Jacques
dB is always referenced to power. The voltage or current forms of the equation
assumes that the impedance is the same for both input and output. In this
particular case, it is not.
The elaborate demonstration you mention took time to build and perform so I
took the time to describe it. I used 1% resistors because they were at hand
and would provide more accuracy to match the calculator results. There are
many ways to skin this "cat", as long as the same conclusion is reached, the
journey, interesting or not, is less important then the destination.
The Collins spec is still correct. There may be a clerical error in the
heading for the balanced and unbalanced paragraphs but the engineering is sound.
Jim
Logic: Method used to arrive at the wrong conclusion, with confidence. Murphy
On Wednesday, October 23, 2024 at 10:35:28 AM CDT, Jacques Fortin
<jacque...@videotron.ca> wrote:
Hi Jim,
I feel that you went to a unnecessary elaborate demonstration to find the
(power wise) attenuation of a DA-121/U.
I cannot understand why you had not replicated the resistor values used in the
DA-121/U because the discussion starting point was to evaluate the DA-121/U
output vs the output setting of the SG connected to it.
In your experiment, it is clear that two DA-121/U connected back to back will
provide a POWER attenuation being the double of what one unit will provide.
But the initial Larry Haney concern was about the VOLTAGE output of a DA-121/U
in the R-390A balanced input (125 ohms load).
The calculated attenuation IN VOLTAGE measured by Larry of the DA-121/U is
correct as being 0.56.
If all the resistances values are EXACT (68 and 100 ohms) the calculated output
voltage across the 125 ohms receiver impedance is 0.5676 of the SG output
setting.
When Larry mentioned "5 dB" below, that is what his AC voltmeter tell him from
it's dB scale, (as being 20 log 0.56).
But what he used is a VOLTMETER, not a POWER meter: calling that -5 dBv will be
correct, but it cannot be translated as a POWER ratio because the input and
output impedances are not the same.
The real POWER attenuation of the DA-121/U is (calculated) 8.896 dB, close
enough from the 9dB you expected as the result of your own experiment.
When you mention an "experiment" about inserting a 50 ohms resistor in parallel
with a 50 ohms coax, it will give the same result as loading the output of the
source SG with a 25 ohms value.
One of the 50 ohms resistors used will receive 3.52 dB less signal (10 log
0.44444) than if it was left alone as the load of the SG.
Not even necessary to involve the 50 ohms coax SWR (or the Gamma, for that
matter) in the demonstration.
About your statement:
" Transmission lines behave differently then DC circuits"
I have a funny conceptual experiment for it, that I used as the introduction to
the course about transmission lines that I gave to my students at the time:
Let's figure that you have a PERFECT 50 ohms coaxial line (no power loss of any
kind inside the line) for which the electrical signal velocity inside is 70% of
the speed of light (Teflon dielectric) and that this "perfect" transmission
line is 210000 km long.
You take an ohmmeter (DC instrument) and connect it to one side of the line,
between the center conductor and the braid:
a) Does the meter reading will be 50 ohms ?
b) If yes, for how much time ?
c) If the other side (the end) of the line is shorted, what will be the
measurement ?
d) And what if the end is open ?
e) And what if the end is terminated in 50 ohms ?
I will provide the answers later.
All the remaining of the course I gave was to explain these answers, and more.
73, Jacques, VE2JFE in Montreal
-----Message d'origine-----
De : r-390-boun...@mailman.qth.net <r-390-boun...@mailman.qth.net> De la part
de Jim Whartenby via R-390
Envoyé : 22 octobre 2024 18:36
À : Larry Haney <larry41...@gmail.com>; r-390@mailman.qth.net
Objet : Re: [R-390] Official specs
Larry
I built a test fixture that is essentially two DA-121's connected back to back.
Photos and drawing are enclosed. This does the conversion from 50 ohms to 125
ohms and then back to 50 ohms. I used 1% resistors to make the attenuator
circuit with the values close to those found here:
https://k7mem.com/Res_Attenuator.html
The closest I could come to the 64.18 ohms result from the attenuator
calculator was 63.9 ohms. This is from the parallel combination of 3 each 237
ohm in parallel with a 1k, in parallel with a 499 ohm resistor. Five resistors
in parallel, all 1% resistors. The result was 63.85 ohms, a 0.5% error. The
sub for the 96.83 ohm resistor is a 100 ohm 1% resistor (3% error) and the sub
for the R-390's 125 ohm impedance was a 121 ohm 1% resistor (3% error). This
is still much better then the 5% resistors used in the original DA-121.
For a test oscillator I used a Helper SM-1000 signal generator and measured the
insertion loss with a Stoddart NM-25T frequency selective voltmeter. The
insertion loss was measured at 10 MHz using two 4 foot BNC RG-58 coax cables
from Pomona Electric. 4 foot of coax from the SM-1000 to the test fixture and
another 4 feet from the test fixture to the NM-25T.
The SG was set for a reading of 30 dB on the NM-25T signal strength meter when
measuring a BNC through connection and then measured 11 dB when the test
fixture was installed in place of the BNC through. The insertion loss for the
test fixture is 19 dB. Dividing this by two since there are essentially two
DA-121s back to back gives an insertion loss of about 9.5 dB for a single
DA-121. This closely agrees with the attenuator calculator findings.
So it seems that your DC circuit calculation do not agree with the RF
measurements. Transmission lines behave differently then DC circuits. You
calculate a 5 dB insertion loss, I measure a 9.5 dB insertion loss.
Here is an experiment that you can try. Insert a 50 ohm resistor in parallel
with the 50 ohm coax. What do you think will happen? Perhaps nothing since
the coax is 50 ohms and the resistor is also 50 ohms? In reality, the coax has
reactive elements, parallel capacitance and series inductance that make up the
coax impedance. Neither of which will dissipate the signal carried on the
coax. The only losses are from the resistance of the conductors that make up
the coax. Adding a parallel resistor will attenuate the signal to the receiver
by 3 dB.
If anyone on this list wants to make their own version of the DA-121, I can
supply the resistor values I used for a token $2 plus postage. Just DM me with
your address and if you want one or two resistor sets.
Regards, Jim
Logic: Method used to arrive at the wrong conclusion, with confidence. Murphy
On Friday, October 18, 2024 at 05:36:08 AM CDT, Larry Haney
<larry41...@gmail.com> wrote:
Hi Jim, I just checked and I only have 1 da-121. As for insertion loss, my
coax is very short and the connections are very good so the loss there would
not be possible for me to measure. Now for the insertion loss due to impedance
mismatch (due to resistance variations) would also not be possible for me to
measure, as I don't have the equipment required for that. But, because the 3
resistors in the circuit are very close to the required values for a perfect 50
ohm match to the sig gen, I am sure that the insertion loss due to that very
slight impedance mismatch is extremely small. I have no way to measure that
loss as I don't have the 3 exact value resistors to compare it to. I could
calculate it, but I believe that would be a waste of time without being able to
measure it.
After all the input you have given me and the research just done, I'm satisfied
with my current measurements and calculations (IE: the output voltage of the
da-121 is 56% of the input voltage when the load is 125 ohms).
My biggest concern about making snr measurements is for those folks that don't
have a recently calibrated sig gen or calibrated rms AC voltmeter to verify
their readings with.
Regards, Larry
On Thu, Oct 17, 2024 at 1:55 PM Jim Whartenby <old_ra...@aol.com> wrote:
LarryNo, just one SG and one 125 ohm load. You should be able to determine the
total loss through two DA-121 attenuators connected back to back with an
o'scope and then divide the loss by two to solve for the insertion
loss.JimLogic: Method used to arrive at the wrong conclusion, with confidence.
Murphy
______________________________________________________________
R-390 mailing list
Home: http://mailman.qth.net/mailman/listinfo/r-390
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:R-390@mailman.qth.net
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
