[minima] Re: Si5351
- From: "Joe Rocci" <joe@xxxxxxxxxx>
- To: <minima@xxxxxxxxxxxxx>
- Date: Thu, 30 Oct 2014 07:30:02 -0400
Mac
I did a quick and dirty look at the noise floor using my Tek 495P analyzer and
it was at least as good as or better than the analyzer's own LO noise.. Maybe
over the weekend I can do some more quantified testing using a crystal notch
filter and a low noise amp to jack up the dynamic range of my measurements.
I'll have to get the rotary encoder working for that so I can dial in the
crystal frequency very closely.
The bigger issue for me right now is the spurious outputs. My belief is that
they're either intermodulation products of the 2 active output channels
produced inside the chip, or possibly reference sidebands from the fractional-n
synthesis. They're definitely not coming from the Arduino board, as I powered
the Si5351 board separately from a lab supply and turned off the power to the
controller board after the chip got loaded from I2C - no change. Maybe if I get
some time tonight I'll look at it with all but one channel turned off; this
will tell if it's intermodulation related. If this turns out to be inherent,
then the idea of one Si5351 to cover both the VFO and the BFO is probably kaput
for anything except a low-expectations receiver. However, even if 2 chips are
required, it's still cheaper than an Si570 and much more flexible.
I've been corresponding with some folks over at Jason's site and also on the
SiLabs forum. I'll report if I get any useful info.
Joe
----- Original Message -----
From: Mac A. Cody
To: minima@xxxxxxxxxxxxx
Sent: Thursday, October 30, 2014 12:06 AM
Subject: [minima] Re: Si5351
Joe,
I obtained one of the Si5341B devices (8 outputs with VCO input) awhile
back, but I have not done anything with it due to other activities.
I've studied the part specifications and I believe that the phase noise
may be significantly higher than that of the Si570. I just can't
quantify exactly how much more, though.
The Si570 specs quote a phase jitter of 0.62 ps typical for the CMOS
part. The Si5351 specs quote a period jitter and cycle-to-cycle jitter
of 50 to 70 ps, typical. These are apples-to-oranges comparisons, but
that is about a 100X difference. Even being generous and saying there
is only a 10X difference, it is still significant. I've downloaded
some of the Silicon Lab application notes on phase noise, but I've not
sorted through them yet. Still, Jason at NT7S.com has been using the
part in some of his QRP experiments. He thinks it might be a usable
part, but hasn't answered the phase noise issue either.
Hope that this helps.
73,
Mac AE5PH
On 10/29/2014 01:06 PM, Joe Rocci wrote:
Hi DuWayne!
I purchased the AdaFruit board with high hopes, but my initial spectrum
analyzer results are a bit disappointing. At first, the 2 active outputs (I'm
using the NT7S 2-channel example code) were crosstalking so that the Ch0 output
was leaking into the Ch1 output only 40 db down. I sort of cured that by
reprogramming to use channels 0 and 2, which have more space between them on
the chip. However, the bigger problem is numerous spurs every couple Mhz only
about 60-70 db down. At first I thought this might be power supply noise from
the Arduino Nano that I'm running it with. I connected a lab supply to the
Si5351 board, but that didn't change anything. Then, thinking it might be some
digital noise from the Nano coming over on the I2C, I held the reset button on
the Nano, which I assume kills all I/O. That didn't help either. I now think
this might be some artifacts from the fractional-N synthesis process, as the
level of the spurs jumps up and down several dB. If this truly is intrinsic in
the part, it's almost a deal-killer for anything but a novelty receiver.
I'd like to hear others' experiences.
Joe
W3JDR
----- Original Message -----
From: DuWayne KV4QB
To: minima@xxxxxxxxxxxxx
Sent: Wednesday, October 29, 2014 11:11 AM
Subject: [minima] Re: Si5351
Just started to play with one of the Adafruit boards. Changed my DDS
board to use the 5351 board instead
The library from Jason seems to work well. Other than getting the board
to run I have not done anything else with it.
Don't know if I will use it on a Minima or will go with some other
design. Here is a shot of what I have so far.
DuWayne KV4QB
On 10/29/2014 8:28 AM, Joe Rocci wrote:
There was some talk here a while back about the Si5351, an equivalent
to the Si570 that has 3 programmable oscillators and only costs a couple
dollars. I just got an application board from AdaFruit and fired it up to see
how it performs. I have some questions about spurious outputs that I'd like to
discuss with anyone who's using this part. If you have any experience with it,
please contact me directly(joeatroccisdotcom)
Joe
W3JDR
--
Of all the things I lost, I miss my mind the most.
Other related posts:
