Well, mailman not only strips photos but also any changes in typeface so
the larger type and bold letters are stripped as well.
Logic: Method used to arrive at the wrong conclusion, with confidence.
Murphy
On Monday, October 7, 2024 at 10:53:50 AM CDT, Jim Whartenby via R-390
<r-390@mailman.qth.net> wrote:
GianniComments in BOLD
On Monday, October 7, 2024 at 02:43:53 AM CDT, Ing. Giovanni Becattini
<giovanni.becatt...@icloud.com> wrote:
Hi Jim and thanks for your patience, but I still don’t understand.
The generator sees 50 // (100+125), 50 // 225 = 52.22 ohm No, this
should be 40.9 ohms not 52.22 ohms!
R in parallel = 1/ (1/50 + 1/ 225) so 1/ (1/50 + 1/225) = 40.9 ohms or
if you prefer
product over sum = (50 X 225) / (50 + 225) = 11,250 / 275 = 40.9 ohms
The result of two resistors combined in parallel will always be a lower
value then the lowest of the two combined resistors.
Going a step further, 40.9 ohms in parallel with the Signal Generator
impedance of 50 ohms results in an impedance of 22.5 ohms so the SG now
sees about half of the expected impedance. You can think of the SG as
being a current source feeding a 50 ohm resistor. If the current source
now sees half the expected impedance, the voltage output of the SG will now
be half of the set voltage.
Now applying the voltage divider rule to the series 100 ohm and shunt 125
ohm resistors, the voltage across the 125 ohm resistor will be the SG
voltage X (125 / 225) = SG voltage X 0.55 We already know that the signal
generator output is half of what the SG attenuator says so 0.5 X 0.55 =
0.27 so the output of the DA-121 is now approximately one fourth of the SG
dial setting.
I checked with the VOM using a 125 ohm terminator instead of the R-390A
and read 52.4 ohm. You must have a wiring error! The DA-121 should read
approximately 40.9 ohms at the SG terminals when the DA-121 output is
terminated with 125 ohms. This is what was calculated above. If you now
put a 50 ohm resistor across the DA-121 input, the resistance of the input
to the DA-121 should measure approximately 25 ohms.
To be sure that I was not tricked by the cables, I made the same test at
100 kHz with 10 mV and that below is what I read, again using the 125 ohm
terminator on the oscilloscope side.
Probably I am doing something wrong, but what?
Your experimental data should closely agree with the math, it does not.
There is at least a simple wiring error or the BNC to TWINAX adapter is not
wired properly. As I mentioned in the email below, one of the TWINAX pins
should be directly connect to the BNC center conductor, the other TWINAX
pin should be directly connect to the shell of the BNC connector. There
should not be any measurable resistance, ideally a short circuit for both
ohmmeter readings. Can you verify this?
Thanks again
Gianni
Regards,
Jim
Il giorno 6 ott 2024, alle ore 17:05, Jim Whartenby<old_ra...@aol.com>
ha scritto:
Gianni
There is something wrong with your measurements. They do not agree with
the mathematical analysis.
Reducing the resistances of the DA-121 with the input resistance of the
R-390 to a single resistance results in the total resistance seen by the SG
of 25 ohms. So the generator output should fall from 10 mV to 5 mV which
you confirm although there is an error of some 14% ((5.7 mV - 5 mV) / 5
mV). But as you say, the resistors are not perfect.
What is apparently the problem is that your adapter from BNC to TWINAX
does not measure correctly. One TWINAX pin should connect to the BNC
center pin and the other TWINAX pin should connect to ground. If this does
not happen, the second voltage divider, the 100 ohm in series with the 125
ohm is not connected to ground. This error would give you the voltage that
you measure.
There is agreement between us that when the 68 ohm resistor is connected
to the SG that the output will fall from 10 mV to about 5 mV. Putting the
two remaining resistors into the circuit results in a series 100 ohm
resistor and a parallel 125 ohm resistor. Applying voltage divider
analysis to this we have (5 mV X 125 ohms) / 225 ohms) which equals 2.28
mV. 2.28 mV divided by 10 mV gives a ratio of 0.23 which is in agreement
with the DA-121 reducing the SG output from 10 mV to 2.5 mV or 4:1.
The above analysis agrees completely with figure 3, the analysis of a
T-pad, which was done in the 1950s. It changes the SG impedance of 50 ohms
to the receiver impedance of 72 ohms with a voltage loss of 4:1 which I
again enclose in this email.
Regards,
Jim
Logic: Method used to arrive at the wrong conclusion, with confidence.
Murphy
On Sunday, October 6, 2024 at 01:46:02 AM CDT, Ing. Giovanni Becattini
<giovanni.becatt...@icloud.com> wrote:
Hi Jim,Setting rge SG to 10 mV I have1) with no terminator oscilloscope
side: 20 mV2) with 50-ohm terminator: 10 mV3) with DA-121 no terminator:
11.4 mV4) with DA-121 and 125 ohm terminator (which simulates the
receiver): 5.7 mV
exactly as I would expect. Now I am going to pickup another generator to
see if it behaves like the 8640.
In the afternoon I tell you the result of the test.
YoursGianni
Il giorno 6 ott 2024, alle ore 00:00, Jim Whartenby<old_ra...@aol.com>
ha scritto:
Giovanni
If you measure twice the voltage with no load on the SG then the actual
voltage when the SG is properly loaded with a 50 ohm termination, what does
the meter read when you put a 25 ohm resistor on the SG output? It should
now read a third of the unterminated SG voltage.
Enclosed is page 51 of the Measurements catalog. Figure 3 shows a T pad
to match 50 ohms to 72. The resistor values are chosen to reduce the SG
output voltage by half at the input to the T pad and to 1/4 at the output
of the T pad when the T pad is terminated with a 72 ohm resistor.
The same is done with the DA-121 but the impedance transformation is now
from 50 to 125 ohms. Can you measure the voltages at the output of the SG
with an oscilloscope? It should be 2X of the SG meter reading with no load
on the SG, 1X with a 50 ohm load and 1/4X of the SG meter at the output of
the DA-121 when the DA-121 is terminated with a 125 ohm non inductive
resistor in place of the R-390A. If you do not terminate the DA-121 with a
125 ohm load then what you report as 0.56 of the SG meter reading would be
correct.
Regards,
Jim
Logic: Method used to arrive at the wrong conclusion, with confidence.
Murphy
On Saturday, October 5, 2024 at 03:14:58 PM CDT, Ing. Giovanni
Becattini<giovanni.becatt...@icloud.com> wrote:
Thanks for replying, I am very intrigued by this theme.
See below please and tell me your opinion.
Il giorno 5 ott 2024, alle ore 20:33, Jim Whartenby via R-390 <
r-390@mailman.qth.net> ha scritto:
Giovanni
I need some clarifications.
1) You said: "It attenuates the signal voltage of 0.56 V, i.e. 5 dB."So
the Signal Generator (SG) meter indicates that the output voltage is 0.56
volts or are you are measuring 0.56 volts at the output of the DA-121/U
when the SG meter reads 1 volt? If so, how are you measuring this
voltage? Is it peak or peak to peak or RMS? The assumption here is that
it is RMS.
I wrote wrongly; I meant that the DA-121 is a voltage divider that,
considered 125 ohm the input impedance of the receiver, multiplies the
siggen voltage x 0.56.
2) The DA-121/U contains two resistors, a 68 ohm resistor in parallel with
the signal generator output and a series 100 ohm resistor to the center pin
of the BNC output connector. You are then adapting the BNC output
connector of the DA-121/U to TWINAX and then connecting it to the balanced
RF input connector on the back of the R-390A, correct? Yes
3) What are the two resistor values in the DA-121 when you measure with
your DMM? How close are they to what is expected? I am guessing that
these two resistors are carbon composition and are a bit off in value. It
is interesting to note that carbon composition resistors will change value
when soldered into a circuit. No, it is not the original, I built it with
new components.
4) When you measure the BNC to TWINAX adapter, one of the TWINAX pins goes
to the center pin of the BNC connector and the other TWINAX pin goes to
ground? Yes Both read close to zero ohms? each other yes, but they are
open to ground.
5) How old are the coax cables used in your measurements? In other words,
how lossy are they? Coax ages so the cable losses will increase and it
will have an affect on your measurements. The coax is 50 ohms? Yes, they
are normal BNC/BNC, 1 meter long, with 50 ohm cable, bought new ready to be
used.
The way I see it, 50 ohms in parallel with 68 ohms = 29 ohms. 29 ohms in
series with 100 ohms = 129 ohms which is approximately your impedance
transformation needed from 50 to 125 ohms. Because of the 68 ohms is in
parallel with the SG output, the voltage at this point should be half of
what the SG meter indicates. I am not sure it is so. The siggen indicated
voltage is in Vrms and it is true when you have a 50 ohm load. If you don’t
have the 50 ohm load, the voltage is double. I am sure of this, I tested
more times with different generators and oscilloscopes.
The second voltage divider of 100 and 125 ohms is again reducing the SG
output voltage by another half so the actual voltage applied to the
receiver is 0.5 X 0.5 or 0.25 times the SG meter reading. In other words,
actual voltage applied to the R-390A receiver is 1/4 of what the SG meter
indicates or 12 dB down. No, I am sure of 0.56. In the doubt, I built a 125
ohm terminator and checked with the oscilloscope. Starting with 10 mVrms I
read 5.7 mVrms because the resistors are not perfect. And thus reduces the
voltage by 5 dB. Do you agree?
So what this means to the original discussion is that the 6.5 microvolt
limit in the R-390A specification is actually 1.6 microvolts that is
applied to the R-390A balance RF input for a 10 dB S+N/N reading when all
of the losses in the test setup are accounted for. So the spec has
simplified the measurement and eliminated all of the above math. Again,
spec is spec and those who wrote it knew what they were doing.
This back of the envelope analysis does not agree with what you have
measured. I am interested in what you find when you have a chance to take
a closer look.
Jim
Logic: Method used to arrive at the wrong conclusion, with confidence.
Murphy
On Saturday, October 5, 2024 at 01:48:09 AM CDT, Ing. Giovanni
Becattini<giovanni.becatt...@icloud.com> wrote:
Hi Jim and thanks for your reply. I read the very interesting document
you pointed out. I did not understand everything, but for my practical
interest it confirms that the impedance matching is mandatory.
I am using an HP8640B as a signal generator. Let’s suppose it is ideally
calibrated. I use also the DA-121/U impedance adapter which shows 50 ohm to
the siggen and 125 to the receiver. It is the fourth type of pad of figure
4 of the article.
My practical question is how to take in account the DA-121?
It attenuates the signal voltage of 0.56 V, i.e. 5 dB. So,
- in volts: the voltage value for the 10 dB S/N I read on the
generator’s scale should be multiplied by 0.56.
- in dBm: the dBm value for the 10 dB S/N I read on the generator’s
scale should be reduced by 5 dBm.
Is this correct?
Thanks
______________________________________________________________
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
______________________________________________________________
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
