"Jeff Liebermann" wrote in message
...
On Mon, 6 Jan 2014 06:28:11 +0000 (UTC), gregz
wrote:

wrote:
"only 3 db", but that's twice the signal. I have mine stacked 12 feet,
but I believe Winegard says either 8 or 10 feet. Mine work swell. +:^] I
got mine just after they were discontinued in 2005/6. Had to email a
number of suppliers until I found the second one. I bet there are some
still in storage somewhere, email different places that sell Winegard,
you may still find one.

John K9RZZ

Twice the signal means twice the voltage, for me.
Greg

Nope. Power is by the square of the voltage:
P = V^2 / R
If you double the voltage, you get 4 times the power.
A 1.414 times increase in voltage will produce twice the power.

I tried to convert the antenna model of the HD-6066P antenna from the
AO .ant format to .nec using 4NEC2 and failed. The plan was to model
the stacked arrangement and see what happens:
http://www.ham-radio.com/k6sti/hd6065p.htm
The .ant file imported without error, the wire tables and images look
correct, but the pattern is more like a point source than a gain
antenna. I'll look at it later to see where I screwed up, but it
would be nice if someone would look at the problem.

I just set up an experiment. I connected my roof antenna to my signal
level meter and read the signal strength of my Channel 10. It was 10 dBmV,
the unit typically used for TV signal strength work.

Next, I connected the same roof antenna to the inport port of one of a pair
of passive splitters connected back-to-back with equal short lengths of the
same 75-ohm cable.

Finally, I connected the output port of this network to the signal level
meter and observed a signal that was approximately 1.25 dBmV less. (A
quarter of a dBmV is about as close as I can reliably read; individual whole
number marks are only a few mm apart.)

Thus, I conclude that the 1 dB nominal loss for a passive splitter -- either
combining or splitting -- is confirmed. Combining two identical suignals
does get you something more than one, alone.

RELATED: When I used identical twin UHF antennas side-by-side, separated by
a free-space half-wave distance to cancel interference from one side, it
worked nicely and showed about the same loss figures as above. That is, my
reading for two antennas combined was about 2 dBmV higher than for either of
the twin antennas alone, thus reflecting the 1dB loss in the combiner.

Combining antennas can be an uncertain business because the phase
relationships change with wavelength; the arrangement that strengthens one
channel may weaken another channel if the respective signals come from
different directions and/or the cable lengths are not matched. It's a
matter of reinforcement or cancellation, depending on phase relationships.

"Sal"
(KD6VKW)

In message , Sal
writes

"Jeff Liebermann" wrote in message
.. .
On Mon, 6 Jan 2014 06:28:11 +0000 (UTC), gregz
wrote:

wrote:
"only 3 db", but that's twice the signal. I have mine stacked 12 feet,
but I believe Winegard says either 8 or 10 feet. Mine work swell. +:^] I
got mine just after they were discontinued in 2005/6. Had to email a
number of suppliers until I found the second one. I bet there are some
still in storage somewhere, email different places that sell Winegard,
you may still find one.

John K9RZZ

Twice the signal means twice the voltage, for me.
Greg

Nope. Power is by the square of the voltage:
P = V^2 / R
If you double the voltage, you get 4 times the power.
A 1.414 times increase in voltage will produce twice the power.

I tried to convert the antenna model of the HD-6066P antenna from the
AO .ant format to .nec using 4NEC2 and failed. The plan was to model
the stacked arrangement and see what happens:
http://www.ham-radio.com/k6sti/hd6065p.htm
The .ant file imported without error, the wire tables and images look
correct, but the pattern is more like a point source than a gain
antenna. I'll look at it later to see where I screwed up, but it
would be nice if someone would look at the problem.

I just set up an experiment. I connected my roof antenna to my signal
level meter and read the signal strength of my Channel 10. It was 10 dBmV,
the unit typically used for TV signal strength work.

Next, I connected the same roof antenna to the inport port of one of a pair
of passive splitters connected back-to-back with equal short lengths of the
same 75-ohm cable.

Finally, I connected the output port of this network to the signal level
meter and observed a signal that was approximately 1.25 dBmV less. (A
quarter of a dBmV is about as close as I can reliably read; individual whole
number marks are only a few mm apart.)

Thus, I conclude that the 1 dB nominal loss for a passive splitter -- either
combining or splitting -- is confirmed. Combining two identical suignals
does get you something more than one, alone.

RELATED: When I used identical twin UHF antennas side-by-side, separated by
a free-space half-wave distance to cancel interference from one side, it
worked nicely and showed about the same loss figures as above. That is, my
reading for two antennas combined was about 2 dBmV higher than for either of
the twin antennas alone, thus reflecting the 1dB loss in the combiner.

Combining antennas can be an uncertain business because the phase
relationships change with wavelength; the arrangement that strengthens one
channel may weaken another channel if the respective signals come from
different directions and/or the cable lengths are not matched. It's a
matter of reinforcement or cancellation, depending on phase relationships.

You missed out step #2, which was to measure the output level of the
splitter alone.

Using your figures, this would have shown a signal loss of 3.625dB (3dB
power split loss and 0.625dB of circuit loss).

When you then added the combiner, you would have 3dB power split loss
and 0.625dB of circuit loss, followed by 3db power combine gain and
0.625dB of circuit loss - so as you measured, a total loss of only
1.25dB.

Despite working in the cable TV industry for 43 years, for some reason
this is an experiment I don't recall ever performing!
--
Ian

"Ian Jackson" wrote in message

You missed out step #2, which was to measure the output level of the
splitter alone.

Using your figures, this would have shown a signal loss of 3.625dB (3dB
power split loss and 0.625dB of circuit loss).

When you then added the combiner, you would have 3dB power split loss and
0.625dB of circuit loss, followed by 3db power combine gain and 0.625dB of
circuit loss - so as you measured, a total loss of only 1.25dB.

Despite working in the cable TV industry for 43 years, for some reason
this is an experiment I don't recall ever performing!

Thanks Ian,

Earlier in this thread, I saw what I thought to be an error in some postings
.... about losses in excess of 3dB in the combiner and a conclusion that
stacking results in less signal, which shouldn't be the case. My little
experiment was meant to demonstrate a signal increase from combining
in-phase signals in a passive device. Put another way, I wanted to show
that a 3dB loss is not inherently present in both directions.

You are correct that I did not make the measurement of the output level of
the splitter alone, since it has been made and documented on many occasions.
A real lab experiment would have measured that and the cable losses, too.
(My 35 year-old Jerrold 747 was within easy reach and "close enough.")

It was my intent to show, when two equal signals (presumptive on my part
that the two outputs of a splitter are equal) are combined, that the result
is the addition of the two, minus ohmic and coupling losses, which I think I
did show.

"Sal"

On 1/7/2014 10:55 PM, Sal wrote:
"Jeff Liebermann" wrote in message
...
On Mon, 6 Jan 2014 06:28:11 +0000 (UTC), gregz
wrote:

wrote:
"only 3 db", but that's twice the signal. I have mine stacked 12 feet,
but I believe Winegard says either 8 or 10 feet. Mine work swell. +:^] I
got mine just after they were discontinued in 2005/6. Had to email a
number of suppliers until I found the second one. I bet there are some
still in storage somewhere, email different places that sell Winegard,
you may still find one.

John K9RZZ

Twice the signal means twice the voltage, for me.
Greg

Nope. Power is by the square of the voltage:
P = V^2 / R
If you double the voltage, you get 4 times the power.
A 1.414 times increase in voltage will produce twice the power.

I tried to convert the antenna model of the HD-6066P antenna from the
AO .ant format to .nec using 4NEC2 and failed. The plan was to model
the stacked arrangement and see what happens:
http://www.ham-radio.com/k6sti/hd6065p.htm
The .ant file imported without error, the wire tables and images look
correct, but the pattern is more like a point source than a gain
antenna. I'll look at it later to see where I screwed up, but it
would be nice if someone would look at the problem.

I just set up an experiment. I connected my roof antenna to my signal
level meter and read the signal strength of my Channel 10. It was 10 dBmV,
the unit typically used for TV signal strength work.

Next, I connected the same roof antenna to the inport port of one of a pair
of passive splitters connected back-to-back with equal short lengths of the
same 75-ohm cable.

Finally, I connected the output port of this network to the signal level
meter and observed a signal that was approximately 1.25 dBmV less. (A
quarter of a dBmV is about as close as I can reliably read; individual whole
number marks are only a few mm apart.)

Thus, I conclude that the 1 dB nominal loss for a passive splitter -- either
combining or splitting -- is confirmed. Combining two identical suignals
does get you something more than one, alone.

RELATED: When I used identical twin UHF antennas side-by-side, separated by
a free-space half-wave distance to cancel interference from one side, it
worked nicely and showed about the same loss figures as above. That is, my
reading for two antennas combined was about 2 dBmV higher than for either of
the twin antennas alone, thus reflecting the 1dB loss in the combiner.

Combining antennas can be an uncertain business because the phase
relationships change with wavelength; the arrangement that strengthens one
channel may weaken another channel if the respective signals come from
different directions and/or the cable lengths are not matched. It's a
matter of reinforcement or cancellation, depending on phase relationships.

"Sal"
(KD6VKW)



In addition to the splitter losses, you have coax and connector loss.
Coax loss probably isn't too bad, but unless you use a high quality
crimping tool, connector loss can easily approach 0.25 to 0.5 db. Even
with a high quality crimping tool, you can get around 0.1 db per connector.

There is also the possibility of a slight phase difference of the
signals coming out of the combiner, which would also affect the output
(splitters/combiners aren't perfect, either). But I wouldn't think this
would show up at such low frequencies unless you have lab-grade test
equipment (microwave frequencies and above are a different story).

--
==================
Remove the "x" from my email address
Jerry, AI0K

==================

"Jerry Stuckle" wrote in message
...

snip

In addition to the splitter losses, you have coax and connector loss. Coax
loss probably isn't too bad, but unless you use a high quality crimping
tool, connector loss can easily approach 0.25 to 0.5 db. Even with a high
quality crimping tool, you can get around 0.1 db per connector.

There is also the possibility of a slight phase difference of the signals
coming out of the combiner, which would also affect the output
(splitters/combiners aren't perfect, either). But I wouldn't think this
would show up at such low frequencies unless you have lab-grade test
equipment (microwave frequencies and above are a different story).

All correct. As I said to Ian, I wanted to show I could create two matching
signals then add them and the passive splitter/combiner output would be
greater than either input, alone. Accuracy within a dB or so was sufficient
to make the point. I wouldn't go to a professional meeting with the
demonstration rig I used last night.

Another experiment I ran (back around 1975) was to take 100 feet of cable
and measure the loss, then repeat the measurement using a different 100 feet
made from ten different pieces. Yup, the loss was about 3 dB more,
indicative of an average 0.3 dB loss per joint, neatly within the range you
specified.

"Sal"

On Thu, 9 Jan 2014 21:08:11 -0800, "Sal" salmonella@food
poisoning.org wrote:

Another experiment I ran (back around 1975) was to take 100 feet of cable
and measure the loss, then repeat the measurement using a different 100 feet
made from ten different pieces. Yup, the loss was about 3 dB more,
indicative of an average 0.3 dB loss per joint, neatly within the range you
specified.

0.3dB per connector at what frequency?

This is more fun:
http://802.11junk.com/jeffl/antennas/connector-loss/index.html
Just take every connector that you can find, put them in series, and
measure the loss. In this case, it was done at 2.4Ghz and 450MHz. End
to end loss at 2.4GHz was 2dB for about 25 adapters or about 0.08dB
per adapter. At 250MHz, the loss was about 0.2dB or 0.008dB per
adapter.

I've done similar demonstrations using two wattmeters at the local
radio club meeting. The results are typically that the adapter string
has the same loss as an equivalent length of small coax cable. I had
a surplus of BNC T connectors, so a strung about 50 of them in series
and obtained similar results.

Bottom line: Connectors and adapters aren't as evil as the data
sheets and literature suggest.


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

In message , Jeff Liebermann
writes
On Thu, 9 Jan 2014 21:08:11 -0800, "Sal" salmonella@food
poisoning.org wrote:

Another experiment I ran (back around 1975) was to take 100 feet of cable
and measure the loss, then repeat the measurement using a different 100 feet
made from ten different pieces. Yup, the loss was about 3 dB more,
indicative of an average 0.3 dB loss per joint, neatly within the range you
specified.

0.3dB per connector at what frequency?

This is more fun:
http://802.11junk.com/jeffl/antennas/connector-loss/index.html
Just take every connector that you can find, put them in series, and
measure the loss. In this case, it was done at 2.4Ghz and 450MHz. End
to end loss at 2.4GHz was 2dB for about 25 adapters or about 0.08dB
per adapter. At 250MHz, the loss was about 0.2dB or 0.008dB per
adapter.

I've done similar demonstrations using two wattmeters at the local
radio club meeting. The results are typically that the adapter string
has the same loss as an equivalent length of small coax cable. I had
a surplus of BNC T connectors, so a strung about 50 of them in series
and obtained similar results.

Bottom line: Connectors and adapters aren't as evil as the data
sheets and literature suggest.

I've always assumed that the loss measured through connectors and
adapters was mainly
(a) because they have unavoidable length (ie not a lot), and
(b) because the impedance match through them is less than perfect (ie
not a lot).
The ohmic contact resistance may also be a tiny tad higher than the same
length of coax (even less).
--
Ian

On 1/10/2014 7:01 PM, Ian Jackson wrote:
In message , Jeff Liebermann
writes
On Thu, 9 Jan 2014 21:08:11 -0800, "Sal" salmonella@food
poisoning.org wrote:

Another experiment I ran (back around 1975) was to take 100 feet of
cable
and measure the loss, then repeat the measurement using a different
100 feet
made from ten different pieces. Yup, the loss was about 3 dB more,
indicative of an average 0.3 dB loss per joint, neatly within the
range you
specified.

0.3dB per connector at what frequency?

This is more fun:
http://802.11junk.com/jeffl/antennas/connector-loss/index.html
Just take every connector that you can find, put them in series, and
measure the loss. In this case, it was done at 2.4Ghz and 450MHz. End
to end loss at 2.4GHz was 2dB for about 25 adapters or about 0.08dB
per adapter. At 250MHz, the loss was about 0.2dB or 0.008dB per
adapter.

I've done similar demonstrations using two wattmeters at the local
radio club meeting. The results are typically that the adapter string
has the same loss as an equivalent length of small coax cable. I had
a surplus of BNC T connectors, so a strung about 50 of them in series
and obtained similar results.

Bottom line: Connectors and adapters aren't as evil as the data
sheets and literature suggest.

I've always assumed that the loss measured through connectors and
adapters was mainly
(a) because they have unavoidable length (ie not a lot), and
(b) because the impedance match through them is less than perfect (ie
not a lot).
The ohmic contact resistance may also be a tiny tad higher than the same
length of coax (even less).

The main loss in a connector is due to the impedance bump at the
connector. This can be easily seen on a TDR (Time Domain Reflectometry)
display.

Some connectors are better than others; for instance, the older F
connectors which are crimped down with a ring are the worst. Next is
the connector where the crimp is a hex crimp - it doesn't give a
consistent impedance around the connector.

The best (and the ones we use) compress the entire base of the connector
evenly, creating a smooth crimp. The end of the coax is evenly covered
by the connector.

The other problem is the technician installing the connectors. I've
seen great ones, and not-so-great ones. There are a lot of chances for
going wrong - for instance, it's easy to screw up the braid when trying
to insert a crimp-on connector under the outer jacket and shield. And
soldering connectors (i.e. PL-259 and N) is almost sure to give you a
huge bump (and loss) because it's almost impossible to solder the shield
without melting the inner insulator to some point. It may not short
out, but that doesn't mean you don't have loss there.

--
==================
Remove the "x" from my email address
Jerry, AI0K

==================

On Fri, 10 Jan 2014 19:36:59 -0500, Jerry Stuckle
wrote:

The main loss in a connector is due to the impedance bump at the
connector. This can be easily seen on a TDR (Time Domain Reflectometry)
display.

Rubbish. Let's pretend that I mix in a 75 ohm coax connector into a
50 ohm system. Depending on the location of this "impedance bump",
the VSWR is no more than 1.5:1 which is generally considered marginal.
That's 0.18dB of mismatch loss.
http://www.microwaves101.com/encyclopedia/calvswr.cfm
If you're doing satellite or microwave DX work, then 0.18dB might be
important. However, for most other applications, it's a trivial
amount.

You might be amused to know that most of my rooftop antennas are fed
with 75 ohm coax and that my favored antenna designs are also 75 ohm.
There are various reasons, but the main one is that coax cable losses
are less at 75 ohms, than at 50 ohms. 50 ohms can handle more power,
but 75 ohms has less loss.
http://www.belden.com/blog/broadcastav/50-Ohms-The-Forgotten-Impedance.cfm
The only problems I have with 75 ohms is finding the proper connectors
and dealing with the pads needed to make my 50 ohm test equipment look
like 75 ohms. (Actually the real reason is that the 75 ohm stuff is
mostly CATV surplus, which tends to be really cheap).

Mo
http://www.qsl.net/n9zia/wireless/75_ohm_hardline.html

Some connectors are better than others; for instance, the older F
connectors which are crimped down with a ring are the worst. Next is
the connector where the crimp is a hex crimp - it doesn't give a
consistent impedance around the connector.

I rip those out wherever I find them, even if they're on the ends of
commercially crimped cables (usually RG-59/u which is another
nightmare). However, the loss mechanism with the old CATV coax and
associated crappy crimp connectors was radiation, not mismatch loss.
The ground connections would fall apart, turning the coax shield into
an impressive antenna.

The best (and the ones we use) compress the entire base of the connector
evenly, creating a smooth crimp. The end of the coax is evenly covered
by the connector.

I've had problems with some of those push-on connectors. I also don't
want to stock a zillion different connector variations from different
vendors. So, I've standardized on the "red" univeral T&B SNS1P6U
RG-6/u connectors:
www.ebay.com/sch/i.html?_nkw=SNS1P6U

The other problem is the technician installing the connectors. I've
seen great ones, and not-so-great ones. There are a lot of chances for
going wrong - for instance, it's easy to screw up the braid when trying
to insert a crimp-on connector under the outer jacket and shield. And
soldering connectors (i.e. PL-259 and N) is almost sure to give you a
huge bump (and loss) because it's almost impossible to solder the shield
without melting the inner insulator to some point. It may not short
out, but that doesn't mean you don't have loss there.

Actually, it's not the crimp job that kills the connection. It's the
stripping of the coax that causes the most problems. I use various
rotary contrivances that have razor blades to make the cuts at the
correct spacing. Those work well initially, but after about 50
connectors, the blades become dull and useless. Of course, nobody has
spare blades or knows how to adjust them. They either continue to use
a dull razor or steal my new stripper.

Oops... dinner... gone.
--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

In message , Jerry Stuckle
writes



The best (and the ones we use) compress the entire base of the
connector evenly, creating a smooth crimp. The end of the coax is
evenly covered by the connector.

In the CATV industry, for F-connectors, isn't Snap-n-Seal now de the
norm?

--
Ian