Re: Signal strength puzzle

Allan Isaacs

Hi jdow

This all started when my long wire developed a fault somewhere along its length and one day I lost loads of signal.

I started experimenting to see if I could work out where the problem was and quickly noticed two computers simultaneously monitoring the SDR had totally different meter readings for the same signal.

I’m not concerned with absolute signal strength but only comparisons using different aerials so you can understand why some 20dB difference in readings is so confusing.


I fixed a longstanding network problem a few days ago that was quite interesting,, but probably unique.

Periodically my audio would start to get distorted accompanied by the buffer time increasing from 20mS to several hundred mS.

I could find nothing within the PC to explain this.

The answer turned out to be simple.

I was using an old wireless router as a network switch (not the wireless part) and I noticed on the back of the case it had the default network address

That is the same as my Sky router address and the firmware in the old router must have been confused and messing up the data flow on a timed basis.

I fitted a new standard switch and now the audio is perfect, never varying by more than a few mS.

Not a standard problem, in fact probably unique…




From: [] On Behalf Of jdow
Sent: 04 November 2020 09:59
Subject: Re: [SDR-Radio] Signal strength puzzle


It's pure magic. Leave it at that if you have to ask that question. It is the way the SDR software is designed. *NOTHING* is calibrated. You are trying to make it calibrated when it is not even close.

It is not network hardware, unless something 100 mps got into the picture. It would be the way the SDRServer was designed. Slightly different bandwidth or other settings can give you different results.


On 20201104 01:10:15, Allan Isaacs wrote:

Hi jdow

When you say “server” do you mean the network chip (KSZ8851SNL) in the Andrus SDR?

It’s decades since I designed data equipment and times have changed (I have a patent in the field US4035601A etc), but I’m afraid I know very little about the precise way modern networks operate. I imagined the Andrus output seen by other devices on the network would be identical so any dropping of accuracy must be down to one of the switches??

My network is 10/100Mbps as far as I understand but what goes on as far as packets and bits is something I’ve never bothered about.

It seems to cope with Netflix.


Allan G3PIY


From: [] On Behalf Of jdow
Sent: 04 November 2020 02:36
Subject: Re: [SDR-Radio] Signal strength puzzle


The server is built to work on slow networks. (Evidence shows it does work on slow networks, at least.) That means some data must get trimmed. One logical trimming is 8 bit data that has an AGC on it in the server. That ruins your "dBm" sort of baseline.


On 20201103 14:26:14, Michael R. Owen wrote:

Dear jdow,

Could you please elaborate on “Due to the nature of the server you will have to calibrate it separately as signal levels get changed as a result of the data trimming done in the server to make it practical.”

I don’t understand what you are saying. 




From: <> On Behalf Of jdow
Sent: Tuesday, November 03, 2020 5:16 PM
Subject: Re: [SDR-Radio] Signal strength puzzle


About what part of "dB" is not "dBm"? The VERY unfortunate thing is Simon labeling his axis as "dBm" when there is nothing in the system to calibrate it. And you confusion is precisely what Simon created mislabeling his signal strength axis.

Simon, PLEASE relabel that axis.


On 20201103 12:05:38, Allan Isaacs wrote:

I’m even more puzzled

Allan G3PIY


From: [] On Behalf Of jdow
Sent: 03 November 2020 18:18
Subject: Re: [SDR-Radio] Signal strength puzzle


You are making an assumption that is not supported by facts here. Te meter reports dB, not dBm. As it happens 0 dB happens when an exactly full scale peak to peak signal enters the A/D converter. If you want readings in dBm then you need to calibrate the scale yourself using a precision (by your definition of precision) source. Due to the nature of the server you will have to calibrate it separately as signal levels get changed as a result of the data trimming done in the server to make it practical.


On 20201103 08:14:20, Allan Isaacs wrote:

Hi Tom

I tried removing the switch in my workshop and feeding the SDR directly to the incoming CAT5 cable from my office (about 80 feet)

The office computer gave me the same results as before with the workshop switch in place


198KHz          -63dBm

252KHz          --84dBm

234KHz          -90dBm

60KHz            -107dBm


I then connected the SDR directly to the workshop computer, results the same as before

198KHz          -49dBm

252KHz          -71dBm

234KHz          -74dBm

60KHz            -95dBm


I then put back the workshop switch and used a second Windows 10 computer in the workshop.

Results exactly the same as the first 198KHz -49dBm etc


I haven’t tried bypassing the office switch yet because it’s in a rats nest of cabling, and a router plus the switch (the latter is brand new and carries two network cameras as well as a printer, the workshop CAT5 cable and a couple other cables not being used).


I’m a bit puzzled. I imagine the CAT5 cable from the SDR carries digital signals to the SDR program so should get decoded exactly the same in two computers. The only difference is I’m using Win 10 in the workshop and Win 7 in the office.

I’ll dig out a Win 7 laptop next and try that. Surely something must account for over 20dB difference in signals??


All signals are solid and changing AGC has no effect on the readings

I have 4 gain settings. On 60KHz the -30dB gain gives -113dBm, -24dB gain gives -107dBm, -12dB shifts the base line down about 10dB with the signal still -107dBm and 0dB gain shifts the baseline down another 10dB with the signal at -108dBm.

Allan G3PIY



From: [] On Behalf Of Tom Crosbie G6PZZ
Sent: 03 November 2020 15:11
Subject: Re: [SDR-Radio] Signal strength puzzle



I’m not familiar with the Andrus. My only thought might be latency in the network, but that’s probably clutching at straws. Can you bypass the second switch and see if it makes a difference?

I can’t think why it should but I’ve seen latency problems but only on much bigger networks operating at the limit of several CAT5 cable lengths with switches in between. Perhaps you have a rogue peripheral on the network?

I wondered if one computer is measuring the incoming signal at a slightly different point in the signal strength variation, e.g. is AGC making a difference? Although 198 is normally rock solid.

Have you tried looking at other stations and on other bands to see if the same phenomenon exists?

I can’t think what else could be causing this.

Good luck!




From: <> On Behalf Of Allan Isaacs
Sent: 03 November 2020 14:19
Subject: [SDR-Radio] Signal strength puzzle


I've been using my SDR which is an Andrus on my Windows 7 x 64 office computer and have recently used it on my Windows 10 x64 workshop computer.
The Andrus (powered by an adjacent 6 volt PSU)  is in my workshop and is fed with an 80m dipole which receives reasonably well over most bands.
Connection to the Andrus is via a local network switch to the workshop computer and via that switch to a second switch to my office computer.
Receiving Radio 4 on 198KHz on my office computer records S9 + 12dB or -62/-63dBm but on the workshop computer it recorded as S9 +24dB and -49/-50dBm.
The software is identical, Sept 2020 on both and all the settings are the same.
Why the difference in signal strength??
Allan G3PIY




Join to automatically receive all group messages.