Si570 temp / freq stability

[4Z4TJ]

I have started operating WSPR with my MOBO and found that the drift is
much higher than with my FT-450.

The housing for the xcvr is quite small and cramped but I did mount a
fan on the rear panel.

If I do not use the fan the temperature as shown on the LCD rises from
about 30 C to 45C, with the fan about 10C less, to 35 C.

Drift reported up to 4 Hz on start up but seems to stabilize and be
less after the rig has been operating for some time. 2 Hz on 14 MHz
is not much in ppm and would certainly not be noticed on SSB. If I am
reading the spec sheet correctly we are well within the tightest
temperature stability that SiLabs promises.

I have some Murata 1608 series thermistors here and was thinking of
gluing one on to the Si570...

To summarize the above rambling: if I keep output power down to about
1.7W and run the fan mean reported drift is -1 while without the fan
and at higher xmit power drift increases to +/- 4 Hz.

Interim conclusions: If WSPR is a prime operating mode, a larger case
should be used with a larger fan and pay some attention to thermal
design layout.

Would welcome some suggestions because I sure do not want to move the
boards to a new case.

Tnx Earl

[Alex Lee]

You can consider using an "oven" for the Si570. There was a
description in the past in the Softrock40 list about a temperature
controlled Si570.

One possibility is to remove/disable the Si570 on the SRv6.3, and
provide an temperature compensated Si570 in its separate little pcb,
and route the Si570 clock with a short coax cable to the SRv6.3.

Another possibility is to request Loftur to take a look into this
issue. With the ATMega32U2 cpu or the sdr-widget cpu, there is enough
spare FLASH memory to include a temperature correction scheme in the
Mobo firmware. Basically the firmware will take into consideration
the Mobo temperature (note that it is a very indirect relationship
between the RD16 heatsink temperature and the temperature at the
Si570) to provide a small correction factor to the Si570 frequency.
The precise "correction" factor can be set/changed via a modified
MoboControl.

73 Alex

[VK5ZAG]

I also seem to remember an "oven" for the Sis570 being
spoken of. The USB synthesizer using an Sis570 even
has a PC connection point for an Oven.
I think someone soldered a thermistor to an Sis570
and reported better temperature stability.

A commercial marine transceiver I use has a bipolar
transistor as a temperature sense and a larger bipolar
transistor as a heater element both bonded to the crystal
with a simple circuit

I would like to see a frequency standard and comparator
with feedback correction though the CPU. This might not be
as difficult as it appears.

With the current interest in WSPR and frequency stability
this might be popular. The need for stability will only
increase with time and newer digital modes.

Regards Eric VK5ZAG.

[Alex Lee]

Hi Eric et al,

George and others have discussed the possibility of a freq standard to
sync the Si570 to:

(1) GPS based - the 1s tick from the GPS can be used to sync the
Si570 freq. This has been done in at least several projects.
However, the problem is having to have an "outdoor" GPS antenna to
have a reliable GPS signal.

(2) Internet based - this is using an "internet" timer server to
derive a sync signal. However, the problem is the latency of the
network. This makes it feasible for coarse timing (as is used by WSPR
to start the 2 min intervals), but not for very accurate applications.

In the end, the "oven" based approach seems to be the simplest.

73 Alex

[Chuck Nolin]

Hello Eric,

I wanted to pass along a circuit that is used on one of my kits that I
have completed a few months ago. The Genesis G59 uses that setup to
keep the si570 at temp. It is an oven. You can check out the schematic
on line at the web page. The Genesis kit page is in Australia.
The circuit is in Phase Eleven, Si570 Quasi Oven Circuit. It is a very
simple, but very reliable
design. You might talk to Tasa about using some of his design to work
into you needs. He is very helpful and one of the earliest designers
of SDR.
I hope you find the information you need.

Chuck Nolin
WA6LTV/4

--
Thank you,
Chuck Nolin
WA6LTV/4

[Alan]

I have used the Khune heater
<http://www.kuhne-electronic.de/en/products/crystal-heater.html>

This is a ready-made module that can be mounted on the Si570 with something
giving some slight pressure to ensure contact. I have found it reduces drift
by 2 or 3 times. Not perfect, but better.
Some have claimed good results with thermistor crystal stabilisers. Si570s
vary quite a lot some have reported very little drift after stabilising the
temperature. I would suggest putting one in a separate oven, there will
always be heat loss to the PCB that cannot be controlled. They will also
drift after a frequency change.
73 Alan G4ZFQ

[Ed]

Hello All,

Regarding Oscillator Frequency Stability,  I have a new product about to come on line that you may be interested in.

 

It has been specifically designed for the Si 570, but can also be used with other TCXO and Crystal Can Oscillators.

It is a small unit the size of a postage stamp. 

I have Si 570 Grade ' B '  parts currently demonstrating Frequency Stability of   0.012 ppm. This performance is about to be further improved.

Keep tuned on the scottie-sdr  web site for the final details.

Regards

Ed GM3SBC