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

[danny_isr]

I'm doing some hobby experiments with AM/FM radio and transmitters. I'm pretty sure the Rigol DSA815-TG is more then enough for that.

Though am thinking in the future to work on 2.4Ghz transivers (for home automation projects). The 815 is limited to 1.5Ghz.
I saw on eBay: HP 8594E 2.9 GHz, and they go for less then $1k (sometimes), but it's older and i'm not sure about their performance and reliability.

I'm new to RF and don't know enough about it to understand if i need the capability to measure beyond 1.5Ghz, when working with 2.4Ghz Transmitters/receivers.

SA is quite an investment after all...

[kd5iku]

The rigol may work for you, but it depends on what you want to see. What kind of experiments are you wanting to do? The 100hz minimum RBW of the 815 is probably too coarse for certain applications. My previous SA had a minimum RBW of 10hz and worked fine for just about anything I threw at it. (I am an amateur radio operator so it is CW and SSB for me.)

Are you planning on designing from scratch the 2.4ghz trx?

[uncle_bob]

Hi

There are a couple of "divides" in the HP empire:

There is the older gear with simple tubes that seem to hold intensity for a long time ( = do not wear out). After that came ever more interesting CRT's that added color. They can get pretty dim pretty fast. The newer E series stuff seems to have that issue taken care of.

Next there is the rather wide range of instruments. Some stuff was made for cost sensitive markets. Other stuff was pretty high spec. Be very careful you don't bump from one to the other.

Finally there is the minor issue of calibration. Some of these beasts (in fact a lot of them) are software calibrated. There's some monster machine that they get attached to and magic numbers fly into the eeprom. It works amazingly well and it is darn cheap to do. The gotcha is that once the last magic machine is gone .. no more cal.

To me the "no more cal" issue is the big one. If you blow the mixer diode on the analyzer (and replace it ... easy to do) you have a perfectly good analyzer again. It no longer is calibrated and it can't be calibrated ... grrrrrr....

Bob

[TheSteve]

You can always look for a deal on an Agilent N1996A. They are a rather unique machine with lots of good stuff that can be "enabled".

[nctnico]

There are also several Advantest spectrum analysers out there which aren't expensive due to a relatively unknown name on the sticker.

[Electro Fan]

There are lots of more experienced SA users in this forum but here are some thoughts from a rookie SA user.

I had similar questions and spent a lot of time studying the choices - especially the Rigol and the HP 856X and 859X models.  I pretty quickly determined that the HP Es made more sense than the As.  Next I determined that while the 9X series looks attractive especially given their prices the 6X was/is a superior spectrum analyzer.  Not that the 9X is bad - I'm sure it's good, but I think when you study the specs and get opinions from knowledgeable spectrum analyzer users and owners you will get consistent feedback regarding the stronger performance and reliability/serviceability of the E series.

One of the things that I wrestled with was how much RBW was desired/needed - this was one of the things that drove me to the 6XE over the 9XE.  In hindsight I didn't need the extra RBW; not that the 1Hz RBW isn't cool but when you run these models at 1Hz RBW you can take the sweep speed from 50ms to over 10 minutes.  I don't often use RBW below 10kHz - but of course it depends on what you are doing.

In any event, even though I "over-spec'd" on RBW I am glad I went for the HP (I wound up with a 8561E, which is the 6.5GHz model).  I probably could have lived with the 2.9GHz (8560E) model but I came across a good deal on the 61.  I have used it for various things ranging from a few hundred MHz with key fobs to 2.4GHz wifi to more recently some ham radio testing. All of these would have been fine with a 2.9GHz SA but I would definitely miss having the ability to work with 2.4GHz wifi (had I gone with the 8560E), and if I decide to go up to 5GHz wifi the 61 will still be good.

Having said all that, I have no doubt that if you can live with the lesser bandwidth of the entry Rigol that having a brand new highly software driven SA will be an enjoyable experience + you can probably add the tracking generator; a tracking generator on the HPs will probably make your choices more scarce and/or more expensive.   On the Rigol the TG doesn't add much to the cost - but it might come down to whether you have use for a tracking generator.  I wanted a TG but so far I haven't missed it; on the other hand I might not know what I'm missing. 

One of the things that took me away from the Rigol though was the superior low noise performance specs on the HPs.  To be sure, I didn't know how far down in the noise I might want to go and I still don't but I like knowing that short of a very big investment the 8561E will likely be as good as it gets.  (HP made some other similarly vintage models with even better specs that are considered the mothers of great SAs but they are very big and heavy.)

Before I found my SA I thought I wanted a 8560EC with the color display.  I think color would be great and the lower power consumption of the LCD would be nice; plus the EC units are newer (or less old).  If you can find an EC it's worth considering; in my case I couldn't find an EC at a price that made sense.

So, those were some reasons in favor of the HP86XE.  On the flip side these models are not real young - they are perhaps middle age if not older age and they might need some tuning or repair.  After coming very close to purchasing a SA through eBay I found my SA at a local cal lab.  It turns out that was a lucky and good thing; the unit had some intermittent problems in the power supply.  It took me several trips to the dealer with the pretty bulky/heavy SA.  The dealer was impeccable in standing behind the unit and while it took several tries the unit was fixed and it's been A-OK ever since (knock on wood).  Had I needed to ship the unit back to a remote seller the shipping costs along with the bumping and banging of shipping probably would have turned the good deal into a not so good deal.  Moral of the story, don't buy a used SA from anyone you aren't sure is highly trustworthy.  But if you can find the model that is right for you from a solid seller the 856XE is a great piece of test equipment.  If you don't find such a deal and can live with the specs of the Rigol that would probably be a good route to go too.

YMMV, EF 

[Mechatrommer]

Though am thinking in the future to work on 2.4Ghz transivers (for home automation projects). The 815 is limited to 1.5Ghz.
I saw on eBay: HP 8594E 2.9 GHz, and they go for less then $1k (sometimes), but it's older and i'm not sure about their performance and reliability.

why buy less speced digital buzzword one, if analog is more capable? i'm not sure whats potential post-processing DSA815 has to offer, but i'll buy cheaper and more speced one anytime even if its analog. but one thing to watch out for used item is that you have to make its not broken OOB.

[Wuerstchenhund]

Have a look at Rohde & Schwarz (R&S), they make some of the best SAs on the market and can often be surprisingly cheap. You didn't say what your budget is but the R&S FSP Series might be a good alternative as although being 'just' a mid-range SA it's performance is pretty close to top end HP analyzers like the 8566, plus they are much smaller, less noisy, less power hungry and much younger (i.e. more reliable).

Also have a look at the R&S FSEx and FSIQ Series which are older high end analyzers and pretty much surpass anything that HP has ever made.

[Pinkus]

I would like to bring one subject which was not mentioned yet: used SAs on Ebay often are not OK. It is pretty easy to destroy parts of the input circuitry or other parts (e.g. mixer) and thus many are defective. This is nothing the self test can bring up and the unit seems to be fully functional, but it is not (and even the seller might not be aware of this).

I purchased three SA through Ebay in the past 10 years, and all of them were defective in some way. Even if a service manual including schematics is available (usually older units only, nowadays you will not be able to get this), sometimes it is only helpful to find the defective device .... only to find out it is a unique or hybrid part available only from the manufacturer for $1000.
Oh - and avoid Japanese stuff. The devices are great - but do not expect any help from the manufacturer. They do not answer to your mails, they won't even send you the regular manual as a PDF and the user group is pretty small. Try to stay with Agilent or Tektronix; at least you will receive help from the user base.

I learned my lesson: A SA is a test gear, I now would only buy either brand new or from somebody nearby where I can drive by and test it (or at least somebody -usually a more expensive reseller- who allows sending it back for a full refund). The risk purchasing from private auctions is way too high.

[Theboel]

I want to share my experience about buying my first SA. an HP 8590L
1. When its arrive its working but after several month I have a problem with a blow tantalum capacitor fortunately many many many HP user are more experience so after some advice from HP group I ended in replace a handful tantalum cap from all over section of my SA.
2. The second problem are the power supply for my SA are not build by HP so when I have problem with power supply I must trace by myself and spend a lot of time.
3. The flexibility to save and read the result of measurement is good as long as You want to work with GPIB to USB adapter (remember to check first are GPIB port installed) and searching around for program, installed it. In my experience its not like installing windows OS in to new PC. 
4. Usually because its ages there is a burn mark in the CRT.
5. its lack of TG (tracking generator) made me. again spend alot of time for sweeper generator.
thats my experience 
 

[Wuerstchenhund]

I would like to bring one subject which was not mentioned yet: used SAs on Ebay often are not OK. It is pretty easy to destroy parts of the input circuitry or other parts (e.g. mixer) and thus many are defective. This is nothing the self test can bring up and the unit seems to be fully functional, but it is not (and even the seller might not be aware of this).

I purchased three SA through Ebay in the past 10 years, and all of them were defective in some way.

I probably bought around 50 to 80 SAs in the last five years or so via ebay (including my own R&S FSP), and pretty much all but one were working absolutely fine. I'm not saying there's not a ton of overpriced broken crap on offer but with a bit of caution there are lots of genuine and fully working SAs available.

Also, not all analyzers are DC intolerant, or have no overload protection on their inputs. Especially newer kit is a lot less sensible (although not necessarily immune) to misuse.

Oh - and avoid Japanese stuff. The devices are great - but do not expect any help from the manufacturer. They do not answer to your mails, they won't even send you the regular manual as a PDF and the user group is pretty small.

I can second that, Anritsu & Co are pretty poor when it comes to support of older kit, at least in Europe and the US. They offen don't even talk to people who bought some of their older devices on ebay.

Try to stay with Agilent or Tektronix; at least you will receive help from the user base.

True, but Tek's SAs have always been pretty poor even back in the good old days when they were still one of the leading T&M manufacturers, plus their support has declined a lot since they were taken over by Danaher.

BTW, R&S is often pretty friendly towards hobbyists, and there's a large user base for R&S gear out there. Just saying, that there's more than HPAK & Tek out there...

I learned my lesson: A SA is a test gear, I now would only buy either brand new or from somebody nearby where I can drive by and test it (or at least somebody -usually a more expensive reseller- who allows sending it back for a full refund). The risk purchasing from private auctions is way too high.

No, it isn't, really. Even if you end up buying a broken device on ebay the buyer is pretty much always right. Just open a "not as described" claim and if the seller doesn't bend over backwards to satisfy you then ebay will just take the money off them and refund you.

I bought test gear in excess of $100k over ebay, and had very little problems, and most of the complaints we had revolved around smaller things like missing accessories.

[Howardlong]

I'm doing some hobby experiments with AM/FM radio and transmitters. I'm pretty sure the Rigol DSA815-TG is more then enough for that.

Though am thinking in the future to work on 2.4Ghz transivers (for home automation projects). The 815 is limited to 1.5Ghz.
I saw on eBay: HP 8594E 2.9 GHz, and they go for less then $1k (sometimes), but it's older and i'm not sure about their performance and reliability.

I'm new to RF and don't know enough about it to understand if i need the capability to measure beyond 1.5Ghz, when working with 2.4Ghz Transmitters/receivers.

SA is quite an investment after all...

I know this may sound negative, but what do you expect to be using an SA for, particularly at 2.4GHz? Is this for product development, or a bit of tinkering? In that part of the spectrum, it is common to use pre-certified modules with integrated antennas. If it is for EMC precompliance, then as well as the cost of your own test gear, there is also the cost of the compliance house itself which can make the cost of an SA seem quite reasonable.

If it's for home tinkering, then I'd question the real need for an expensive SA other than as a trophy item for the bench: you may well find it sits there gathering dust for months on end while you figure out what you want to use it for.

If you're doing AM/FM narrowband stuff for transmit and receive, I'd seriously consider that instead of purchasing an SA, get a communications/radio test set. Some of them have SAs and even TGs as well, although those SAs are not specifically designed for EMC usage profiles, and often don't go up beyond 1GHz.

Comms test sets have some very useful facets though for AM and FM, in particular signal generators and transmitter testers for measuring receiver sensitivity and transmitter power output and modulation.

Some examples:

Marconi 2955A/B: No SA, to 1GHz CRT
HP 8920A SA/TG option 102, to 1GHz CRT
HP 8920B SA/TG option 102, to 1GHz CRT
HP 8924C SA/TG option 102, to 1GHz CRT
HP 8935 E6380A/E6381A SA/TG  + return loss (via external bridge) all standard, to 1GHz, 1.7 to 2GHz, electroluminscent
Agilent E8285A SA/TG option 102, to 1GHz, 1.7 to 2GHz, electroluminscent
Anritsu MT8802A SA option 7, to 3GHz, FM only, no AM, colour LCD

This is quite a useful site: http://www.amtronix.com/diff.htm

[nctnico]

BTW, R&S is often pretty friendly towards hobbyists, and there's a large user base for R&S gear out there. Just saying, that there's more than HPAK & Tek out there...

That also brings Advantest into the picture because they where taken over by R&S. Even the newer spectrum analysers from R&S still have an Advantest feel to them because of the way the UI works.

[Wuerstchenhund]

If you're doing AM/FM narrowband stuff for transmit and receive, I'd seriously consider that instead of purchasing an SA, get a communications/radio test set. Some of them have SAs and even TGs as well, although those SAs are not specifically designed for EMC usage profiles, and often don't go up beyond 1GHz.

Comms test sets have some very useful facets though for AM and FM, in particular signal generators and transmitter testers for measuring receiver sensitivity and transmitter power output and modulation.

Some examples:

Marconi 2955A/B: No SA, to 1GHz CRT
HP 8920A SA/TG option 102, to 1GHz CRT
HP 8920B SA/TG option 102, to 1GHz CRT
HP 8924C SA/TG option 102, to 1GHz CRT
HP 8935 E6380A/E6381A SA/TG  + return loss (via external bridge) all standard, to 1GHz, 1.7 to 2GHz, electroluminscent
Agilent E8285A SA/TG option 102, to 1GHz, 1.7 to 2GHz, electroluminscent
Anritsu MT8802A SA option 7, to 3GHz, FM only, no AM, colour LCD

That's a good tip! I would also add the R&S CMU200 and CRTU-RU to the list, which both come with a very good spectrum analyzer as standard that fully covers 10MHz to 2.7GHz. They also come with one or two RF generators, and the CMU200 can also have an optional 20 tone audio generator/analyzer. The good thing with CMU200 and CRTU-RU is that the UI is pretty simple and with normal controls you'd find on a standalone instrument while with many other comms testers you have to wade through lots of cell phone related settings to make it work. In addition, CMU200 and CRTU-RU have a single large band (many other testers have split up their bandwidth, often with gaps in between).

That also brings Advantest into the picture because they where taken over by R&S. Even the newer spectrum analysers from R&S still have an Advantest feel to them because of the way the UI works.

That's not correct. R&S had previously partnered with Advantest to sell some of their gear in Europe (and Advantest sold R&S gear in Asia), but that partnership was limited to specific products only (R&S' own kit served the high end, Advantest the mid-range and entry level), and pretty much ended 2004 when R&S started to serve the mid-range and entry-level market with their own products

Advantest is not part of R&S, never has been, it's a separate company that aside from that partnership has no relation to R&S. There's also no technological (development) partnership between R&S and Advantest, each company was and still is developing their products independently of each other.

As to the UI, I can't say there's much similarity between the UI of a R&S standalone SA and an Advantest SA. The menu structure and screen layout is completely different. Some of the R&S handhelds (FSH) may share more UI similarities but if so then they are purely coincidental.

And unless it's one of the Advantest SAs supported by R&S then all the listed negatives re. poor support from Asian brands applies to Advantest gear as well.

[nctnico]

I'm currently owning my second Advantest SA. When I was at an EMC lab I had no problem operating their high-end R&S spectrum analyser. It looked different at first glance but it's basic operating procedures where exactly the same as the Advantest SAs I have used. The similarities where striking!

[Wuerstchenhund]

I'm currently owning my second Advantest SA. When I was at an EMC lab I had no problem operating their high-end R&S spectrum analyser. It looked different at first glance but it's basic operating procedures where exactly the same as the Advantest SAs I have used. The similarities where striking!

Well, pretty much all the interfaces of older SAs are somewhat similar, even much more so than with older scopes. I'm sure it would have been similar easy for you to operate a Tek, HP or any other SA.

You don't happen to remembed which models that were?

[tautech]

To add to that above, Siglent have a SA that exceeds the 815 specs
http://siglentamerica.com/pdxx.aspx?id=1546&T=2&tid=227

Thread
https://www.eevblog.com/forum/testgear/siglent-ssa3000x-spectrum-analyzers/

[Wuerstchenhund]

To add to that above, Siglent have a SA that exceeds the 815 specs
http://siglentamerica.com/pdxx.aspx?id=1546&T=2&tid=227

Considering that there's not much user feedback for these units yet, and that Siglent is still struggling with getting its act together in regards to writing proper firmware for its 2000 Series scopes, I'd say buying one of them now is a pretty huge gamble.

The Rigol DSA815 may be older but it's at least a somewhat mature and proven instrument that pretty much right now works as advertised. The new Siglent SA may or may not work as advertised now, but it's a lot of money to find out if this is another half-finished banana product.

I'd wait at least until others have taken the plunge and reported back.

[jwm_]

There is also the agilent E7495B which is very cheap on ebay compared to similar gear that is marketed as a spectrum analyzer even though it has one built in.

[mojoe]

To add to that above, Siglent have a SA that exceeds the 815 specs
http://siglentamerica.com/pdxx.aspx?id=1546&T=2&tid=227

Considering that there's not much user feedback for these units yet, and that Siglent is still struggling with getting its act together in regards to writing proper firmware for its 2000 Series scopes, I'd say buying one of them now is a pretty huge gamble.

The Rigol DSA815 may be older but it's at least a somewhat mature and proven instrument that pretty much right now works as advertised. The new Siglent SA may or may not work as advertised now, but it's a lot of money to find out if this is another half-finished banana product.

I'd wait at least until others have taken the plunge and reported back.

Yes, I have also been waiting for someone to do a review.

[G0HZU]

I'm doing some hobby experiments with AM/FM radio and transmitters. I'm pretty sure the Rigol DSA815-TG is more then enough for that.

Though am thinking in the future to work on 2.4Ghz transivers (for home automation projects). The 815 is limited to 1.5Ghz.
I saw on eBay: HP 8594E 2.9 GHz, and they go for less then $1k (sometimes), but it's older and i'm not sure about their performance and reliability.

I'm new to RF and don't know enough about it to understand if i need the capability to measure beyond 1.5Ghz, when working with 2.4Ghz Transmitters/receivers.

SA is quite an investment after all...

If you only ever see yourself doing basic spectrum analysis where you are looking for obvious issues then you could choose almost any of the analysers listed so far. But I think many people will have a better user experience with something like the Rigol (or maybe the new Siglent if it turns out to be better than the Rigol). This is because they have a fresh modern display and modern features/connectivity and a growing community of users. The RF performance of these analysers isn't going to be that great but most users seem content with them.

My requirements are more strict as I have worked in RF labs all my working life. So I have a much easier time choosing because I can't/won't tolerate anything with poor RF performance.

So with something like a $2000 budget I would probably try and buy a big old HP8568B (or a scruffy example of a 22GHz 8566A/B) and then hopefully have enough change to buy an Agilent E4406A signal analyser to sit alongside it. I'm not saying you should do the same but this is probably the best bang for buck combo if you can live with the size/weight/power/noise of these big machines.

The next tier down is the HP8560, HP8561 and HP8563 in A or E format. These analysers have all made it into our design labs at work and I'd be happy using any of them. But these still fetch >=$2000 typically so you probably won't have change left to also buy the E4406A.

Everything below represents stuff that I would not want to use and I can say the same for my colleagues at work.

HP859x (including the HP8594E) These are a poor relation of the 8560/1/3 A and E models and they have very limited log detector range, an awful display, fairly poor phase noise and limited BW options. These are/were designed for field trials rather than for serious lab use. They aren't that much cheaper (used) than the 8560/1/2/3 range so for this reason these HP859x analysers really should only be considered if they are very cheap and if you are desperate. However, they are apparently very popular with ham/hobby users.

R&S FSP3/7. After Wuerstchenhund tried to convince me (on an earlier thread) that these little analysers were better than the HP8566/8 I went over to the test/ATE area where we had several FSP7 models that were sat unloved on the shelves. It's not difficult to see why these analysers are not popular. They take ages to boot up, the display is gloomy and the far out phase noise performance is really poor. The noise floor performance is poor at low frequencies and the linearity suffers badly down here too. Also the default span/RBW settings seem to be bizarre and the user interface is annoying to use. I can't recommend this analyser even though it is going to be technically superior to the Rigol and the HP859x series in many ways. I'm not sure who chose to buy them at work but I think they were mugged.

If you dive even lower then you get down to the world of the RF test set. These are OK if you just want to have a casual look around but they aren't serious instruments. So that includes the CRTU, the CMW200 and also the various HP test sets listed above. To me these analysers are the equivalent of an old eastern bloc made motor car and not worth the air volume they consume on a test bench.

I do find it puzzling why people on Eevblog are often hyper critical (snobby?) about scopes and DMMs yet they are happy to spend big money on mediocre (often new or nearly new) spectrum analysers. I can understand the popularity of the Rigol 815 with its modern display and modern features but I can't understand why people buy the HP859x series or the test sets.

If you did buy the Rigol you could always make a simple external mixer/downconverter to let you look at 2.4GHz signals.

[G0HZU]

Maybe I should also say that I often use the cheap combo of the HP8566B and the E4406A here at home. I think I paid about £520 for the E4406A and, about 5 years ago, I paid £1000 for the HP8566B as an ex rental machine. At the time this was a very low price indeed.

These are both big and noisy machines and the 8566B weighs nearly 50kg so this combo is best suited to someone with lots of space and a sturdy workbench and a willingness to put up with lots of fan noise.

I think you will probably be very happy with the Rigol 815. But I can't recommend the HP8594E. Maybe you could try and find a cheap HP8560A/E that goes to 2.9GHz. The 8560A/E is a decent analyser in terms of its RF performance although the CRT version of this analyser will look very dated next to the Rigol and it won't have the modern connectivity of the Rigol either.

[uncle_bob]

Hi

Ok, here's one feature we have not yet beat on very much: The tracking generator option. For a variety of reasons these get confused with network analyzers. Just for the record, a network analyzer measures the signal applied to the device as well as the signal going through the device. A tracking generator applies a signal and measures only the output. Yes it's a subtle point, but it does impact accuracy.

For most of us, peaking up a filter with a tracking generator works pretty well. It's a handy thing to have for high level stuff. The more fancy things you try to do with one, the more you start wishing for a real network analyzer. It is a pyramid. Most of what you do with a spectrum analyzer does not use a tracking generator. The tracking generator is fine for a lot of sweep stuff. The network analyzer gets pulled out less often than the tracking generator.

Yes, I'm quite sure there is at least one person on the thread who uses a network analyzer 24 hours a day / 7 days a week and who never uses a spectrum analyzer. What I'm talking about is normal home use in a home lab. In the same lab, the scope gets used a lot more than the spectrum analyzer. The DVM gets used more than the scope. The noise figure meter may not get used a lot at all. The 3048A could sit back there for a long time ...

Bob

[G0HZU]

Some spectrum analysers can operate in a kind of VNA mode although this type of VNA combo analyser is quite rare. My big old Advantest 4172 spectrum analyser has the factory  tracking gen and impedance measuring options fitted so it can mimic a VNA when used with an external bridge. It can measure the filter passband response and also the filter impedance.

HP brought out the HP4396A in the early 1990s that was closer to a real spectrum/VNA combo and these old analysers are still popular today in A and B format. They still fetch decent money today even though they only cover up to 1800MHz.

[Bud]

I have a HP 8594E , it is a good piece of equipment and I can recommend it. There is nothing wrong with the display.

Putting rigol and HP in one line is even not a bad joke, it is utter nonsense.