Re: [softrock40] Si580 Group buy Output levels, Temp Stability Options
- --- mikecol2 <mikecol@...> wrote:
> Hi Tom,Mike,
> Thanks so much for helping with the group buy of the
> I had a couple of comments on the order
> The new group buy web page shows the LVDS phase
> noise performance but
> the parts you bought are the CMOS version. Not sure
> which way the
> group buy is going.
> My input would be to buy the 1st Option Code: B or N
> code which is the
> 3.3V LVDS (High or Low OE not important to me). The
> spec sheet shows
> the LVDS phase noise performance which is excellent.
> The CMOS version
> probably isn't that great since the edge speeds are
> 1ns typical ( the
> LVDS and others is 350 ps max ). Also to get an
> indication of phase
> noise performance SI has an AN291 describing the
> Output Buffer
> Options. As shown in Figure 7 CMOS has the poorest
> jitter performance
> (translation to phase noise). The figure also shows
> that CML and
> LVPECL would offer the best phase noise performance.
> I would still
> choose LVDS because of it's ease of use, low power
> consumption, and
> very common standard in newer designs.
> One other consideration would be the 2nd Option
> Code. My choice would
> be The code B option which is 31.5ppm total
> stability (compared to the
> code A option which is 61.5ppm). Since we are using
> these for our
> radio frequency control, I think I would be worth an
> addition cost to
> get the higher stability. I don't know if you have
> priced the
> difference but it sure would be worth considering.
> As far a initial freq. I don't care that much. I
> like that most have
> wanted it to be for 20M (i.e. 56Mhz++). If I had to
> select (for my
> most used mode SSB) I would select 57Mhz. Of course
> there will be
> many opinions on this.
> Thanks for the consideration of LVDS output and high
> stability in the
> part selection.
> What do others think? Don't you want the better
> phase noise
> performance and stability??
> Thanks, Mike Collins KF4BQ
Thanks for your excellent comments.
My goal in doing this was to add to the softrock
experience at minimum cost and simple implementation
with reasonable performance.
I read AN291 in my research and chose CMOS because it
seemed the easiest to use and is typically used below
160 MHz. Besides, I don't know how to interface LVDS
to the softrock. Also, CMOS draws less power which is
not a big issue. The noise performance of the CMOS
version is much better than any other chip that I am
aware of by quite a bit.
I am certainly not against the lowest possible noise
and best stability.
A couple of questions to consider. Will anyone notice
the slightly less noise? I know from my experience
that the stability is not a problem as I can copy
PSK31 all day long.
Since the demand is so high, maybe I could offer a
cheap version and a more expensive version for those
who want or need it.
But, I will try to do whatever the group wants.
So, sound off.
73, Tom KM5H
- ----- Original Message -----From: k5nwaSent: Thursday, January 10, 2008 7:27 PMSubject: Re: [softrock40] Si570 PCBAt 01:41 AM 1/9/2008, you wrote:
Here's the schematic for an Si570 PCB I'm working on:
The board includes a PIC16LF88 MCU, with a 3.3V LDO regulator, and has
connectors for a rotary encoder and two pushbuttons. It's mostly
surface-mount, and being single-sided will be easy to make at home. Size is
1.325" x 2.15".
Is that a ICSP connector to program the chip? Or is it some other standard.Yes, it's for an ICD 2 debugger/programmer.73, Leon