开云体育

Hi Erik,
I guess this is an issue with the resolution.I measured at 10 MHz 25 Hz error, but at 300MHz isn't 750Hz (25Hz x 30) .


Please see the pictures. Is? not possible to adjust exactly 10 MHz because of the i.c. specification. Useful to remember as you just described in the tracking gen. topic.

The 10MHz adjustment works for high band also, I measured at 800MHz? only 100Hz error.
Of course I had the same results testing Tinysa with other measuring system, spectrum analyzer with rubidium reference.

See the spec


Its not the TCXO that is uncertain but a limitation of the PLL setting so the resolution does not depend a lot on the frequency
--

Yes I wrote, ?? because of the i.c. specification.
Anyway It's a little limitation , I tested at various frequencies and I make a correction table.

I posted these results earlier -? my measurements using a GPS disciplined oscillator as the frequency reference (1 ppb accuracy).? The error was small and consistent across the range:

TinySA Frequency Output Measurements	?	?
S/N SA-2009074	?	?
?	?	?
Frequency (MHz)	Error (Hz)	Error (ppm)
50	-17.5	-0.35
150	-53	-0.35
200	-70.8	-0.35
400	-139.8	-0.35
600	-208.9	-0.35
800	-278	-0.35

Hello Bruce,
thank you for your table.
When all the data are in the minus side, adjusting the 10MHz one? step lower(is the resolution), it is possible to reduce the error upper 400 MHz? 100 Hz or less
and have some data centered or positive.
I use calibrated rubidium so can measure/generate at 4GHz with less of 1 Hz error but, with these low frequencies, I have the same result of your good Gps.
So, I am thinking of an harmonic mixer to go in the microwave region with Tinysa, would be again a very interesting handy instrument.
Sandro

Sandro,
Be aware the problem is not an offset but a resolution issue. The output frequencies are on a fixed grid. So the maximum error will always stay the same.
The TCXO itself is rather accurate. Try to use your frequency counter to measure the error in the cal output as the cal output is directly coupled to the TCXO and does NOT have the frequency resolution problem?
The TCXO runs on 30MHz and the lower cal output frequencies are created using a divider instead of a PLL
The cal output is like to 10MHz reference clock of other spectrum analyzers.
If the cal output is offset you can correct it using the CONFIG/EXPERT CONFIG/MORE/CORRECT FREQUENCY where you can enter the actual frequency of the 10MHz cal output
--

On Wed, Feb 17, 2021 at 10:34 AM, Erik Kaashoek wrote:

The TCXO runs on 30MHz and the lower cal output frequencies are created using a divider instead of a PLL
The cal output is like to 10MHz reference clock of other spectrum analyzers.
If the cal output is offset you can correct it using the CONFIG/EXPERT CONFIG/MORE/CORRECT FREQUENCY where you can enter the actual frequency of the 10MHz cal output
--

?

Erik,

I measure my 30 MHz. cal output and it was 45 Hertz off.? So I measured the 10 MHz. cal output and it was 15 Hz. off.? I went to CONFIG/EXPERT CONFIG/MORE/CORRECT FREQUENCY and entered 9999.985 K and did not notice any difference in the 10 MHz. cal output.? So I thought maybe this is just a divide by 3 of the TCXO and this offset adjustment does not have an effect.? So I went to the Low Output mode and? set it to 10 MHz. and i was still 15 Hz. off.

Could you please provide further details on how to use this feature or point me to a video that explains proper procedure.?

Regards - Roger

The cal output is hard coupled to the TCXO. The correction factor is used only for the low and high input frequency calculations