Exclusive! Video Tektronix RSA306 USB Real-Time Spectrum Analyzer Review

[Andy2]

Hey Shahriar,

all your Videos are awesome!

Maybe you could do a comparision of low cost USB Analyzers, that would be great.
(Tek RSA306, Signalhound BB60, Aaronia Spectran V4 and V5)
Does not have to be all in the same video, just for example measure the same signals or so and
compare saved results.

Cheers!

And just to add to Shahriar's load, it'd be nice if he could include the budget Signal Hound SA44b. The phase noise performance of this analyser is significantly better than the Tek 306 and the BB60 and a truckload better than the Rigol models.

https://www.signalhound.com/sigdownloads/SA44B/AppNote-4.pdf

At 1GHz, the figure at 10KHz is -104 dBc and at 100KHz is -110 dBc.
At 10 MHz, the figure at 10 KHz and 100 KHz is around -118 dBc.

Still a mile away from a decent hardware analyser of Ye Golden Age, but quite respectable and useful, especially at the price.  It's a pity the supplied software is so wobbly. Apparently the re-write is almost ready.

[Laidukas]

K5OBS also has a "World Exclusive" review of this device on his YouTube channel although the review on The Signal Path is more extensive. I would like a new spectrum Analyzer, I really would, but when you consider the cost of this device plus the cost of the new laptop that would be needed to drive it the stand-alone units from Rigol are still a viable option.

I would be interested to see a side by side review of this unit together with the Signal Hound and (say) a $3000 Rigol device.

In my opinion the selling point of this device is real time signal analysis + you get bonus spectrum analyzer which appears to be a little bit limited by spurious. I do not think that it is fair to compare this Tek (or Signal Hound) to conventional spectrum analyzer as it is very limited in real time signal analysis and can not compare with all measurements that Tek is capable of doing. With conventional RIGOL spectrum analyzer you can sweep on a signal that is not intermittent and do some basic measurements, that's it. of course Conventional spectrum analyzer has benefit of tracking generator which is very important.
At work i use R&S FSL series with tracking gen. Yes FSL is has better specifications but it is so limited in signal analysis and measurements are only basics. At home i have old HP8753B VNA. And now after seeing this Tek analyzer on a review i come up to a conclusion that the best mix is signal analyzer like this Tek + VNA. VNA replaced the need of spectrum analyzer tracking generator and Tek gives impressive real time signal analysis for the price. I now really starting to like the idea of buying this Tek to my home lab and in combination of VNA i will cover all measurements that I need.       

[G0HZU]

In my opinion the selling point of this device is real time signal analysis + you get bonus spectrum analyzer which appears to be a little bit limited by spurious. I do not think that it is fair to compare this Tek (or Signal Hound) to conventional spectrum analyzer as it is very limited in real time signal analysis and can not compare with all measurements that Tek is capable of doing.

I agree but I do think it's worth comparing these analysers on a hobby forum because some people won't be able to decide which analyser suits their needs.

The Tek is the clear winner in terms of being able to do real time signal analysis over a 40MHz wide IF BW  (anywhere up to about 6GHz) but it would be a poor choice for someone who wanted to (critically) measure the spurious performance of a ham transmitter or any kind of test signal at V/UHF. The spurious performance is pretty woeful and the phase noise is very high.

The appearance of the Tek RSA306 is definitely going to influence what we see appearing in the next few years in terms of small (but powerful) signal analysers at maybe $4000-6000. I'd expect to see similar sized units appearing but with internal battery power possibly charged via USB. This increase in power budget would open the door for respectable performance in the RF section.

[electronic_eel]

Why didn't they put the frontend from the MDO4000B series into these boxes? They don't have the problem with these spurious tones there.

Maybe the box would have become a little bit bigger or they would have needed a snap on battery pack - but I don't think that would hurt the usability.

[Rasz]

teardown:



so it is basically a glorified overengineered SDR
just wow, they packed so much shit into it, so many filters and stages, and at the end got tons of spurious tones? WTF Tek?

USRP B200/N200 delivers same(or better quality signal) at fraction of a price.

This RSA306 is a good deal only if you
-must buy tek
-like the software

 

[Laidukas]

USRP B200/N200 delivers same(or better quality signal) at fraction of a price.

This RSA306 is a good deal only if you
-must buy tek
-like the software

Does USRP B200/N200 allow to measure absolute values? Or just relative? Does it have all in one software solution ready to use after install? Or you need to spend week/month gathering information from bunch of sources?  Joining software packages, fetching info between them to achieve one type of signal demodulation?

Is there software in general that gives technical parameters usable for electronics engineer? Or it just spits out final result - demodulated data?

Can you recommend good introduction/overview source for USRP B200/N200 ?

I watched few videos on youtube, indeed it is wonderful piece of equipment but serves not the same purpose as RSA306. And it looked form my that you need to be quite advanced in computers to use USRP B200/N200 in any more advanced way.   

[G0HZU]

so it is basically a glorified overengineered SDR
just wow, they packed so much shit into it, so many filters and stages, and at the end got tons of spurious tones? WTF Tek?

There are various reasons why it has so many spurious terms. It looks like Tek decided to aim for a ballpark 50dB SFDR and for this reason there are numerous aspects to the design that limit it to this level of performance.

I definitely wouldn't describe it as an over engineered SDR. It has been engineered to work from USB power and to tune up to 6GHz and also it has been engineered to have a (relatively) low cost layout. The penalty for all this is relatively poor RF converter performance when compared to other options. But it does offer the 40MHz RTBW and the Tek Signal Vu engine in a suitable PC.

One thing I noticed in your youtube link image to the KF5OBS video. It is frozen at 14:52 in the video and one of the large inductors in the 140MHz diff filter after the last amp and before the ADC looks to be badly burned. This is best seen at 8:52 in the video.

I'm trying to fathom how this could have happened if the analyser is USB powered. To do this with DC would take an extremely high current.

It more likely would have been done with RF but I'm not sure how several watts of RF could get to that point in the analyser. Either by the converter path or by the LF bypass mode. Maybe someone fed a huge RF power into it or maybe someone at Tek tried to connect a high power 2m (144MHz) transceiver here to listen to the 140MHz IF and accidentally hit the PTT transmit button? Or maybe Tek did the damage during development because it is a preproduction model and may have been zapped by a big external power supply somehow? Or maybe they spilt something corrosive on it or used a predamaged part? Either way, it would have made a lot of smelly smoke... It may have been in the hands of various Tek and non Tek testers before this review but still seems to work though.

 

[Laidukas]

One thing I noticed in your youtube link image to the KF5OBS video. It is frozen at 14:52 in the video and one of the large inductors in the 140MHz diff filter after the last amp and before the ADC looks to be badly burned. This is best seen at 8:52 in the video.

I'm trying to fathom how this could have happened if the analyser is USB powered. To do this with DC would take an extremely high current.

It more likely would have been done with RF but I'm not sure how several watts of RF could get to that point in the analyser. Either by the converter path or by the LF bypass mode. Maybe someone fed a huge RF power into it or maybe someone at Tek tried to connect a high power 2m (144MHz) transceiver here to listen to the 140MHz IF and accidentally hit the PTT transmit button? Or maybe Tek did the damage during development because it is a preproduction model and may have been zapped by a big external power supply somehow? Or maybe they spilt something corrosive on it or used a predamaged part? Either way, it would have made a lot of smelly smoke... It may have been in the hands of various Tek and non Tek testers before this review but still seems to work though.

Now this is impressive How it was possible to fry that thick wire inductor in the end of RF stage and not do damage anything to that point? Even with down conversion bypass...

Maybe someone accidentally bumped the inductor and they needed to change it. Then used simple hot air station.  It is possible to notice that inductor on the left and resistors on the right  has excessive heat signatures. If you do not preheat these type of boards with hot plate and try to solder only with hot air station it ends in fried component that you need to solder in, because this type of boards sucks heat very effectively and soldering pads do not reach melting point so easily.  Add RoHS solder with its high melting point and way you go ...

[G0HZU]

Yes, it's a weird one isn't it?

Perhaps it was predamaged. Maybe the coating on the wire is very prone to heat damage from unskilled use of a poorly set up rework station as you suggest.
I would expect it to be fairly easy to remove/rework because it is part of a balanced filter (no direct heat path to the groundplane) so the part should come off easily for a skilled operator. But maybe an impatient tech at Tek had an informal attempt to rework it themseleves and then ignored the subsequent carnage.

However, it 'looks' like classic RF damage. but this doesn't add up for the reasons we have already listed.

At my place of work we do something called passive immunity testing on a new design where the harshest test we do is to feed RF from a high power sweep generator into each RF port with the equipment in an unpowered state. The sweep is often up to many GHz and the equipment is observed with a thermal camera as the sweep is slowly done and repeated at ever increasing power levels up to quite a few watts. Once you get up into the UHF/uW region it can be surprising how far into the equipment the resonant hotspots develop (and how narrowband the resonances can be) and also how well other parts of the equipment survive this kind of deliberate 'abuse'. Usually this is done under very controlled conditions and I've done this testing myself on my own designs. It is easy to cook individual parts deep within the unitwith this method but I doubt Tek would go to these lengths on a product like this. Also, I wouldn't expect to see that amount of damage. It looks like that part has been exposed to a lot of energy in one form or another...

 

[Laidukas]

I would expect it to be fairly easy to remove/rework because it is part of a balanced filter (no direct heat path to the groundplane) so the part should come off easily for a skilled operator.

Yes, you are right, I agree.


O.T. It seams that G0HZU is building serious gear at work

[G0HZU]

Yes, you are right, I agree.

Maybe some engineer in a lab at Tek hastily took it off (maybe for test access purposes) and then lazily refitted the same rework/heat damaged part. So it saw two lots of overheating. If it's a preproduction unit then all kinds of engineering fudgery may have happened to it in its short lifetime. Damage from unskilled rework seems to be the most likely cause

[Rasz]

Does USRP B200/N200 allow to measure absolute values? Or just relative? Does it have all in one software solution ready to use after install? Or you need to spend week/month gathering information from bunch of sources?  Joining software packages, fetching info between them to achieve one type of signal demodulation?

Is there software in general that gives technical parameters usable for electronics engineer? Or it just spits out final result - demodulated data?

Can you recommend good introduction/overview source for USRP B200/N200 ?

-absolute, but probably not in the same ballpark of accuracy compared to dedicated SA?
-no software, it spits pure I/Q samples, you can do whatever you like with this data stream, be it Gnuradio, Matlab, Labview

examples



Gnuradio will let you build receivers/test benches for many modulations in minutes (all gui, basically like lego)

this can go up all the way to MIMO WIFI transceivers/3G base stations/DVB transceivers, all by simply drag&dropping modules in gui

all at 1/5-1/8 of a price.
You can even get 120MHz RF BW usrp X300 cheaper than this tek SA, so BW/hardware itself is really not a strong feature of RSA306.

RSA306 is tek selling you $5K of software with free hardware dongle as a bonus.

[videobruce]

The £2,340 equates ti $3664 in the US?

[Mr Simpleton]

all at 1/5-1/8 of a price.
You can even get 120MHz RF BW usrp X300 cheaper than this tek SA, so BW/hardware itself is really not a strong feature of RSA306.

RSA306 is tek selling you $5K of software with free hardware dongle as a bonus.

I had to check Ettus webshop to see for myself! The cheapest X300 is roughly 1.5 times MORE than the Tek RSA306, and then you need all sorts of add-on to get it running (daugher boards, interface boards).  I guess it would land at 2x the price for a Tek! Now don't get me wrong, the Ettus is pure magic and is THE thing to have if you develope SDR stuff. The Tek is an instrument intended for measurement, not develope applications.

Lets compare apples to apples 

[Lukas]

Lets compare apples to apples 

IMO the comparison isn't that unfair since the RSA306 way more resembles a SDR than a spectrum analyzer. Its just marketing and software that turns this thing into a spectrum analyzer. Someone could come up with some software and call the USRP a spectrum analyzer with integrated RF signal source.

[PA0PBZ]

Am I the only one wondering what it would take to feed the software with some otherwise obtained signal? 

[Rasz]

X300 is roughly 1.5 times MORE than the Tek RSA306

my bad, I somehow remembered "around 5K" from the review YT clip :/

[dthung]

X300 is roughly 1.5 times MORE than the Tek RSA306

my bad, I somehow remembered "around 5K" from the review YT clip :/

The Ettus B200 is listed around USD $675 in Ettus site. It has driver support for GNURadio and Matlab (I believe).

[Rasz]

X300 is roughly 1.5 times MORE than the Tek RSA306

my bad, I somehow remembered "around 5K" from the review YT clip :/

The Ettus B200 is listed around USD $675 in Ettus site. It has driver support for GNURadio and Matlab (I believe).

I meant the tek. Somehow I forgot the 3 at the beginning, and remembered $4900


Btw its not all matlab labview and text based outputs, any SDR can do pretty nice things with gnuradio, for example


Im scared to think how much you would have to pay for SA that does this at those BW spans at fluid 60 frames per second.

[dthung]

Honestly SDR + GNURadio, e.g. RTL Dongle/HackRF/BladeRF, is THE actual disruptive product. The Agilent VSA 89600 system: software starts at $10K and 2nd compatible HW starts at another $10K in xBay. They can't even transmit the signal. 

[MySuggestions]

Thanks for a great review on this SA.

One of the reasons why Tektronix has to be aggressive in the price structure for a USB based SA, is probably the German based company AARONIA. They have USB based SP for many years and very aggressive pricing.
http://www.aaronia.de/

Yes i think you got the point.
And the Aaronia V5 offers a much higher frequency range (1Hz to 20GHz), onboard I/Q signal generator, 40GHz power meter, -170dBm sensitivity...
But the main point is that it is a real hand held unit without the need of a PC thanks to the color display 
The only problem: Not on the market (jet) 

[Laidukas]

I never used AARONIA spectrum analyzers, but from one person that did used I hear that they can hardly be called real time spectrum analyzers as they sweep very slowly, especial with bigger spans. This is major disadvantage. Maybe someone working with AARONIA can comment on that?

[MySuggestions]

I think you are mixing things up. I am speaking about the new SPECTRAN V5. For sure this is a real time analyzer.

There is a video from Analog Devices showing the unit in an early development stage at

[kaz911]

I think you are mixint things up. I am speaking about the new SPECTRAN V5. For sure this is a real time analyzer.

There is a video from Analog Devices showing the unit in an early development stage at

V5r2 is supposed to ship end of march (2015) - r1 never shipped. Preorders started in 2011..... (take THAT kickstarter) But up to 20 GHz - sensitivity -170dBm(Hz) at a wide range (range not published)

[MySuggestions]

V5r2 is supposed to ship end of march (2015)

Perfect Can't wait to get my hands on one of those.
I have just seen that they show the V5 V2? at the EMV fair in Stuttgart (end of March). I think i will go there. Anybody interested in a review?