Author Topic: Spectrum Analyzer - Rigol DSA815  (Read 615621 times)

0 Members and 1 Guest are viewing this topic.

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Spectrum Analyzer - Rigol DSA815
« Reply #125 on: December 17, 2012, 06:26:02 pm »
Its very likely videobruce will be making the more extensive review of the 815s RF characteristics!  tequipment has the best price I've seen in the 'net, not to mention free shipping.  Either way, should anything snafu it all goes back to Rigol in Ohio.

Please give us a detailed review when able, since you have an existing SA to compare it against.
Best Wishes,

 Saturation
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #126 on: December 17, 2012, 09:38:21 pm »
Just to highlight the phase noise issue have a look at the two plots below.

One is from the Rigol from somebody who tested it on a narrow span of 2kHz. You can see that the analyser has a lot of difficulty resolving any detail on a narrow span like this because it is limited by the 100Hz RBW filter and also there is a lot of stopband noise caused by the noisy synthesiser used in the Rigol. This noise manifests itself as a raised noise floor and it limits the visible signal to noise ratio on narrow spans like this.

The Rigol has 30dB attenuation and no preamp selected but this would still allow a typical DANL of about -85dBm with a 100Hz RBW filter so the noise you see is phase noise from the synthesiser.


The plot below it is taken from a very old 1500MHz HP analyser (this model dates back nearly 35 years)  and you can clearly see how much cleaner the response is.
Now this was a top class HP analyser in its day but the margin by which it outclasses the Rigol here in terms of signal to noise and resolution on a narrow span is quite marked. It's actually better than this by several dB but the 134MHz sig gen I used introduces some of the noise you can see.



« Last Edit: December 17, 2012, 09:59:03 pm by G0HZU »
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #127 on: December 17, 2012, 10:47:57 pm »
As a further comparison here's another old analyser from about 25yrs ago  (1800MHz model made in Japan)

This time I've put in a basic 10MHz crystal oscillator as a test signal and I've set the attenuator to 30dB and the preamp is off to match the Rigol settings.

The span is 1kHz.

Again you can see far more detail and far more dynamic range because the RBW is lower and the synthesiser noise is much lower. You can see some tiny 100Hz sidebands on the oscillator signal caused by the 50Hz PSU. The Rigol can't display this info despite it having modern DSP at its disposal.

However, if you aren't too concerned about looking on very narrow spans for modulation issues or 100Hz (120Hz?) PSU ripple on signals or carrier noise then I guess this phase noise issue isn't that relevant :)





 

Offline videobruce

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
Re: Spectrum Analyzer - Rigol DSA815
« Reply #128 on: December 18, 2012, 04:38:45 pm »
Quote
Its very likely videobruce will be making the more extensive review of the 815s RF characteristics!
Man, you are really going out on a limb there.  ;D
Quote
if you aren't too concerned about looking on very narrow spans for modulation issues or 100Hz (120Hz?) PSU ripple on signals or carrier noise then I guess this phase noise issue isn't that relevant
I would say probably not. To sum it up, the phase noise problem pretty much concerns the noise floor level, correct?
« Last Edit: December 18, 2012, 06:56:40 pm by videobruce »
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #129 on: December 18, 2012, 07:31:30 pm »
Yes the noise effectively acts as a mask that hides information on the narrowest spans.

I suppose one other example of how it would limit the performance is if you tried to do a swept response of a crystal filter designed for narrow SSB or CW using the tracking generator.

The true stopband performance of the filter would not be displayed because the analyser would hide it with phase noise. You could still measure the passband reasonably well and prove the filter wasn't grossly faulty though.
« Last Edit: December 18, 2012, 07:34:03 pm by G0HZU »
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Spectrum Analyzer - Rigol DSA815
« Reply #130 on: December 19, 2012, 02:18:54 am »
 :-BROKE Sorry, I think I forgot to post an edit I made to that post "... more extensive review ... because you have another SA to compare it against."

In a query  IP2 = SHI on the manual: Is that suppose to mean it isn't good? If so, what should I be looking for?

Purely in theory, IP2 calculated by harmonics, SHI = second harmonic intercept, are higher by ~ 5dB compared to IM, IM used for estimating IP3. :-//

http://rf-mw.org/nonlinearity_effercts_nonlinearity_effercts_second_and_third_order_intercept_points.html#top

I see G0HZU point fairly clearly, but here other amateur radio users demonstrate how they can push the noise floor down further or compensate for it. 

Here via signal averaging:



or here showing the effect of turning on the preamp, removing the 30dB attenuation and further changing RBW:







Quote
Its very likely videobruce will be making the more extensive review of the 815s RF characteristics!
Man, you are really going out on a limb there.  ;D
Quote
if you aren't too concerned about looking on very narrow spans for modulation issues or 100Hz (120Hz?) PSU ripple on signals or carrier noise then I guess this phase noise issue isn't that relevant
I would say probably not. To sum it up, the phase noise problem pretty much concerns the noise floor level, correct?

« Last Edit: December 19, 2012, 06:35:01 pm by saturation »
Best Wishes,

 Saturation
 

Offline videobruce

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
Re: Spectrum Analyzer - Rigol DSA815
« Reply #131 on: December 19, 2012, 12:54:52 pm »
Golly. Thanks for all of that.
I don't bother with FM broadcast, but I was surprised to find out the FCC has allocated a wider bandwidth for these so called "HD" radio stations. I just assumed they found a way to piggyback the additional ability within what they already had, similar to ATSC TV.

That green line with the letters "DL" in those screen shots, what does that stand for?
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #132 on: December 19, 2012, 01:09:02 pm »
Quote
I see G0HZU point fairly clearly, but here other amateur radio users demonstrate how they can push the noise floor down further or compensate for it. 



Yes but the phase noise issue affects narrow spans the most. The solutions you posted up above are on wider spans.
As a rough rule of thumb the phase noise issue gets less significant by 6dB every time you double the distance from the signal. At a 1kHz offset it is highly significant as you can see in my earlier post.

You can see it spoils the noise floor on a 2kHz span. No amount of messing with the attenuation or preamp will fix this on a 2kHz span. The dynamic range in the 2kHz plot is limited by phase noise and not the settings for the preamp or attenuator.

Hope this helps :)
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Spectrum Analyzer - Rigol DSA815
« Reply #133 on: December 19, 2012, 03:46:43 pm »
Thanks GOHZU, again, you've pointed that out very clearly; that's at least one other of this SAs limitations.  But, like the Rigol 1052e is to DSOs,  a buyer of this SA has to decide if this is enough for its intended application.   I think most of the video youtube posts on the 815 have been by ham radio users, as they've taken to this unit favorably but other applications, detailed EMC testing perhaps? It won't be as applicable but maybe EMC pretesting?  On T&M World recently:

http://www.tmworld.com/electronics-products/electronic-product-reviews/other/4398946/2/Product-review--Rigol-DSA815TG-spectrum-analyzer

http://www.edn.com/electronics-products/electronic-product-reviews/other/4398963/Test-Drive--Rigol-DSA815TG-spectrum-analyzer

I think Rigol is  on to something here, as its exploiting a niche that's been filled until now mostly with used eBay SAs that are affordable by the ham community, or by EE looking for gear for more field level testing.  You now have a new unit that offers some competition to used units, with several caveats as you've noted.  The ARRL has about 150,000 members,  and would constitute at least one market segment, not sure how many of them would want a new SA or have never been able to get one until now.  By comparison, the US Bureau of Labor estimates there are ~ 300,000 EEs in the USA, so the ham community alone as another market for T&M gear other than EE's is sizeable.


Quote
I see G0HZU point fairly clearly, but here other amateur radio users demonstrate how they can push the noise floor down further or compensate for it. 



Yes but the phase noise issue affects narrow spans the most. The solutions you posted up above are on wider spans.
As a rough rule of thumb the phase noise issue gets less significant by 6dB every time you double the distance from the signal. At a 1kHz offset it is highly significant as you can see in my earlier post.

You can see it spoils the noise floor on a 2kHz span. No amount of messing with the attenuation or preamp will fix this on a 2kHz span. The dynamic range in the 2kHz plot is limited by phase noise and not the settings for the preamp or attenuator.

Hope this helps :)
« Last Edit: December 19, 2012, 06:29:02 pm by saturation »
Best Wishes,

 Saturation
 

Offline videobruce

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
Re: Spectrum Analyzer - Rigol DSA815
« Reply #134 on: December 19, 2012, 04:27:13 pm »
I couldn't get an answer from this Guy at Rigol for these questions;

1. Max input level, there are two figures 20dBm and 30dBm. Supposedly, 20 was what the unit was tested to, 30 is the "damage level" according to the spec sheet. Is this figure based on total levels from all signals within the full bandwidth of the scope, or just the specific range that is currently displayed?

2. Assuming it is within the specific span you set, if you are in the FM broadcast band with many stations at high levels, that number has to be exceeded isn't it?

3. A unusual feature of the tracking generator this "Power Sweep" where the output of the generator increases across the frequency range under test. Why and how is this used?
 

Offline olsennTopic starter

  • Frequent Contributor
  • **
  • Posts: 993
Re: Spectrum Analyzer - Rigol DSA815
« Reply #135 on: December 19, 2012, 05:41:22 pm »
Quote
1. Max input level, there are two figures 20dBm and 30dBm. Supposedly, 20 was what the unit was tested to, 30 is the "damage level" according to the spec sheet. Is this figure based on total levels from all signals within the full bandwidth of the scope, or just the specific range that is currently displayed?


For the DSA815 the max input level before (potentially) damaging the device is 20dbm. From my understanding (but I could be wrong about this) that is now only the combined input accross the 9kHz - 1.5GHz range of the device, but even frequencies outside this range as well. I always leave a 30db attenuator on the input just in case.
 

Offline videobruce

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
Re: Spectrum Analyzer - Rigol DSA815
« Reply #136 on: December 20, 2012, 01:02:54 pm »
On the front of the scope under the input jack it states "Max+20 dBm/50 Vdc", in the specs it states "CW RF Power / RF attenuation=30db / +20 dBm (100mW)", then "Max. Damage Level / +30 dBm".

My IFR 2399A states "+30 dBm, 50Vdc" on the front panel, but in the specs it states "+30 dBm max input with 30 db attenuation, 50 Vdc"

My utter confusion is, can I assume those specs are for a single signal generated on the bench, as opposed to connecting it to an antenna for broadband monitoring?
« Last Edit: December 20, 2012, 01:05:03 pm by videobruce »
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #137 on: December 20, 2012, 01:40:11 pm »
Connecting your analyser to an external antenna is a risky business because it won't be designed for this.

You run the risk of ESD damage from nearby lightning storms and also you run the risk of someone using a high power CB radio nearby. But they would have to be running very high power and be driving past just a few metres away in order to couple 30dBm into your antenna.

However, if you had a longwire antenna fitted to the analyser and someone (maybe you?)  was transmitting high power on 3.5MHz then they could be 100 to 200 metres away and still toast the analyser.

Basically you should consider fitting an external diode based limiter to the front of the analyser to minimise the risk of damage if you want to regularly connect a large antenna to it.

I seriously doubt you need to worry about multiple broadcast stations damaging your analyser unless you live very close to a transmitter mast. It's local CB/Ham radio across 2-30MHz or ESD damage eg usually via lightning storms that are the biggest risk. A mobile phone is unlikely to harm it unless it is right next to the antenna or you connect directly somehow. A nearby phone mast shouldn't be a risk unless it is your next door neighbour because the path loss at 900MHz+ will be so high.

Also, the true damage threshold level of your analyser will depend on frequency and duty cycle of the damaging signal(s).

So there can never be a precise figure given for this. But your Rigol has some very small SMD devices in the front end and these will probably be rated for 30dBm average power in a thermal sense. i.e. 1W is going to make the tiny little input attenuator get very hot if it is set to 10dB attenuation or more.

That's why the Rigol has a crude overload detector by the input connector. This broadband detector will sense a big signal (eg +20dBm?) and it will redirect the input signal to a dummy load and prevent it from hitting the vulnerable attenuator.

But this ability is probably only limited to two or three watts before the protection circuit itself gets damaged. Plus it wont be 100% foolproof so the damage could still happen if you hit it with a Watt or more of RF anywhere from LF to many GHz.
« Last Edit: December 20, 2012, 01:49:01 pm by G0HZU »
 

Offline videobruce

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
Re: Spectrum Analyzer - Rigol DSA815
« Reply #138 on: December 20, 2012, 01:46:50 pm »
It's not continuously connected and living in a city, there is little chance of lightning strikes. Besides, I wouldn't be using it when any lightning was evident.  ;)
No long wire and "CB' is basically dead for the most part.

Quote
Basically you should consider fitting an external diode based limiter to the front of the analyser
Any sources for this, or is this a home brew deal?
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #139 on: December 20, 2012, 02:01:47 pm »
You can buy a limiter or make one yourself.

A commercial limiter rated to maybe 2GHz should protect for strikes up to about 10W average power. It should reflect the power back and only leak about +16dBm into the analyser.

If you want to make one yourself you can use cheaper silicon diodes and this will give pretty good protection up to maybe 5W up to maybe 500MHz.

Obviously the limiter doesn't mean you can now deliberately put 1W to 5W into your analyser. It would just be like an insurance policy against you overdriving the analyser by accident.

A good limiter will only introduce a fraction of 1dB insertion loss but the downside of the limiter is that it will introduce IMD and harmonic distortion if you feed signals to tha analyser above about +3dBm because of the limiting action.
« Last Edit: December 20, 2012, 02:06:12 pm by G0HZU »
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #140 on: December 20, 2012, 02:13:37 pm »
The other aspect to this is the Rigol appears to use commercially available parts in the front end.

I do RF design for a living and use parts like this all the time. I did look at the closeup images of the insides of the Rigol and I was able to identify manuf part numbers for nearly all the SMD parts used in the RF signal path. I have even identified the ones that have been sanded by Rigol to hide the part number. eg the little SMD switches, the SMD mixer, the SMD 5 bit RF attenuators and the SMD amplifiers.

They are all cheap parts so as long as the PCB doesn't get thermally toasted by silly amounts of RF power then a repair should only be $20 - $30 in parts or less.

If Rigol are helpful here they may even produce a damage overload repair kit. HP/Agilent do this for some of their instruments if the relevant parts required are small and cheap.
« Last Edit: December 20, 2012, 02:27:46 pm by G0HZU »
 

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4787
  • Country: us
  • Doveryai, no proveryai
    • NIST
Re: Spectrum Analyzer - Rigol DSA815
« Reply #141 on: December 21, 2012, 04:58:47 pm »
That's great insight again GOHZU, another plus for users should Rigol tech support be unhelpful today, or the future.   Some years ago there was an Instek sub-$1000 SA that didn't generate as much interest as the Rigol, due to design limitations.

http://www.tequipment.net/InstekGSP-730.html

The other aspect to this is the Rigol appears to use commercially available parts in the front end....

Best Wishes,

 Saturation
 

Offline videobruce

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
Re: Spectrum Analyzer - Rigol DSA815
« Reply #142 on: December 21, 2012, 05:12:41 pm »
The two calls I have made to Rigol have been unhelpful. The same guy answered the phone both times. He knew about as much as I did about the SA.  :--
One would think I was dealing with a typical CE importer.

Quote
Some years ago there was an Instek sub-$1000 SA that didn't generate as much interest as the Rigol, due to design limitations.
Never was aware of that. It appears to be current. What "design limitations", other than no TG?
« Last Edit: December 21, 2012, 05:16:51 pm by videobruce »
 

Offline PA0PBZ

  • Super Contributor
  • ***
  • Posts: 5121
  • Country: nl
Re: Spectrum Analyzer - Rigol DSA815
« Reply #143 on: December 21, 2012, 07:29:59 pm »
What "design limitations", other than no TG?

I see minimum RBW 30 KHz, that alone would keep me from buying it.
Keyboard error: Press F1 to continue.
 

Offline pauln

  • Contributor
  • Posts: 28
  • Country: au
Re: Spectrum Analyzer - Rigol DSA815
« Reply #144 on: December 22, 2012, 07:49:35 am »
A question re the VSWR option.  I have seen several YouTube videos use an non Rigol directional coupler and the VSWR measuring option on the DSA815TG.  I thought you had to buy their coupler in order to get the option license as well?  So can this option be activated without buying Rigols accessory?
 

Offline videobruce

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
Re: Spectrum Analyzer - Rigol DSA815
« Reply #145 on: December 22, 2012, 12:01:28 pm »
AFAIC, I 'm not paying $440 for a bunch of ones & zeros (software activation). That's outrageous, ridiculous and just plain greedy. >:(  $50 yes, $100 maybe, but 1/3 the cost of the SA, forget it. This isn't the first time I have seen this. It's poor marketing and poor customer relations. It really turns me off to these companies.
I did voice my opinion when I called, but received a excuse that since the SA was priced the way it was, spending the additional $$ still brought it under all the others. While technically true, it still doesn't change the fact you are paying 1/3 of the price of the device for some limited function software "key".

I have a separate RLB (Return Loss Bridge): Eagle RLB150x5 that I use with my Aeroflex 2399A. Of course it doesn't provide VSWR measurements, but at least I can see the plot on the scope.
One thing I will admit, their VSWR adapter, which is a RLB is far cheaper than the Eagle. But, the Eagle is made here, not China. ;)

For the cost of their software activation (you can d/l the software, one of which is a 350MB package for free) and a RLB, you can buy a decent SWR meter.

Quote
So can this option be activated without buying Rigols accessory?
AFAIK, yes. They are listed separately.
« Last Edit: December 22, 2012, 12:11:58 pm by videobruce »
 

Offline jimeagle

  • Newbie
  • Posts: 7
  • Country: us
  • The only constant is change
    • WB6JNN Webpage
Re: Spectrum Analyzer - Rigol DSA815
« Reply #146 on: December 23, 2012, 04:54:03 am »
If you'd like to see how a standard $85 Mini-Circuits Directional Coupler can be used to determine Return Loss (VSWR in terms of dB), check out my videos on You Tube at https://www.youtube.com/jimeagle1.  I have several videos showing how the Rigol DSA815-TG can be used for Ham Radio or even commercial use.

Jimeagle1
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #147 on: December 23, 2012, 02:09:20 pm »
Hi Jim
Nice video with the coupler. That ZFSD-20-5 coupler has a very impressive spec (eg for directivity) and covers the whole range of the Rigol and is very good value for $85.


For people who maybe just want to do measurements up to 144MHz the low cost option is to make a transformer based coupler based on the classic Sontheimer design. The biggest cost for a homemade one will be in the choice of connectors and the time taken to box it all up but it is possible to make a high performance 20dB coupler over a fairly restricted range for just a few dollars if you build it from mostly salvaged parts...:)
« Last Edit: December 23, 2012, 02:13:23 pm by G0HZU »
 

Offline videobruce

  • Frequent Contributor
  • **
  • Posts: 447
  • Country: us
Re: Spectrum Analyzer - Rigol DSA815
« Reply #148 on: December 23, 2012, 03:17:23 pm »
How does that compare to a true RLB such as the one from Eagle?
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3012
  • Country: gb
Re: Spectrum Analyzer - Rigol DSA815
« Reply #149 on: December 23, 2012, 03:54:46 pm »
For measuring return loss the spec on a decent return loss bridge should be better than a coupler. But a coupler is arguably more versatlie.

I'm not familiar with eagle products but you can make a pretty decent return loss bridge using a few precision resistors and a homemade balun transformer for a few dollars.

This would be good enough for use up to 144MHz.

The alternative is to use a transformer based coupler like a Sontheimer coupler but you need to choose a coupler that has very high directivity if you want to get close to the performance of a well made return loss bridge.

« Last Edit: December 23, 2012, 03:57:24 pm by G0HZU »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf