Pages: [1]   Go Down

Author Topic: Testing a LPF with a spectrum analyzer, but no tracking generator?  (Read 19927 times)

2E0ILY

  • Member
  • Posts: 237

I am building a Softrock Ensemble RXTX SDR transceiver from a kit. It's nearly done and part of the build instructions are to make a low pass filter. The RXTX can be built for one of four pre defined sets of bands and I am building mine for 30, 20 and 17 meter operation. For those bands the filter in the attachment is to be built. I have a spectrum analyzer available,(HP8568B), but no tracking generator for it. Can I use the incrementing output of my signal generator, (Marconi 2019A), using the up and down arrow keys which goes 80KHz to 1040MHz to feed the filter input, and somehow follow it with my spectrum analyzer? I don't doubt the instructions for the filter, I am just curious to learn and actually "see" it working.  Thanks.
Logged
Best regards, Chris Wilson.

WB6BYU

  • Member
  • Posts: 20896
    • Practical Antennas
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #1 on: August 20, 2012, 12:30:27 PM »

You can step the signal generator and record the output level as measured on the spectrum
analyzer, then plot the results by hand.  But that doesn't give you the nice display.

Lacking a tracking generator, one other approach is to use a wide-band noise source as the
signal.  Many of noise bridges used a zener diode as a noise generator with a couple stages
of amplification.  A household appliance that wipes out the RF spectrum with noise could also
be used.  I've run a gross check on filter performance by inserting it in the antenna lead of
my receiver and tuning across the passband listening to the background noise level.

VK2TIL

  • Member
  • Posts: 695
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #2 on: August 20, 2012, 02:06:26 PM »

The signal generator/spectrum analyser setup will work very well; although it's slow compared to swept methods it doesn't take long.  I have done a lot of this kind of measurement with an SA or a power meter.

Try, say, 2 MHz steps from zero to 50 MHz; this should give a good picture of the filter response but you can use, say, 1 MHz steps for a closer look.  Measuring to a frequency a long way above the filter cut-off will show any irregularities in the upper response; just because the skirt slopes down from cut-off doesn't mean that it won't rise at higher frequencies.  Go to 100 MHz if you wish; you may see some interesting things.

Excel is useful to enter and plot the data; if you don't have Excel I think there are freeware programs that do the same, otherwise just a pencil and some squared paper.

Your attachment didn't appear in your post; I assume that the filter is 50 ohms in & out impedance, otherwise you would have to terminate it in its correct impedance.  But an output LPF will be 50 ohms so the 50-ohm generator and analyser can be used directly.

I encourage your desire to "see it working" and the kind of measurement you propose is a great way to true understanding.  You have nice test gear and it will tell you a lot.
Logged

2E0ILY

  • Member
  • Posts: 237
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #3 on: August 20, 2012, 04:08:04 PM »

Thanks fr the replies, I was thinking of making or buying a noise source. I would buy the matching tracking gen for my SA, but I am in the UK and all the ones I can find are in the US, and I have just been quoted $300 just to ship one to me from the States that's on a well known auction site right now. That seems a hell of a price just for the shipping. then plus duty and VAT and the shipping comapanies "fees"..... :(

My sig gen doesn't have a sweep function as such, i can only sweep it by holding the frequency set button down. I have access to a synthesised function gen that sweeps, but it only goes to 20MHz, so not much use. I'll have a play based on your suggestions, thanks for the help!
 I couldn't see a way to actually attach the circuit diagram to a post, sorry, but it is a 50 ohm device. It's at http://www.gatesgarth.com/filter.jpg
.
« Last Edit: August 20, 2012, 04:13:28 PM by BASIL »
Logged
Best regards, Chris Wilson.

VK2TIL

  • Member
  • Posts: 695
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #4 on: August 20, 2012, 04:30:26 PM »

You can set the frequency increment on the 2019; the procedure is in the manual.

Set increment to, say, 2 MHz then set frequency to 2 MHz.  Each press of the Up button will then increase the frequency by 2 MHz so you record filter output at 2 MHz, press Up button, record filter output at 4 MHz, press Up button, ... until you reach the max frequency you want to record.

It doesn't take long.
Logged

KA5IPF

  • Member
  • Posts: 1824
    • homeURL
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #5 on: August 20, 2012, 07:31:39 PM »

If using the Sig Gen and the SA start by establishing a zero reference point using a straight piece of coax. Then insert the filter.

I would start below the designed operating range and step in small steps to see the loss and then when the cutoff is reached increase the step size see the rejection.

Nice SA BTW.

Clif
Logged

AF6WL

  • Member
  • Posts: 224
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #6 on: August 20, 2012, 08:20:31 PM »

A bit of a sledge hammer to crack a nut, but you could get a USB to GPIB adapter, cable both the source and SA and write a little program to step, measure and plot.

So more simply just set the SA sweeping with max hold and use the GPIB to step the source.
« Last Edit: August 20, 2012, 08:30:00 PM by AF6WL »
Logged

TANAKASAN

  • Member
  • Posts: 933
Logged

G3RZP

  • Member
  • Posts: 2254
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #8 on: August 21, 2012, 02:23:27 AM »

If you put the spec. an. into 'max hold' and set it sweeping the range you are interested in, you can then sweep the signal generator. Done for long enough, it gives similar results to a noise generator - which method does need a fair amount of amplification if you are going to see a 50dB filter stop band. It's not a brilliant method, and can hide the dips in the stopband, but it is fine for a start.

Another method is to go the US for Dayton, having made arrangements to have a tracking generator shipped to you while you are there. I went to San Diego on business and brought mine home as checked baggage. As 'obsolete electronic equipment for a museum', I wasn't charged any import duty either!
Logged

AA4HA

  • Member
  • Posts: 2689
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #9 on: August 22, 2012, 03:23:17 PM »

You can make a noise source but there are no guarantees it will be linear. I just use a variable signal generator and graph paper.
Logged
Ms. Tisha Hayes, AA4HA
Lookout Mountain, Alabama

WB8VLC

  • Member
  • Posts: 1155
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #10 on: August 23, 2012, 11:19:13 AM »

The Marconi 2019 has a one touch frequency increment feature that is standard.

  If you hold the up or down frequency key in then the frequency will automatically step up or down in a preset increment then wait one second on frequency and repeat incrementing or decrementing until it reaches a preset maximum or minimum frequency after which it will then roll back to a pre programmed lower or higher frequency.

There is also a method to eliminate the 1 second hold time on a stepped frequency but since I don't have that model gen anymore I don't remember the exact method to set it up.

These features should be in the 2019 manual on how to set the min and max frequencies and the increment step and how to make the generator perform a pseudo sweep.


Mike
Logged

2E0ILY

  • Member
  • Posts: 237
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #11 on: September 09, 2012, 02:22:08 PM »

Using the frequency increment feature it works quite well, but to be honest it's still a bit of a PITA. I am going to have to bite the bullet and get the matching tracking gen for the analyzer. It's not so much the price of the generators, it's just that they are all in the US and I am in the UK, so carriage, duty and VAT are a killer. Many thanks for the detailed replies, and sorry for the delay in replying, sadly my father in law is very ill so I have been unable to devote as much time to "play" as I would have liked. Cheers gentlemen.
Logged
Best regards, Chris Wilson.

VK2TIL

  • Member
  • Posts: 695
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #12 on: September 09, 2012, 02:58:04 PM »

Sorry to hear about your father-in-law; hope he recovers.

If you are shopping for a HP8444A TG, make sure it has Option 59; that is required to make it work with your 8568B.  Make sure that there is a Option 059 sticker on the rear panel.

The 8444 was originally designed for the 8554 plug-in for the 141; option 59 was introduced later to make it work with the later 8558, 8568 etc analysers.

Pity we are on opposite sides of the world; I have two 141s, several plug-ins and an 8444A Opt 59 gathering dust in the store-room.  I would be happy to see them go to a good home.
Logged

2E0ILY

  • Member
  • Posts: 237
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #13 on: September 09, 2012, 03:18:18 PM »

Sorry to hear about your father-in-law; hope he recovers.

If you are shopping for a HP8444A TG, make sure it has Option 59; that is required to make it work with your 8568B.  Make sure that there is a Option 059 sticker on the rear panel.

The 8444 was originally designed for the 8554 plug-in for the 141; option 59 was introduced later to make it work with the later 8558, 8568 etc analysers.

Pity we are on opposite sides of the world; I have two 141s, several plug-ins and an 8444A Opt 59 gathering dust in the store-room.  I would be happy to see them go to a good home.


Hmmm, I wonder how much the shipping would be, if you feel it worth the hassle you might find out, I would definitely like one at the right price :) Thanks for the kind words re my fater in law, sadly his illness is terminal, but he lived a "full and interesting" life, and is not a young man. Thanks.
Logged
Best regards, Chris Wilson.

VK2TIL

  • Member
  • Posts: 695
RE: Testing a LPF with a spectrum analyzer, but no tracking generator?
« Reply #14 on: September 09, 2012, 04:12:26 PM »

I don't think that shipping from VK would be a proposition; these things are heavy.

In addition, I can't guarantee that it works.  It was working perfectly when it was stored-away some years ago but you never know.

The 141 systems aren't fully operational (the 141s and the various RF sections were working but both 8552 IF sections failed so it was all put-away and I "went digital") so I can't test it.

It would be a pity if you paid hundreds of $$$s for shipping and it didn't work.

I would like someone local to take the whole lot.

I'm surprised at what they are worth or, at least, what is being asked for them.
Logged
Pages: [1]   Go Up