Scope and Generator Spectrum Analysis / General Chatting / Forums

Forums

4hv.org :: Forums :: General Chatting

« Previous topic | Next topic »

Scope and Generator Spectrum Analysis

1 2 next

Move Thread

LAN_403

Hazmatt_(The Underdog)

Fri Jan 26 2007, 07:23AM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

I was just wondering if any of you have set up your O-scope and a sweeping generator to give you a sprectrum of a signal in frequency domain. I have a 10MHz sweep function generator and an analog scope that I could setup to see a spectrum, theoretically, If I knew how to set this sort of thing up. I'm just not sure how to do it. 10MHz is probably all the bandwidth I really need for observations, so If anyone has done it, don't hesitate to spill your guts. I'd like to hear what you've done and some theory would be nice too.

Steve Ward

Fri Jan 26 2007, 05:36PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Seems that you would need to trigger the scope from the sig-gen (need a sync pulse at the start of the sweep), and set the time base for the length of time it takes for 1 sweep to occur. Seems like a really neat trick (i hadnt thought about it until you asked). If that doesnt answer your question (which i suspect it may not), then i think we might need to know more about the capabilites of your equipment.

Oh, and make sure your sweep is slow enough for high Q circuits that might exhibit a long transient response (im assuming you only want the steady state response).

I might have to play with this on my own some time.

Steve Conner

Fri Jan 26 2007, 06:02PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

To be pedantic, that combination can only display the transfer function of a system, not "a spectrum of a signal". The other Steve just assumed that you knew that and meant to say transfer function.

If a generator has a sweep function, then it will almost always have a trigger output to sychronise a scope to the sweep, or a trigger input to synchronise the sweep to a free-running scope, or marker pulses, or whatever. The sweep really isn't much use otherwise.

One nice add-on to this setup is a log detector to make your scope screen read out in dB.

Sulaiman

Fri Jan 26 2007, 07:00PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Assuming you can set your 'scope for X-Y input (lissajous etc.) you need;
A relaxation oscillator to produce a sawtooth waveform feeds the X-axis and a Volteage-to-frequency convertor
The output frequency goes to the device under test
The output of the device under test can go directly to the Y-axis input OR it can go to the Y-axis via a rectifier.

The sawtooth generator can be a 555 astable using a constant current charging source
The VCO could be any suitable.

By the time you've built and CALIBRATED everything you'll have a nice piece of kit.
A used SWEEP GENERATOR off eBay would be easier.
e.g. item# 120078223034 270081683800 120078239519 140077282623

...

Fri Jan 26 2007, 07:13PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

When I was working at ortel they did something like this:

They took their scope, and set it for XY. Then they had a fancy sweep function generator's output fed into the filter they were testing. the output of the circuit was the y-axis of the scope. The sweep output of the gen went into the x-axis. It would make a line that moved across the front of the scope at about 1hz that started out small, peaked at the resonant frequieny of what they were testing, and then fall back down.

They used it for about 10 minutes, and then decided that is was easier to use the scope normally, and watch the sine wave as they varied the frequiency...

BTW, if you wanted to make one of your own, you would jut need a slow sawtooth generator (icl8038 comes to mind, but really since it is running at a constant frequiency you might make it out of descrete components) and a fast VCO (take your pic, max038 would have been ideal if they still made it). You feed the sawtooth generator into the vco and the x-axis of the scope, and the output of the vco into the circuit.

Hazmatt_(The Underdog)

Fri Jan 26 2007, 09:47PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

I'm going to start looking at some of the DIY analyzers in more depth, and get some things together, like sawtooth generators and things. With the ICL8038 (which I have a few of) I can hook him up to my DIY counter as a tracking oscillator counter, if I'm not mistaken.

The "Plan" is to use my good scope (2430A with Tekmate) for the basic output of the Tesla coil like Terry does. The antenna is done, and I'm going to make some Impedance probes too.
Then, I will want to setup my BK 2125A scope and the BK4017A sweep generator as an analyzer.

It sounds like I'm going to have to make some HF mixers, but that's okay, I can use ARRL examples and design guidelines to make those.

Steve Ward

Fri Jan 26 2007, 09:58PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

To be pedantic, that combination can only display the transfer function of a system, not "a spectrum of a signal". The other Steve just assumed that you knew that and meant to say transfer function.

Doh! Yes, i was simply assuming that you had some circuit and wanted to know H(w) (transfer function). But now its clear to see he actually needs something like a fast fourier transform (FFT) for his application.

Hazmatt_(The Underdog)

Fri Jan 26 2007, 10:10PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

nono...that's okay. Transfer functions are very important too. I need H(w) for my analysis, so its okay.

Bored Chemist

Sat Jan 27 2007, 12:26PM

Registered Member #193 Joined: Fri Feb 17 2006, 07:04AM
Location: sheffield
Posts: 1022

I think what you are talking about used to be a pretty standard bit of kit for checking the frequency response of speakers, Hi Fis etc. It's a useful thing if you do that sort of work.
Somebody will need to check my maths on this, but I think you can use much the same sort of kit to do a real frequency analysis. (Believe it or not, people did this sort of thing in the analoge domain before they had computers.)
Say you have some signal that you want to analyse and you multiply it by a 1 KHz sine wave and average the result. For most frequencies you get nothing- for every occasion where the product is positive it's negative half a milisecond later. On the other hand, if the input signal is also at 1KHz then the signals multiply together and you get a non-zero product.
If I'm righht then all you need is a mulitplier circuit and you have the basis of a swept frequency spectral analyser.
Multiply the input signal that you want analysed by the swept frequency from the generator, low pass filter the product and use the scope to plot that filtered product against the "sweep" output from the sig gen.
There is a potential problem with phasing- I guess ideally you need two generators 90 degrees out of phase and 2 multipliers and 2 scoope traces to give the real and imaginary parts of the analysis. It would also be possible to combine them to get the magnitude of the frequency response.

Sulaiman

Sat Jan 27 2007, 02:22PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

For a TC it should be straightforward;
Set your Sweep Generator to give the X-deflection and source signal;
drive the primary with the source signal and directly monitor the TC topload signal
with a capacitive pickup on the 'scope y-axis.

The Y-axis signal will be the +ve and -ve envelope of the frequency response
so set Y-axis in centre of screen and use ac coupling.

Should be able to lash it up in a minute or two.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob