Page 3 of 11

Re: Feedback for version 1.0a here

Posted: Sat Mar 23, 2019 8:32 am
by hb9ari
I'm not able to go on 1000 MHz and over. I'm using a RSP1 and was able to go over 1000 MHz with the older version.
What i'm doing wrong ?
Thank you for your help.
73 QRO, Rudi, HB9ARI

PS Seems to be related to an out of range index

Re: Feedback for version 1.0a here

Posted: Sat Mar 23, 2019 12:42 pm
by K9AQ
Thank you Steve!

I build and experiment with high power solid state amplifiers and I have been looking for an affordable spectrum analyzer for many years. I even used a surplus HP analyzer that must have cost many thousands of dollars when it was new, but nothing gave me what I needed. I needed good resolution at a small span so that I could see the IMD products, a large span so i could see (and measure the values) of the harmonics, and I also needed a way to evaluate the performance of the low pass filters.

I am very happy to say that your newest release of the software will do everything that I need. Thanks! I can take the Rigol spectrum analyzer off of my Christmas list.

My newest project is a dual LDMOS 2.8KW 160-6M amplifier. The timing of the software release is perfect. The amplifier is ready for testing with the spectrum analyzer. I will post screen snapshots when I complete the testing.

This morning I did a quick test of the ability to measure harmonics using an Elecraft XG3 signal generator. The placement of the marks takes some getting used to. I found the fastest way was to enable peaks and then use the "next" buttons to place the marker on the peak that I wanted. ... ew&id=1592


Re: Feedback for version 1.0a here

Posted: Sat Mar 23, 2019 3:19 pm
by ja8scd
Hi Steve, thank you very much for your releasing the new version.

While I was measuring an airband BPF, RSPduo showed bit different operation than the other two models. I am attaching three screen shots showing the results with RSP1A, RSP2 and RSPduo.

Yellow Line: Noise generator is directly connected to RSP Ant port.
Blue Line: Noise generator is connected through an airband BPF.
While Line: Ant port is terminated with a 50ohm load.

1) spurts at 24MHz x n are so conspicuous compared to those observed with RSP1A and RSP2.
Also, noise floor of RSPduo seems quite noisy. Is there any reason you can think of, or is this totally dependent on the RSPduo hardware?
2) As the default, the BC/FM and DAB are enabled although those button ares not inverted. By clicking them twice respectively, the filters will be removed.

Best regards,

Yas - JA8SCD Tokyo

Re: Feedback for version 1.0a here

Posted: Sat Mar 23, 2019 4:49 pm
by fastputt
Thank you Steve for a wonderful program and all the time and effort you put into this!

I have one problem: I use windows 7 pro and your program loaded without problem and works great to the extent I can use it. However,
the Control Panel does not load completely. Everything to the right of the word "frequency' is missing straight down the entire panel. Please tell me what I did wrong and how to fix it. I tried reloading the program twice with same result. Never had this happen before.

Thank you in advance for your help. 73, Dale K8ETI

Re: Feedback for version 1.0a here

Posted: Sat Mar 23, 2019 6:07 pm
by respero
Very cool, promises to be quite useful!

One feature that would be nice for a future release is a resolution bandwidth (RBW) annunciator.

I'm not seeing the expected "processing loss" from
none-blackmann.png (66.43 KiB) Viewed 9129 times
windowing: Noise level for Blackman (yellow) is 1 dB above None (white) , expect 2.38 dB from the factor of 1.73 bandwidth penalty for the Blackman window.

Re: Feedback for version 1.0a here

Posted: Sun Mar 24, 2019 2:05 am
by DaveR
DaveR wrote: I've been looking a bit more at the problem, and it's not as straightforward as it first looked, eg starting at 50MHz centre frequency:

At 50Mhz, max span is 20MHz (stops after 1 sweep at 50MHz).
At 51MHz, max span is 10MHz (stops after 1 sweep at 20MHz).
At 55MHz, max span is 10MHz (stops after 1 sweep at 20MHz).
At 59MHz, max span is 2MHz (stops after 1 sweep at 5MHz).
At 60MHz, max span is 1MHz (stops after half a sweep at 2MHz and above).
At 65MHz, max span is 10MHz (stops after 1 sweep at 20MHz).
At 70MHz, max span is 20MHz (stops after a quarter of a sweep at 50MHz).
At 80MHz, max span is 20MHz (stops after 5MHz at 50MHz).

These results are repeatable over and over, with both an RSP1A and an RSPduo. PC is running 64bit Win7 with 16GB and i7-4770K. I'll try it next on a Win10 laptop to see if the same thing happens.
Hi Steve - I got exactly the same behaviour on the laptop as above, with the exception that the trace usually stopped after two to four sweeps instead of one or less, so the bug doesn't appear to be machine or OS specific. I was running other programs on the Win7 PC at the same time, whereas the Win10 laptop was only running RSP-SAS by itself - this might account for the higher number of sweeps achieved before they stopped, but probably doesn't shed any light on the root cause of the problem.


Re: Feedback for version 1.0a here

Posted: Sun Mar 24, 2019 3:01 pm
by nixiefreqq
GREAT JOB! the improvements are fantastic!

only suggestion would be the addition of "sweeps per second" display. this would help when fiddling with the step rate on the signal generator.

attached are some return loss sweeps of a homebrew 222Mhz ground plane vertical on the "indoor range".

Re: Feedback for version 1.0a here

Posted: Sun Mar 24, 2019 6:00 pm
by K9AQ
I am getting more familiar with the changes in this release and it is doing everything that I hoped it would. There are some quirks but I have worked around them. I have experienced your "know problems" of intermittent display freezing and gain setting changing.

I am struggling with the functionality of the markers. This is not as easy to use as the "cursors" were in your first release. Using the enable peaks functions helps, but sometimes I would just like to quickly point to the crest of a peak and put a marker there.

Thanks again for all your hard work.


Re: Feedback for version 1.0a here

Posted: Sun Mar 24, 2019 7:01 pm
by K9AQ

I have found that it can take a lot of time for the peaks to calculate the frequency and db difference. I have seen it take over a minute. Is that normal? I running the software on an older laptop so that my be the problem.

I just used the software to see what the harmonics look like on my dual LDMOS amplifier. This screen snapshot shows the harmonics out of the PA pallet running 1.5KW, with no low pass filter. You can see that the second harmonic is down 31db and the 3rd is down 13.8. This is about what I would expect. My diplexer will easily reduce these to the FCC requirements.

Thanks again!

K9AQ ... ew&id=1609

Re: Feedback for version 1.0a here

Posted: Sun Mar 24, 2019 7:15 pm
by G0HZU
I recently bought an RSP1A and downloaded the SA software last night. Thanks for providing this!

I've had an initial play and it looks good and the best news is that it runs OK on my little EeePC netbook. However, I have a couple of wishful suggestions for a future release.

Can it have a 1Hz noise marker function? Ideally, this would be done the same way it is done in a regular lab analyser.
Can you add a Flat Top window function?

Also, the spurious free dynamic range behaviour is a bit strange with a two tone test. Can the user have more control over what is happening at the front end? I've only had a few fumblings with it so far but I seem to get good SFDR performance with the ref level set at -20dBm. But if I start and stop the analyser and then start again the SFDR collapses to just a few dB. The fix appears to be to select 0dBm ref level and then go back to -20dBm. The SFDR with the ref level set at -10dBm is much worse than with it set at -20dBm so something strange is also happening here.