EEVblog Electronics Community Forum

Products => Test Equipment => Topic started by: Dan Moos on August 18, 2020, 02:11:58 am

Title: How to not blow up spectrum analyzer
Post by: Dan Moos on August 18, 2020, 02:11:58 am
Just won a HP 8590d on eBay. Price was right. Anyway, I know that the front end of these things are easily damaged. I suspect I should get an attenuator. Am I correct? I'm new to rf. I just want to be safe.
Title: Re: How to not blow up spectrum analyzer
Post by: 0culus on August 18, 2020, 02:23:43 am
DC block too.
Title: Re: How to not blow up spectrum analyzer
Post by: joeqsmith on August 18, 2020, 02:36:10 am
I treat my SAs the same as my scopes.  I don't normally do anything additional beyond wearing a wrist strap and make sure I discharge what ever I am hooking them to.  For the older SAs, I would normally leave the 10dB attenuator selected. 

If I need more protection, say I am working with higher voltages, I will use some combination of Citel GDTs, diode clamps and attenuators.     If I am looking at higher power devices (amplifier for example) I have a small 50 ohm load with a tap built-in.   
Title: Re: How to not blow up spectrum analyzer
Post by: DaJMasta on August 18, 2020, 02:38:44 am
A DC block can be a good idea to have onhand, unless you're working with really high power signals you shouldn't really need an attenuator, and leaving it connected all the time just removes your low noise/low signal performance.


Proper spectrum analyzers should be pretty durable in terms of desktop generators and signals.  If you're connecting up to homebrew then the external stuff can be there to make sure, and if you're dealing with broadcast then you want a beefy attenuator... but they are designed to be failure resistant, at least.  The specs say it can tolerate 1W continuous (+30dBm) and 25V DC, which at least for desktop stuff where a signal generator's maximum output is probably +20dBm or less, you really don't need extra protection equipment.

The most important thing is to know the devices you're measuring and being careful/using appropriate protective equipment when you think there's a risk.
Title: Re: How to not blow up spectrum analyzer
Post by: Dan Moos on August 18, 2020, 02:53:07 am
Ok, good info.

The thing was advertised as "working, but not calibrated" . Got it for $300 plus $98 shipping. I bid the  minimum, and was the only bidder. Being the only bidder on an instrument that was cheaper than every other similar one I saw concerns me. As a hobbyist, how concerned about not being calibrated should I be? And out of curiosity, how much does a cal for such a thing usually cost?
Title: Re: How to not blow up spectrum analyzer
Post by: 0culus on August 18, 2020, 03:12:24 am
It may be "close enough for government work" as it were...you aren't going to be doing anything that requires traceable cal I'm betting, so just see how it performs when you get it. The main thing now is to hope that the seller packs it right.
Title: Re: How to not blow up spectrum analyzer
Post by: Electro Fan on August 18, 2020, 03:22:51 am
DC block too.

+1

DC block lives on the SA
Title: Re: How to not blow up spectrum analyzer
Post by: Dan Moos on August 18, 2020, 03:23:41 am
"Close enough for government work" is all I need.

Is adding a tracking gen to one of these expense/difficult/with the hassle?
Title: Re: How to not blow up spectrum analyzer
Post by: Dan Moos on August 18, 2020, 03:26:10 am
DC block seems to be a recurring theme here. I presume  you mean more than  just a coupling cap. Any specific suggestions?
Title: Re: How to not blow up spectrum analyzer
Post by: 0culus on August 18, 2020, 03:28:42 am
DC block seems to be a recurring theme here. I presume  you mean more than  just a coupling cap. Any specific suggestions?

I use Mini-circuits DC blocks. They make a range of them...I have some 18 GHz ones that I always use on my 8566B. A true dc-block is indeed just ac coupling the signal to protect spectrum analyzers that can't be dc coupled at all. However, quality ones will have a nice flat frequency response across the specified bandwidth. Definitely one thing to buy new and not used.
Title: Re: How to not blow up spectrum analyzer
Post by: 0culus on August 18, 2020, 03:30:37 am
"Close enough for government work" is all I need.

Is adding a tracking gen to one of these expense/difficult/with the hassle?

Honestly? Tracking generators aren't that useful anymore when you can purchase a NanoVNA v2 for under $100 and then you also get to have vector measurements.
Title: Re: How to not blow up spectrum analyzer
Post by: Dan Moos on August 18, 2020, 03:36:52 am
On tracking gens...Remembering I already bought the HP, let's say I want to look at the freq. response of a filter I've built with it. I have a few sig gens. Tracking gen not needed? Can the HP be synced to the sweep trigger pulse of my Siglent 40 MHz arb gen for instance? Does it need to be?
Title: Re: How to not blow up spectrum analyzer
Post by: 0culus on August 18, 2020, 03:42:27 am
Depending on the connections available on your other stuff, you could probably cobble something together with a few mixers and things. FWIW, I have an old version of this: https://www.dsinstruments.com/store/products/tg6000-rf-tracking-signal-generator-usb/ (https://www.dsinstruments.com/store/products/tg6000-rf-tracking-signal-generator-usb/)

It works very well, and the new ones don't need an external CW source like mine does. It should work on any spectrum analyzer that provides a 1st LO output (most do).
Title: Re: How to not blow up spectrum analyzer
Post by: Dan Moos on August 18, 2020, 03:49:17 am
Ha! That thing is $700! I have only $400 into the SA!

Is there no way to sync my Siglent to the sweep on the HP? Some combo of using the sweep trigger signal on the Siglent arb gen, and setting the  span/rbw on the HP?
Title: Re: How to not blow up spectrum analyzer
Post by: tkamiya on August 18, 2020, 01:16:16 pm
Lab I use charges somewhere around $200 to $250 to calibrate 859x series.

To check filters, etc, you don't necessary need a tracking generator.  Output of a sweep generator goes to filter and its output goes to SA.  Since they are not synchronized, it will take a while to fill the screen.  But it can be done.

Better, yet....  if you are going to do that, it'll be easier to do sweep generator to filter to a detector into an Y axis of oscilloscope.  Then drive the X axes by ramp signal from sweep gen.