EMRFD Message Archive 12912

Message Date From Subject
12912 2016-06-06 07:19:49 ka3j12 SI570 vs Si5351 Controversy

There has been much discussion and analysis of these two interesting components.  The core question has been is the cheaper, lower power, lower performing Si5351 suitable for "quality" receivers?  Well, Elecraft apparently thinks so.  They used the Si570 in the KX3 but switched to the Si5351 in the recently released KX2 (see block diagram - page 62 of owner's manual).  Wayne has also stated that although the KX2's receiver does not perform quite as well as the KX3, he still expects it to show up among the top 20 rigs in Sherwood Engineering rankings  http://www.sherweng.com/table.html  Hmmm...


My simple homebrew handheld rig that I've use nearly everyday at the beach for several years uses a Si570. I think it's one of the best sounding rigs I've ever used.  However, I think it may be time to give the Si5351 a spin in my next rig. 


Ron (KA3J)

 

12913 2016-06-06 08:09:09 pete VE3HOH-W3 Re: SI570 vs Si5351 Controversy
I am wondering if the 5351 could be used to improve the phase noise in a Drake TR7.

The Si 570 would be my variable LO of choice, but the TR7 has a lot going in reference to the main oscillator.

Just a thought.


[😉]

________________________________
12914 2016-06-06 08:20:36 Ashhar Farhan Re: SI570 vs Si5351 Controversy
i suspect that both the chips use almost exactly the same technology with one difference : the Si5351A needs an external crystal while the Si570 has an in-built crystal. To that extent, the reason for higher phase noise of Si5351A could be the cheap quality crystals and the in-built oscillator. Instead, it might make sense to add a simple outboard low noise oscillator as the reference to the Si5351. This is just conjecture of course, if someone could measure it, it would be swell....

- f

12915 2016-06-06 08:30:06 Cecil Bayona Re: SI570 vs Si5351 Controversy
Or there is internal variations of the device that makes it noisier by coupling internal signals to the oscillator, or greater coupling between the output and the internal signals, the only way to find out is to try an experiment with a really clean LO as the source to the device to eliminate one possibility.


12917 2016-06-06 12:27:27 aa0zz Re: SI570 vs Si5351 Controversy
I would bet the other way.  Judging from the completely different way the programming parameters have to be calculated and set up in the two parts, I'd say they are two different animals completely.  
73,
-Craig, AA0ZZ
12918 2016-06-06 13:06:37 Dana Myers Re: SI570 vs Si5351 Controversy
12919 2016-06-06 14:37:42 aa0zz Re: SI570 vs Si5351 Controversy
Good analysis.

A while back (~2010), when I was first getting going on the Si570, I also took a close look at the Si598 because it looked like it had some advantages over the Si570.  At that time, the documentation for the Si570 did not show anything about the "Freeze M / New Freq" capability while the Si598 that came out a bit later did.  This is important because without it "crud" is sprayed around when sending parameters.  Because of this difference I was all set to move to the Si598 but then I had a talk with the engineers at Silicon Labs and they assured me that the Si570 had the same feature but that it wasn't documented in the latest spec sheet yet.  In that same conversation they told me that the fact that the reference clock of the Si570 at 114 MHz vs and the Si598 was at 56MHz made the Si570 much superior in performance. The fact that the Si5351 must use an external clock at either 25 MHz or 27 MHz fits the same pattern in terms of performance - ie, less than the Si570 or even the Si598. 
By the way, back in 2010 the Si570 in single quantities cost $25 from SiliconLabs while the SI598 cost $13.   Today the Si570 is down to $15 and some versions are even available from DigiKey.  Still a long way from the $1.50 for the Si5351 (plus $1.50 for a crystal) so, if the lower performance of the Si5351 is adequate, it can be a good alternative. I'm still partial to the Si570.  For small quantity, home-built radios, the $10 difference is peanuts. An Si570 is also a LOT easier to solder on a board than an Si5351! 
73,
-Craig, AA0ZZ

12920 2016-06-06 18:08:14 kb1gmx Re: SI570 vs Si5351 Controversy
there are many things going on at the same time.

Multiplication factors for the reference to the DCO , higher multiplication is poorer.

Division factors, higher is better but its not a 1:1 comparison as the loops are different 
and have different frequencies and ranges.

One things there is no way to compare is that multiple oscilators on one die will
have cross talk and higher noise as the die it self has resistance (think poor 
common ground).  Its not only physical pin adjacency its the die it self. That 
part can be helped if the load at each pin causes small current spikes rather 
than large that may propagate on the die.

While the Si5351 is not as good as SI570 one has to evaluate the use and what that 
much jitter means in terms of noise and bandwidth. also if the output is divided 
by 4 (typical of SDR and phasing radios) the net jitter improves.  The calculation for that
is non trivial as your adding the output of I and Q to get signal. Radios are all about 
the interaction of signals often some are out of your control (the other guy).
 
While the si5351 would not be my first choice its still pretty good.  Then again I've been 
playing with one time programmable oscillators which vary from terrible to at least as 
good as 5351.  In simple radios even terrible works better than some crystal oscillator
I've seen with discrete parts.

Good, bad, terrible are all random and not quantified and are at best value judgement
and what works in one design may fail or be noticeably lacking in something else.
The biggest thing is many jumping on a bandwagon they know not its direction or
or intent based on "its not as good as....".


Allison
12925 2016-06-07 00:47:06 Roelof Bakker Re: SI570 vs Si5351 Controversy
Hello Dan,

Your 13 dB estimated difference in sideband noise for the Si570 and
the Si5351 is quite close to what I found.

At 10 kHz from the carrier it is about 14 dB on 7 MHz and 12 dB on
28 MHz.

If you have not seen it yet, PERSEUS spectrum noise plots +/- 100
kHz from the carrier of a good (not excellent) X-tal oscillator, the
Si570 and the Si5351 can be found here:

http://ndb0.home.xs4all.nl/Oscillator Noise

73,
Roelof, pa0rdt
12926 2016-06-07 09:06:11 Dana Myers Re: SI570 vs Si5351 Controversy
12927 2016-06-07 09:11:25 Dana Myers Re: SI570 vs Si5351 Controversy
12928 2016-06-07 09:33:30 Dana Myers Re: SI570 vs Si5351 Controversy
12929 2016-06-07 10:08:42 Alberto I2PHD Re: SI570 vs Si5351 Controversy
12930 2016-06-07 16:07:22 David Wilcox Re: SI570 vs Si5351 Controversy
I almost hesitate to butt in on this topic, but can a 70 year old with "average" ears hear the "phase noise" that seems to be the hot topic now a days?

I probably don't know "phase noise" from background noise on my home brew kits and Elecraft radios but am interested in what it all means to the average cw QRPer.

Dave K8WPE

12931 2016-06-07 18:15:16 Tayloe, Dan (Noki... Re: SI570 vs Si5351 Controversy

It means a big signal up the band that would normally be outside the wide of a superhet filter would get “mixed into” your passband as “noise” by mixing with your LO noise sidebands.  Phase noise increases the power of your LO 1 KHz, 2 KHz, 5 KHz, 10 KHz out from the center frequency of your LO.

 

Thus, it can mix strong, “out-of-passband” junk into your received passband.  We are talking sidebands that are 100+ dB down, so it has to be a rather large signal (or the junk of many lesser signals combined like a field day) for this to make a noticeable difference.

 

-          Dan, N7VE

 

12932 2016-06-07 19:31:15 David Wilcox Re: SI570 vs Si5351 Controversy
Thank you.  Will wiki it too for more information....... So many reasons why the bands can be noisy and they seem to add together with all the stray RF in a computerized world.  And then there is the 10 kw AM station 2 miles away with 4 antennas pointing the signal in my direction......... 3rd harmonic on 3930 KHz right in the middle of the local 75 meter net frequency...... It's a wonder I can hear anything.  

Dave K8WPE

12933 2016-06-07 22:35:10 Lasse Moell Re: SI570 vs Si5351 Controversy
You can easily hear poor phase noise, as it will degrade you selectivity! Having spent a fortune on very sharp filters, but having a poor LO with high phase noise close to carrier will make all signals to appear wide.

/Lasse SM5GLC

8 juni 2016 04:22:10 +02:00, skrev David Wilcox djwilcox01@yahoo.com [emrfd] :
 



Thank you.  Will wiki it too for more information....... So many reasons why the bands can be noisy and they seem to add together with all the stray RF in a computerized world.  And then there is the 10 kw AM station 2 miles away with 4 antennas pointing the signal in my direction......... 3rd harmonic on 3930 KHz right in the middle of the local 75 meter net frequency...... It's a wonder I can hear anything.  

Dave K8WPE

12934 2016-06-08 10:40:47 kb1gmx Re: SI570 vs Si5351 Controversy
Thats the whole point.

You spend time and money for good filter and then cheap out on a poor LO.
The SI570 is not that costly compared to a expensive filter.

For more basic radios with simple filters of 3-4 crystals its likely hard to hear 
the difference.  Why even a narrow filter using that few poles will have soft
skirts and be hard to discern from a wider LO.

Farhan said it.  You will not know if you don't build and test it.  The rest is 
speculation.


Allison
12936 2016-06-08 16:44:32 vasilyivanenko Re: SI570 vs Si5351 Controversy
Thanks Allison, Farhan and the others who promote the build and measure philosophy
 
Every 4 months or so, the SI570 vs Si5351 question exacerbates here, or some other thread; or on my email chains and yet another round of subjective debate erupts. Hope to see some measures and perhaps a fresh topic at some point.  

Best!
12942 2016-06-09 09:11:53 Andy Cutland Re: SI570 vs Si5351 Controversy
The Soldersmoke podcast addressed this very issue. I can't remember which one off the top of my head but I think it was within the last 9 months. Bill and Pete discussed it at length. Worth checking the site and having a listen I reckon.

73's
De Andy
--------------------------------------------