RSP Spectrum analyser software available shortly

[vk7jj]

Hi again Steve, looking at your screen dumps makes me actually feel elated ... of course it's your hard work and our pleasure but it sure is pleasure!

Brief (selfish) thoughts in response

- thanks for the info on noise floor, that's excellent, very satisfying. And it's really good to be able to look below the noise floor, great incentive to minimise one's personal contribution to it.

- don't worry your head about that darned tracking generator one bit. Leave it till people exploit the current functionality with and without white noise.

- likewise the DC spike, much much better to put up with it and not have to wonder if you're staring at spurious stuff

- I wonder if the struggle for wider than 10MHz bandwidth really worth it given the obvious issues you've mentioned that will inevitably compromise things, plus I wonder how much processing power is required for 10MHz, as quite a few ham shack owners who would be your loving target are using older CPUs

- delighted that you feel happy with the project, that means a lot in terms of boding well and me feeling less guilty, it sure is nice to feel happy about the coding and structure, and you should already feel proud about the handsome GUI.

Enough!

Best hopes for more clean coding! 73, Phil VK7JJ

[someYguy]

Very good! As Phil said, no worries regarding the tracking generator. I use a broadband noise source and it works very well. Adjustable trace averaging (smoothing) perhaps with adjustable attack & decay such as SDR# uses would be most helpful when using a noise generator.

As to the DC spike and possibly using Low IF, I currently use Low IF with SDR# for spectrum plots and it works very well especially at narrow frequency spans, with little or no unwanted images. So that is probably a good option.

Thanks again Steve for all of your hard work and what looks to be a wonderful & useful application .

[someYguy]

I just thought of something else which would be helpful - direct access to the RSPx device control panel. Perhaps just a simple button to bring it up?

[G4HPW]

Hi Steve
Its great news that you are making progress and that the various suggestions made were of use and have been implemented.
I am certain your hard work will be very much appreciated by lots of people.

Just getting some spectrum analyser ability will be enough for many users with a tracking generator not being missed at all. At work we only had a few analysers with the tracking gen option fitted as most of the time it was not required.

As for the spike, if you know it's there and what it is it can easily be ignored and should not be much of a problem.

I am looking forward to the release in what ever form it takes with much interest to compare it with my Anritsu analyser as what you are achieving in terms of cost against performance looks amazing.

Regards John / G4HPW

[SteveAndrew]

- I wonder if the struggle for wider than 10MHz bandwidth really worth it given the obvious issues you've mentioned that will inevitably compromise things, plus I wonder how much processing power is required for 10MHz, as quite a few ham shack owners who would be your loving target are using older CPUs

Not so much of a problem as I thought it would be. As I'm not demodulating anything, I don't have to be concerned about maintaining a constant data stream at 10MHz. I'm just grabbing enough to feed to the Fourier function, ignoring any more streaming data until the current block has been processed and displayed. My PC starts to stutter at 5MHZ sampling with SDRuno, but a see no evidence of that (so far) with the spectrum analyser running at 10MHz. I'm in the process of trying a different approach to wide-band sweeps that looks promising. Fingers crossed.

As Phil said, no worries regarding the tracking generator. I use a broadband noise source and it works very well. Adjustable trace averaging (smoothing) perhaps with adjustable attack & decay such as SDR# uses would be most helpful when using a noise generator.

Trace filtering and averaging are already done, although it does take time for the trace to settle, but a small price to pay for a pretty clean trace.

As to the DC spike and possibly using Low IF, I currently use Low IF with SDR# for spectrum plots and it works very well especially at narrow frequency spans, with little or no unwanted images. So that is probably a good option.

Yes, I think that may the way to go. I might have a play and see how it looks.

I just thought of something else which would be helpful - direct access to the RSPx device control panel. Perhaps just a simple button to bring it up?

Unfortunately the panel can only be displayed and used if I use EXTIO.DLL to control the device. Essentially my code is doing the same thing as the DLL, so I'm able to do everything the panel does. But thanks for the thought.

I am looking forward to the release in what ever form it takes with much interest to compare it with my Anritsu analyser as what you are achieving in terms of cost against performance looks amazing.

I'd be most interested in seeing how they compare. Even a low-end analyser can be eye-wateringly expensive, which is one of the reasons I started this project in the first place.

One again, many thanks to you all for your suggestions and kind words of encouragement.

73's - Steve

[arvopl]

Very cool! I'll be able to use mine with a noise bridge for an antenna analyzer!

[NK7Z]

Is this software available now?

[someYguy]

Thanks again Steve. Looking forward to it.

[vk7jj]

I wonder if a modified RSP could be used to provide the tracking signal.

Confessing my utter ignorance of the architecture but supposing it's a single conversion superhet providing direct conversion in zero IF mode, I don't know how the IF offset is generated - whether digitally in the passband or via a second LO. If via a second LO then could the other secondary unwanted mix product be filtered and buffered and made available at an external connector in some future iteration of the hardware and thus provide an RSP that could automatically track it's own output as it was tuned?

[JayDubya]

A few things come to mind...

1) I have a Signal Hound SA and tracking generator. Broad band (>10MHz) is achieved by simply re-tuning the SA to the next band segment and adding that data to the accumulated data. In other words the Signal Hound can sweep a huge spectrum but takes a few samples evenly spaced for each frequency hop. It's a little slow for huge swaths but is quite functional to show a general idea of how a bandwidth is being used. So with a code tweak, >10 MHz should be a breeze even for the slowest of computers. My Signal Hound and tracking generator run perfectly fine using an old Dell Latitude laptop single core CPU. Although I must admit it works faster on a little more current device. (Yes I know how much work goes into a 'tweak'.)

2) This sounds like a good opportunity for the RSP folks to modify a receiver into a tracking generator.

3) Seems like the LO DC spike could be calibrated out of the equation by having a 'no input base line' cal button that would effectively disconnect any input and then mathematically remove the spike from subsequent input connected sweeps.

I know how much work this might involve but it's worth a shot.....

THANK YOU SO MUCH FOR MAKING THE RSP EVEN BETTER!!!

John KB0NE