Author Topic: Presenting my new software for the SMA Spectrum Analyzer 138MHz-4.4GHz (Read 62840 times)

Bicurico · « **on:** July 24, 2016, 09:04:53 am »

Hello,

On eBay, a cheap redesign of the amateur NWT4000 devices is sold as "SMA Spectrum Analyzer 138MHz-4.4GHz". Bascially it uses only one ADF5430 chip instead of two and hence cannot be used for antenna tuning (which was the original NWT series purpouse), as the signal being output cannot be read back in at the same time (a second ADF5430 chip would be needed).

The included software (WinNWT) is a freeware made by the original author of the NWT series of devices (NWT500), which is not particularily nice for pure spectrum analisys.

I read the documentation, did some hacking and ended up programming my own spectrum analyzer software, more for fun and learning, than anything else.

I came to this forum, following links regarding the Siglent SSA3021X and as Douglas Adams used to say: "I may not ended up where I wanted, but I ended where I needed to be."

Anyway, check my forum:

http://vma-satellite.blogspot.pt/2016/06/sma-spectrum-analyzer-138mhz-44ghz.html
http://vma-satellite.blogspot.pt/2016/06/the-travellers-guide-to-spectrum.html
http://vma-satellite.blogspot.pt/2016/06/138m-44g-usb-sma-signal-sourcesignal.html

The software is free, but I am distributing it myself, because I don't want Chinese eBay sellers making profit with it, when I had all the work and had to purchase two units to get it done.

Regards,
Bicurico

Fraser · « **Reply #1 on:** July 24, 2016, 09:42:29 am »

Thank you for your hard work on this. I own several such units and shall take a look at your web pages 👍

Fraser

technogeeky · « **Reply #2 on:** July 24, 2016, 11:23:22 pm »

Nice work!

I might just add this spectrum analyzer to my short-list of things I can't yet afford.

Forgive me if this was in the videos, but why did you need to buy two of them in order to develop the software?

I understood why you wanted two TV tuners to do what you did.

In any case, great work!

Bicurico · « **Reply #3 on:** July 25, 2016, 02:19:57 am »

Hello.

I broke the first one just a few days after receiving it by connecting a cable from my satellite receiver. I did not pay attention and it carried live voltage to control the LNB. I already started my software development and did not want to live without it anymore, so I bought a second one.

Simultaneously, I bought 5 IAM-81008 as this IC was likely to be the culprid. After swapping it out, the device was working again, so now I have two.

One thing I am considering is the possibility to use both simultaneously to double the sweep speed. But for this I definitely need to implement some calibration routine first.

Also, having two of them allowed me to see that they have slightly different performance. Some calibration algorithm is definitly required for best results.

Regards

Alex Eisenhut · « **Reply #4 on:** July 25, 2016, 02:55:33 am »

Just a correction, the chip part number is ADF4350.

technogeeky · « **Reply #5 on:** July 25, 2016, 04:23:21 am »

Quote from: Bicurico on July 25, 2016, 02:19:57 am

Hello.

I broke the first one just a few days after receiving it by connecting a cable from my satellite receiver. I did not pay attention and it carried live voltage to control the LNB. I already started my software development and did not want to live without it anymore, so I bought a second one.

Simultaneously, I bought 5 IAM-81008 as this IC was likely to be the culprid. After swapping it out, the device was working again, so now I have two.

One thing I am considering is the possibility to use both simultaneously to double the sweep speed. But for this I definitely need to implement some calibration routine first.

Also, having two of them allowed me to see that they have slightly different performance. Some calibration algorithm is definitly required for best results.

Regards

Couldn't you use two of them to do whatever the NWT4000 can do? I don't have any interest in "antenna" tuning anyway, but I was just curious to know.

Thanks,
-tg

Bicurico · « **Reply #6 on:** July 25, 2016, 06:55:11 am »

@Alex: Thanks! That was a wild typo...

@tg: You could in theory, but WinNWT would not work, as it is not prepared to control two separate devices. I have thought about that, too, especially since two "SMA Spectrum Analyzer 138MHz-4.4GHz" cost less than the NWT4000. But beware: the original NWT device apparently features 10 bit resolution, while this "SMA Spectrum Analyzer 138MHz-4.4GHz" only has 8 bit resolution. There seems to be a difference, so don't just go for the lowest price!

Because I am not into antenna tuning, either, the "SMA Spectrum Analyzer 138MHz-4.4GHz" is really OK for the price it is sold: 50-60 Euro.

Regards

technogeeky · « **Reply #7 on:** July 25, 2016, 09:30:00 am »

Quote from: Bicurico on July 25, 2016, 06:55:11 am

@Alex: Thanks! That was a wild typo...

@tg: You could in theory, but WinNWT would not work, as it is not prepared to control two separate devices. I have thought about that, too, especially since two "SMA Spectrum Analyzer 138MHz-4.4GHz" cost less than the NWT4000. But beware: the original NWT device apparently features 10 bit resolution, while this "SMA Spectrum Analyzer 138MHz-4.4GHz" only has 8 bit resolution. There seems to be a difference, so don't just go for the lowest price!

Because I am not into antenna tuning, either, the "SMA Spectrum Analyzer 138MHz-4.4GHz" is really OK for the price it is sold: 50-60 Euro.

Regards

I was thinking along the lines of feeding a selected portion of the input to the output so you cloud view it with an oscilloscope.

I really don't know what I'm talking about, so if this doesn't make sense, just ignore it.

Bicurico · « **Reply #8 on:** July 25, 2016, 11:25:44 am »

You cannot output part of the input. You can select a given frequency to output a reference signal and then:

1) measure it directly
2) measure it using for instance the antenna you want to tune or the cable whose attenuation you want to determine.

Actually, using my software as it is, you could use the first session to generate the signal and the second session to measure.

And that should be viable in WinNWT as well.

Regards

technogeeky · « **Reply #9 on:** July 25, 2016, 06:16:54 pm »

Quote from: Bicurico on July 25, 2016, 11:25:44 am

You cannot output part of the input. You can select a given frequency to output a reference signal and then:

1) measure it directly
2) measure it using for instance the antenna you want to tune or the cable whose attenuation you want to determine.

Actually, using my software as it is, you could use the first session to generate the signal and the second session to measure.

And that should be viable in WinNWT as well.

Regards

I think I'm going to wait until you release the software to buy one of those, but I can't imagine getting anywhere near that good of a spectrum analyser for anywhere near that price.

Bicurico · « **Reply #10 on:** July 25, 2016, 07:09:39 pm »

Well, 50-60 Euro is not much money, indeed...

The spectrum is kind of slow and precision is almost non-existant at this point. I need to figure out how to implement a suitable calibration routine. Fortunately I own a professional calibrated signal source plus several high end TV analyzer/field meter with spectrum analyzer, so I have somthing to work with.

Regarding the "release": you can have the software whenever you want - just send me an e-mail and you get a copy.

Regards

Bicurico · « **Reply #11 on:** August 21, 2016, 11:16:51 am »

Hi,

Just published a new post on my blog: did some reflectometer measurments!

Cool! It actually works.

Regards

TurboTom · « **Reply #12 on:** August 21, 2016, 06:11:07 pm »

Unfortunately, you cannot really do narrow band work with this device. Since it's just a plain down-converter with a log detector at the IF output of the mixer, you won't detect any signal when you're right on the frequency (IF 0). You will find the signal mirrored to the left and right of the actual frequency. This won't really matter if you look at large spans of several or even hundreds of megahertz. But what matters in this case is that the PLL output of the device (which is also used as the LO signal for the mixer) is more or less a square wave (at the lower frequencies easily checked with a fast oscilloscope). At least it carries a whole lot of powerful harmonics. Since these harmonics also result in mixing products, the spectrum will be very ugly with lots of mirror and "ghost" signals that make it very difficult to get a clue of the actual input signal. I'ts a nice toy to start tinkering in the microwave frequency range for little money. But if you want to do any half-way serious measurements, do yourself a favor and buy, steal or borrow a real spectrum analyzer.

Anyway, congratulations on your work, Bicurico, your software looks superb and appears to be much easier to use than the original WinNWT software (that even doesn't support the device completely...).

All the best,
Thomas

Bicurico · « **Reply #13 on:** August 21, 2016, 06:27:31 pm »

Thanks Thomas.

I actually do own a lot of professional equipment, mainly TV field meter, encoders and modulators, apart from lots of receivers, dishes, etc. My collection of field meters has more than 20 devices...

I just purchased an SSA3021X, as well.

Why am I writing the software and spending time with this cheap device, if I could just use my R&S, Rover Instruments, Kathrein, Spaun, etc field meter?

The answer is simple: it has been a steep learning curve! I know now more about spectrum analysis than I would by just reading textbooks.

And yes, you are right: the spectrum does not look very good for longer cables. But activating the average function, I could get a readable spectrum for a 100m cable. I put a new entry on my blog.

Also, you are right: I am mostly interested in TV applications and hence wide spans our several megahertz are ok.

Regards

Bicurico · « **Reply #14 on:** August 22, 2016, 01:26:47 pm »

Hi,

I was struggeling to provide an easy download, which could be updated with the latest release.

Found out i could just attach it here...

This version has a bug in the activation code routine corrected. If you already downloaded from TinyUpload, please use this attached version, instead.

Send me an e-mail with the system code and you get an activation code for free.

Regards

nctnico · « **Reply #15 on:** August 22, 2016, 01:45:34 pm »

Just question: does the generator mode support stepped sweeping? Go through a frequency range and wait for a while until going to the next frequency?

Bicurico · « **Reply #16 on:** August 22, 2016, 01:52:44 pm »

Not in the current release.

It could be implemented, though.

Like, you specify:

Start Frequency (MHz):
End Frequency (MHz):
Step size (MHz):
Pause (ms):

This would start at Start Frequency, Pause, Increment, Pause, Increment, ... -> while Current Frequency <= End Frequency

Is that what you want/need?

What would you want to use it for?

Note that you CANNOT connect the SMA Output connector to the SMA Input connector! It will not work, as you only have one ADF4350: while the device is generating the output signal, it will not read the input signal. When it does read the input signal, it won't be outputting anything. Effectively you would need two devices or, better, purchase the NWT4000.

Regards

nctnico · « **Reply #17 on:** August 22, 2016, 02:07:51 pm »

Yes!
Long story: Many years ago I build an ADF4350 board which could do a stepped sweep as you describe but I hacked the board badly during testing (in the end my spectrum analyser turned out to be faulty) so I would need to build another one and see if I can find/polish the firmware but even then it is command line only. All in all it is just easier to buy a readily made unit from Ebay and be lazy

I just like to have a device which is somewhat useful beyond 3GHz (that is as far as my RF generator goes).

Bicurico · « **Reply #18 on:** August 22, 2016, 03:27:22 pm »

Hi,

I just had some time and the way the software is structured, it was easy to implement the new Sweep Generator tab.

The pictures show the tab in action and the result using a professional spectrum analyzer using MAX HOLD on the current trace.

Regards

nctnico · « **Reply #19 on:** August 22, 2016, 06:22:53 pm »

Great! I ordered a SMA spectrum analyser from Ebay so I have to wait until it arrives to test your changes.

bitseeker · « **Reply #20 on:** August 23, 2016, 12:52:15 am »

Bicurico, does your software also work with the boards/boxes that use the single ADF4351 (starts at 35 MHz instead of 138 MHz)?

Bicurico · « **Reply #21 on:** August 23, 2016, 07:09:20 am »

I only own two of the 138MHz-4.4GHz versions.

I *think* that the 38MHz-4.4GHz version with the ADF4351 *should* just work with the current software as it is, though you won't be able to select frequecies lower than 138MHz, because I verify user input.

I could let the user device if the software should allow 38MHz-4.4GHz or 138MHz-4.4GHz, but I have no way of testing it. Any volunteer is welcome: just run the software and tell me if indeed it works with the 38MHz-4.4GHz version. If so, I implement a selector on the SETUP TAB.

Regards

TurboTom · « **Reply #22 on:** August 23, 2016, 08:32:19 am »

I've got the 35MHz version of the "box" but currently it's at home and I'm on work. I can try it on friday. But I've got another idea on how to improve the gadget for TG/scalar network analyzer mode which currently isn't possible due to the zero IF situation:

I think it should be working if we use another double-balanced mixer on the TG output (just the same as on the input) and feed a sine signal (maybe 10...100kHz) on the second mixer input (or back-feed it to the IF output, depending on the possible frequency ranges) so the mixer outputs only the two sidebands wich can be easily analyzed by the input section. This won't provide really narrow bandwidth but for matching an antenna or input filters, it should be sufficient. And it can be easily implemented as an add-on. To compensate for the mixing losses, an amplifier may be required though. Maybe Bicurico could implement a "normalization" function in his software for this application (if it will ever be tried out).

Cheers,
Thomas

Bicurico · « **Reply #23 on:** August 23, 2016, 08:52:37 am »

TurboTom: I might be saying something stupid (I am an amateur when it comes to RF electronics - well, to electronics in general)...

Wouldn't it be easier to just use two SMA Spectrum Analyzer devices? One for Input, one for Output?

Also, I was using the BG7TBL noise source (until I fried it by accident - the PCB touched a cable connector lying around and made a short-circuit...

- a replacement has been ordered), and the results weren't that bad.

This is to say that I would not know how to implement your suggested hardware modification.

My software does lack one big feature: calibration! I need to think how to do that.

Regards

Bicurico · « **Reply #24 on:** August 30, 2016, 10:49:37 pm »

Hi,

I published a new version of my software!

Change log:

- Corrected dBm calibration for improved precision
- Implemented new function: Powermeter - not sure if it is correctly implemented
- Implemented new graphics engine
- Spectrum/Waterfall in main image can be saved as PNG
- Introduced change log
- Cleaned up GUI (Setup screen)
- Cleaned up some code
- Added a SET button on Sweep Generator (less prone to crash due to open RS232 port)

Download it from here http://www.filedropper.com/smasimplespectrumanalyzer-2016-08-30

More information on my blog.

Regards,
Vitor


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Presenting my new software for the SMA Spectrum Analyzer 138MHz-4.4GHz (Read 62840 times)

Share me