DIY Invisible Dog Fence Detector

All things pertaining to wireless and RF links

Moderator: phalanx

Post Reply
adiadidas15
Posts: 5
Joined: Thu Jul 06, 2017 9:53 pm

DIY Invisible Dog Fence Detector

Post by adiadidas15 » Thu Jul 06, 2017 10:23 pm

Hello everyone,
I'm a mechanical engineer by day, by I have gotten into the field of robotics as a hobby. This has led me to take up my first major electrical component design: a circuit to provide a 0-5v DC signal indicating proximity to an off-the-shelf invisible dog fence made by Pet Safe. Using an oscilloscope, I have determined that the fence operates at 10.8kHz. I have designed and breadboarded the following circuit to detect the fence. I'd like to turn this into a PCB, but before I do so, I was hoping to get some advice from more experienced electrical engineers. Are there any glaring mistakes I've made in the design?

I have actually created two identical breadboard versions of this schematic, but their performance is substantially different. I'm hoping the jump to PCB will fix some of the discrepancy. One of the hard parts initially was finding the right combo of inductor and capacitor to give a close resonance match. Would I be better off incorporating a trim capacitor? The best combo I could find from Mouser components uses a large 100mH inductor. Would I be better served using a smaller inductor? I noticed that the dog collar that came with the fence uses much smaller inductors.

Thanks in advance for any help you can offer!

[Here's where I wanted to attach my schematic as an image or PDF file, but apparently I need administrator approval to do so...]

-Chris

adiadidas15
Posts: 5
Joined: Thu Jul 06, 2017 9:53 pm

Re: DIY Invisible Dog Fence Detector

Post by adiadidas15 » Fri Jul 07, 2017 6:32 am

I'm not sure if this will work, but here is a link to my PDF schematic from Google Drive: https://drive.google.com/file/d/0B3Vt8F ... p=drivesdk

jremington
Support Volunteer
Posts: 2107
Joined: Fri Jun 15, 2007 9:41 pm
Location: Eugene, Or

Re: DIY Invisible Dog Fence Detector

Post by jremington » Fri Jul 07, 2017 8:53 pm

For such low frequency signals, breadboards should work fine and a PCB isn't likely to "fix" any issue.

Do you have any idea what the voltage across the LC combination is when near the fence, and what the actual resonant frequency is? That would be essential design information.

By all means, use trim caps to get the LC circuit tuned correctly.

If you don't have a sine wave audio signal generator, signal generation software is available for free, for smart phones and PCs, and would help a lot. Cut up an earphone cable and attach it to a loose coil or 10 mH inductor to act as a transmitting antenna, as an aid to tuning the LC circuit.

uomo_di_hart
Posts: 3
Joined: Mon Jul 24, 2017 7:50 am

Re: DIY Invisible Dog Fence Detector

Post by uomo_di_hart » Tue Aug 01, 2017 8:26 am

In stage 2 you have a gain of -30 not +30.

adiadidas15
Posts: 5
Joined: Thu Jul 06, 2017 9:53 pm

Re: DIY Invisible Dog Fence Detector

Post by adiadidas15 » Tue Aug 01, 2017 8:51 am

Thanks to both of you for your feedback. Several weeks ago I set out to get some screen shots from my USB oscilloscope to show the output of each stage in my circuit and share in this thread. However, I ended up redesigning the circuit in the process. I hope to share my updated design and oscilloscope measurements over the next week.
There are two aspects to my current design that I want to improve:
1) changing motor speeds of my brushed DC motors will saturate my sensors regardless of the sensor's location on the robot. I am hoping that a steel sheet metal shield will mitigate this problem.
2) the sensor can only detect the fence from ~18" away. I'd like to have some more range, maybe 3ft. This where I wonder about my choice of inductor size. As I said in my original post, my circuit uses a large 100mH inductor. I assumed a large inductor would be more sensitive to the RF because the cross sectional area is relatively large and more loops means more pickup? But I also know inductors naturally resist change in current flow, so perhaps a smaller inductor would work better? The dog collar utilizes much smaller inductors but I am sure it has better range than my sensor.

Sent from my XT1575 using Tapatalk

jremington
Support Volunteer
Posts: 2107
Joined: Fri Jun 15, 2007 9:41 pm
Location: Eugene, Or

Re: DIY Invisible Dog Fence Detector

Post by jremington » Tue Aug 01, 2017 1:51 pm

a large inductor would be more sensitive to the RF because the cross sectional area is relatively large
.
Correct. The larger the cross sectional area, the better.

As for motor electrical noise, use inductors and capacitors as close to the motor terminals as possible to suppress electromagnetic radiation from the wiring. Here is one tutorial: https://www.pololu.com/docs/0J15/9

adiadidas15
Posts: 5
Joined: Thu Jul 06, 2017 9:53 pm

Re: DIY Invisible Dog Fence Detector

Post by adiadidas15 » Tue Aug 01, 2017 2:54 pm

jremington wrote:
a large inductor would be more sensitive to the RF because the cross sectional area is relatively large
.
Correct. The larger the cross sectional area, the better.

As for motor electrical noise, use inductors and capacitors as close to the motor terminals as possible to suppress electromagnetic radiation from the wiring. Here is one tutorial: https://www.pololu.com/docs/0J15/9
If two inductors have the same cross sectional area but different inductance values, which would you expect to produce a stronger signal?

I had assumed that the motor noise was caused by the magnetic field emanating from the motor coils. But this didn't totally fit with the my observation that noise only occurred when I changed the motor speed. If the motor coils were the cause, I'd expect noise to scale with speed. But I could hold my motors at max speed and have little noise. The tutorial seems to focus on noise caused by the power lines feeding the motors, which I wouldn't have expected. But this makes more sense with my observation because the current change in the power lines (to vary motor speed) will cause a change in flux in the inductor's magnetic field. I hope the suggestions in the tutorial work because they are easier than making enclosures! Thanks for the continued help!

Sent from my XT1575 using Tapatalk

jremington
Support Volunteer
Posts: 2107
Joined: Fri Jun 15, 2007 9:41 pm
Location: Eugene, Or

Re: DIY Invisible Dog Fence Detector

Post by jremington » Tue Aug 01, 2017 5:06 pm

If two inductors have the same cross sectional area but different inductance values, which would you expect to produce a stronger signal?
That is hard to say, because you need to have a resonant LC circuit and the peak response will depend on several factors such as the coil resistance, coil capacitance, capacitor quality, etc.

adiadidas15
Posts: 5
Joined: Thu Jul 06, 2017 9:53 pm

Re: DIY Invisible Dog Fence Detector

Post by adiadidas15 » Tue Aug 01, 2017 5:46 pm

Yes, I was also assuming that I would adjust the capacitor to keep the resonant frequency the same: just having two different LC combos with the same resonance. It sounds like you're saying the inductance difference may not cause a straightforward result, there are other factors that may be equally important.

Sent from my XT1575 using Tapatalk

Post Reply