At 11:37 PM 5/3/2008, you wrote:
On the Si570 Group Buy page at:
<http://www.softrockradio.org/>http://www.softrockradio.org/
The specs there imply the LVDS version of the Si570 has 20ppm
stability vs. 50ppm for the CMOS Si570. This is a significant
advantage to the LVDS part in my opinion.
There are some studies of phase noise by Mike Collins KF4BQ in the
Files area. If my whiskey-soaked memory serves, the LVDS Si570 does
outperform the CMOS version in the tests, but it is arguable the
differences will be noticeable on HF in my opinion.
Given the LVDS part is only $2 more than the CMOS part, and it is
cleaner and significantly more stable. One would wonder why the CMOS
part would ever be used with a SRLite. Comments?
73's David
Both parts are available in 20ppm, it's just that the CMOS is being
bought in the 50ppm flavor to begin with, when Tom was thinking of
buying the LVDS I pushed for the 20ppm for better performance, many
other agreed so when the LVDS chips were bought they were 20ppm
parts. Sometime back in March I mentioned temperature stabilizing the
part like Flex does it with a thermistor, Lyle K0LR posted
experimental results 5 hours later and it worked extremely well, once
we receive the final board we need to try and see how well can we
insulate the chip and temperature stabilize it.
The difference in noise is supposed to be around 25 db which is a
significant amount, but in order to achieve it one must be very
careful with the design, the counter circuity must be designed
carefully, good grounding, very clean power, I would put the
SI570 on it's own voltage regulator that is used by it alone.
Tony will have to give you his reasons but I can guess, we humans are
creatures of habit. That is why many companies offer free samples to
get you to try their parts, I assume it went something like this;
* Tony is hearing good reports on the list on the performance of
the chip, having a variable frequency crystal has tremendous
advantage so he looks into it.
* Tom probably has both chips available, but after a discussion
he doesn't have LVDS to CMOS translator chips so he tries out the
CMOS version which is easier to interface if you don't have the
FIN1002 chips, it also cost less so that is a factor.
* He tries out the CMOS and lo it works well, so he designs it in
because that is what he has tried and it works.
* He comes out with the CMOS version of the crystal board, but
people with the LVDS version feel left out and ask about it.
* By now I have FIN1002 chips for the buy, Tony buys 10 to try them out.
* Modifies the final design to go either way.
* Tony offers the SI570 chip for future buys, offers the CMOS
version, why? It's $2 cheaper, and remember Tony likes to make his
products inexpensive, and $2 is quite a bit extra.
Since many have the LVDS and Tony made the design be able to go with
either chip the answer is simple buy the kit without the SI570 and
use the one you already have, to people that don't have a chip the
solution is also simple, buy the whole thing from Tony.
Next trick is to convince Tom to buy a cheaper version that has
20ppm, how can it be cheaper? This part is strange, the price of the
part varies quite a bit by what the default startup frequency is. It
should not make a difference but it does, so find the frequency that
cost the least. It really doesn't matter what the frequency is since
we program the chip to a new frequency first thing after power is applied.
Back to my cave.
.
Cecil
K5NWA
www.softrockradio.org www.qrpradio.com
"Blessed are the cracked, for they shall let in the light."
