EMRFD Message Archive 7577

Message Date From Subject
7577 2012-07-12 07:01:59 Reginald Beardsle... Low cost, software defined, spectrum analyzer?
I recently bought a Softrock Ensemble II RX kit and have been studying the design and datasheets in preparation for building it.

It seems to me that w/ suitable changes the design could form the basis of an inexpensive spectrum analyzer for 0-300 MHz that would be particularly easy to construct w/ little other test equipment (i.e. a DMM and a clean, fast rise time square wave source).

Basic idea is a wideband front end w/ an Si570 w/ USB controlled divider chain feeding an I-Q stream to the sound card. The output would be a series of overlapping spectra which would then be spliced and averaged in software.

Does anyone know of prior work on such an idea?

Have Fun!
Reg
7581 2012-07-12 17:15:16 RW Mail List Re: Low cost, software defined, spectrum analyzer?
Hi Reg,

Don’t know about a spectrum analyser. But Rocky SDR software at one stage had a feature where one of the softrocks would work as a VNA, and you were able to measure the impedance of your antenna. Don’t know if the latest versions are able to function as a VNA.

Roderick Wall, vk3yc.

7583 2012-07-13 13:49:05 Ken Frazer Re: Low cost, software defined, spectrum analyzer?
Conceptually, George Steber, WB9LVI, did something similar. See "Experimenter's RF Spectrum Analyzer" October 2008 QST.

Regards,
Ken
KE5JCB

7584 2012-07-13 19:01:28 ehydra Re: Low cost, software defined, spectrum analyzer?
Reginald Beardsley schrieb:
> I recently bought a Softrock Ensemble II RX kit and have been studying the design and datasheets in preparation for building it.
>
> It seems to me that w/ suitable changes the design could form the basis of an inexpensive spectrum analyzer for 0-300 MHz that would be particularly easy to construct w/ little other test equipment (i.e. a DMM and a clean, fast rise time square wave source).
>
> Basic idea is a wideband front end w/ an Si570 w/ USB controlled divider chain feeding an I-Q stream to the sound card. The output would be a series of overlapping spectra which would then be spliced and averaged in software.
>
> Does anyone know of prior work on such an idea?
>

Looks like the same as:
http://www.hamcom.dk/VNWA/HAMRADIO%202012%20DG8SAQ%20VNWA.mp4

- Henry

--
ehydra.dyndns.info
7588 2012-07-13 19:02:14 Reginald Beardsle... Re: Low cost, software defined, spectrum analyzer?
LoL!

A VNA is the next item on the list.

All the SDR software I've seen produces a spectrum display over a limited range. It's not much work to collect overlapping spectra and splice them together.

FWIW W/ a 192 KS/s sound card, it would take 2000 steps to go from 0-300 MHz. The Si570 settling time for large steps is 10 mS so w/ 40 mS of sampling time at each step it would take 1 minute to get a spectrum w/ 25 Hz resolution.
That would have poor statistics, so it would be better to average into 250 Hz bins. Longer sweep times will permit arbitrarily fine frequency sampling and high system S/N.

For both the spectrum analyzer and VNA, the hard part is designing a 0-300 MHz frontend that's easy to build w/ limited test gear.


Have Fun!
Reg

7612 2012-07-16 08:42:23 kb9fko Re: Low cost, software defined, spectrum analyzer?
Sounds like fun.

It does remind me of figure 7.64 in my copy of EMRFD. Same idea but even more bare bones.

Best regards,

Drew
kb9fko

7615 2012-07-16 21:19:49 Reginald Beardsle... Re: Low cost, software defined, spectrum analyzer?
The more I dig into the Softrock Ensemble II design the more intrigued I am. It's *really* simple.

For a dedicated build I have in mind using an MSP430 & USB for the ADC. There's a new version w/ two 24 bit, 1 MS/s ADCs which would allow grabbing a bigger chunk of spectrum at each step in LO frequency.

Have Fun!
Reg

7619 2012-07-17 04:53:54 John Greusel Re: Low cost, software defined, spectrum analyzer?
You'll be very surprised by it's performance if you haven't put it together yet. It stands up pretty nicely to many commercial radios.

John
KC9OJV
 



________________________________
7625 2012-07-18 06:38:49 Russ Ramirez Re: Low cost, software defined, spectrum analyzer?
"You'll be very surprised by it's performance if you haven't put it
together yet. It stands up pretty nicely to many commercial radios.

John
KC9OJV"

Just ordered a Softrock Ensemble II RX. Hard to beat something like this
for $70 (with shipping) in terms of playing with SDR and its usefulness as
a receiver or as an integral part of a test equipment project.

Russ
K0WFS


[Non-text portions of this message have been removed]
7629 2012-07-23 16:08:51 Russ Ramirez Re: Low cost, software defined, spectrum analyzer?
Just happened across this today.

http://ossmann.blogspot.co.uk/2010/03/16-pocket-spectrum-analyzer.html

It would take a bit more work than described to get this into the HF bands,
but it's pretty cool even if you just used it around 70 and 33 cm.

Russ
K0WFS


[Non-text portions of this message have been removed]
7637 2012-07-24 22:40:10 Reginald Beardsle... Re: Low cost, software defined, spectrum analyzer?
That is cool! I'd seen an audio version, but not an RF version.

The USB TV Tuner dongles may actually be one of the best approaches for a software defined SA. There's a lot of activity in the SDR arena using those. So much so that the desirable units have gone up in price from $10US to $20+ :-(

http://hackaday.com/2012/03/30/working-software-defined-radio-with-a-tv-tuner-card/

The Realtek unit tunes 64-1700 MHz. So w/ a good LO & mixer, a software defined 0-1600 MHz spectrum analyzer is pretty cheap. The 8 bit A/D is somewhat limiting, but that can be dealt w/ in software by summing adjacent bins in the FFT and using a long sampling time.

Calibration is the big puzzle. The best I can think of so far is an Si570 followed by a high speed divider to guarantee that the output is truly square w/ ~100 ps rise time. Then run a scan measuring the amplitude of the harmonics for a series of different fundamental frequencies and using those to compute an amplitude correction. There are other issues as well such as spurs and leakage. So not trivial, but not intractable. I find the prospect of getting a 100 ps rise time square wave from the divider to a connector more daunting than anything else especially since I have nothing to check the rise time with.

Meanwhile, I'm *almost* ready to start assembling my Ensemble II. I got a soldering station and have a bench DMM arriving tomorrow. Now all I need is the courage to start soldering SMD parts ;-)

Surely is a different world from when I last played w/ radios.

Have Fun!
Reg




7638 2012-07-25 23:01:34 Russ Ramirez Re: Low cost, software defined, spectrum analyzer?
" Meanwhile, I'm *almost* ready to start assembling my Ensemble II. I got a
soldering station and have a bench DMM arriving tomorrow. Now all I need is
the courage to start soldering SMD parts ;-) "

In spite of having a nice temperature controlled Hakko unit, I opted to go
with silver solder paste and one of these units:

http://www.ebay.com/itm/850B-SMT-SMD-HOT-AIR-REWORK-STATION-SOLDER-STATION-/160837555794?pt=LH_DefaultDomain_0&hash=item2572aa5652#ht_3333wt_1037

The nice thing about hot air is that is useful for several other
applications (heat shrink tubing being the obvious one) and it gets the job
done in a parallel fashion, i.e. soldering a whole side of a chip at a time
or the whole cap or resistor. With the passive components, you hold down
the part with one hand and apply the heat with the other. It's also a good
investment because SMT is here in a big way.

Russ
K0WFS


[Non-text portions of this message have been removed]