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

[G0HZU]

will look again, but I thought the 856X did not have signal/tracking generators (I am not sure that is correct though). 

At work we have a couple of old HP8560A analysers that have the internal tracking gen option fitted.

However, the 'tracking' performance of the internal TG isn't very good and this caused issues/confusions to any new users. The other common issue with the HP8560/1/3 portable analysers at work is that the button membrane on the front panel is quite fragile and so most of our older analysers (eg the CRT versions) in this range suffer from this deaf/intermittent button issue and it can be very frustrating to use the analyser once it gets like this. But they are still very nice analysers in terms of the RF performance they offer. But if you don't need this performance then you would be better off with the Rigol IMO.

You can also attach an external TG to many of the HP856x analysers in the form of the HP 85644/5 unit. I've got one of these here and it works very well as a TG with my old HP8566B. But these TG units are normally very expensive so not really an option. However, I was lucky to buy a very clean one for just £175 on ebay (faulty, turned out to be a blown fuse due to wrong voltage selection) but that was a rare fluke. They normally cost several thousand dollars on ebay.

[Wuerstchenhund]

Tuning duplexers is a big part of what I need to be able to do so A tracking generator would be necessary.

Many R&S SAs like the FSP KE5FX mentioned have an option called "external generator control" which allows to use an external RF generator (i.e. R&S SME/SMIQ, but also some non-R&S models like the Agilent ESG Series) as tracking generator, which is quite neat. Retrofitting the option on the FSP if it doesn't already have it isn't overly expensive either (I paid less than GBP250 for the parts from R&S, installation takes less than half an hour).

Considering that the presence of a TG often raises the price by 50% or so, ext gen control might be a decent alternative. Plus the RF performance of many TGs isn't as good as a standalone RF generator either.

As to the FSP itself, be aware that there are two variants out there, the older ones (1064.4495.xx) have a beige/grey front bezel and run WindowsNT on a slow Pentium3 controller with PATA hard disk, while the later ones (1164.4391.xx) have a beige/blue front bezel run WindowsXP on a Pentium-M based controller with SATA hard drive. The newer variants also have some other improvements, i.e. in the RF sections. Should you opt for an FSP then I'd strongly recommend looking at the later models.

The FSP UI isn't that bad, many things just require a few more button presses than on many other analyzers, but main functions have their own front panel buttons.

[razberik]

Why are in fact TGs so rare and so expensive ?
I have never seen such a R&S TG unit alone. Google images gives no answer of unit itself. Contrary to HP units for 859x series.

[uncle_bob]

Hi

If the intent here is to work on radio *receivers*, especially hand held ones, a tracking generator as a signal source is likely to be of very limited value. They simply are not designed to provide calibrated signals at very low levels. The vast majority of them are not shielded well enough for low level work. They are not set up to provide modulated signals for receiver work. Unless they are fully synthesized to an external reference it's unlikely they are accurate enough for receiver  alignment. Simply put - they aren't up to the job.

Bob

[KJDS]

Why are in fact TGs so rare and so expensive ?
I have never seen such a R&S TG unit alone. Google images gives no answer of unit itself. Contrary to HP units for 859x series.

If a business wants test equipment to tune filters then it will by a VNA, not a spec ana.

[razberik]

So who is the main customer of devices with built-in TG ?  Who is willing to pay extra costs for TG ?
I am asking for better understanding the market.

[uncle_bob]

So who is the main customer of devices with built-in TG ?  Who is willing to pay extra costs for TG ?
I am asking for better understanding the market.

Hi

1) General repair / troubleshoot shop that needs to track down system faults. Think stuff like cable TV head end repair.  They head out in the truck and don't really know what they will find when they get there.

2) Bench level design on things that don't have very tight specs. Also where cost of equipment constraints rule out a network analyzer. Think in terms of three guys in a garage ....

3) People who bought a demo unit. The field demo gear is often loaded with options. That way they have fewer models to stock. After the discount, the "tracking generator was free".

(Yes there *is* overlap between numbers 2 and 3)

Bob

[Wuerstchenhund]

Why are in fact TGs so rare and so expensive ?

As other said they are rare because TGs are of limited value for anything that very basic tasks, and with rarity comes a price premium on the 2nd hand market if you really need/want it.

Also, for many of the tasks a TG is used there are other (cheaper?) alternatives, i.e. a wideband noise source which works with any SA.

I have never seen such a R&S TG unit alone. Google images gives no answer of unit itself. Contrary to HP units for 859x series.

With R&S they are even more rare as for HP is simply because many who require a coupled RF source rather go for the combination of Ext Generator Control option (something that HP/Agilent never offered) plus a separate RF generator.

[razberik]

Thank you for your explanation.

It just annoys me when I see beautiful Rohde & Schwarz beast with many GHz's and there is only a black plug with label RF OUT.
I much more like the way of Advantest almost 30yrs ago. If there was no TG, there was no hole & plug in front panel. Well, I am playing with my R3361A here around. It last for few days since I repaired it. And I like it !

[Wuerstchenhund]

It just annoys me when I see beautiful Rohde & Schwarz beast with many GHz's and there is only a black plug with label RF OUT.

It's not just you. You'd think a company like R&S could come up with something better than a black piece of plastic behind the (usually beige) front panel cut-out, but no.

[uhfradio]

Hi

If the intent here is to work on radio *receivers*, especially hand held ones, a tracking generator as a signal source is likely to be of very limited value. They simply are not designed to provide calibrated signals at very low levels. The vast majority of them are not shielded well enough for low level work. They are not set up to provide modulated signals for receiver work. Unless they are fully synthesized to an external reference it's unlikely they are accurate enough for receiver  alignment. Simply put - they aren't up to the job.

Bob

I am trying to better understand these comments.  I thought that the various radio test sets (and their tracking generators) like a HP 892x were specifically designed, at least in part, for receiver alignment.  They do provide a modulated signal.  And they are capable of SINAD, so I would assume they were somewhat accurate at low levels.  I also thought that they were routinely used by radio technicians for repeater site maintenance, duplexer tuning, and hand held alignment.  But the comments above indicate that this is not the case?

[uncle_bob]

Hi

If the intent here is to work on radio *receivers*, especially hand held ones, a tracking generator as a signal source is likely to be of very limited value. They simply are not designed to provide calibrated signals at very low levels. The vast majority of them are not shielded well enough for low level work. They are not set up to provide modulated signals for receiver work. Unless they are fully synthesized to an external reference it's unlikely they are accurate enough for receiver  alignment. Simply put - they aren't up to the job.

Bob

I am trying to better understand these comments.  I thought that the various radio test sets (and their tracking generators) like a HP 892x were specifically designed, at least in part, for receiver alignment.  They do provide a modulated signal.  And they are capable of SINAD, so I would assume they were somewhat accurate at low levels.  I also thought that they were routinely used by radio technicians for repeater site maintenance, duplexer tuning, and hand held alignment.  But the comments above indicate that this is not the case?

Hi

The tracking generator on a *spectrum analyzer* doe not do any of these things. Thus a spectrum analyzer with a tracking generator is not very useful at all for receiver work. On the other hand, a proper signal generator *does* give you the stuff you need to do receiver tests. A communications test set (there are lots of them out there, going back at least 50 years) combines a signal generator with a few other things to let you do multiple tests with a single box. The original idea was to have one box in the truck when you went out on a service call or one box on the bench to work on things. A few decades back, cell phones came along and the internals of a communications test set changed a bit. It's still the same idea, just slightly different bits and pieces.

Bob

[uhfradio]

I see, thank you for clarifying.  That does agree with my understanding and previous research.  So yes, a Rigol 815 +TG (for example) is not very useful, for receiver work.  But a communication test set with SA and TG is useful as the RF generator is in fact designed to provide TG functions as well as low level output, modulation, etc.

[uncle_bob]

I see, thank you for clarifying.  That does agree with my understanding and previous research.  So yes, a Rigol 815 +TG (for example) is not very useful, for receiver work.  But a communication test set with SA and TG is useful as the RF generator is in fact designed to provide TG functions as well as low level output, modulation, etc.

Hi

Yes, that's correct as long as you get the right test set. (= it does indeed have both SA and TG functions).

Bob

[nctnico]

I see, thank you for clarifying.  That does agree with my understanding and previous research.  So yes, a Rigol 815 +TG (for example) is not very useful, for receiver work.  But a communication test set with SA and TG is useful as the RF generator is in fact designed to provide TG functions as well as low level output, modulation, etc.

Every communication test set I have seen only offers a very limited span on the spectrum analyser part. This is no problem for testing communication devices but it will be of no use for EMC testing.

[Wuerstchenhund]

Every communication test set I have seen only offers a very limited span on the spectrum analyser part. This is no problem for testing communication devices but it will be of no use for EMC testing.

Most of them have, yes, however there are some exceptions like the R&S CMU200 and CRTU-RU, or the Agilent E7495A/B which pretty much offer zero to full span (I *believe* the same is true for the Anritsu MT88xx cell phone testers if the SA option is installed, but I don't know for sure).

The CMU200/CRTU has the better RF performance but the E7495 has some nice features (i.e. a wideband noise source for simple network analysis which is a good replacement for a TG) and is portable.

[uncle_bob]

I see, thank you for clarifying.  That does agree with my understanding and previous research.  So yes, a Rigol 815 +TG (for example) is not very useful, for receiver work.  But a communication test set with SA and TG is useful as the RF generator is in fact designed to provide TG functions as well as low level output, modulation, etc.

Every communication test set I have seen only offers a very limited span on the spectrum analyser part. This is no problem for testing communication devices but it will be of no use for EMC testing.

Hi

There are a *lot* of testers out there. In many cases they have multiple soft configurable options. There are some that do quite well over a 400 KHz to 1 GHz sort of range. You *do* have to be very careful to not just get the right model, but to get one that is both "unlocked" and calibrated over the full range.

Bob

[radar_macgyver]

What about the USB-connected instruments from Signal Hound? Some models meet the OP's requirements (frequency range and price), although from all I've gathered thus far, their RF performance for the models other than the BB60C suffer from leakage and harmonics that are removed in software. Some reviews complained about the software, but that's been updated significantly in the recent past with the release of their Spike software.

They also sell signal sources with a sync input that can be used as a tracking generator.

Has anyone used these? Care to comment about their performance? I'm specifically interested in their phase noise spec. Shahriar did a nice review, but didn't look at the phase noise in detail (or maybe I just missed that part?)

The Tek RSA306B seems nice, but priced outside the OPs budget.

[uncle_bob]

What about the USB-connected instruments from Signal Hound? Some models meet the OP's requirements (frequency range and price), although from all I've gathered thus far, their RF performance for the models other than the BB60C suffer from leakage and harmonics that are removed in software. Some reviews complained about the software, but that's been updated significantly in the recent past with the release of their Spike software.

They also sell signal sources with a sync input that can be used as a tracking generator.

Has anyone used these? Care to comment about their performance? I'm specifically interested in their phase noise spec. Shahriar did a nice review, but didn't look at the phase noise in detail (or maybe I just missed that part?)

The Tek RSA306B seems nice, but priced outside the OPs budget.

Hi

I would not count on really good phase noise on a device like that. I simply is no something they focus on. If it is not a critical spec, they didn't spend the extra money to do a good job of it. That said, if -130 dbc / Hz at 100 KHz is "good enough" it may be able to hit that.

Bob

[hendorog]

Has anyone used these? Care to comment about their performance? I'm specifically interested in their phase noise spec. Shahriar did a nice review, but didn't look at the phase noise in detail (or maybe I just missed that part?)

The Tek RSA306B seems nice, but priced outside the OPs budget.

Hi

I would not count on really good phase noise on a device like that. I simply is no something they focus on. If it is not a critical spec, they didn't spend the extra money to do a good job of it. That said, if -130 dbc / Hz at 100 KHz is "good enough" it may be able to hit that.

Bob

SH published an old whitepaper showing phase noise measurements for the SA44B:
https://signalhound.com/sigdownloads/SA44B/AppNote-4.pdf

So it looks cleaner than the Rigol 815 spec. (-80 dBc @10kHz)

[radar_macgyver]

I would not count on really good phase noise on a device like that. I simply is no something they focus on. If it is not a critical spec, they didn't spend the extra money to do a good job of it. That said, if -130 dbc / Hz at 100 KHz is "good enough" it may be able to hit that.

I should clarify, the reason I asked about phase noise is that the larger instruments achieve decent phase noise since they have YTOs to generate their swept LO. The Signal Hound uses monolithic oscillators so I figured it would be similar to what the Rigol can achieve (which, from all the teardown videos I've seen, does not contain a YTO, but seem to use discrete LOs instead - see Dave's teardown video at ~30:35).

Per the spec sheet of the BB60C, at 1 GHz it's -93 dBc/Hz at 100 kHz offset; that's worse than the DSA815's -100 dBc/Hz at 100 kHz. I guess it's wise to get an RTSA for signal analysis and a swept SA for things like phase noise, just like G0HZU suggested.

SH published an old whitepaper showing phase noise measurements for the SA44B:
https://signalhound.com/sigdownloads/SA44B/AppNote-4.pdf

So it looks cleaner than the Rigol 815 spec. (-80 dBc @10kHz)

Thanks, that clears things up - those plots look pretty good, especially with an external OCXO. I'll stop threadjacking now.