This morning, I put my 17 Khz vlf transmitter (petsafe wireless fence)
on spectrum lab running on my laptop. What I saw has put me in a
tailspin.

I have the schematic and the screen captures showing the output
spectrum for anyone who wants to email me, aballen at colby dot edu.

Using spectrum labs built in vlf receiver and a short antenna wire, I
was able to display the output spectrum of the transmitter without any
problems. BUT, I'm at a loss to explain the output spectrum.

Around 18 Khz, there is a peak. Exactly 15 Hz below and above the peak
is a smaller set of peaks, presumably these are the sidebands that the
receiver recovers. BUT, there is a peak every 60 Hz that slowly gets
weaker as the spectrum moves away from the (center) largest peak. Each
one of these smaller peaks has upper and lower sidebands as well.
There are many of these peaks, not just a few.

I wish I could attach the gif's and the schematic diagram of the
transmitter, but I think binary attachments are forbidden here.

For those who didn't read the original post, here's a summary. I
purchased a petsafe PIF-300 wireless fence to make an invisible fence
to contain our dog. The unit has a 17 Khz transmitter that runs in the
house from ac power. The small battery powered receiver is mounted on
the dogs collar and gives an audible warning and a correction IF the
dog tries to go past the transmitters range.

The url for the product is at:

http://www.petsafe.net/pet_containme...tant_fence.htm

Here's my question...

Why would they use a transmitter with such a wideband output and
containing so many spurs? Is there any kind of simple receiver that
can properly recover the transmitter data?

As I said before, send me an email and I'll be happy to forward the
schematic and the screen captures showing the output spectra.

A

The product is an Instafence (also called Instant Fence or wireless
fence). The manual for the thing is available on petsafe.net, look for
the PIF-300 product (which also includes the collar). It transmits a
short range signal at 18.7 kilohertz using large air coils as
antennas. It is designed to contain a dog within a 90 foot radius and
to issue mild shocks if/when the pet attempts to leave the area. It
does not work with a buried wire loop, it uses the transmitter to
define the dogs play area.

For anyone interested in purchasing one, they are horribly expensive
(list price $280) and they have a special battery for the collar,
which is only used by petsafe, available at 6 dollars each. The
batteries don't last long. The range is only 90 feet, so it doesn't
allow the pet to move very much.

It appears to be grossly over-engineered as well, an ASIC feeds one
each amplifier separately and the amps appear to run at very low
power. No doubt a much cheaper transmitter could be made, this one is
also physically large as well!

For anyone interested in the petsafe IF100 vlf transmitter, you can
look on www.petsafe.net for the manufacturers info-

Or, check out pictures, schematic and field strength documentation at:

https://gullfoss2.fcc.gov/cgi-bin/ws...cc_id=KE3IF100


Here are some notes based on my personal observations (before I
returned the device).

SW2 is hardwired, and is not adjustable as shown in the schematic. For
some reason, this switch was called 'mode' back in 1998 when petsafe
submitted the documentation to the FCC. The schematic shows it as a
switch though. When looking at the schematic, the position of the
switch in the schematic is shown correctly.

Switching S3 to ground selects low power transmit.

The pot, R33 is the power control.

The ASIC is marked as follows:
Supertex
500-029
204T770
0225

It is impossible to tell how many turns are on each loop (transmit
antenna). But, each of the 4 loops has the same size wire and has a dc
resistance of 4.5 ohms.
Vdd for the ASIC is 4.85 volts.

I have spectral output for each of the output lines (X, Y, Z). The
spectral display shows the same signal on each output although the
same output signal on all 3 output lines, although there are different
peak amplitudes (some of the amp stages are driven harder than others,
but all have the same basic input signals).

There are 3 audio amp chips, they are each the LM 1875T. They can run
20 watts output, but the output power is much lower than 20 watts for
each stage. This is based on the V+ being very low (16.8v) and the
observation that the 1.8A wall wart (at 14v ac) runs very cool. Also,
although the heat sinks for each stage are very small, they are barely
warm to the touch.

If anyone wants to see the spectral output of the ASIC that feeds the
amps, send me an emmail to:

K Y 1 K at pivot dott nett.

Enjoy,

A

Albert wrote in message ...

It appears to be grossly over-engineered as well, an ASIC feeds one
each amplifier separately and the amps appear to run at very low
power. No doubt a much cheaper transmitter could be made, this one is
also physically large as well!

Far more things than you can imagine are designed like this. On short runs
it isn't cost effective to come up with an efficient design. Same applies to
a lot of software!
--
N

Thanks N.

Well, the audio amp chips might be cheaper than some chips that are
less powerful. I know this happens sometimes.

Also, the manufacturer attempts to get a somewhat circular (3D) power
distribution so the orientation of the transmitter is non critical
(and is a circle). So, three transmit stages might be necessary-I
notice each stage has slightly different voltage gain, no doubt to try
and achieve a circular coverage pattern. So, three stages might
actually be necessary.

But, ASIC's are very expensive to develop and this one doesn't seem to
be doing much. Maybe it's cheaper to have an ASIC than to build a
larger circuit board with discrete components?

GL.

A







On Wed, 06 Apr 2005 03:24:05 GMT, "NSM" wrote:


Albert wrote in message ...

It appears to be grossly over-engineered as well, an ASIC feeds one
each amplifier separately and the amps appear to run at very low
power. No doubt a much cheaper transmitter could be made, this one is
also physically large as well!

Far more things than you can imagine are designed like this. On short runs
it isn't cost effective to come up with an efficient design. Same applies to
a lot of software!

Albert wrote in message ...

But, ASIC's are very expensive to develop and this one doesn't seem to
be doing much. Maybe it's cheaper to have an ASIC than to build a
larger circuit board with discrete components?

"When you only have a hammer, everything looks like a nail".
--
N