开云体育

I am assembling a portable RF bench setup with a nanaoVNA and 3 tinySAs among other things. I would very much like to be able to use a GPSDO (e.g. RFzero) to synchronize all of them.

With sufficient flash space, a one second tick sent over USB would probably be "good enough" in consideration of the other limitations.

Please consider this.

Thanks,
Reg

The frequency stability of the tinySA is better than the resolution of the resolution filters.
It is sufficient to calibrate the frequency once
--

On Sat, Jul 24, 2021 at 12:27 PM, Reginald Beardsley wrote:

I am assembling a portable RF bench setup with a nanaoVNA and 3 tinySAs among other things. I would very much like to be able to use a GPSDO (e.g. RFzero) to synchronize all of them.

With sufficient flash space, a one second tick sent over USB would probably be "good enough" in consideration of the other limitations.

Most higher end lab equipment uses an external 10 MHz. reference? input to sync the time base.? This requires additional circuitry in the test equipment for this function.? The NanoVNA and TinySA are low cost products that do not have this capability.

However they can achieve greater frequency accuracy if external frequency sources like a GPSDO are used to "calibrate" them. The TinySA has this feature in the current firmware and the NanoVNA-H and -H4 can be updated to a recent DiSlord version that has this feature.

Erik did a TinySA video on this subject >>?

Roger

A good bit of hobby kit takes a reference e.g. VNWA and SDRplay RSP2.

But if the FW can take pps ticks via USB then it can discipline the local oscillator. No HW changes are needed. Just room for a bit of SW to control the frequency based on the pps tick. USB latency adds a modest amount of complication, but nothing hard. A 1 pps tick is about as low an overhead as I can imagine.

Problem at present is there is no room left in an STM32F072 part. Hence no room for other developers without disabling a unit by removing existing functions.

My point in posting is this:

There are a lot of ancillary functions as well as increased accuracy available if the clocks of multiple units are synchronized closely. This does not need to be a costly feature. Count cycles per tick and adjust the clock appropriately. That's not a difficult task.


Have Fun!
Reg

On Sun, Jul 25, 2021 at 03:04 PM, Reginald Beardsley wrote:

Count cycles per tick and adjust the clock appropriately. That's not a difficult task.

Reg, can you send me the code for measuring the frequency so I can try to integrate it?
?
--