FOX TCXO


Alan G4ZFQ
 

A couple of shots showing the FOX TCXO looking at the 27MHz output.
One shows the first 2 minutes of warmup from initial switch-on.
The other shows it warm but dithering, up/down a Hz or so, seemingly having no hysteresis. (This was part of my WSPR on 2m tests, made 2m WSPR useless.)

Just the one sample, has anyone else seen this?

73 Alan G4ZFQ


DC1MAK
 

Hi Alan,
I also use the Fox TCXO in all my QRP-Labs and own projects...
What equipment did you use 
for this measurements ?

This evening i will have a look on the output of my 
synths...

Several times i thought about a 27MHz source disciplined by 10 MHz ( GPS or Rubidium)

I'll post my experiences this evening

73 de micha DC1MAK


Graham, VE3GTC
 

Yes, I have seen this behaviour too.

I have been messing about with the QRP-Labs si5351a synthesizer board to
see if I could get a precise (i.e. steady, unchanging) signal from it,
actual frequency (accuracy) was less important.

the design works quite well whether GPS disciplined or not. Using a
standard crystal or oscillator it is (of course) subject to
environmental variations.

I tried the FOXX924B-27.000 TCXO as noted by Hans in the documentation
and as tried by others. Works OK but I did notice the "discrete
dithering" as you have observed. Couple that with GPS disciplining and
the actual frequency output jumped around in response. It seemed as
though the occasional discrete jump in the TCXO and the GPS trying to
correct where "fighting" each other for a bit. I can see where at VHF
this would be an issue.

However, one thing I did notice which seemed to make a big difference in
stability and this is termination of the synthesizer output.

I was testing using a qrp-labs si5351a synthesizer and prog-rock
monitoring the desired frequency on a frequency locked Icom R75
receiver. My "antenna" was just a simple short piece of wire
conveniently soldered to the synthesizer board output. this was all set
up on the workbench at the far end of the basement. I noticed that the
frequency I was measuring on the receiver was varying every time I was
near the workbench. It would change quickly - up, down, whatever, and
return to near normal after I left it alone. It was a bit of a "light
bulb" moment when it occurred to me that there was enough stray
capacitance and inductance in my body that was being coupled into the
short piece of wire antenna and that was effecting the
synthesizer/progrock to the point that it was causing these changes.

A bit of research on the subject and I found some reference to others
experiencing cross talk between the outputs of the si5351a and within
the documentation there are comments regarding the configuration of the
output drive of the si5351a and output impedance.

I am uncertain at what level Hans' software sets the si5351a to but I
would guess likely at 8ma level which (I think) corresponds to an
impedance of about 85 ohms.

In any case, I found that if I terminated the output of the synthesizer
module ( I used 50 ohms as it was handy) then the output become much
more stable and of course current consumed by the modules also increased.

You may find this document of interest:

https://www.sitime.com/api/gated/AN10028-Probing-Oscillator-Output.pdf

A quick search for similar documents on the subject will find others.

Those are simple qualitative observations.

I found a couple of much higher spec'ed TCXO 0.5 PPM which I have yet
to try. I have put the progrock and si5351a's aside for the time being.

cheers, Graham ve3gtc

On 12/11/2018, "Alan G4ZFQ" <alan4alan@gmail.com> wrote:

A couple of shots showing the FOX TCXO looking at the 27MHz output.
One shows the first 2 minutes of warmup from initial switch-on.
The other shows it warm but dithering, up/down a Hz or so, seemingly
having no hysteresis. (This was part of my WSPR on 2m tests, made 2m
WSPR useless.)

Just the one sample, has anyone else seen this?

73 Alan G4ZFQ





Alan G4ZFQ
 

I also use the Fox TCXO in all my QRP-Labs and own projects... > What equipment did you use> for this measurements ?
To monitor the 27MHz I used a FCDPro+ SDR and HDSDR. Any reasonably stable RX with Argo, SpecLab or even HDSDR set to mono input will do the same thing.
My GPS reference is simple, cheap, shown here https://sites.google.com/site/g4zfqradio/u-blox_neo-6-7 a u-Blox GPS which has a configurable output up to about 15MHz. For 27MHz I use the 3rd harmonic of 9MHz, maybe +/- a few Hz.

73 Alan G4ZFQ


Alan G4ZFQ
 

Graham,

However, one thing I did notice which seemed to make a big difference in
stability and this is termination of the synthesizer output.
Those shots were taken when the Synth module was installed in the U3S, I guess unknown load value? But constant.

You may find this document of interest:
https://www.sitime.com/api/gated/AN10028-Probing-Oscillator-Output.pdf
My "probe" was a small length of wire coiled under my custom U3S box which has the display at the top and back removed. So it was about 25mm behind the synth module. No measurements apart from frequency.

73 Alan G4ZFQ


freefuel@...
 

Hi Graham, what does the termination do for crosstalk between the outputs? 

-Justin N2TOH


HF
 

Hi Alan,
Yes, very similar to my observations.  Now that I have the Taitien TCXO installed, I get much more reliable decodes and almost always zero drift reports.  Another thing I observed is that the SNR reported in WSJT-X is now more stable.  I've also noticed that the Taitien's centre frequency drifts a little bit as the ambient temperature varies between 3 and 20 C.  I hope to have a chance to quantify that drift later this week.
Cheers
Halden NR7V


Graham, VE3GTC
 

Alan,

I would think that load on the synthesizer in a U3S would or at least
should be constant.

When the U3S drifts under transmit we always jump to the conclusion that
it is heat related as the TX warms up a bit but at a stretch - is this
always so?

Does the U3S actually provide a constant load to the synthesizer during
periods of not in TX mode and when in TX mode?

I know that is a bit of stretch and I have nothing more than a few
qualitative observations but this behaviour seems to fit in with some of
the comments I had seen regarding cross talk between output channels and
the effect on termination on that cross talk.

I would think that the si5351a being a digital device that once
programmed to an output frequency then that should be the output. I
don't know enough about the internals of the si5351a but might there be
something inside (a VCO or similar perhaps) that is affected by changes
in load?

Just a WAG on my part of an observation that made me go "hmmmmm ?!?!"

cheers, Graham ve3gtc

On 12/11/2018, "Alan G4ZFQ" <alan4alan@gmail.com> wrote:

Graham,

However, one thing I did notice which seemed to make a big difference in
stability and this is termination of the synthesizer output.
Those shots were taken when the Synth module was installed in the U3S, I
guess unknown load value? But constant.

You may find this document of interest:
https://www.sitime.com/api/gated/AN10028-Probing-Oscillator-Output.pdf
My "probe" was a small length of wire coiled under my custom U3S box
which has the display at the top and back removed. So it was about 25mm
behind the synth module. No measurements apart from frequency.

73 Alan G4ZFQ







Graham, VE3GTC
 

Justin,

Sorry, I don't have an answer to that question.

I had found an online posting (something I think to do with the BITX 40
Raduino maybe) where someone had noted cross talk between the three
outputs and someone else did some measuring of the device with the
conclusion that proper termination made a difference. Unfortunately I
cannot find that posting at the moment.

cheers, Graham

On 12/11/2018, "freefuel@gmail.com" <freefuel@gmail.com> wrote:

Hi Graham, what does the termination do for crosstalk between the outputs? 

-Justin N2TOH





Hans Summers
 

Hi Graham

I believe it's as I've said a few times. The U3S load on the Si5351A in TX and non-TX is not the same. 

During TX the Si5351A must drive the PA. It must drive the BS170 gate impedance. Which itself is lower as the frequency gets higher. The Si5351A also needs more power as the frequency increases, as CMOS takes power when it switches. 

During non-TX the Si5351A does (by default) operate Clk1 at the operating frequency. But that doesn't drive anything. So the Si5351A is consuming less power. 

More power during TX than non-TX. Means the Si5351A, as it is dissipating more heat during TX, increases its temperature. That effects the parameters of the oscillator section of the Si5351A, which pulls the frequency.

Hence the "park mode" which operates the Si5351A Clk1 at your chosen "park frequency" during non-TX. That keeps the Si5351A working hard in non-TX; because you can set the park frequency much higher than the operating frequency. So the Si5351A consumes more power due to the CMOS parts switching more often (higher frequency). Though Clk1 isn't driving any load. 

So the trick is to find a frequency, higher than the operating frequency, at which the higher current due to the higher frequency, is EQUAL to the Clk0 current consumption when it is driving the BS170 PA gate impedance. Then being no change in current consumption between TX and non-TX, there is no temperature change of the Si5351A silicon. Therefore no change in oscillator component characteristics, and hence no drift of the reference crystal frequency.

This is my understanding of the situation after all my observations of the U3S and Si5351A.

73 Hans G0UPL 

On Tue, Dec 11, 2018, 20:44 Graham <planophore@... wrote:
Alan,

I would think that load on the synthesizer in a U3S would or at least
should be constant.

When the U3S drifts under transmit we always jump to the conclusion that
it is heat related as the TX warms up a bit but at a stretch - is this
always so?

Does the U3S actually provide a constant load to the synthesizer during
periods of not in TX mode and when in TX mode?

I know that is a bit of stretch and I have nothing more than a few
qualitative observations but this behaviour seems to fit in with some of
the comments I had seen regarding cross talk between output channels and
the effect on termination on that cross talk.

I would think that the si5351a being a digital device that once
programmed to an output frequency then that should be the output. I
don't know enough about the internals of the si5351a but might there be
something inside (a VCO or similar perhaps) that is affected by changes
in load?

Just a WAG on my part of an observation that made me go "hmmmmm ?!?!"

cheers, Graham ve3gtc



On 12/11/2018, "Alan G4ZFQ" <alan4alan@...> wrote:

>Graham,
>
>> However, one thing I did notice which seemed to make a big difference in
>> stability and this is termination of the synthesizer output.
>
>Those shots were taken when the Synth module was installed in the U3S, I
>guess unknown load value? But constant.
>
>> You may find this document of interest:
>> https://www.sitime.com/api/gated/AN10028-Probing-Oscillator-Output.pdf
>
>My "probe" was a small length of wire coiled under my custom U3S box
>which has the display at the top and back removed. So it was about 25mm
>behind the synth module. No measurements apart from frequency.
>
>  73 Alan G4ZFQ
>
>
>
>
>
>
>




HF
 

Hi all

To summarize: Since the Si5351a's load changes at TX startup, its own power dissipation changes substantially as others have noted.  In other words, it warms up.  So, when the internal oscillator's load capacitance changes in response to the increased temperature, it pulls the crystal's resonance frequency.  When using an external oscillator instead of the original crystal, that changing capacitance is isolated from the oscillating circuit, eliminating the effect. 

I would like to offer an alternative theory that I think also fits the data:  The changing temperature of the Si5351a at TX onset puts some heat into the copper PCB traces that go over to the crystal.  The crystal then warms up a little bit and generates the observed chirp.  The PCB itself also moves some of this heat to the crystal, causing it to drift frequency a little bit during the entire TX.  Additionally, heat from the PA transistor could find its way over to the crystal, also contributing to drift during the TX.

Regardless of whether the internal capacitances are changing in response to the temperature rise, or the heat actually reaches the crystal, the solution is to use Park Mode to reduce the Si5351a's temperature changes.  I added a capacitor load to Clk1 on my U3S to mimic the load of the PA in an attempt to make Park Mode's dissipation more closely mimic that of TX mode.

If the latter theory is contributing part of the drift or chirp, thermally isolating the crystal could reduce the effect.  Once can lift the crystal off the PCB adding very thin wires to connect it.  Those wires would reduce heat transmission from the Si5351a.  One could glue a chip of metal to the crystal to slow down the heat gain from the PA.  One could wrap the crystal with foam to reduce the heat it receives from the PA.

If the former theory is dominant, maybe the effect could be reduced by programmatically minimizing the Si5351a's internal load capacitance and adding external NP0 capacitors on the PCB to give the crystal the load it expects.

Any thoughts on this?

Halden NR7V


KEN G4APB
 

Hi Hans,
can you provide some actual real life examples of say Tx freq and recommended Park freq?
I have never found this very effective but maybe I just never found an "ideal" park frequency. What is the highest freq I can set Park to in the U3S?

73 ken G4APB


Alan G4ZFQ
 

I would think that the si5351a being a digital device that once
programmed to an output frequency then that should be the output. I
don't know enough about the internals of the si5351a but might there be
something inside (a VCO or similar perhaps) that is affected by changes
in load?
Graham,

All output frequency variations I see are reflected by a move of the 27MHz crystal frequency. Use a TCXO, there is no movement, no output variation.
I feel that frequency changes alter the loading on the crystal, we are talking a very small effect.
My observations on 2m WSPR seemed to show the crystal frequency moving in a cyclic way with the WSPR modulation. I think the changes are too quick for temperature variations. One thing I did was to glue an aluminium block to the Si5351 with no noticeable effect.
I must check again using a quicker frequency monitoring system. If I can produce identical wavy lines for each WSPR transmission... with a fixed load on the Si5351.. and very carefully check current at differing frequencies but it's probably not easy to do this.

73 Alan G4ZFQ


Alan G4ZFQ
 

All output frequency variations I see are reflected by a move of the 27MHz crystal frequency.
I've done a few more observations and I think I'll stop now.
U3S with my particular synth module with a 27MHz crystal (and a 8x8x3mm aluminium block glued to the Si5351) Some of this confirmed by using the same module in a VFO with 50 ohm load:- Note these effects are not seen with a suitable TCXO.
Big changes in output frequency seem to have two effects, a quick change in frequency, maybe a Hz or so, followed by a slower drift.
When the U3S goes into calibration mode the 27MHz can change upwards by a Hz or so depending on the frequency of the preceding transmission. This would seem to limit the accuracy of the GPS calibration.

Small changes in frequency seem to have no effect. At 144MHz my synth is not particularly stable. The cyclic variation, +/- a few Hz, I saw on WSPR is there without modulation. 10x8x3mm aluminium glued to the crystal seemed to make no difference.

The VFO uses a different GPS adjustment, the 6.75MHz is on Clock 2 permanently, although it does move with large frequency changes I guess the huff-puff averages it better. My checks with PPS show the VFO settling rather closer to nominal (a few 10ths of a Hz at 27MHz) than the U3S. I'm fairly certain the clock 0 output moves less on big changes of frequency when GPS mode is enabled (but no PPS).

73 Alan G4ZFQ