First Spectrum Analyzer , new Rigol, used HP on ebay ?

[KJDS]

I've never used the Rigol, but I'll comment on a lot of the others.

If all you want to do is to have a look at output power and check that the modulation is looking reasonable, then an E4406A offers great value for money. There were hundreds dumped on the second hand wholesale market a few years ago and I still have a few left, but they're all 12 bit. It's easy enough to check a few frequencies where harmonics and other predicted spurious are, but not much use for wideband work.

The HP859X are also common on the second hand market. These rarely come up with the TG option, but the full service manual (apart from the PSU) is available and they are fairly easy to work on.

The HP856X are far better instruments in terms of RF performance, but are also rarer and higher cost. I keep a 22GHz 8563A on my bench for serious work

Anritsu make some good products, I also keep an MS8604A transmitter tester on my bench for 8.5GHz coverage and this has the added advantage of a built in 10W attenuator, so far less care is needed in everyday use than most spec anas

Also mentioned on this thread has been the Marconi 2955 radio test set. It's a great bit of kit to have lying around, a basic sig gen, frequency counter, mod meter, power meter and simple audio analysis all in one box and they are readily available second hand though do fetch good prices. The 2945, which is smaller, lighter, lower power and has a spectrum analyzer is far better instrument but even though tens of thousands have been sold new, they are very rare second hand.

I don't mind the R&S stuff but I've not used them often enough to comment in depth

[Howardlong]

thanks for all this information !
Part of the issue i have is my lake of knowledge. My background is digital design, and Radio/RF is a new hobby for me.
Beyond wanting to work on FM/AM/Filters maybe some Ham the main reason of buying SA is educational for me. And i will probably know better what will be most useful for me after owning one.

I am fairly sure from your desicription that it'll spend 99.9% of its life gathering dust after an intial hour or so of switching on.

So you don't think for building/tuning AM/FM radio SA will be useful  ?  Or for educational purposes of RF in general ?
not sure i follow ...

No, for building and AM/FM radio, a signal generator and SINAD meter would be far more useful. What I have been suggesting is a test set with signal generator, SINAD meter, transmitter tester AND a spectrum analyser, some with TG too.

[cncjerry]

My first SA was an Advantest R3131a with TG.  Nice analyzer, good to 3Ghz, portable, quiet, etc.  I wouldn't buy an SA without a TG but having written that, a cheap noise generator will also suffice if your SA has a way to subtract A-B, etc to normalize the curve.  With a noise generator it is also helpful to have peak detection or some kind of infinite persistence.

Thinking my battery is about ready to go in the Advantest, I recently bought an HP 8568B locally off Craigs List with the 85685 RF preselector and TG.   It was too good of a deal, has lots of buttons, knobs, etc, very customizable, GPIB programming that can automate all kinds of tests, etc.  This SA stack makes the Advantest look like a toy but it is loud, very loud.  I was looking at a low pass filter last night and the HP has much more resolution than the Advantest.

Both of these SAs were in perfect condition and cost around 1/2 the Rigol ($750 for the Advantest, $850 for the HP stack, maybe a little more for the TG I added).   The Rigol has had a ton of bug reports so make sure they don't impact your expected use-cases. The one in particular that put it out of the running for me (prior to picking up the HP) was the 10Mhz issue with the TG.  That would drive me crazy.

As someone else mentioned you need to watch your power input and DC levels. I use an external HP switchable attenuator, HP 33321-60045, costs less than $25 off eBay. I made a simple Arduino control box for it.  In addition to the SA you will need all kinds of adapters, cables, reflection bridges, power splitter/combiners, directional couplers, etc. to be able to do productive work.  I would allocated about $150 for the kit of parts, which is cheap, because a good reflection bridge would cost way more than that amount.

Back to bugs on the Rigol, yes, all of the lab-grade analyzers have bugs.  The difference in them is that HP, Agilent, Tek, etc fixed them all over time.  I haven't looked lately, but some of the bugs on the Rigol have been around quite a while and the last I read was that the 10Mhz TG bug wasn't going to be corrected though they might have fixed it in later models.

Lastly, if my HP monster breaks, I have the capability to fix it.  The other night I powered on my HP 4815a Impedance meter, for example, and the amplitude meter was way off.  It took me 5 minutes to pull a card from a parts unit I have to get it working again.  I checked a few components and found the problem right away.

Jerry

[chris_leyson]

Just for the record I use a 7L12, cost me $200 and the RF section is still reasonably well calibrated, she was well looked after by a retired CATV engineer. The RF and IF sections are good, shame about the rest of the crap that takes up useful "RF space". Grrr.

[TheSteve]

thanks for all this information !
Part of the issue i have is my lake of knowledge. My background is digital design, and Radio/RF is a new hobby for me.
Beyond wanting to work on FM/AM/Filters maybe some Ham the main reason of buying SA is educational for me. And i will probably know better what will be most useful for me after owning one.

I am fairly sure from your desicription that it'll spend 99.9% of its life gathering dust after an intial hour or so of switching on.

So you don't think for building/tuning AM/FM radio SA will be useful  ?  Or for educational purposes of RF in general ?
not sure i follow ...

No, for building and AM/FM radio, a signal generator and SINAD meter would be far more useful. What I have been suggesting is a test set with signal generator, SINAD meter, transmitter tester AND a spectrum analyser, some with TG too.

If that is really what he is looking for then I'd recommend an HP 8920B or 8921A. They would need to have the proper options though. And again he would be dealing with older hardware.

[woodchips]

This is an interesting thread, along with the one talking about Advantest SAs.

Sold all my SA, other than a couple of 7L14s that arrived in a pile of Tek 7000 stuff, years ago, never thought I would need to look at RF again but everything seems to be radio now. Did wonder if getting something to look at transmitted signals from mobile phones, WiFi and all the rest would be useful. Thing is, how do you find out if the weather transmitter is transmitting without one?

This will be a 99.9% unused analyser, so not going to spend lots on it, but what? Did fancy an 8566 and bought the manuals from Manuals Plus before they closed, cost far more for shipping than the four large folders did! Reading the threads and what gets me is everyone, at some point, talks about software, calibration or even just operating the analyser. Yes, it is useful for doing analysis but I can't really see if it is really that important, if you want to process the spectrum then upload to a computer? So going through the various models, really only from Tek and HP, can't stand Anritsu and never used any others, and I end up at a HP 8555A. I already have a 7L14 but looks like more than 1.8GHz is wanted, besides it has just gone pop and looks a nightmare to fix. I had the 141 based analysers years ago doing EMC testing, and others, and found them to be easy to drive and reliable. Also had an 8568A which was a real struggle to repair.

Question. If time isn't that important, just what benefits do the newer software based instruments have being that they can become unrepairable pretty quickly? The 141 is simple, don't blow the RF stuff up and it can always be repaired. Have to point out that my bench has four 7000 mainframes on it!

[chris_leyson]

Woodchips just reminded me, HP8558 in a 182 or 853 mainframe and an HP8444A for the tracking gen. Nice

[trojanrabbit]

I have access to the Rigol (with the TG) here at work, and for what it is, it's fine.  You have to be careful when using the TG and understand its limitations.  You only have a 0dBm to -20dBm range to work with and it's fairly intolerant of mismatch.  But even with its limitations, the TG is a must have.  The analyzer itself gets its share of use for bench sniffing when we don't want to tie up the EMC Chamber.  I mutter obscenities under my breath every time I look at the front panel of the R&S ESPI-3 I use as the daily driver here and see a hole cover where the TG output should be.  Not upgradeable of course.

/first post

[Wuerstchenhund]

I mutter obscenities under my breath every time I look at the front panel of the R&S ESPI-3 I use as the daily driver here and see a hole cover where the TG output should be.  Not upgradeable of course.

Of course it's upgradeable. The ESPI is basically a FSP, and the TG (option FSP-B9) is pretty much a simple plug-in module that can be retrofitted to existing instruments.

However, considering that the TG only covers up to 3GHz, the more attractive (and a lot cheaper) option is to retrofit the External Generator Control option (FSP-B10) which allows to use an external RF generator (doesn't have to be R&S, others like the Agilent ESG Series is also supported) as TG.

[danny_isr]

So i went a head and ordered the 815-tg, should be shipped in 2 days. Will post few pics and my first impression.

Question, what "must have" connectors/attenuators/Antenna i need ? and a good place to buy from ?


thanks Dan

[trojanrabbit]

I mutter obscenities under my breath every time I look at the front panel of the R&S ESPI-3 I use as the daily driver here and see a hole cover where the TG output should be.  Not upgradeable of course.

Of course it's upgradeable. The ESPI is basically a FSP, and the TG (option FSP-B9) is pretty much a simple plug-in module that can be retrofitted to existing instruments.

However, considering that the TG only covers up to 3GHz, the more attractive (and a lot cheaper) option is to retrofit the External Generator Control option (FSP-B10) which allows to use an external RF generator (doesn't have to be R&S, others like the Agilent ESG Series is also supported) as TG.

One would think it would be upgradeable, I will have to search my emails because 4 years ago I had to send it out for a display replacement (at $6K including cal) and I seem to remember it wasn't an option, maybe because I have an older unit that has the old style PS2 keyboard connector on the front panel instead of the USB ports.  Of course I could be confused and/or wrong.

No matter, I ended up getting an Agilent generator for RF immunity testing and I wrote a Q&D LabVIEW app to check loss in my test station cables, which is what I need it for.  Slow but it works.

[apelly]

Question, what "must have" connectors/attenuators/Antenna i need ? and a good place to buy from ?

Some guys helped me out here: https://www.eevblog.com/forum/testgear/new-spectrun-analyser-what-accessories/msg576456/#msg576456

[danny_isr]

Question, what "must have" connectors/attenuators/Antenna i need ? and a good place to buy from ?

Some guys helped me out here: https://www.eevblog.com/forum/testgear/new-spectrun-analyser-what-accessories/msg576456/#msg576456

great, exactly what i was looking for ! 

[Wuerstchenhund]

I mutter obscenities under my breath every time I look at the front panel of the R&S ESPI-3 I use as the daily driver here and see a hole cover where the TG output should be.  Not upgradeable of course.

Of course it's upgradeable. The ESPI is basically a FSP, and the TG (option FSP-B9) is pretty much a simple plug-in module that can be retrofitted to existing instruments.

However, considering that the TG only covers up to 3GHz, the more attractive (and a lot cheaper) option is to retrofit the External Generator Control option (FSP-B10) which allows to use an external RF generator (doesn't have to be R&S, others like the Agilent ESG Series is also supported) as TG.

One would think it would be upgradeable, I will have to search my emails because 4 years ago I had to send it out for a display replacement (at $6K including cal) and I seem to remember it wasn't an option, maybe because I have an older unit that has the old style PS2 keyboard connector on the front panel instead of the USB ports.  Of course I could be confused and/or wrong.

I don't think you're wrong (maybe the support person has given you incorrect information) but you can definitely retrofit the TG in these instruments. Not that it's worth doing, though, the TG module is priced pretty typical for R&S (i.e. $$$) ;-)

No matter, I ended up getting an Agilent generator for RF immunity testing and I wrote a Q&D LabVIEW app to check loss in my test station cables, which is what I need it for.  Slow but it works.

As I said, I can recommend having a look at the B10 option (Ext Gen Control), which works really well and makes the RF generator behaving pretty much like a TG (it's all controlled from the FSP/ESPI). Plus it's simple to retrofit (a PCB module plus two connectors) and the parts aren't that expensive.

[trojanrabbit]

I mutter obscenities under my breath every time I look at the front panel of the R&S ESPI-3 I use as the daily driver here and see a hole cover where the TG output should be.  Not upgradeable of course.

Of course it's upgradeable. The ESPI is basically a FSP, and the TG (option FSP-B9) is pretty much a simple plug-in module that can be retrofitted to existing instruments.

However, considering that the TG only covers up to 3GHz, the more attractive (and a lot cheaper) option is to retrofit the External Generator Control option (FSP-B10) which allows to use an external RF generator (doesn't have to be R&S, others like the Agilent ESG Series is also supported) as TG.

One would think it would be upgradeable, I will have to search my emails because 4 years ago I had to send it out for a display replacement (at $6K including cal) and I seem to remember it wasn't an option, maybe because I have an older unit that has the old style PS2 keyboard connector on the front panel instead of the USB ports.  Of course I could be confused and/or wrong.

I don't think you're wrong (maybe the support person has given you incorrect information) but you can definitely retrofit the TG in these instruments. Not that it's worth doing, though, the TG module is priced pretty typical for R&S (i.e. $$$) ;-)

No matter, I ended up getting an Agilent generator for RF immunity testing and I wrote a Q&D LabVIEW app to check loss in my test station cables, which is what I need it for.  Slow but it works.

As I said, I can recommend having a look at the B10 option (Ext Gen Control), which works really well and makes the RF generator behaving pretty much like a TG (it's all controlled from the FSP/ESPI). Plus it's simple to retrofit (a PCB module plus two connectors) and the parts aren't that expensive.

Thanks for your input.  Yeah, 4 years ago the TG option was quoted at $4K (I was looking for a new analyzer).  The ESPI will have to go in for cal soon anyway (or replacement), we're thinking about joining a Satellite Lab program so we can "self-certify", if one of the requirements is that we need to upgrade to a "full-compliance" receiver it would not surprise me.

[Wuerstchenhund]

Thanks for your input.  Yeah, 4 years ago the TG option was quoted at $4K (I was looking for a new analyzer).  The ESPI will have to go in for cal soon anyway (or replacement), we're thinking about joining a Satellite Lab program so we can "self-certify", if one of the requirements is that we need to upgrade to a "full-compliance" receiver it would not surprise me.

You're welcome. Should you ever consider retrofitting the TG to your ESPI let me know and I can find out for you what the exact P/Ns are that you need . As I said TGs modules are generally expensive but it should not cost anywhere near $4k.

[uhfradio]

This was a very helpful thread.  I have been wrestling with the same question as the OP.

In my case I have three low power (50W) UHF repeaters (all with duplexers) and a handful of HTs to maintain for private/non critical use.  I have been debating the Rigol 815 + TG vs something like the HP 8920A.  In my case I will only be working between 400MHz and 500MHz.  I want to be able to tune duplexers, and align the repeaters and HTs.  So I need to be able to check RX sensitivity, desense testing, check/align narrow/wide band as well as tone modulation, check/align freq drift etc.

If all I wanted to do was tune duplexers, I would not even think about it and just go the Rigol 815 + TG route.  But if I want to do other things then it gets more complected.  Technically the 815 + TG can be used for sensitivity measurements, but you have to use attenuators to get the signal generator way down there plus it won't do SINAD, although there are other methods.  You can also check modulation, but again attenuators, voice modulation, plus a rough min/max hold method instead of FM modulation and deviation meter.  With the 815 you need a sampler, etc.  Of course no real o-scope capabilities I believe.  Of course it has a nicer screen and a warranty

I have been looking at the HP 8920A.  As I understand it it has SA, TG, o-scope, deviation meter, SINAD, FM modulation, tone (with the correct option), and better RBW.  But it is more expensive.  I see them around for $2200-$2500 vs the $1500 Rigol 815 + TG.  Of course the 815 cost starts going up when you add in a sampler, variable attenuator and various fixed attenuators for sensitivity and repeater and HT TX testing.  My understanding is that the 8920A is capable of handling 60W continuous and 100W intermittent (very intermittent I would assume).  Since I would be under 50W in all cases (and in a very low RF environment) , I don't think I would need to worry about samplers and attenuators and could just hook up the 4W HT and even the repeaters directly to the service monitor.  Correct?

Decisions decisions....

Edit: How much difference does the firmware really make?  Although I suspect that depends on what bugs/enhancements they add and if it applies to how one uses it.  I have seen some 8920A's with rev A.17 but I found a user guide for A.18 and above...

[TheSteve]

For repeater/radio work you're better off with the 8920A(or 8921A or 8920B). The trick is to make sure you get one with the proper options(high power input, spectrum analyzer etc). Also some go 400 kHz and up and some start at 30 MHz. May as well try to find a version that goes lower so you can also align HF rigs. They have quite the fan in them and don't over heat very fast even with 100 watts in. You also get a proper signal gen and can do return loss if you buy a return loss bridge(815 would need one too).

and yes, I do own an 8921A

[uncle_bob]

For repeater/radio work you're better off with the 8920A(or 8921A or 8920B). The trick is to make sure you get one with the proper options(high power input, spectrum analyzer etc). Also some go 400 kHz and up and some start at 30 MHz. May as well try to find a version that goes lower so you can also align HF rigs. They have quite the fan in them and don't over heat very fast even with 100 watts in. You also get a proper signal gen and can do return loss if you buy a return loss bridge(815 would need one too).

and yes, I do own an 8921A

Hi

A basic spectrum analyzer isn't of much use for setting up an HT or other radio receivers. Any of the 89xx series will do a much better job. Most of the options are firmware issues. There are places around that can turn one into an amazingly broadband device ...(Yes I have one as well).

Bob

[uhfradio]

Although one thing I just noticed.  If I am reading the specs right, the HP 89XX has a min RBW of 300Hz, while the Rigol 815 has 100Hz.  Some claim that even 100Hz is not good enough, so I am wondering about that 300Hz RBW for the 89XX.

[uncle_bob]

Although one thing I just noticed.  If I am reading the specs right, the HP 89XX has a min RBW of 300Hz, while the Rigol 815 has 100Hz.  Some claim that even 100Hz is not good enough, so I am wondering about that 300Hz RBW for the 89XX.

Hi

For some things, a resolution bandwidth of over 5 Hz is simply not adequate. For a lot of work an RBW of 1 KHz would do just fine. Setting up radios is in the "1 KHz is fine" range.

Bob

[KE5FX]

The trend is toward wider bandwidths in modern communications work -- up to hundreds of MHz, in fact.  The analysis bandwidth of some of the newer signal analyzers is in the 1 GHz range.   3 MHz or even 10 MHz isn't enough to take in a whole WiFi signal, and it only gets worse from there.  But if you don't know that you need wideband and/or real-time analysis capabilities, the answer is that you don't.

Meanwhile, resolution bandwidths down to 1 Hz or even less can be very useful in low-level RF work with stable sources.  This is one thing that the E4406A really excels at.  Second place in this category goes to the 856xE series portable analyzers -- which are also the quietest in terms of their own LO noise floor.  But if you don't know that you need extremely narrow resolution bandwidth or low PN, the answer is that you don't.

What that means is that you should take the same advice that would/should be given to someone who's just learning to play guitar for the first time.  Buy something dirt cheap and use it to learn on.  Then, when (if) you're ready to move on from that first guitar, you'll know what to look for in your next one.

Possible mistakes:

- Buying an 8568 or 8566 if you don't have plenty of room on your bench and actively enjoy playing with test equipment.

- Paying too much for an 8590 series model.  These are warmed-over incarnations of much older instruments, released by HP when Tektronix caught them flat-footed with no portable instruments.

- Turning down a great deal on an 856xE-series model.  These are among the best SAs ever built in the analog era, even though they were sold as portables and not "lab grade" benchtop models.  I prefer the display and UI on the 8566/8568 but performance-wise, the 856xE series is superior.

Also keep an eye out for an LCD-based R&S FSP in good shape.  I can't stand their user interfaces myself, but you may not agree.  Their performance is in the 8560E class in most respects, and they are a more modern design.  They do take forever to boot, but their fan isn't stupidly loud like the 8566/8568, so you won't mind leaving them on.  The biggest drawback is that if anything goes wrong with an off-brand spectrum analyzer (read: anything other than HP or Tek, and yes, that includes R&S and Advantest and other well-respected names) you are hosed.

I don't know anything about Rigols, but I will say that some of the low-cost Asian offerings are starting to look pretty competent.  They will not perform up to HP standards in pretty much any respect, but they have other advantages that you shouldn't ignore off because a bunch of spectrum analyzer snobs on EEVBlog told you to.   To go back to the earlier metaphor, you will not find it useful as a beginning musician to listen to Eric Clapton and David Gilmour argue about whose guitar is better.

[uhfradio]

Yea, thanks.  That is always the thing, specs are just numbers, sometimes it is hard to know how much a particular number really matters.  I am aware that I do not necessarily need the "top RF performance".  I am aware of the danger of chasing specs and having capabilities one neither needs nor will use.  Of course that is the challenge, figuring out what specs matter and what level you need them at when just getting started.

I will look again, but I thought the 856X did not have signal/tracking generators (I am not sure that is correct though).  Tuning duplexers is a big part of what I need to be able to do so A tracking generator would be necessary.

One of the key things I want to do is be able to tune some duplexers.  One of the ones I have is a PD526 band pas/reject.  My understanding is that you can get the reject on those down around -100dB.  The 89XX specs say that the SA display range is only 80dB.  So the notch would potentially be below the display range.  I suppose you might be able to move the ref down (move the trace up) by 30dB and be able to see the bottom, but I am wondering if these systems are limited in tuning deep notches?

Edit:  This is similar to a comment earlier in this thread about the HP859X and "the display is cramped and only shows 80dB range (of which only 70dB is claimed as valid as a log display in the specs)"  I am concerned if this will impact duplexer tuning below 80db using a HP892X.

[TheSteve]

Adjusting the reference level does get around that limitation for duplexer tuning. You'll have no problems tuning them.

[uhfradio]

Thank you for the feedback.  I do think a 892XA/B with the right options is a better way to go (assuming you are looking for one piece of equipment) if you want to both tune filters and do basic radio work.

Now to justify the cost.....