EMRFD Message Archive 4923

Message Date From Subject
4923 2010-08-26 17:26:25 k5myj SI-570?
Any comments on using a SI-570 as a LO?

Bob Macklin
Seattle, Wa.
4924 2010-08-26 19:30:44 kb9fko Re: SI-570?
I picked up one of the Si570 based LO kits recently. I hope to get to it sometime this fall and give it a try. Might be interesting for an R1 type receiver or for cobbling together an even more modular receiver approach like that in "Simply Getting on the Air From DC to Daylight". Might also be an interesting alternative signal source to try out some of the interesting ideas KK7B has been presenting in CQ VHF (and discussing in the Yahoo vhfkits group.) I don't remember what the current consumpti
4925 2010-08-26 19:51:12 Dave - WB6DHW Re: SI-570?
I use the Si570 in my UHFSDR board, which covers from 1.75 to 700 MHz. The
later Softrocks use the Si570. I have noticed for the last 2 weeks, at least,
that Silicon Labs does not let you order free samples of the Si570 online like
they used to. Also, the Si598 is very similar with a max frequency limit of 512
MHz for less money. I am heading up a club build which will use the TAK-40 with
a Si570 or Si598 as LO to cover all HF bands.

Dave - WB6DHW
<http://wb6dhw.com>

4926 2010-08-26 19:56:17 Bob Macklin Re: SI-570?
My interest is to design a special purpose 2M FM receiver. Different than a scanner or 2M HT.

And I don't want to have to buy a bunch of special order crystals.

Bob Macklin
K5MYJ
Seattle, Wa.
"Real Radios Glow In The Dark"


----- Original Message -----
4927 2010-08-27 09:03:25 ajparent1 Re: SI-570?
4928 2010-08-27 09:08:52 Scott Robson Re: SI-570?
I also plan a general coverage DC RX in the near future using the SI570.

I picked up a couple from KM5H. I have the LVDS 'BBC' version. I think this
part requires 3.3V and will work up to about 250 MHz. Can anyone verify that
for me?

Scott (K6AUS)



4929 2010-08-27 11:06:07 Bob Macklin Re: SI-570?
Why is the SI-570 so hard to find?

Mouser has 16 ON ORDER!

DigiKey shows them but no stock

Bob Macklin
K5MYJ
Seattle, Wa.
"Real Radios Glow In The Dark"
----- Original Message -----
4930 2010-08-27 12:05:09 k5nwa Re: SI-570?
They are generally custom ordered to a specific startup frequency so most
vendors do not stock them.

On Fri, 27 Aug 2010 13:05:38 -0500, Bob Macklin <macklinbob@msn.com> wrote:

>
> Why is the SI-570 so hard to find?
>
> Mouser has 16 ON ORDER!
>
> DigiKey shows them but no stock
>
> Bob Macklin
> K5MYJ
> Seattle, Wa.
> "Real Radios Glow In The Dark"
> ----- Original Message -----
4931 2010-08-27 12:52:21 Dave - WB6DHW Re: SI-570?
You can order them from the SiLabs website. Unfortunately, the Sample
button doesn't come up, so I guess you have to pay for them now. Also,
check the new Si598.

Dave - WB6DHW
<http://wb6dhw.com>

Bob Macklin wrote:
> Why is the SI-570 so hard to find?
>
> Mouser has 16 ON ORDER!
>
> DigiKey shows them but no stock
>
> Bob Macklin
> K5MYJ
> Seattle, Wa.
> "Real Radios Glow In The Dark"
> ----- Original Message -----
>
4932 2010-08-27 12:57:56 Dave - WB6DHW Re: SI-570?
You can go to the silabs website and enter the part number and get the
specs. The C grade LVDS device goes to 280 MHz. I think KM5H only gets
the 3.3V versions.

Dave - WB6DHW
<http://wb6dhw.com>

Scott Robson wrote:
> I also plan a general coverage DC RX in the near future using the SI570.
>
> I picked up a couple from KM5H. I have the LVDS 'BBC' version. I think this
> part requires 3.3V and will work up to about 250 MHz. Can anyone verify that
> for me?
>
> Scott (K6AUS)
>
>
>
>
4933 2010-08-27 13:01:02 Bob Macklin Re: SI-570?
I have not been able to figure out how. But for the time being I have 3.

Bob Macklin
K5MYJ
Seattle, Wa.
"Real Radios Glow In The Dark"
----- Original Message --
4934 2010-08-28 06:15:06 iami78 Re: SI-570?
> I see it as inexpensive and good for all but most critical
> receiver/transmitter applications. Its a surface mount part
> but the number of connections are only eight so it is not
> hard to accommodate or closely spaced.
>
>
> Allison
>

Hi Allison!

I wonder why SI570 wouldn't be good for the critical too?
Looking at the specs and measurements made by many hams, it
seems to beat most of the DDS and PLL techniques out of the table, hi.

73 de Janne OH1SDR
4936 2010-08-28 17:13:49 kb1gmx Re: SI-570?
It's a good part but there are two parameters that
me be important to attaining very high performing systems.

One is Drift and Accuracy, drift is both power and temperature dependent and compared to a cheap crystal decent. Accuracy is
a calibration issue and can be zeroed out.

The other is jitter or FM noise, for a lot of applications
the spec is very good but I don't know how bad it would look
at the 3rd or 4th harmonic.

I've tested it in a filter based radio of the K2 caliber and
its very good. I don't know if it would made the grade for
a serious SDR or a ultra high performance radio like the
Triad.

But it's cheap and better than most VFOs, and most simple PLLs.
There are a few DDS chips I think can do better is they have
a good high quality clock but using them is not trivial.

The biggest thing is keeping the digital side (microprocessor)
from contributing it's noises to the board layout and with simple
attenti
4937 2010-08-29 05:54:13 Vojtech Re: SI-570?
DDS still wins in agility. It is very easy to modulate a multi tone FSK digital mode without any additional hardware. I did this with the AD9834 in ATS-3a,b and ATS-4 transceivers. Even BPSK31/63 is possible, if one provides amplitude modulation of the PA. I wonder how the output of the Si570 looks like during frequency change and how fast the frequency change is (how well it is suitable to QSK).

73, Vojtech OK1IAK
4938 2010-08-29 07:30:58 George N2APB Re: SI-570?
DDS also still wins out when spectral purity is required, such as when used
in reflectometer-based antenna analyzers and VNAs. Can¹t have energy
introduced beyond the fundamental (e.g., spurs, clocks, odd harmonics) as it
introduces spectral energy that ultimately gets integrated with the
fundamental and skews the measurements.

73, George N2APB


4941 2010-08-29 12:06:58 w7zoi Re: SI-570?
Hi Allison and gang,

My preliminary measurements on a Si570 based synthesizer are about the same as yours. With all of the discussion on the subject, I thought I should put my results up for grabs, so I have put them on my web page in the "designs and experiments" section. Just navigate to w7zoi.net.

I built the same kit that most of the folks seem to be using. I am certainly grateful for the assistance found on Jack Smith's web page! (Cliffton Labs) Jack, in his review, comments that this is definitely a kit for the advanced builder. This experience convinced me that I am definitely a beginning kit builder.

I have not looked at the drift in any detail, but a quick look indicated that it is probably fine for casual applications. The overall noise level is low and the spurs are lower than found with a simple DDS. I don't see that harmonics would be any problem beyond the usual 20Log(N) degradation that we normally expect. I'm glad to hear that you have used the part in a K2 level rig. The biggest problem I see for use in ham rigs is that the parts consume a lot of current.

73, Wes
w7zoi


4942 2010-08-29 12:33:06 w4zcb Re: SI-570?
Wes, Tnx the details of your evaluation of the 570. I was unaware of
the device until it showed up here on EMRFD and finding it useful
enough, I'm trying to communicate with the "couple of retired guys" to
put one in my shop as well. I am curious, whether you bought the CMOS
or the ECL unit, and it was my impression that the lower frequency
limit of the package was 10 MHz. Were you dividing the output to get
that 5 MHz info?

Regards
W4ZCB


----- Original Message -----
4944 2010-08-29 13:05:44 Dave - WB6DHW Re: SI-570?
It can actually be programmed down to 3.5 MHz. I use it in my UHFSDR,
which uses both outputs from the LVPECL version in a divide by 2 LVPECL
circuit to produce I and Q signals. By the way, the LVPECL will
directly drive the LO input of a DBM.

Dave - WB6DHW
<http://wb6dhw.com>

w4zcb wrote:
> Wes, Tnx the details of your evaluation of the 570. I was unaware of
> the device until it showed up here on EMRFD and finding it useful
> enough, I'm trying to communicate with the "couple of retired guys" to
> put one in my shop as well. I am curious, whether you bought the CMOS
> or the ECL unit, and it was my impression that the lower frequency
> limit of the package was 10 MHz. Were you dividing the output to get
> that 5 MHz info?
>
> Regards
> W4ZCB
>
>
> ----- Original Message -----
>
4945 2010-08-29 14:41:21 w7zoi Re: SI-570?
Hi Harold, et al,

Yep, I used the CMOS version. I got it with the kit.

The system worked down to around 3.5 MHz, as I recall, but not much further.

It is so easy to divide these kind of things down. While some folks might really use the sine wave output of a DDS, I think that most of us in this Yahoo Group are going to be more than happy with a square wave. Most of the mixers that I use end up squaring the LO waveform anyway.

Have fun.

73, Wes
w7zoi


4946 2010-08-29 14:42:18 w4zcb Re: SI-570?
> It can actually be programmed down to 3.5 MHz. I use it in my
> UHFSDR,
> which uses both outputs from the LVPECL version in a divide by 2
> LVPECL
> circuit to produce I and Q signals. By the way, the LVPECL will
> directly drive the LO input of a DBM.
>
> Dave - WB6DHW
> <http://wb6dhw.com>

Interesting Dave. I had read where the upper end went way beyond
specs, had no realization it went both ways. I intend to get the ECL
version myself. Would also be interested in seeing something of your
divide by 2 quadrature. I have some very quiet ECL SP8402 dividers
from Zarlink just itching to do the job.

W4ZCB
4947 2010-08-29 14:51:57 Bob Macklin Re: SI-570?
If you are going to use it with a quaderature detector you need the square wave to clock the registers.

Bob Macklin
K5MYJ
Seattle, Wa.
"Real Radios Glow In The Dark"
----- Original Message -----
4948 2010-08-29 15:23:36 k5nwa Re: SI-570?
At 04:42 PM 8/29/2010, you wrote:
>
> > It can actually be programmed down to 3.5 MHz. I use it in my
> > UHFSDR,
> > which uses both outputs from the LVPECL version in a divide by 2
> > LVPECL
> > circuit to produce I and Q signals. By the way, the LVPECL will
> > directly drive the LO input of a DBM.
> >
> > Dave - WB6DHW
> > <<http://wb6dhw.com>http://wb6dhw.com>
>
>Interesting Dave. I had read where the upper end went way beyond
>specs, had no realization it went both ways. I intend to get the ECL
>version myself. Would also be interested in seeing something of your
>divide by 2 quadrature. I have some very quiet ECL SP8402 dividers
>from Zarlink just itching to do the job.
>
>W4ZCB

That was in the early days when the chips were not locked down to the
range they sold it for. Nowadays when you buy the low frequency
device it will not allow you to set it to the higher bands even
though the chip is probably capable of functioning way higher. Also
the CMOS versions have always had way lower frequency capability, I
bought 10 of unrestricted LVDS versions early on and now I'm glad I did.


Cecil
k5nwa
< www.softrockradio.org > < www.qrpradio.com >
< http://parts.softrockradio.org/ >

Never take life seriously. Nobody gets out alive anyway.
4950 2010-08-29 16:31:34 Dave - WB6DHW Re: SI-570?
Here is the schematic, etc, for the UHFSDR.
<http://wb6dhw.com/UHF-SDR/production/UHFSDRproduction.zip>

Dave - WB6DHW
<http://wb6dhw.com>

4951 2010-08-29 16:57:23 kb1gmx Re: SI-570?
4952 2010-08-29 19:23:41 k5nwa Re: SI-570?
A positive temperature coefficient thermistor
glued to the chip and connected across a stable
power source will stabilize the chip very quickly
and hold it fairly stable. I helps also to have
the chip covered in Styrofoam package so it's not
so exposed to varying room air temperatures.

At 06:57 PM 8/29/2010, you wrote:
>
>
4953 2010-08-29 22:43:32 Eric Söderman Re: SI-570?
What I read in the softrock group is that the lower limit is 3,5 MHz.
The softrocks normally use the CMOS version.

Eric SA5BKE

4954 2010-08-30 00:40:07 Roelof Bakker Re: SI-570?
Hello all,

> A positive temperature coefficient thermistor
> glued to the chip and connected across a stable
> power source will stabilize the chip very quickly
> and hold it fairly stable. I helps also to have
> the chip covered in Styrofoam package so it's not
> so exposed to varying room air temperatures.

The German Magazine "Funkamateur" sells a kit which incorporates this
feature.
The following link is to a article in German, but the circuit diagram
shows how it has been done:

http://www.box73.de/catalog/pdf/BX-026.pdf?osCsid=61f0gcgv9rkc1u0vovm8rmpqi6

On an other note, I am using a Si570 CMOS in Tony Parks RX LF Ensemble
with the frequency divided by 16 and performance is comparable with a high
end direct sampling SDR which has a very low noise oscillator.

I have a K5BCQ kit which has to be build yet, to use it in a more
conventional manner.

73,
Roelof Bakker, pa0rdt
4957 2010-08-30 07:36:01 kb1gmx Re: SI-570?
Hi Wes,


Couple of things..

I don't work for Silicon Labs. I do a lot of digital work
and analog so I tend to view this part more from the
system view.

The Si570 was designed as a variable rate clock source for
systems with digital cores that maybe doing things in the communications realm or any other task. For that use there
is no comparison. That said it was never designed as a VFO
or DDS replacement though it's low phase noise makes it
very useful for that.

I have two of the same kits and found below 10mhz operation
to be non-optimum. Note that the si570 CMOS is spec'd for 10
to 160mhz so below 10mhz is questionable. When I need less than
10mhz I add a /10 counter and it does two things. Extends the
reach downward to under 1mhz and also cleans up the output.
A /100 would allow getting to 100khz and allow .01hz resolution.
Of the two I have the one is used with a added dividers to get
/10, /100 and output protected with a 6db pad as a general
use signal high accurate and stable signal source.


The modulation on start up I have not observed. My method of
testing is to fire up at 50.125mhz and listen to the resulting carrier on the 6M radio and it chips initially then settles to
a clean tone almost instantly but there is some startup drift
(small but noticed). Thermal drift is slower but behaves like
any quartz source.

The spec for startup is 10mS but that does not include
pumping in some 60 bits of housekeeping and frequency
programming data and that takes a finite amount of time
depending
4964 2010-08-31 05:27:19 hanssummers2000 Re: SI-570?
Hi all

I agree that the Si570 is much easier to use than a DDS.

From the datasheets, an Si570 has about 20dB worse phase noise than a top end DDS (e.g. AD9910) driven by a high quality 1GHz reference (not easy, of course!)

For HF use, the spurs of a high-end DDS are not a problem. With earlier generations of DDS they are, but not the later ones. So the Si570 will not offer any performance advantage in terms of spurs, at HF. I think at VHF the spurs output of even a 1GHz reference clock DDS will start to degrade so the Si570 would start to look better at VHF.

So I still think that for the purist, a good DDS (and a lot of care and attention) will outperform an Si570.

73 Hans G0UPL
http://www.hanssummers.com





4965 2010-08-31 05:49:43 Bob Macklin Re: SI-570?
"I agree that the Si570 is much easier to use than a DDS"

Yes it is a hell of a lot easier to solder to a PCB than a 44 pin TQFP!

Bob Macklin
K5MYJ
Seattle, Wa.
"Real Radios Glow In The Dark"
----- Original Message -----
4978 2010-08-31 19:01:12 AD7ZU Re: SI-570?
Has anyone measured the Si570 duty cycle?  The data sheet specs 45 - 55% for
both the Si570 and Si598
A 5% asymmetric clock driving a switching mixer in a phasing rx / tx will result
in significant quadrature error.

just curious

Randy




________________________________
4980 2010-08-31 20:04:29 Tayloe Dan-P26412 Re: SI-570?
It will have no effect if you are doing a divide by four. In that case
only the repeatability of the rising edge or the falling edge (depending
on which you use to clock with) is used. The exact duty cycle is not
important.

- Dan, N7VE

________________________________

4993 2010-09-01 15:39:19 kb1gmx Re: SI-570?