Thetis S meter readings

[Pez]

Hi all,

In Thetis, I have noticed that the S meter reads incorrectly at low signal levels, and does not drop much below S3, regardless of the actual noise floor, which is extremely low at my QTH. Even when connected to my Marconi service monitor for RX tests, or with no antenna connected at all, or even when connected to a dummy load, the S meter just won't drop below about S3.  

Interestingly, the noise floor on the bandscope does visually show the correct noise floor level. As an example, I am seeing a noise floor on the scope of -136dBm, yet the S meter is still sitting there at S3. I was expecting to see the meter follow the HF "standard" but it currently isn't doing that:

For clarity, there is definitely not a -109dBm (S3) noise floor here. 

I have performed S meter calibration in Thetis previously, but it really didn't help this behavior.

Other software (SDR Console, Spark SDR, SDR++Brown) all read correctly without any calibration, as do my traditional rigs, so I am wondering if there is a Thetis issue, or perhaps there is a different Thetis calibration method that I have missed?

Thanks in advance for any help. 73

[Duncan Clark]

Hi Pez,

Try:

https://community.apache-labs.com/viewtopic.php?f=6&t=2463&p=2999&hilit=s+meter#p2999


Duncan

In message <edf18750-43ad-4f3e...@googlegroups.com>, Pez
writes

>Hi all,
>
>In Thetis, I have noticed that the S meter reads incorrectly at low signal
>levels, and does not drop much below S3, regardless of the actual noise
>floor, which is extremely low at my QTH. Even when connected to my
>Marconi service monitor for RX tests, or with no antenna connected at
>all, or even when connected to a dummy load, the S meter just won't
>drop below about S3.  
>
>Interestingly, the noise floor on the bandscope does visually show the
>correct noise floor level. As an example, I am seeing a noise floor on
>the scope of -136dBm, yet the S meter is still sitting there at S3. I was
>expecting to see the meter follow the HF "standard" but it currently isn't
>doing that:
>

>s9.png
>
>For clarity, there is definitely not a -109dBm (S3) noise floor here. 
>
>I have performed S meter calibration in Thetis previously, but it really
>didn't help this behavior.
>
>Other software (SDR Console, Spark SDR, SDR++Brown) all read
>correctly without any calibration, as do my traditional rigs, so I am
>wondering if there is a Thetis issue, or perhaps there is a different
>Thetis calibration method that I have missed?
>
>Thanks in advance for any help. 73
>

>--
>You received this message because you are subscribed to the Google
>Groups "Hermes-Lite" group.
>To unsubscribe from this group and stop receiving emails from it, send
>an email to hermes-lite...@googlegroups.com.
>To view this discussion on the web visit
>https://groups.google.com/d/msgid/hermes-lite/edf18750-43ad-4f3e-94a8-9ed2c3490e94n%40googlegroups.com
>.
>
>[ A MIME image / png part was included here. ]
>

--
Duncan Clark
G4ELJ

[Pez]

Thanks for that link Duncan. An interesting, technically very accurate, but slightly frustrating read. IMHO. 

Obviously, every other ham radio in the known universe measures S meter levels incorrectly, but Theitis is correct. We should all start our SSB S meters scale at 2 or 3 from now on... hihi (I joke)...  I understand these concepts rather well, however despite how "accurate" the technical facts are (I'm not challenging them at all - it is fact), I argue that the end result offers very little value to 99% of end users. But, each to their own. I love so many things about Thetis and have great respect for everyone involved in it's development, but the S meter does frustrate me. 

I would love an option in Thetis to change the S meter behavior to more closely reflect traditional radio S meter behavior (maybe we can call it "ITU S meter"?). Then it could display just the same readings as most other SDR software packages, and the also the same way all the traditional rigs do. Regardless how technically "wrong" that is, I think that many people would like that option to exist. Just my 2c... 

I can see now this issue has been discussed before on the apache forums so I won't flog the dead horse, and there is no offence to anyone intended by my comments. 

Best 73

[Pez]

One other thing to add: Given that changing the bandwidth in Thetis impacts the S meter level (as it is designed), I would assume the other SDR packages I referred to, are using noise power density (dBm/Hz), or similar. This would be independent of the bandwidth to calculate the S meter value. (Changing bandwidth has no significant impact on the S meter in these other programs). I understand dBm/Hz is not necessarily the "correct" way to do this. This topic seems to be argued till the cows come home. I find it very interesting, but I definitely don't want to start any wars like I have just seen in some other forums!. :)

73 

["Christoph v. Wüllen"]

I guess there is confusion because there are two different "dBm" values.

a) the signal in the panadapter is a power *density*, which has, in
principle, the unit "dBm per Hz". It is usually normalized by
multiplying with the width of one pixel of the screen (in Hz)
As such, the noise floor in the panadapter should change with the
width of the panadaper. However, since you cannot change this
easily by a factor of 10, you won't notice it.

b) the S-meter reading (in dBm) is an integral of the signal,
multiplied with the filter curve, over the whole frequency
range. For "brick wall" filters, the S-meter reading you should
expect is the noise floor times the filter width.

Clearly, the noise floor does not depend on the filter, but the
S-meter dBm reading jumps by 10 dB if you switch from
a 250 Hz CW-filter to a 2500 Hz SSB-Filter. This *must* be.

c) Your table is correct (below 30 MHz), but the "dBm" is the
integrated signal.


And, every other ham radio does it hopefully correctl, since what you
measure in a traditional radio (behind the filter) is the power
that comes out, and this is an integral of the power density.

Note the notion of "S9" changes for the 6m band, here -93 dBm integrated
power density should be S9, while it is -73 dBm below 30 MHz.

> To view this discussion on the web visit https://groups.google.com/d/msgid/hermes-lite/06dda241-7437-4cd4-a457-c5c1c9dd5893n%40googlegroups.com.

["Christoph v. Wüllen"]

> Am 28.07.2023 um 14:37 schrieb Pez <perrin...@gmail.com>:
>
> One other thing to add: Given that changing the bandwidth in Thetis impacts the S meter level (as it is designed), I would assume the other SDR packages I referred to, are using noise power density (dBm/Hz), or similar. This would be independent of the bandwidth to calculate the S meter value. (Changing bandwidth has no significant impact on the S meter in these other programs).

If you go from 250 Hz filter to 2500 Hz filter and the S-meter does not change,
the program is broken.

Period.

You also expect this change for a traditional analog
radio when there is white noise at the input.

Christoph.

[KG4ODA]

If the table you posted is what Thetus reports as S value versus dBm input then it's precisely correct.

Proper value of each S unit = 6 dB. Icom and Yaesu use 3 dB per S unit which is incorrect and is likely done for "marketing" (mah receiver is so quiet).

If receiver will not go below -109 dBm then it means you are at the noise floor of 12bit ADC and will need a preamp to go below that at the expense of strong signal tolerance. If you are seeing -136 dBm in spectrum display it means the resolution bandwidth is set about 400 times narrower versus your RX filter passband. Assuming your RX filter BW is 2800Hz then RBW is set somewhere around 7 Hz. This behavior is normal and your radio is working as it should.