 1.712 * PI * SQR( L * C ))

This also means that if you want to compute values of C to use in dropping the frequency, the total capacitance in Farads (the added C plus the tweeter's) would be...

1 / ( F^2 * L * 29) = C in Farads

/Steve.

[bouwe...@gmail.com]

I hope I'm not to late, I was following this very closely about 3 weeks ago, and then was pulled away due to school finals

Is everything still on track with "beta testers" or build this using kits, either way, It looks very promising from all the interest. I hope we can actually get this done, and not let it die like last time. 

thanks Steve. 

[Steve Gibson]

Hi,

We already have a ready-to-go design with complete parts list and schematic.  People are waiting for things to settle down, since I keep coming up with useful improvements.  But the LEAST we'll have is something that anyone with basic electronics understanding and construction skills should be able to build.  And the circuit is SO SIMPLE that a printed circuit board won't be necessary... just a 1/10th inch perf board... and not much of that!  :)

/Steve. 

[coo...@gmail.com]

What does a person use to open .p7s or .fzz  ???

[mihai.a...@gmail.com]

You quoted the answer...

[Kindanyume]

Ahh true... what I would have done to have that back in when I was
starting in this area... it could have saved me infinite time, which I
could have spent chasing girls of course :)~~

[Thomas Trostel]

A .fzz file is opened with the Fritzing program available at frizzing.org

I stumbled across it about a year ago and thought it was pretty nifty … esp for free!

From their description:

ABOUT FRITZING

Fritzing is an open-source hardware initiative to support designers, artists, researchers and hobbyists to work creatively with interactive electronics. We are creating a software tool, a community website and services in the spirit of Processing and Arduino, fostering an ecosystem that allows users to document their prototypes, share them with others, teach electronics in a classroom, and layout andmanufacture professional pcbs.

[bigbear...@yahoo.com]

Hello Steve and group members, I have arrived late but this was because of need.  Two of my neighbors now have dogs. One has two German Shepards and the other has THE DOG FROM HELL. This dog runs out of the house barking... Just to bark. It's loud and constant. Sometimes when a person gets very excited or angry their voice rises and they loose their voice momentarily, it changes pitch and sometimes breaks up. Well this dog barks like that, always. So loud and constant it looses it's voice but that does not stop her from continuing barking. She will run around the house to find something in particular to bark at, but will continue barking anyway. The two German Shepards have a regular schedule. Every hour from 7:00 am to 11:00 pm they will come out and bark for 20 minutes, and one howls like a siren, loud and long. I need a break.

I have tried building other circuits which failed to produce the expected/desired results (not powerful enough) before discovering your group. I built the 2.2.0 and it's currently breadboarded. I went for the stealthy frequency and chose a 820uh inductor which is producing 23kHz @ 32vpp. This seems low according to other members reports, I was expecting 70v. I plan to put the tweeter(s) close to their properties but hidden in trees and camouflaged for additional stealth but not blocked.

I have read this entire groups posts and thank you all for you ideas.

Before making a PCB I am testing it on the dogs breadborded. I have designed the PCB though so will include it with this post.

It used to be so peaceful here, hope I get that back. Pete

[bouwe...@gmail.com]

pete, have you have got this working with successful results, the one i build when steves first come out did not work at all, if you did please let me know, barking dogs still drive me insane 

thanks. 

[greenmax]

HI,

I have a neighbor who's dog is very loud. I have tried talking to him and also complained to the city, but to no avail. I find this project very interesting and am handy with electronics.

I just want to know if someone has built this device and successfully shut up a dog from a distance of 20 feet. I went through a lot of posts and could see that people were experimenting a lot. I also noticed that on the grc website Steve says that this design does not work as a bark deterrant. Can someone confirm this assertion.

https://www.grc.com/tqc/thequietcanine.htm

[rob...@gmail.com]

I want to build one of these but DigiKey no longer sells two of the components or doesn't appear to.  The components I can't find are:

• 445-2855-ND
• 445-3778-1-ND

Are there alternatives or is there another version?

[Steve Gibson]

Hi!  SORRY for my delayed reply.

The 0.22uF ceramic cap choice (445-2855-ND) can be pretty much anything.

I would recommend: 399-13990-1-ND

For the 1mH inductor (445-3778-1-ND), I would recommend this replacement:  811-3489-ND

I hope that's useful.   All the best!

/Steve.                                 

--
You received this message because you are subscribed to the Google Groups "Portable Sound Blaster" group.
To unsubscribe from this group and stop receiving emails from it, send an email to portable-sound-bl...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/portable-sound-blaster/332d34b8-90c6-4ebe-953a-77fc62577f02%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

/Steve.

[anar...@gmail.com]

I just wanted to point out an old  bad assumption in an old post of yours

The RMS or peak current thru the piezo speaker at the resonant frequency is the best indicator of output power as the real mechanical energy is a series resistance in the speaker and not the energy stored in the inductor or capacitance.  For series resonance, I²*ESR.=Pd  ( at some f)

As you know, it is counter-intuitive yet true,  that for impedance matching, max power transfer theorem states the impedances must be matched for this Max output power,  to the speaker, yet efficiency can never be greater than 50%. 

So in the end for sonic reasons, we tend to go for lower drive impedance (<0.1) and higher loads (8ohm) which are not at maximum power transfer but at lower current distortion levels.  But this is irrelevant we don't care about distortion none of the harmonic energy is transferred in the high Q shape filter.

The choice of L and it's DC winding resistance control both the output power and frequency.  So we want higher L/R in every choice here but lower L will actually raise the frequency and raise the current which mechanically increases force on the piezo.  So pick whatever frequency works best then choose the lowest DCR inductor for that frequency and don't worry about  L increasing stored energy, if that reduces current the reduces energy more by II²  than L increasing.  E(L) = 0.5 I² L