Re: Waterfall display slightly offset from actual signal strength display

Srinivas Chennupaty

Ron _ Fascinating explanation! Thank you!! the technology outstrips my capabiilty :-) 

On Sunday, May 23, 2021, 06:02:53 AM PDT, Neil Smith G4DBN <neil@...> wrote:

Hi Ron, how wide are the FFT bins you are using?   On the DUO at max zoom on the main waterfall, with a bin size of 11.7 Hz, the peak is around 6Hz off.  On the IF windows at 2.9 Hz bins, it is around 1.5 Hz off. The reason is just that the peak for my 10.00000000MHz signal is "somewhere" in a bin that happens to be from (perhaps) 10000000 Hz to 10000011.7 Hz instead of 9999994.65 to 10000005.35 Hz, when it would appear to peak at 10.000000 exactly.

IF window, showing peak at about 1.5 Hz HF of the 10.000600 Hz point (I'm using CW offset of 600 Hz). Sometimes, you'll find the peak is in the middle of a bin, but on average it will be around 6Hz off from the true frequency simply because that's how the maths work. (or "how the math works" for those who use the singular shortened form of "mathematics").

You should be able to calculate the bin edges and make any necessary adjustments.  However, if you really want to know the exact frequency, and your radio is locked to a precision 10 MHz reference, you can use Spectrum Lab on a long integration FFT (262144 bins perhaps) to check your PC sourd card clock calibration against a precise tone generated from your reference, then use Speclab to look at the actual frequency:

That is monitoring drift of an OCXO at 1kHz offset and 512k bins in Speclab. The DUO shows it at 1006 Hz offset approximately.

Neil G4DBN

On 23/05/2021 13:09, Ron Hunsicker via wrote:
Srinivas:  I see something similar with my S3. 

So we are not comparing apples and oranges:  I have the OCXO version and it runs 24/7, so it should have stabilized.  Also, I have a GNSS antenna attached and am advised that I have a lock. 

I just tuned to WWV on 10 MHz.  When I expand the main waterfall, the red arrow head is centered to the right of the carrier (higher in frequency) by something less than 10 Hz, but more than 5 Hz.  When I expand the IF waterfall, it appears similarly offset to the right.

I think that this is an interesting observation rather than an annoying or concerning one.  I have to keep reminding myself that the S3 is a radio, not a laboratory instrument.  Consider that, at 10 MHz, 7 Hertz is 7 parts in 10 million; less than a part per million.  (By all accounts, less than 1 part per million of almost anything will not cause cancer.)

Remember slide rule dials?  Remember tuning with a piece of white cardboard with small marks on it noting the known (We hoped!) locations of BBC, Radio Moscow, HCJYB, and VOA stations and interpolating between these marks?  The mark was probably 5 KHz wide, so, at 10 MHz, 5 parts in 10,000; 1 part in 2000. 

And we were happy.  Modern radios have spoiled me!

Ron Hunsicker

Join to automatically receive all group messages.