S/N ratio question - have I got this right?
 

Someone check my math, please? I'm using
a program called "Cool Edit Pro" which can
give me very accurate readings of tone levels
coming from my FT-857D into my soundcard.

I'm trying to develop my own way of evaluating
my new vertical's (Cushcraft R7000) performance
compared to my dipole. I'm starting by tuning
700 hz off of WWV's 10 mhz carrier and recording
the tone I get in USB mode with my AGC off. I
switched antennas during the recording and took
2 samples:

1st example Cushcraft R7000
Noise @ 2k = -39.27db
Signal @ .7k = -7.26db
S/N differance = 32.01db
S/n Ratio = 14.662

2nd example 40m ladderline fed dipole
Noise @ 2k = -30.11db
Signal @ .7k = -8.864db
S/N differance = 21.246db
S/n Ratio = 10.622

For some reason my program gives me the back-
ground and signal levels in negative DB. I used
this formula to come up with a S/N ratio:

S/N = 20 log10(-dbs/-dbn)

Can someone check my math? My guess is
that my R7000 is better at WWV for this test.

--
73's es gd dx de Ken KGØWX
Grid EM17ip, Flying Pigs #1055,
Digital On Six #350,
List Owner, Yahoo! E-groups:
VX-2R & FT-857

Ken Bessler wrote:

Someone check my math, please? I'm using
a program called "Cool Edit Pro" which can
give me very accurate readings of tone levels
coming from my FT-857D into my soundcard.

I'm trying to develop my own way of evaluating
my new vertical's (Cushcraft R7000) performance
compared to my dipole. I'm starting by tuning
700 hz off of WWV's 10 mhz carrier and recording
the tone I get in USB mode with my AGC off. I
switched antennas during the recording and took
2 samples:

1st example Cushcraft R7000
Noise @ 2k = -39.27db
Signal @ .7k = -7.26db
S/N differance = 32.01db
S/n Ratio = 14.662

2nd example 40m ladderline fed dipole
Noise @ 2k = -30.11db
Signal @ .7k = -8.864db
S/N differance = 21.246db
S/n Ratio = 10.622

For some reason my program gives me the back-
ground and signal levels in negative DB. I used
this formula to come up with a S/N ratio:

S/N = 20 log10(-dbs/-dbn)

Can someone check my math? My guess is
that my R7000 is better at WWV for this test.


CoolEdit always references with 0 as the maximum signal level; i.e.,
32768/-32767 are max for 16 bit recordings. This viewing mode is
selected with View - Vertical Scale Format - Sample Values. You may
also select Decibels, Normalized Values, etc.

I have CoolEdit 2000. Is yours a newer one since the company was bought
out?

When you did the level check, how was that done? Did you select Analyze
- Statistics and run it against a marked area in the recording for an
average? If so, how long was the average? Did you run a filter first
to isolate the signal from the "noise"? Or did you run the FFT function?

tom
K0TAR

"Ken Bessler" wrote in message
news:ZrB8e.2850$mS1.499@okepread04...
Someone check my math, please? I'm using
a program called "Cool Edit Pro" which can
give me very accurate readings of tone levels
coming from my FT-857D into my soundcard.

I'm trying to develop my own way of evaluating
my new vertical's (Cushcraft R7000) performance
compared to my dipole. I'm starting by tuning
700 hz off of WWV's 10 mhz carrier and recording
the tone I get in USB mode with my AGC off. I
switched antennas during the recording and took
2 samples:

1st example Cushcraft R7000
Noise @ 2k = -39.27db
Signal @ .7k = -7.26db
S/N differance = 32.01db
S/n Ratio = 14.662

2nd example 40m ladderline fed dipole
Noise @ 2k = -30.11db
Signal @ .7k = -8.864db
S/N differance = 21.246db
S/n Ratio = 10.622

For some reason my program gives me the back-
ground and signal levels in negative DB. I used
this formula to come up with a S/N ratio:

S/N = 20 log10(-dbs/-dbn)

Can someone check my math? My guess is
that my R7000 is better at WWV for this test.

--
73's es gd dx de Ken KGØWX
Grid EM17ip, Flying Pigs #1055,
Digital On Six #350,
List Owner, Yahoo! E-groups:
VX-2R & FT-857

Signal to noise ratio is the difference between the signal and the noise.
i.e. In your
first example: -8.864-(-30.11) = 21.264 dB.

Not sure what you mean by quoting levels in "dB", since it must be
referenced to a
specific value. Assume it is "dBm".

A sky wave signal should be averaged of a long period to estimate the actual
signal level.
Also the RF gain must be backed off so the receiver is not saturating from
the input signal.

Your "20log10" is the expression used when calculating the dB ratio between
two voltages.
Assume you mean "Log to the base 10".

Hope this helps.

73,

Frank

Tom Ring wrote:

CoolEdit always references with 0 as the maximum signal level; i.e.,
32768/-32767 are max for 16 bit recordings. This viewing mode is
selected with View - Vertical Scale Format - Sample Values. You may
also select Decibels, Normalized Values, etc.


Sorry, I should have said 0dB, and then that samples are referenced to
integer values which can be viwed, etc.

This damned sinusitus. I can begin to see why people who are really
sick, as opposed to what I have, get to the point where they want to
give up trying to continue to live.

tom
K0TAR

"Tom Ring" wrote in message
.. .

CoolEdit always references with 0 as the maximum signal level; i.e.,
32768/-32767 are max for 16 bit recordings. This viewing mode is selected
with View - Vertical Scale Format - Sample Values. You may also select
Decibels, Normalized Values, etc.

I have CoolEdit 2000. Is yours a newer one since the company was bought
out?

When you did the level check, how was that done? Did you select
Analyze - Statistics and run it against a marked area in the recording
for an average? If so, how long was the average? Did you run a filter
first to isolate the signal from the "noise"? Or did you run the FFT
function?

tom
K0TAR

I have 2.1 and I'm still learning. I didn't do a level check
- I was only wanting to compare 2 antennas, not establish
an actuall DBv reference.

Ken KG0WX