| Duo |
My latest and greatest project is a home made spectrum analyzer.
I've built the basic prototype using a TV tuner board, a 555 sweep generator, some op-amps to achieve DC level shifting and amplification, and a simple transistor RF amp and diode detector.
The display is a general purpose oscilloscope.
Here is a picture of the analyzer displaying the output of a nearby oscillator running on a battery at about 100MHz. The analyzer is picking it up through the air.
More on it later, I will be doing a great deal of work on this to make a full-blown analyzer capable of operating from 0 - 500MHz or so. |
|
|
| DigitalJunkie |
| Count me as extremely interested! :att'n: |
|
|
| AR2 |
Cool,
I am one more in the line.:smash: |
|
|
| jacquesl |
| You’ve got some nice equipment there man |
|
|
| gbyleveldt |
Heehee, Jacques, ek is mal oor jou Avatar. Dit moet suig om in Kaapstad te bly :D :D
Sorry guys for being off topic... |
|
|
| jacquesl |
Thanks man, nah, I’m not in Cape Town but near it, the only that suck is some down time from koeberg once in a while, power outage
EDIT: Huh, didn’t you typed that in Eng a couple of seconds ago? :scratch: |
|
|
| gbyleveldt |
| Didn't want to distract the thread too much... But now I have :smash: |
|
|
| Duo |
Here is a hand-drawn image of my working prototype schematic.
There are no embellishments as of yet; no log amp in the video section, no filters, no bandwidth limiting, no sweep rate control, no nothing really.
It does, however, function very nicely in the bands available to the TV tuner. I've only got it set up on a 12V power supply and TV tuners usually like around 30V tuning range, so later I'll make a 30V supply and driver circuit to allow the tuner sweep to go all the way.
The extras I -did- add were spectrum width and center frequency controls. This at least allows me to move the band a bit and set my span.
I'll get back with more after I've developed the additions. I do plan to make some heterodyne circuits to allow this to tune in ranges outside of the tuner's bands; such as audio frequency, AM and HF radio bands, and perhaps something into the microwave area.
 |
|
|
| infinia |
Hi
Nice work so far.
You can possibly use a single chip log amp for use of your AGC/det function. Use an opamp for some baseband/DC gain after this.
Is the tuner unmodified (ie PLL still closed loop) for close to the carrier (that is for audio) performance? I forgot the last IF freq for these tuners is it a SAT/TV tuner? |
|
|
| Duo |
The last IF of a north american TV tuner is 45.75MHz.
I am going to look at what log amp chips I can get as samples.
The tuner itself is not modified. For audio performance, I'll be using a balanced modulator. |
|
|
| infinia |
Hi
I see a modulator for audio. you can have a wireless audio analyzer, hahaha.
Last time I looked ADI had some cool log amps. |
|
|
| Duo |
| You see a modulator for audio? Where? |
|
|
| infinia |
Right here what you said. "I see" aka "I understand"
| quote: | Originally posted by Duo
For audio performance, I'll be using a balanced modulator. |
|
|
|
| Duo |
| Ah, sorry I get you now. ;) |
|
|
| pinkmouse |
| Didn't Hickman do something similar in The Analogue Circuit Cookbook? |
|
|
| Duo |
| I'm not sure about him specifically, but I know that this topology is very popular among amatuer radio operators because of its simplicity, low cost, and high versatility. |
|
|
| pinkmouse |
| I wasn't knocking your project at all, just re-reading my post, and it may have seemed that way, sorry. It's just - you know how you sometimes have a feeling in the back of your mind that you've seen something before? :) |
|
|
| Duo |
| hehe don't worry, I didn't think you were knocking my project. |
|
|
| DigitalJunkie |
Lately I've been looking over this setup (perhaps the same one Pinkmouse is talking about?)
http://www.science-workshop.com/
Here's a site with some more info,etc.
http://www.nitehawk.com/rasmit/sa50.html
I was tempted to start ordering the parts of that kit,But maybe I'll see where this project goes first. ;)
I mostly need a spectrum analyser for my recent RF projects..But they're way too spendy for a poor geek like me..Unless I can build something suitable myself,I've got a well-stocked junk box! :D |
|
|
| Duo |
Those are the same two sites I've been looking at for ideas.
Most of what I'm doing is just out of my own head, but there are only so many ways of doing these things, so mine looks about the same as the other's. |
|
|
| infinia |
hi
If you guys want to get serious with this SA stuff.
1) Get familar with the tuner you are gonna use. ie learn I2C to tune it. Do Not open the PLL loop. Most cheat by running it open loop with all the badness that this implies ie freq drift, FM/PM noise,low PSRR.
2) A lot of the work will be in the freq. span circuitry. Now the LO PLL loop will try to fight it at lower span rates. Then the tricks come in here. |
|
|
| Duo |
Haha, the tuner I'm using didn't have a PLL in the first place. However, I simply selected that one because I could get going with only the implementation of a simple sweep generator.
I do agree, and in the future this probably will be I2C controlled.
I have two AD8307 chips on sample order now; those should be here in about a week.
The other thing I may do is instead of using a tuner with I2C; design a PLL that is voltage/current controlled like the old days. This eliminates the necessity for computer control. Part of the reason for this is that I have a 10MHz time base accurate to six digit precision which will be disciplined in the future by a GPS receiver. (I know I could also use this time base with an I2C chip as well.)
We'll see what happens as this goes along. For now I'm waiting for the AD633 chips to show up. These are wideband balanced modulators which I intend to use in implementing the audio frequency input function. The 633 functions to 2MHz. I will use the 1MHz output of my timebase with the AD633 to be modulated by audio as double-sideband-supressed-carrier (DSBSC), then filter off one sideband (the lower one), and use this again in another heterodyne circuit to get it into the low 100MHz region so it can be sent directly to the analyzer.
More later. |
|
|
| FastEddy |
Off topic: Anyone got recommondations for PC based oscilloscope, preferably for use with a Mac laptop via USB ports ...?
:confused: |
|
|
| Duo |
My tuner is a Sharp... hehe
I'm not looking to spend money on this project. This is purely a junkbox effort, I'm not in need of a high end SA, just something capable of basic work. |
|
|
| panomaniac |
Very cool project!
A year or so ago I was seriously trying to do this, too. Was inspired by the "Poor Man's Spectrum Analyzer". Also bought some stuff off eBay.
The problem for me was that I wanted to do audio band and above, maybe out to 3Mhz - for working on Class-D stuff. But no tuner would cover that band.
Asked the guys at science-workshop about it, they couldn't help, as they didn't know any tuners that would work for that range.
So I'm VERY intersted to know how you will get the lower frequencies. That would rule! |
|
|
| Duo |
If you perform AM modulation of a carrier with audio, you get a carrier and sidebands as output. Each sideband is the carrier frequency plus or minus the audio frequency. The sidebands literally exist on either side of the carrier.
If you look at one of these sidebands on the spectrum analyzer, you find that it is actually showing the spectrum of the audio that you modulated the carrier with.
If you modulate SSBSC (Single Side Band Supressed Carrier), you end up with nothing but a sideband, which could be of a frequency that any TV tuner will accept.
What you've basically done here is converted the audio band up to a radio frequency band that the analyzer can look at.
My intention with the AD633 chip is exactly as above. It is a doubly balanced modulator which will accept frequencies as low as DC and as high as about 2MHz. I can modulate a 1MHz carrier with audio, then upshift that to maybe 50 or 100MHz for the tuner, and be on my way.
(At least in theory ;)) |
|
|
| Duo |
The next design issue I face in the prototype right now is that of a suitable IF bandpass filter. This is because the output IF of a TV tuner is 6MHz wide, which is useless if you want any sort of resolution from the analyzer.
I can't decide whether to try a crystal ladder filter, a storebought filter of some sort, or who knows what.
Finding a crystal with a fundamental around 45MHz is probably going to be very difficult now that I think of it. |
|
|
| panomaniac |
| Cool! Let us know how you get on with it. I want one. |
|
|
| infinia |
Hi Duo
Are you replacing the TV tuners filter? The tuners usually have 45MHz SAW filter. The AD8307 data sheet has some usefull info on a tuned matching circuit for the IF input that has good additional LF rejection. I think its possible to use overtone crystals for a filter but this narrow band filter would be project unto itself I think. Maybe add another conversion to DC with digital LP filters would be the most flexible design. |
|
|
| jackinnj |
Ever see this in Radio Electronics -- September 1989:
 |
|
|
| jackinnj |
From the manual of the Tektronix 5L4N:
Variable Resolution
The variable resolution circuit consists of four amplifier
stages, each containing a crystal filter with variable
bandwidth from 10 Hz or less to 3 kHz. Automatic
bandwidth variation, as a function of sweep rate and span,
is a feature of each stage. Gain compensation maintains a
constant output level as the bandwidth is changed. The
signal level at TP404 is 2 V p-p for full screen display in
10 dB/DIV mode.
The first stage consists of an operational amplifier
driving the input of a crystal filter. Gain is about two. T320
provides about a 4:1 current gain to drive the crystal in
narrow bandwidth. Resonant frequency of the circuit is
250 kHz. Adjustment C328, in series with the crystal, sets
the resonant frequency of the crystal. Adjustment C324
neutralizes the effects of crystal parallel capacitance and
is adjusted for response symmetry at 20 dB down.
A parallel resonant circuit, in the output of the crystal,
consisting of L324, C324, plus stray circuit capacitance to
ground, is tuned to a center frequency of 250 kHz. The Q
of the circuit determines the response bandwidth. R325
sets circuit Q and is adjusted so the bandwidth, at 1.2 dB
5-3
Circuit Description — 5L4N
down, is about 3 kHz when the resolution bandwidth is set
for maximum. The output load for the crystal, and
consequently the bandwidth of the filter, is determined by
the shunt load of the photo-resistor-LED 1C (U338) in
series with gain adjustment R330. The resistance of this
photo resistor determines the actual operating bandwidth
of the stage and varies from about 200 k n at maximum
bandwidth, to about 250 Q for minimum bandwidth. The
resistance of the photo resistor is a function of the current
through its LED, which in turn is set by the front panel
RESOLUTION control or the automatic resolution circuit.
As the resolution bandwidth decreases, the load on the
crystal filter decreases the output voltage by an amount
proportional to the current through U338 and R330. This
generates a compensating current out of Q335 that is
summed with the output current of Q330 to maintain a
constant input drive current for second operational
amplifier stage Q350 and Q360. Gain from the first to
second stage is determined by the ratio of R356 to R328,
so the drive level to the next crystal is a 2 V peak-to-peak
signal for full screen deflection in the 10 dB/Div mode.
The gain of the remaining two stages is unity, therefore,
this 2 V peak-to-peak signal level is maintained through
the fourth stage.
As previously stated, control of the bandwidth is
accomplished by changing the current through the LED's
for the photo resister IC's. The LED of each package is in
series to the +30 V supply. Each stage is decoupled to
prevent interaction. Current is determined by the voltage
input to operational amplifier U340B. The voltage from the
resolution control circuits (at pin S) is transferred into an
exponential current to allow summing the voltages
proportional to the span and sweep rate. This provides a
voltage and current that automatically controls the resolu-
tion. Q340A is thermal compensation for Q340B.
The output stage contains a two pole bandpass filter
with a bandpass about 6 kHz at 250 kHz center frequency.
This reduces the level of wide band noise that is generated
by the VR circuit and passed on to the display function
amplifier circuit. Transistor Q420 provides the isolation
between the VR circuit and the function IF and provides
gain. The output is normalized at approximately 0.7 V
peak-to-peak for a full screen display level at 10 dB/Divby
R424. |
|
|
| Duo |
Infinia: I have not planned on modifying the tuner at all at this point. I was simply going to add more filtering on its IF output to allow for narrow enough bandwidth to get good resolution. Perhaps it may make it easier to design a filter if I downconvert the IF to a lower frequency. I'm not interested in getting into digital filters here. This is not intended as a high end analyzer. I will do what I have to in the analog domain to get decent operation, I know it has been done before.
jackinnj: I haven't seen that specific article but it seems to go along with what I'm doing. Where did you get that manual for the Tek analyzer. Do you have that analyzer? |
|
|
| jackinnj |
I have 2 of them, one was a junker that was missing one of those proprietary Tektronix chips -- turns out that there was an identical chip in a Tektronix timebase.
You need the patience of Job to adjust the 5L4N.
If I get around to it I'll scan the schematic -- why don't you make life a little complicated and use a 2nd IF -- then the variable resolution would be much easier. |
|
|
| Duo |
| How do you mean second IF? As in a downconverted one as I mentioned? |
|
|
| jackinnj |
There have been acouple of articles in QST and QEX -- this is a dual conversion spectrum analyzer -- if you get the IF down to 100 or 250 kHz then you can easily make the crystal filters
http://www.arrl.org/tis/info/pdf/9808035.pdf
fwiw, there is an HP 3581A "Wave Analyzer" going on EBay for $10 -- this is an incredible instrument for that price -- for a spec analyzer all you need is an HP 3581A and some rudimentary programming skills with microprocessors.
if you see an HP 3586 selective level meter -- it can also be pressed into service as a spec analyzer -- another incredible instrument which goes for a song on EBay. |
|
|
| Duo |
| I would love to get that 3581A, but I cannot afford it right now. |
|
|
| jackinnj |
| quote: | Originally posted by Duo
I would love to get that 3581A, but I cannot afford it right now. |
If it goes for $10 - $12 perhaps cigarettes and beer are too expensive in Canada :) |
|
|
| Duo |
| Exactly what do you mean by that? |
|
|
| Duo |
Does anyone recognize these?
These are mixers I believe. Probably equivalent to the 10514A series but as a plug in brick instead of with BNC connectors. Apparently the 10514A is good from 200KHz to 500MHz.
I would at least like to know what the R X and L designations mean as far as LO, RF, and IF would be used.
 |
|
|
| Duo |
| I would love that thing, especially because I have a plugin for it, but I could not afford it right now. |
|
|
|
