[R-390] Noise Floor question

Wed Mar 29 14:53:02 EST 2006

Perry,

You may want to take a look at the noise floor measurements I did as a 
comparison of the R-390 and R-390A and the description of my method of 
measuring noise floor in the March Electric Radio.   The article is really 
on a method for better audio on the non-A R-390 but had to be broken in half 
for a two part article where the first part is mostly just on the R-390. 
The April half will get down to business and talk about audio performance.

Be very careful when someone gives a noise floor for a receiver and no 
information on how the testing was done.  Noise floor numbers are often give 
as the minimum signal level in -dbm, db below a MW in power, that a receiver 
can successfully copy.  The results are highly dependent on the test method, 
the receiver's IF bandwidth and the type of signal you want to copy.

All my tests are done to measure the noise the minimum strength CW signal 
that could be copied over through a receiver's internally generated noise 
even though most of the receiver I've lately tested, like my R-390's and 
SX-28's, I use for mostly AM reception.

But for comparing receiver noise floor and intermod performance I've found 
those that perform better using my CW method also perform better for AM when 
compared to other receivers.  So when comparing noise floor numbers make 
sure the tests were all done under the same conditions using the same 
method, or better yet the same test fixture or equipment.

I test a receiver, like the R-390's in my ER article, using a narrow CW IF 
bandwidth like one kc or less and for a carrier 3db above the noise.  I 
didn't use .1kc because I don't like that bandwidth for CW, it rings too 
much for my ears.  This was the method suggested in some Ham Radio and QST 
articles back in the 1970's for testing high performance CW receivers I used 
to build my fixture.

I use this method because when I built my fixture over 20 years ago  I was 
mostly using  CW and want to compare the date on noise floor and intermod 
for  the receivers I've tested in the past.  But to properly measure the AM 
performance of a receiver the tests should be run with a 5kc or wider 
bandwidth and a carrier of 10db over the receiver's internal noise.  This is 
the method used for most of the data you will see on the R-390's.  So the 
numbers in my article for noise floor will appear to be much lower than 
those quoted for most R-390 testing.  If I was starting over again I would 
run my tests using both the CW and AM methods.

The equipment to measure noise floor isn't hard to obtain or build.  You 
just need a well calibrated signal generator and attenuators.  But it all 
has to be very well shielded and carefully constructed.  You don't want any 
signal leaking around the attenuators due to a poor set up or cabling.

I was once testing a 75A4 I was working on and getting unbelievable low 
noise floor results.  That darn noise flood was below 160dbm which was 
impossible.  I had the bottom cover off the receiver and was using a digital 
frequency counter on my older HP signal generator.  I then put the bottom 
back on the receiver and the noise floor came up around 10db.  Then I 
unplugged the frequency counter form the raw output of the generator and 
retested the receiver.  Then the numbers were reasonable and match those of 
other tests using my home brew fixture.  The moral is a sloppy setup and an 
unshielded receiver's RF section can ruin your test results.

Also be especially careful of noise floor data in rice box adds with no 
explanation of their test method.  In all my testing because, I'm using a 
home brew fixture whose calibration could be questioned, I like to only 
compare my radios using this same fixture to be sure I'm doing performance 
comparison tests under the same conditions.

Incidentally, I've found good performing R-390's or R-390A's have a noise 
floor that is far below what is required on any HF band, 160 through 10 
meters.  They are also very good candidates for use with converts for VHF 
and UHF reception.  They also have outstanding intermod performance numbers 
along with frequency stability and selectivity.  The R-390 just complete 
amazes me for a 1949 design, just imagine what it's introduction must have 
done to the moral at the competing old line military receiver manufacturers. 
No wonder a lot of them sure faded away fast.

I guess that's why we all collect, use and love these receivers.

Bill N6PY

----- Original Message ----- 
From: "Perry Sandeen" <sandeenpa at yahoo.com>
To: <r-390 at mailman.qth.net>
Sent: Wednesday, March 29, 2006 10:11 AM
Subject: [R-390] Noise Floor question

In ads for the R390A the term of "noise floor close the theoretical limit" 
yada yada yada. is
often mentioned.  A search on google didn't yield anything comprehensible.

If we follow Roger Ruszkowski's method for achieving a 30db ratio, then by 
logical definition we
are lowering the noise floor by a ratio-metric method.

I know the receiver has a very low "noise floor" but how would us mere 
mortals measure it?  Short
the antenna input, terminate the audio out put with a 600 ohm resistor turn 
up the RF gain and
audio controls and then measure the AC voltage on the termination resistor 
with a micro-voltmeter?

If one did Chuck Ripple's audio improvements are we lowering the noise floor 
or just making a
better sounding audio output or both?

Inquiring minds ever seek knowledge for improving our radios.  Regards, 
Perrier

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
_____________________________________________________________
R-390 mailing list
Home: http://mailman.qth.net/mailman/listinfo/r-390
Help: http://mailman.qth.net/faq.htm
Post: mailto:R-390 at mailman.qth.net
Unsubscribe: http://mailman.qth.net/mailman/options/r-390 