Poor Man's Spectrum Analyzer?
John D. Seboldt
Anyone have experience with WA2PZO's "Poor Man's Spectrum Analyzer,"
marketed by "Science Workshop" at Bethpage, NY
<http://www.science-workshop/com/>? I've heard of this idea for some
time, and found them on the web recently. Seems like a useful setup
within the limits of its component parts.
John Seboldt K0JD
Joseph Consugar
John D. Seboldt wrote:
I built this and had good experiences with it. My main problem
was not being able to tell what frequency it was tuned to and the non-
linearity in the tuning of the tuner. However, these are problems that
can be and have been overcome by others.
If you are going to build this, I recommend getting the book along
with it. Lots of good examples and ideas from other builders.
Joseph
Gary Coffman
On 22 Jan 1998 22:46:09 GMT, rohr...@pconline.com (John D. Seboldt) wrote:
>Anyone have experience with WA2PZO's "Poor Man's Spectrum Analyzer,"
>marketed by "Science Workshop" at Bethpage, NY
><http://www.science-workshop/com/>? I've heard of this idea for some
>time, and found them on the web recently. Seems like a useful setup
>within the limits of its component parts.
>
>John Seboldt K0JD
I built one, including tracking generator option. It isn't lab quality, but
it is a useful instrument.
Gary
Gary Coffman KE4ZV | You make it |mail to ke...@bellsouth.net
534 Shannon Way | We break it |
Lawrenceville, GA | Guaranteed |
Mark J Sandford
Mike wrote:
Gary Coffman wrote:
> >John Seboldt K0JD
I think you mean the following which works fineLike your SIG!
Does anyone know if the URL
http://www.science-workshop/com/
is correct or working? I tried it a couple of days ago, and again just
now, and get the same result. No DNS entry.Best Regards,
Mike
CEO, Analog & Digital Design
Automated Production Test
http://www.csolve.net/~add/home.htmHosting Jonathan Ramsey's Pascal TCP/IP for DOS:
http://www.csolve.net/~add/zips/tcp.htm
http://www.science-workshop.com
Michael Black
It's
http://www.science-workshop.com
not
www.science-workshop/com
The latter is not a proper address. Every internet address has to end
with a three letter domain name, such as .org .edu .com .net and so on.
Michael VE2BVW
Mike
Mike wrote:
>
> Mark J Sandford wrote:
>
> [... snip due to NTTP posting limit]
[...]
Never mind - I didn't spot the "/" before the end. That won't work!
Mike
DavidC wrote:
>
> Mike:
>
> Look closely, what is wrong with this picture? :-)
>
> Try http://www.science-workshop.com/
>
> 73, DavidC> > Mike
Yes, you are absolutely correct! Since I always double-check URL's before
posting them, I assumed others did also. There is enough trouble with bad
links, a simple typo should never be a concern. Says me >)
So who would expect to find a "/" where a "." should be?
Mike
Mark J Sandford wrote:
[... snip due to NTTP posting limit]
> > Does anyone know if the URL
> >
> > http://www.science-workshop/com/
> >
> > is correct or working? I tried it a couple of days ago, and again just
> > now, and get the same result. No DNS entry.
> >
> > Best Regards,
> > Mike
> I think you mean the following which works fine
> http://www.science-workshop.com
[... snip HTML dupe of same post]
Hi Mark,
Thanks for solving my problem. But something is strange!
Your URL works fine.
When I go there, I get
http://www.science-workshop.com/
which is the same as the original posting. But if I try that, it doesn't
work. I've used URL's with "/" at the end, and they worked fine. Guess
I'll just have to watch for that. Thanks!
BTW - do you think you could turn off the HTML?
Mike
Gary Coffman wrote:
> On 22 Jan 1998 22:46:09 GMT, rohr...@pconline.com (John D. Seboldt) wrote:
> >Anyone have experience with WA2PZO's "Poor Man's Spectrum Analyzer,"
> >marketed by "Science Workshop" at Bethpage, NY
> ><http://www.science-workshop/com/>? I've heard of this idea for some
> >time, and found them on the web recently. Seems like a useful setup
> >within the limits of its component parts.
> >John Seboldt K0JD
> I built one, including tracking generator option. It isn't lab quality, but
> it is a useful instrument.
> Gary
> Gary Coffman KE4ZV | You make it |mail to ke...@bellsouth.net
> 534 Shannon Way | We break it |
> Lawrenceville, GA | Guaranteed |
Like your SIG!
Does anyone know if the URL
http://www.science-workshop/com/
is correct or working? I tried it a couple of days ago, and again just
now, and get the same result. No DNS entry.
Best Regards,
Mike
DavidC
Mike:
Look closely, what is wrong with this picture? :-)
Try http://www.science-workshop.com/
73, DavidC
> Does anyone know if the URL
>
> http://www.science-workshop/com/
> Mike
dan
John D. Seboldt wrote:
>
> Anyone have experience with WA2PZO's "Poor Man's Spectrum Analyzer,"
> marketed by "Science Workshop" at Bethpage, NY
> time, and found them on the web recently. Seems like a useful setup
> within the limits of its component parts.
>
> John Seboldt K0JD
Hi,
I have looked this over after putting up a page
on 'how the spectrum analyzer works.'
http://www.lakeweb.com/dbLabs/spectrumAnalyzer/intro.shtml
I had questions like:
What happens when you sweep through the IF? The article
doesn't mention dual conversion. If it is just a detected
output, it won't have much dynamic range. Pretty important
for a spectrum analyzer.
Once we are done with the move we are making, I plan on
building a new one that could easily be duplicated using
the AD606. It would mean good dynamic range and accuracy
at a very low cost.
I will post here as I get the pages up.
Thanks, Dan.
Michael Black
Science Workshop is not the only source for this sort of thing, though
they may be the only company that is selling kits. I thought the idea of
using a tv tuner as the front end of a "spectrum analyzer" had originated
in an article in one of the ham magazines, and then Science Workshop took
the idea and did some improvements, but I'm not certain.
There have been a number of articles about the same basic concept over the
years. There have been some very simple ones, while others have been a
lot more serious. It probably is worth looking for those articles, though
I don't have any exact references handy. I've seen the idea in Radio
Electronics, Electronics Now (which used to be Radio Electronics),
Communications Quarterly, and I've seen references to articles in Ham
Radio and QST.
Once you have the basic idea, then you can start improving it if you feel
up to it. There was a recent article in Communications Quarterly that
took the Science Workshop analyzer as a foundation and made significant
improvements, especially in the area of dynamic range.
There was also a very good article in Ham Radio, in 1977 I believe, by
Wayne Ryder. It only covered a hundred MHz, which is small compared to
these recent projects, but large at the time. This was before the idea of
using tv tuners as the front end so that part was built from scratch. On
the other hand, it did make for a better analyzer. It also featured a
narrower bandwidth than many we are seeing these days, and some details
about the tradeoffs. There was also an attempt at making the tuning
linear (or at least more linear), and the IF strip was a log amplifier.
Michael VE2BVW
Mike
dan wrote:
>
> John D. Seboldt wrote:
> >
> > Anyone have experience with WA2PZO's "Poor Man's Spectrum Analyzer,"
> > marketed by "Science Workshop" at Bethpage, NY
> > <http://www.science-workshop.com/>? I've heard of this idea for some
> > time, and found them on the web recently. Seems like a useful setup
> > within the limits of its component parts.
> >
> > John Seboldt K0JD
>
> Hi,
>
> I have looked this over after putting up a page
> on 'how the spectrum analyzer works.'
>
> http://www.lakeweb.com/dbLabs/spectrumAnalyzer/intro.shtml
>
> I had questions like:
> What happens when you sweep through the IF? The article
> doesn't mention dual conversion. If it is just a detected
> output, it won't have much dynamic range. Pretty important
> for a spectrum analyzer.
>
> Once we are done with the move we are making, I plan on
> building a new one that could easily be duplicated using
> the AD606. It would mean good dynamic range and accuracy
> at a very low cost.
>
> I will post here as I get the pages up.
>
> Thanks, Dan.
Dan,
Hurry up with the move - I've been thinking of an AD606 version, with a
2-4 GHz first oscillator and up-converting to get rid of problems
sweeping through the IF.
I was thinking of using the AD DDS chip (don't have the number handy),
and phase-locking the first oscillator for close-in work.
Of course, you need a companion tracking generator, so it would have to
be designed in from the beginning.
Shielding will be a problem - 95 db dynamic range is plenty, but keeping
all those spurs out will be fun. I am thinking of small, fully-shielded
compartments, with the vco's completely separated and shielded from the
phase detector and prescaler.
Then have two separate buffers, one to go back to the phase detector, and
one to go to the mixer through appropriate high-pass filters. This
hopefully will help keep the phase-locking spurs out of the mixer signal.
I don't think the electronics will cost that much - it's the shielding
and filtering to get clean signals that will be tough!
dan
Mike wrote:
>
> dan wrote:
> > I have looked this over after putting up a page
> > on 'how the spectrum analyzer works.'
> >
> > http://www.lakeweb.com/dbLabs/spectrumAnalyzer/intro.shtml
> >
>
> Hurry up with the move - I've been thinking of an AD606 version, with a
> 2-4 GHz first oscillator and up-converting to get rid of problems
> sweeping through the IF.
>
I was thinking 1.x to 2.x and then harmonics mix to go
higher. I have a couple of 2-4 yigs. But that would really
push on the cost.
> I was thinking of using the AD DDS chip (don't have the number handy),
> and phase-locking the first oscillator for close-in work.
If the ad606 had a RF log output you might be able to
DSP 50 MHz at a time. It would also lead to very accurate
frequency processing at low cost.
>
> Of course, you need a companion tracking generator, so it would have to
> be designed in from the beginning.
>
I was going to keep this in mind.
> Shielding will be a problem - 95 db dynamic range is plenty, but keeping
> all those spurs out will be fun. I am thinking of small, fully-shielded
> compartments, with the vco's completely separated and shielded from the
> phase detector and prescaler.
Those inexpensive thin sheet metal boxes for PC mounting,
(Allied sells them), is what I would hope would do the
job.
>
> Then have two separate buffers, one to go back to the phase detector, and
> one to go to the mixer through appropriate high-pass filters. This
> hopefully will help keep the phase-locking spurs out of the mixer signal.
>
> I don't think the electronics will cost that much - it's the shielding
> and filtering to get clean signals that will be tough!
>
Also the display. I was considering digital by the time
it gets there. The display on the TDS scopes is inexpensive
and very easy to look at. The chips are dirt cheep now for
doing monochrome VGA. TTL might work but won't look as good.
More later, my best, Dan.
John Miles
Some thoughts:
1) You want to avoid spurs in the design stage, not just by throwing
shielding at the problem. :-)
2) Instead of rolling your own analyzer from scratch using junk like TV
tuner modules, another alternative might be to start with the guts from
an older 491 or 141T plugin, and add computer control and display
capabilities to it. These units had perfectly good mixers, LOs,
filters, and log detectors, but lousy displays and awkward user
interfaces.
Dead examples of this type of gear can be found fairly inexpensively.
And computerization is just what these old "classics" are crying out
for. I remember seeing a quote awhile back that I found especially
poignant -- something to the effect of "The microprocessor saved the
spectrum analyzer, just as it killed the oscilloscope."
3) For the display side of the question, IMHO it doesn't make any sense
at all to homebrew this, when you can build a simple parallel-port
interface to dump the output of the log detector to a convenient laptop
or other PC. You can handle almost all display and control features in
software with minimal hassle, and end up with a very competent
instrument with the ability to automatically configure its controls, run
automated tests, save and print traces, etc. Life is too short not to
use off-the-shelf components whenever possible.
4) If you have your heart set on building the RF section from scratch,
don't overlook the vast array of junked sweep-generator plugins for the
older Wiltron and HP mainframes available at almost any flea market.
These things can often be had for as little as $100-$200 or so, and
contain YIGs, fixed-frequency microwave oscillators, mixers, filters,
UHF/microwave gain blocks, and tons of miscellaneous spare parts. With
some creative scrounging and a good set of crystal filters salvaged from
a 70s-vintage instrument like an HP8552, you could build a PC-driven
spectrum analyzer that would compete reasonably well on many fronts with
some very modern, very expensive gear. Challenges associated with
building an analyzer around parts from a gutted sweep-generator module
would include LO stabilization for narrow span widths, and IF
filtering. Neither of these are trivial problems, which is why I think
starting with the subassemblies from an older spectrum analyzer is a
better overall approach.
-- jm KE5FX
------------------------------------------------------
Note: My E-mail address has been altered to avoid spam
------------------------------------------------------
Mike
dan wrote:
>
> Mike wrote:
[...]
> Hi Mike!
>
> I was thinking 1.x to 2.x and then harmonics mix to go
> higher. I have a couple of 2-4 yigs. But that would really
> push on the cost.
Q: Where did you get them and how much do they cost?
> > I was thinking of using the AD DDS chip (don't have the number handy),
> > and phase-locking the first oscillator for close-in work.
>
> If the ad606 had a RF log output you might be able to
> DSP 50 MHz at a time. It would also lead to very accurate
> frequency processing at low cost.
I'd like to hear more on this. Do you mean sampling the output from 0 to
50 MHz? Wouldn't that mean more spurs to eliminate, and where are you
going to get a 100 or 200 MHz sampler? And how are you going to get it
into the PCI bus to work on it?
Sounds interesting - if the added noise can be overcome.
> > Of course, you need a companion tracking generator, so it would have to
> > be designed in from the beginning.
> >
> I was going to keep this in mind.
So, how are you going to get the frequency offset? For fairly wide
sweeps, the tracking doesn't have to be so good. But what if you need to
use narrow bandwidths for better s/n, or whatever. It's going to be fun
generating an offset frequency equal to the first IF, between 2 - 4 GHz,
and keep it in a narrow IF. Any ideas?
> > Shielding will be a problem - 95 db dynamic range is plenty, but keeping
> > all those spurs out will be fun. I am thinking of small, fully-shielded
> > compartments, with the vco's completely separated and shielded from the
> > phase detector and prescaler.
>
> Those inexpensive thin sheet metal boxes for PC mounting,
> (Allied sells them), is what I would hope would do the
> job.
Got to find ones with close-fitting lids - perhaps with a way to mount
microwave mesh gasket or contact fingers. Maybe even a box-within-a-box
for the really tough parts.
[...]
> > I don't think the electronics will cost that much - it's the shielding
> > and filtering to get clean signals that will be tough!
> Also the display. I was considering digital by the time
> it gets there. The display on the TDS scopes is inexpensive
> and very easy to look at. The chips are dirt cheep now for
> doing monochrome VGA. TTL might work but won't look as good.
OOooow - that is a good point. Actually, I hadn't got that far, and was
kind of thinking along the lines of a plug-in card for the ISA bus.
Obviously, that would suffer from CPU clock harmonics and all the rest of
the crud on the CPU motherboard. Anyway, the plan was to somehow get the
data into the CPU for analysis and data logging. But that may turn out to
the toughest problem of all.
> More later, my best, Dan.
Best Regards,
Mike
John Miles wrote:
[... snip first part due to NTTP posting limit on my ISP]
> These things can often be had for as little as $100-$200 or so, and
> contain YIGs, fixed-frequency microwave oscillators, mixers, filters,
> UHF/microwave gain blocks, and tons of miscellaneous spare parts. With
> some creative scrounging and a good set of crystal filters salvaged from
> a 70s-vintage instrument like an HP8552, you could build a PC-driven
> spectrum analyzer that would compete reasonably well on many fronts with
> some very modern, very expensive gear. Challenges associated with
> building an analyzer around parts from a gutted sweep-generator module
> would include LO stabilization for narrow span widths, and IF
> filtering. Neither of these are trivial problems, which is why I think
> starting with the subassemblies from an older spectrum analyzer is a
> better overall approach.
>
> -- jm KE5FX
All very good suggestions, John. Some of these things should even come
with manuals and schematics, which would help.
Any good ideas for architecture approaches?
John Miles
Mike wrote:
>
> John Miles wrote:
>
> <snip>
>
> All very good suggestions, John. Some of these things should even come
> with manuals and schematics, which would help.
>
> Any good ideas for architecture approaches?
>
> Best Regards,
>
> Mike
> CEO, Analog & Digital Design
> Automated Production Test
> http://www.csolve.net/~add/home.htm
>
> Hosting Jonathan Ramsey's Pascal TCP/IP for DOS:
> http://www.csolve.net/~add/zips/tcp.htm
Well, my old HP8551B has a fairly simple conversion scheme -- the first
LO tunes from 2 to 4 Gc (it's too old to be GHz :) and works into a 2 Gc
1st IF stage, for coverage from 10 Mc to 2 Gc on the LO fundamental. A
more stable, modern LO, together with a decent mixer, would allow you to
extend the lower limit as low as a few kc. No spurious responses are
possible with this architecture for any signal below 2 Gc, which is
nice.
The 8551B follows the 2 Gc IF with a conversion down to 200 Mc, where
the majority of the IF gain occurs. This is followed by a final
conversion down to 20 Mc, where a series of crystal filters in the
display unit provide the various resolution bandwidth options.
For dedicated HF/low-VHF work, I think the HP8553B plug-in for the 141T
converts 100 kHz - 110 MHz up to 200-300 MHz? Something like that,
anyway. Again, with such a relatively high first IF, it doesn't take
much filtering to eliminate spurs. A high first IF for an HF analyzer
also lets you keep the LO tuning range well under an octave, simplifying
things considerably if you aren't using a YIG.
-- jm
Mike
John Miles wrote:
> Mike wrote:
[...]
>> Any good ideas for architecture approaches?
> Well, my old HP8551B has a fairly simple conversion scheme --
> the first LO tunes from 2 to 4 Gc (it's too old to be GHz :)
> and works into a 2 Gc 1st IF stage, for coverage from 10 Mc
> to 2 Gc on the LO fundamental. A more stable, modern LO,
> together with a decent mixer, would allow you to extend the
> lower limit as low as a few kc. No spurious responses are
> possible with this architecture for any signal below 2 Gc,
> which is nice.
That seems to be the approach we're thinking of. Now, how do
you get a tracking generator to follow the first LO without
causing humongous (RFI, feedthrough, crosstalk, interference -
you pick the word!)
> The 8551B follows the 2 Gc IF with a conversion down to 200
> Mc, where the majority of the IF gain occurs. This is
> followed by a final conversion down to 20 Mc, where a series
> of crystal filters in the display unit provide the various
> resolution bandwidth options.
Presumably the first IF is actually 2.2 GHz, or the input range
is really 0-1.8 GHz and the first IF is 2 GHz.
Obviously, the second LO wants to be above the first IF so it
cannot appear as a spur. There may be some IMD products with
the presence of two mixers and all these local oscillators
floating around, but that will be the fun part!
Presumably, a sharp low-pass filter at the input will help keep
stuff above 2 GHz out of the spectrum, and sharp high-pass,
bandpass, notch, or combination of all three will be needed to
keep 200 MHz input signals out of the first IF due to first
mixer feedthrough.
The trick will be to find out where these filters have to go,
and what kind of performance they need.
That gets us to 200 MHz, where there are plenty of low-noise
amplifiers available cheap. Also, the hams have a lot of
experience with 2-meter repeaters, so getting good 3rd
intercept performance should be a piece of cake.
> For dedicated HF/low-VHF work, I think the HP8553B plug-in
> for the 141T converts 100 kHz - 110 MHz up to 200-300 MHz?
> Something like that, anyway. Again, with such a relatively
> high first IF, it doesn't take much filtering to eliminate
> spurs. A high first IF for an HF analyzer also lets you keep
> the LO tuning range well under an octave, simplifying things
> considerably if you aren't using a YIG.
My heart wants a yig - lowest phase noise. But I understand
they are a devil to keep stray magnetic fields out, and you
have to keep the noise on the dc control voltage down to
microvolt levels. That would mess up a fast PLL loop, so the
sweeps would end up slow.
Of course, the latest Tek and HP spectrum analyzers seem to be
able to handle 1 Hz resolution at 40 GHz. So we know it's
possible - but I'd sure like to know how they do it!
Once the signal is available at 200 MHz, everything is a piece
of cake after that. 95 db dynamic range log amps, DSP signal
processing, software for analysis and data logging, ...
Then, the next step is to extend the full sweep to 20 GHz -
easy. Just multiply everything by 10 :)
Well, my ISP seems to have hung - I don't know if this will
cause a duplicate post, but I'll try.
Peter Bennett
On Sat, 24 Jan 1998 18:28:08 -0500, Michael Black <blac...@CAM.ORG>
wrote:
>It's
> http://www.science-workshop.com
> not
> www.science-workshop/com
>
>The latter is not a proper address. Every internet address has to end
>with a three letter domain name, such as .org .edu .com .net and so on.
Not necessarily - it could end with a two letter country code, like
.ca for Canada, .uk for United Kingdom, etc.
--
Peter Bennett VE7CEI
GPS and NMEA info and programs: http://vancouver-webpages.com/peter/index.html
Newsgroup new user info: http://vancouver-webpages.com/nnq
John Miles
Mike wrote:
>
> John Miles wrote:
>
> > Mike wrote:
>
> [...]
>
> >> Any good ideas for architecture approaches?
>
> > Well, my old HP8551B has a fairly simple conversion scheme --
> > the first LO tunes from 2 to 4 Gc (it's too old to be GHz :)
> > and works into a 2 Gc 1st IF stage, for coverage from 10 Mc
> > to 2 Gc on the LO fundamental. A more stable, modern LO,
> > together with a decent mixer, would allow you to extend the
> > lower limit as low as a few kc. No spurious responses are
> > possible with this architecture for any signal below 2 Gc,
> > which is nice.
>
> That seems to be the approach we're thinking of. Now, how do
> you get a tracking generator to follow the first LO without
> causing humongous (RFI, feedthrough, crosstalk, interference -
> you pick the word!)
Typically the tracking generator doesn't "follow" the first LO per se
-- it makes use of the actual first LO signal. The idea is to bring
the first LO signal out of the analyzer chassis somewhere (the 8551B has
this jack on the rear panel, but the 8554 and Tek 492 have it on the
front panel.) The tracking generator mixes the first LO signal with a
fixed-frequency oscillator representing the first IF frequency (this
oscillator is usually manually tunable over a very narrow range, to
account for variations in the actual IF frequency). The resulting
signal is run through a filter and a levelled amplifier and used as the
tracking generator's output.
So tracking generators are really trivial beasts. I have an HP8444A
generator, for example, that connects to an HP8554 analyzer plugin for
the 141T. It accepts the first LO signal at 2.05-3.3 GHz, plus the 3rd
LO signal at 500 MHz, directly from the front panel of the 8554. At
narrow sweep rates, the 500 MHz 3rd LO is the one that's swept, rather
than the 1st LO YIG. (This seems to be done a lot in PLL-stabilized
instruments, since their stabilizer circuits are designed to keep the
YIG running at a fixed frequency and sweeping the YIG at the same time
you're trying to keep it phase-locked would pose an obvious problem.)
At any rate, the ~500 MHz 3rd LO is mixed with a manually-tweakable 1.55
GHz internal oscillator within the 8444, to yield the ~2.05 GHz signal
that's mixed with the 1st LO signal to regenerate the tracking signal.
>
> > The 8551B follows the 2 Gc IF with a conversion down to 200
> > Mc, where the majority of the IF gain occurs. This is
> > followed by a final conversion down to 20 Mc, where a series
> > of crystal filters in the display unit provide the various
> > resolution bandwidth options.
>
> Presumably the first IF is actually 2.2 GHz, or the input range
> is really 0-1.8 GHz and the first IF is 2 GHz.
>
> Obviously, the second LO wants to be above the first IF so it
> cannot appear as a spur. There may be some IMD products with
> the presence of two mixers and all these local oscillators
> floating around, but that will be the fun part!
The 2nd LO in the 8551B is at 1.8 GHz, which is also where the manual
rates the end of the first band's coverage. However, the scale goes to
2 GHz, with no spurs visible at 1.8, presumably because of the
analyzer's limited sensitivity and the impressive S12 isolation of the
cavity-based, grounded-grid 1st IF amplifier.
> Presumably, a sharp low-pass filter at the input will help keep
> stuff above 2 GHz out of the spectrum,
Theoretically, yes... however in practice the 8551B uses harmonics of
the 1st LO to tune up to 12 GHz, and they don't bother switching in a
filter on the lowest (1.8 GHz) band. So spurs from signals above 2 GHz
are a real possibility. Not much up there, though, so this isn't
usually a big problem. There is an identifier feature that allows you
to check whether or not a given signal is a spurious response, by
temporarily shifting the LO fundamental a given amount.
> and sharp high-pass,
> bandpass, notch, or combination of all three will be needed to
> keep 200 MHz input signals out of the first IF due to first
> mixer feedthrough.
>
Again, they don't bother with this in the 8551B. Any signals at 2 GHz
or even strong signals at 200 MHz will cause baseline lift,
unfortunately.
> The trick will be to find out where these filters have to go,
> and what kind of performance they need.
>
It's not all that big a problem to begin with, certainly if the budget
is there then you can take steps to minimize the problem, but I used the
old 8551B for several years without being overly-bothered by it.
> That gets us to 200 MHz, where there are plenty of low-noise
> amplifiers available cheap.
Nowadays this is true even at the 1st IF of 2 GHz. The 8551B uses some
sort of exotic tube as its first IF amp, which I've seen in the
Alltronics catalog for something like $300. If it ever dies, the thing
to do would obviously be to replace it with a $1.98 MMIC and stripline
filter. I would expect that 1.8 GHz feedthrough spur to become visible
in such a case, though. A small price to pay to save $300!
> Also, the hams have a lot of
> experience with 2-meter repeaters, so getting good 3rd
> intercept performance should be a piece of cake.
>
Generally you don't want to be piping too big a signal into the front
end of these older mixers, or you will see spurious mixing products all
over the place. 1 mw (!) input to the mixer is the specified maximum of
the 8551B, and I think 10 mw is the specified maximum of the 8554
plugin. The newer analyzers (e.g. Tek 492) have much more robust mixer
modules that can take up to a wat, plus this latter instrument uses two
balanced Schottky diode elements instead of just one, for better LO
isolation and (presumably) better immunity to spurious mixer products at
high signal levels. I shudder to think of what it will cost if I ever
fry THAT puppy. :-)
> > For dedicated HF/low-VHF work, I think the HP8553B plug-in
> > for the 141T converts 100 kHz - 110 MHz up to 200-300 MHz?
> > Something like that, anyway. Again, with such a relatively
> > high first IF, it doesn't take much filtering to eliminate
> > spurs. A high first IF for an HF analyzer also lets you keep
> > the LO tuning range well under an octave, simplifying things
> > considerably if you aren't using a YIG.
>
> My heart wants a yig - lowest phase noise. But I understand
> they are a devil to keep stray magnetic fields out
Not at all. In my experience they're generally decent with no external
shielding at all, but most instruments seem to enclose them in an
external mu-metal shield that makes the whole issue a moot point. At
wide sweeps, you don't care about a little stray magnetism. At narrow
sweeps, you've got to stabilize it with a PLL anyway.
> and you
> have to keep the noise on the dc control voltage down to
> microvolt levels. That would mess up a fast PLL loop, so the
> sweeps would end up slow.
Not usually a problem. The 492, at any rate, solves this problem at
narrow sweep spans by bypassing the YIG's main tuning coil with a hefty
(several hundred uF) capacitor. This effectively puts a stop to random
noise on the main tuning line. The FM tuning coil, which has a much
lower sensitivity to noise, is then driven by the PLL. The sweep rates
are still typically pretty slow, but this is due to the fact that you're
usually running at a narrow resolution bandwidth if you're running a
narrow sweep. I haven't checked, but I'll bet the HP 141T plugins (and
newer instruments as well) do much the same thing.
>
> Of course, the latest Tek and HP spectrum analyzers seem to be
> able to handle 1 Hz resolution at 40 GHz. So we know it's
> possible - but I'd sure like to know how they do it!
>
Dunno, either. I don't have any experience with the newer gear at all,
but presumably the LO synthesizer is much cleaner, and DSP is used to
achieve the super-narrow resolution bandwidths. I am actually not sure
what use those really narrow bandwidths and span ranges are, anyway.
They sure look good on paper, though!
> Once the signal is available at 200 MHz, everything is a piece
> of cake after that. 95 db dynamic range log amps, DSP signal
> processing, software for analysis and data logging, ...
>
Right, the big problems are all effectively solved with boilerplate
techniques at that point.
> Then, the next step is to extend the full sweep to 20 GHz -
> easy. Just multiply everything by 10 :)
>
Right, or the first LO, anyway. They all seem to rely on simple
harmonic mixing for high-band coverage, which can really confuse the
picture with spurs on the older/cheaper instruments. The 492 with
option 1 has a very nice multi-octave tunable YIG filter that is kicked
in on the 1.8 GHz - 21 GHz bands, to serve as a tracking preselector.
Michael Black
You're right, and since I'm in Canada I should know better. In trying to
convey that an address has to end in a well defined suffix, I used
examples that were on my mind, and let the error slip in.
Michael VE2BVW
Steve
John Miles wrote:
>
>>>>> stuff snipped <<<<<
> Nowadays this is true even at the 1st IF of 2 GHz. The 8551B uses some
> sort of exotic tube as its first IF amp, which I've seen in the
> Alltronics catalog for something like $300. If it ever dies, the thing
> to do would obviously be to replace it with a $1.98 MMIC and stripline
> filter. I would expect that 1.8 GHz feedthrough spur to become visible
> in such a case, though. A small price to pay to save $300!
>
> > Also, the hams have a lot of
> > experience with 2-meter repeaters, so getting good 3rd
> > intercept performance should be a piece of cake.
> >
>
> Generally you don't want to be piping too big a signal into the front
> end of these older mixers, or you will see spurious mixing products all
> over the place. 1 mw (!) input to the mixer is the specified maximum of
> the 8551B, and I think 10 mw is the specified maximum of the 8554
> plugin. The newer analyzers (e.g. Tek 492) have much more robust mixer
> modules that can take up to a wat, plus this latter instrument uses two
> balanced Schottky diode elements instead of just one, for better LO
> isolation and (presumably) better immunity to spurious mixer products at
> high signal levels. I shudder to think of what it will cost if I ever
> fry THAT puppy. :-)
Truer words have not been spoken. All the Helwett-Packard analyzers I’ve
ever used, 8566 swept-type through the latest FFT based Dynamic Signal
Anlayzers aand Vector Analyzers did not use any amplification ahead of
the mixer. Of course an attenuator was ahead of the mixer. HP literature
goes inot great detail on the dynamic range considerations of the mixer
and the use of any amplification ahead of it. These HP Application Notes
would be an excellent inof resource for anyone looking to build there
own SA. As far as I know, they can be obtained for free from your local
HP sales rep.
Best I can recall, none of these units have a 95 dB dynamic range (at
least at the specifed measurement accuracy specs). I dig out the books
and verify this.
>
> > > For dedicated HF/low-VHF work, I think the HP8553B plug-in
> > > for the 141T converts 100 kHz - 110 MHz up to 200-300 MHz?
> > > Something like that, anyway. Again, with such a relatively
>
>>>stuff snipped<<<<<<<
Mike
John Miles wrote:
>
> Mike wrote:
> >
> > John Miles wrote:
> >
> > > Mike wrote:
[...]
> > That seems to be the approach we're thinking of. Now, how do
> > you get a tracking generator to follow the first LO without
> > causing humongous (RFI, feedthrough, crosstalk, interference -
> > you pick the word!)
>
> Typically the tracking generator doesn't "follow" the first LO per se
> -- it makes use of the actual first LO signal. The idea is to bring
> the first LO signal out of the analyzer chassis somewhere (the 8551B has
> this jack on the rear panel, but the 8554 and Tek 492 have it on the
> front panel.) The tracking generator mixes the first LO signal with a
> fixed-frequency oscillator representing the first IF frequency (this
> oscillator is usually manually tunable over a very narrow range, to
> account for variations in the actual IF frequency). The resulting
> signal is run through a filter and a levelled amplifier and used as the
> tracking generator's output.
[...snip very nice discussion on HP eq]
John,
Thanks for the excellent info on your equipment. We know where to come
for future questions like "how did they do it?" :)
I'm surprised at the missing filters for the first and second IF. You'd
think for a few dollars added cost they would have been part of the
design specification.
I'm still curious about the isolation needed for a tracking generator. Do
you think it can be part of the analyzer design, or does it have to be in
a separate module?
I know single-shielded coax is only good to roughly -60 to -80 db, that
double-shielded is better, and hardline is best.
Using these as a guide, do you think sufficient isolation could be
achieved with shielding and filtering, or would a tracking generator
simply have to be put in a separate box?
John Miles
Mike wrote:
>
> John Miles wrote:
> >
> > Mike wrote:
> > >
> > > John Miles wrote:
> > >
> > > > Mike wrote:
>
> [...]
>
> > > you get a tracking generator to follow the first LO without
> > > causing humongous (RFI, feedthrough, crosstalk, interference -
> > > you pick the word!)
> >
> > Typically the tracking generator doesn't "follow" the first LO per se
> > -- it makes use of the actual first LO signal. The idea is to bring
> > the first LO signal out of the analyzer chassis somewhere (the 8551B has
> > this jack on the rear panel, but the 8554 and Tek 492 have it on the
> > front panel.) The tracking generator mixes the first LO signal with a
> > fixed-frequency oscillator representing the first IF frequency (this
> > oscillator is usually manually tunable over a very narrow range, to
> > account for variations in the actual IF frequency). The resulting
> > signal is run through a filter and a levelled amplifier and used as the
> > tracking generator's output.
>
>
> John,
>
> Thanks for the excellent info on your equipment. We know where to come
> for future questions like "how did they do it?" :)
>
> I'm surprised at the missing filters for the first and second IF. You'd
> think for a few dollars added cost they would have been part of the
> design specification.
>
> I'm still curious about the isolation needed for a tracking generator. Do
> you think it can be part of the analyzer design, or does it have to be in
> a separate module?
>
> I know single-shielded coax is only good to roughly -60 to -80 db, that
> double-shielded is better, and hardline is best.
>
> Using these as a guide, do you think sufficient isolation could be
> achieved with shielding and filtering, or would a tracking generator
> simply have to be put in a separate box?
>
> Best Regards,
>
> Mike
> CEO, Analog & Digital Design
> Automated Production Test
> http://www.csolve.net/~add/home.htm
>
> Hosting Jonathan Ramsey's Pascal TCP/IP for DOS:
> http://www.csolve.net/~add/zips/tcp.htm
I don't see any real reason why it (the generator) couldn't be in the
same box. You'd definitely want to use a good coax grade internally
(everybody uses hardline, it seems), and a well-shielded BNC terminator
for the front-panel jack when not in use.
I think the reason for the lack of notch filters for the IFs is that
instead of IF feedthrough, you'd have a "deaf" spot, which would be
worse.
-- jm
Mike Mccann
On Sat, 24 Jan 1998, Mike wrote:
> Gary Coffman wrote:
>
> > On 22 Jan 1998 22:46:09 GMT, rohr...@pconline.com (John D. Seboldt) wrote:
> > >Anyone have experience with WA2PZO's "Poor Man's Spectrum Analyzer,"
> > >marketed by "Science Workshop" at Bethpage, NY
> > ><http://www.science-workshop/com/>? I've heard of this idea for some
> > >time, and found them on the web recently. Seems like a useful setup
> > >within the limits of its component parts.
>
> > >John Seboldt K0JD
>
> > I built one, including tracking generator option. It isn't lab quality, but
> > it is a useful instrument.
>
> > Gary
> > Gary Coffman KE4ZV | You make it |mail to ke...@bellsouth.net
> > 534 Shannon Way | We break it |
> > Lawrenceville, GA | Guaranteed |
>
> Like your SIG!
>
> Does anyone know if the URL
>
> http://www.science-workshop/com/
>
> is correct or working? I tried it a couple of days ago, and again just
> now, and get the same result. No DNS entry.
>
> Best Regards,
>
> Mike
> CEO, Analog & Digital Design
> Automated Production Test
> http://www.csolve.net/~add/home.htm
>
> Hosting Jonathan Ramsey's Pascal TCP/IP for DOS:
> http://www.csolve.net/~add/zips/tcp.htm
>
>
The site is working. Try www.science-workshop.com.
73
Mike KB2FMG
Douglas Dwyer
Ive come a bit late to this but I wonder if it would be possible to
construct a cavity resonator with mechanical tuning and high Q to use as
a mechanically swept filter followed by an amp/detector. Perhaps part of
the swept volume of the cavity could include dust iron core to extend
the lower frequency range.
dan
> > Hi Mike!
> >
> > I was thinking 1.x to 2.x and then harmonics mix to go
> > higher. I have a couple of 2-4 yigs. But that would really
> > push on the cost.
>
> Q: Where did you get them and how much do they cost?
>
Surplus. I bought an old 86230B for one. I got a couple
of more from ‘cooldave’ but they have shielding missing.
Scary but for $25 each I will hope not to go too wrong.
> > > I was thinking of using the AD DDS chip (don't have the number handy),
> > > and phase-locking the first oscillator for close-in work.
> >
> > If the ad606 had a RF log output you might be able to
> > DSP 50 MHz at a time. It would also lead to very accurate
> > frequency processing at low cost.
>
> I'd like to hear more on this. Do you mean sampling the output from 0 to
> 50 MHz? Wouldn't that mean more spurs to eliminate, and where are you
> going to get a 100 or 200 MHz sampler? And how are you going to get it
Ø into the PCI bus to work on it?
I was thinking of the AD9058. No over sampling here.
So far it is just a thought. The idea is to use a log RF
output into the ADC so you could get lots of dynamic
range with only 8 bits. But the AD606 doesn’t have
an RF output, just a limiting output.
>
>
> So, how are you going to get the frequency offset? For fairly wide
> sweeps, the tracking doesn't have to be so good. But what if you need to
> use narrow bandwidths for better s/n, or whatever. It's going to be fun
> generating an offset frequency equal to the first IF, between 2 - 4 GHz,
Ø and keep it in a narrow IF. Any ideas?
Even with my hp8558/8444 you have to constantly adjust the
trim when the bandwidth is narrow. With a microprocessor
to help, this might be overcome.
>
> > Those inexpensive thin sheet metal boxes for PC mounting,
> > (Allied sells them), is what I would hope would do the
> > job.
>
> Got to find ones with close-fitting lids - perhaps with a way to mount
> microwave mesh gasket or contact fingers. Maybe even a box-within-a-box
> for the really tough parts.
>
> > > I don't think the electronics will cost that much - it's the shielding
> > > and filtering to get clean signals that will be tough!
>
With the analyzer I built in ’86 I am still very surprised
at how clean it was. I could very slightly see the 32Mhz
log IF and that was when I tried. Otherwise, there were
no other spurs.
> 1) You want to avoid spurs in the design stage, not just by throwing
> shielding at the problem. :-)
>
Yes but shielding is required.
> 2) Instead of rolling your own analyzer from scratch using junk like TV
> tuner modules, another alternative might be to start with the guts from
> an older 491 or 141T plugin, and add computer control and display
> capabilities to it. These units had perfectly good mixers, LOs,
> filters, and log detectors, but lousy displays and awkward user
> interfaces.
>
I am not looking at building a one time analyzer but
reproducible product. My first analyzer used a black
box from a pay TV receiver.
> 3) For the display side of the question, IMHO it doesn't make any sense
> at all to homebrew this, when you can build a simple parallel-port
> interface to dump the output of the log detector to a convenient laptop
> or other PC. You can handle almost all display and control features in
> software with minimal hassle, and end up with a very competent
> instrument with the ability to automatically configure its controls, run
> automated tests, save and print traces, etc. Life is too short not to
Ø use off-the-shelf components whenever possible.
Yes and ideally the analyzer processor would be done on
one (or more) board and then you could go to a PC or
use it in a self contained analyzer.
>
>I'm still curious about the isolation needed for a tracking generator. Do
>you think it can be part of the analyzer design, or does it have to be in
>a separate module?
The only trouble with single conversion is the LO will
be the first IF frequency of the spectrum analyzer. In the
hp 8444 this is remedied by doing two conversions.
It is about impossible to keep a single conversion LO from
feeding through. Otherwise it should be no trouble including
a tracking generator. It is almost becoming a standard
feature of a spectrum analyzer.
*************************
FRANK VAN VLOTEN wrote:
Dan,
When you've finished your move, you may like to consider using the
new AD8307, which is probably a lot cheaper (even for you in the
U.S.) than the AD606. Its quoted input frequency is 500MHz, higher if
you read their very interesting application note that goes with the
specs, and dynamic range quoted is 92dB for +-0.4dB linearity
deviation. Nice to be able to talk about 141Ts and so on, but here at
the bottom end of Africa you roll your own, which is exactly what I'm
doing, with a long persistence CRT for slow scan narrow bandwidth
applications, and all that goes with it.
---------------------------------------------
Dan wrote:
I have down loaded the datasheet for this product and I am impressed
at the cost/performance. I have already jigged and tested the AD606
but it looks like I will have to do the same for the AD8307.
I am following this thread but we are in the middle of a move so
I may be scarce now and then.
Best, Dan.
Chris Muriel
On Wed, 28 Jan 1998 11:04:49 -0800, dan <d...@lakeweb.com> wrote:
>> > Hi Mike!
>> >
>> > > I was thinking of using the AD DDS chip (don't have the number handy),
>> > > and phase-locking the first oscillator for close-in work.
>> >
>> > If the ad606 had a RF log output you might be able to
>> > DSP 50 MHz at a time. It would also lead to very accurate
>> > frequency processing at low cost.
>>
>> I'd like to hear more on this. Do you mean sampling the output from 0 to
>> 50 MHz? Wouldn't that mean more spurs to eliminate, and where are you
>> going to get a 100 or 200 MHz sampler? And how are you going to get it
>Ø into the PCI bus to work on it?
>
> I was thinking of the AD9058. No over sampling here.
> So far it is just a thought. The idea is to use a log RF
> output into the ADC so you could get lots of dynamic
> range with only 8 bits. But the AD606 doesn’t have
> an RF output, just a limiting output.
>
>>
>FRANK VAN VLOTEN wrote:
>
>Dan,
>
>When you've finished your move, you may like to consider using the
>new AD8307, which is probably a lot cheaper (even for you in the
>U.S.) than the AD606. Its quoted input frequency is 500MHz, higher if
>you read their very interesting application note that goes with the
>specs, and dynamic range quoted is 92dB for +-0.4dB linearity
>deviation. Nice to be able to talk about 141Ts and so on, but here at
>the bottom end of Africa you roll your own, which is exactly what I'm
>doing, with a long persistence CRT for slow scan narrow bandwidth
>applications, and all that goes with it.
>
>---------------------------------------------
>
>Dan wrote:
>
>I have down loaded the datasheet for this product and I am impressed
>at the cost/performance. I have already jigged and tested the AD606
>but it looks like I will have to do the same for the AD8307.
>
>I am following this thread but we are in the middle of a move so
>I may be scarce now and then.
>
>Best, Dan.
Whilst you're looking at AD8306,check also the AD8309 -it also
has 95 dB dynamic range but over -75dBm to +20dBm.
They're similar devices -the 8309 uses10 stages of successive
detection versus 7 stages for the 8306.
The log accuracy spec is better with the 8306 if you're looking for
precision measurements.
AD8306/8309 should be priced around $9 to $10 in small
quantities.
The AD606 was designed for ultrasound & for RSSI usage in
communications systems (digital cellular base stations).
Hope this helps a little..
Chris Muriel (G3ZDM).
John Mackesy
I'm a poor man, into HF/Lo VHF spectrum analysis, and I found the answer to
be the Singer SSB-50 ALFRED RF Spectrum Analyzer. It's early-'70s, cost
US$60.00, works like a dream. It's used in conjunction with 2 xcvrs, 1 has
500 Khz IF out, the other 30 Mhz IF out.
There's just one thing I've wondered about, though. What the hell does
ALFRED stand for?
John Mackesy VK3XAO
Douglas Dwyer <ddw...@ddwyer.demon.co.uk> wrote in article
<PV7eoEAA...@ddwyer.demon.co.uk>...
RCrusoe
In article <01bd2c3a$83cb3560$139e0ccb@default>, "John Mackesy"
<ma...@melbpc.org.au> writes:
>
There's just one thing I've wondered about, though. What the hell
>does
ALFRED stand for?
Well, back in the NAV, we used FRED as an acronym. The RED stood for Ridiculous
Electronic Device. The letter "F" had the usual Naval meaning. The computer at
"Beep" school was called FRED, as a mattter of fact. If you can think of what
A and L might have stood for, we might have a plausible guess by putting it all
together.
73 EVRBDY DE AB5VH (Robinson)
From the Beautiful Lower Rio Grande Valley
In the Great State of TEXAS
dit diddy-dit dit.........
Dennis Erwin Thurlow
> US$60.00
If that's the very, very, poor man's cost, I'm screwed.
I was hoping for that $1.98 MMIC chip.
--
geo...@trader.com
Elder - Westminster Presbyterian
TRA #2693
N9VEM
John Mackesy
You get a few things with it - CRT display, 500 Khz calibrator, two-tone
calibrator, amazing resolution and frequency response, and it runs off
115/230V 48 -1000 hz. The manual has pages and pages of REALLY INTERESTING
schematics.
The best part of all - you just plug it in and use it.
Dennis Erwin Thurlow <geo...@trader.com> wrote in article
<34D05C...@trader.com>...
Paul Rampelbergh
Hi,
Where can I find spec of AD8306, AD8309?
What's the band name?
Mike
Chris Muriel wrote:
>
> On Wed, 28 Jan 1998 11:04:49 -0800, dan <d...@lakeweb.com> wrote:
>
> >> > Hi Mike!
> >> >
[...snip all]
Hi all,
Thanks for the excellent suggestions - I just wanted to let you know my
ISP has installed new software and changed to some kind of new connection
to Bell Canada - all the newsgroups are hosed and I may be off the thread
for a while.
But it looks like there are some very good ideas - please keep it up and
I'll catch you later in Deja News :)
JohnsonHE
>
>Hi,
>
>Where can I find spec of AD8306, AD8309?
>What's the band name?
You can download any Analog Devices data sheets by what they call
"Fax-on-Demand". An automated 24 hour a day delivery service. Simply call
1-800-446-6212, and answer by touch-tone the quextions asked. In a few seconds
after hanging up, the Fax will be delivered to the phone number you specify.
W4ZCB
Chris Muriel
On 29 Jan 1998 23:27:26 GMT, rampel...@infoboard.be (Paul
Rampelbergh) wrote:
>Hi,
>
>Where can I find spec of AD8306, AD8309?
>What's the band name?
>
Ah, embarrassment.
They're not fully production released yet (I just checked our
external website which is www.analog.com).
You could try calling 1-800-ANALOGD or 1-781-937-1428
& asking for prelim data for a start.
73 de G3ZDM
Luis Yanes
> 2) Instead of rolling your own analyzer from scratch using junk like TV
> tuner modules, another alternative might be to start with the guts from
> an older 491 or 141T plugin, and add computer control and display
> capabilities to it. These units had perfectly good mixers, LOs,
> filters, and log detectors, but lousy displays and awkward user
> interfaces.
And someone has done the A/D interface to PC and adquisition program yet?
I'm sure this analizer will be more powerfull and nice to use with
a computer display and plenty of storage.
If there is some info about this I'm interested that someone point me
in the right direction, before buit my own from scratch.
73's de Luis
mail: me...@esi.us.es
Ampr: eb7gwl.ampr.org [44.133.41.18]
AX25: (sorry shut down)
http://www.esi.us.es/~melus/
JohnsonHE
>
>Hi,
>
>Where can I find spec of AD8306, AD8309?
>What's the band name?
>
>On Wed, 28 Jan 1998 20:06:11 GMT, (Chris Muriel) wrote:
>
>> Whilst you're looking at AD8306,check also the AD8309 -it also
>> has 95 dB dynamic range but over -75dBm to +20dBm.
>> They're similar devices -the 8309 uses10 stages of successive
>> detection versus 7 stages for the 8306.
>> The log accuracy spec is better with the 8306 if you're looking for
>> precision measurements.
>> AD8306/8309 should be priced around $9 to $10 in small
>> quantities.
>> The AD606 was designed for ultrasound & for RSSI usage in
>> communications systems (digital cellular base stations).
>>
>> Hope this helps a little..
>> Chris Muriel (G3ZDM).
>
problem stems from the fact that somebody here is Dyslexic. The correct numbers
for these devices is 8036 and 8039. Both are preliminary, the 8036 is available
from their Fax on Demand service, Phone number 800-446-6212. The Faxcode for it
is 1836.
The part number 8039 is not yet on the Fax-on-demand but can be obtained from
central applications, phone number 781-937-1428.
Very interesting chip, does 15 dB better than my AD640's, and is 20 percent of
the price.
W4ZCB
JohnsonHE
My turn to be embarrased. There is a typo (LARGE PRINT!) in the Analog Devices
data sheet for the 8306. The correct numbers are indeed 8306 and 8309, the
dyslexia is apparently on the part of AD!
The rest is correct including phone numbers of AD support for data sheets.
W4ZCB
Al Williams
Well, my poor man's spectrum analyzer works like this:
I have a Kenwood TS570D controlled by a PC, I sweep the VFO (using commands
sent from a VB program). As the VFO moves, I read the S meter reading from
the radio using the program and plot a histogram of the signal strength vs.
frequency.
It is farily slow because the Kenwood's VFO doesn't sweep really fast, but
it works -- mostly a novelty, I'm afraid.
73,
Al Williams WD5GNR
http://www.al-williams.com/wd5gnr
Douglas Dwyer wrote in message ...
Eric F. Richards
In rec.radio.amateur.homebrew Gary Coffman <ke...@bellsouth.net> wrote:
> On 22 Jan 1998 22:46:09 GMT, rohr...@pconline.com (John D. Seboldt) wrote:
>>Anyone have experience with WA2PZO's "Poor Man's Spectrum Analyzer,"
> I built one, including tracking generator option. It isn't lab quality, but
> it is a useful instrument.
Hi, Gary,
I have a couple questions on it:
o If I wanted to use it as a bandscope, do I need any of the tuners?
will it take a 10.7 MHz I.F. directly?
o What is its power requirement and do I need to have it isolated from
the receiver that I'll pull the I.F. out of? (In other words, any problem
running off the same 12v supply I'm using now?
Thanks much!
--
Eric F. Richards, KB0YDN
efr...@alumni.cs.colorado.edu
"The weird part is that I can feel productive even when I'm doomed."
- Dilbert
Gary Coffman
On 20 Mar 1998 18:54:59 GMT, "Eric F. Richards"
<efr...@flatland.dimensional.com> wrote:
>In rec.radio.amateur.homebrew Gary Coffman <ke...@bellsouth.net> wrote:
>> On 22 Jan 1998 22:46:09 GMT, rohr...@pconline.com (John D. Seboldt) wrote:
>>>Anyone have experience with WA2PZO's "Poor Man's Spectrum Analyzer,"
>[...]
>> I built one, including tracking generator option. It isn't lab quality, but
>> it is a useful instrument.
>
>Hi, Gary,
>
>I have a couple questions on it:
>
>o If I wanted to use it as a bandscope, do I need any of the tuners?
> will it take a 10.7 MHz I.F. directly?
The tuner contains the mixer and VCO which is swept to create the
spectrum display. So you have to have the tuner. If you took your
receiver's IF directly to the spectrum analyzer's IF, there'd be no
sweep.
The way this thing works is that a sawtooth waveform is generated
to drive the tuner VCO. This causes the frequency sweep. The
sawtooth has a DC offset which is adjustable, and the magnitude
of the sawtooth is also adjustable. This lets you set center frequency
(with the DC offset), and dispersion (span) with the sawtooth magnitude
control.
The VCO is mixed with incoming signals in the tuner's mixer and
a 10.7 MHz IF is extracted. This is fed to a log amp and envelope
demodulator. Selectable filters are provided to set the resolution
appropriate for a given span. The larger the span, and the faster the
sweep, the broader filter you have to use. For panadaptor purposes,
you want to use a small span and a relatively slow sweep in conjunction
with a relatively narrow filter.
>o What is its power requirement and do I need to have it isolated from
> the receiver that I'll pull the I.F. out of? (In other words, any problem
> running off the same 12v supply I'm using now?
There shouldn't be a problem powering it off an existing supply.
You want to couple to the IF in such a way that you don't load it.
A small value capacitor should work.
Realize you have to provide an oscilloscope for display. With
mine, I used a 14 inch Wavetek XY display I happened to have
instead of tying up one of the bench oscilloscopes. I built the
spectrum analyzer into the Wavetek's case and tapped power
from its power supply. That worked great because I could
see the spectrum from across the room on that big display.
Sweep linearity isn't great with this spectrum analyzer. A
marker generator can be useful to tell you where you are.
For a fixed use like a panadaptor, you could make up a
calibrated gradicule for your display and set and forget
the center frequency and span controls.