Feedback for version 0.9a here

Information and discussions on Spectrum Analysis
Kees_01
Posts: 24
Joined: Fri Jun 15, 2018 7:05 am

Re: Feedback for version 0.9a here

Post by Kees_01 » Mon Jun 25, 2018 6:50 am

Hi Steve,
Kees_01 wrote:FYI. I have checked the UNO_HEX file for the AD9851. I read in the Trackgen-box "Trackgen V1.1-Uno-AD9851" and I measure also 30 Hz when the AD9851 DDS is switched on.
I have to correct what I wrote.

With your HEX file "Trackgen_Uno_AD9851.hex" I had only checked if I got "Trackgen V1.1-Uno-AD9851" and had not measured the default frequency (1 MHz). At that moment I did not notice yet that the default frequency was changed. But the 30 Hz I had measured was with the previous HEX file DDS_tracking_Uno.hex. For all clarity: I have measured 1 MHz as well with Trackgen_Uno_AD9851.hex. Sorry for this misunderstanding.

Reason: No reason

K9AQ
Posts: 37
Joined: Sat Jan 14, 2017 1:20 am

Re: Feedback for version 0.9a here

Post by K9AQ » Mon Jun 25, 2018 8:04 pm

Steve,

I made the changes to the DDS_tracking.h file and complied the sketch for an Arduino Mega and a AD9850

This is the confirmation from the serial monitor:
Starting setup
Pins set for Mega2560
DDS is AD9850
Setup complete

I am sorry to report that there is still no output from the DDS.

I again uploaded the other Arduino AD9850 test sketch, which is using the exact same pin settings, and I get an output.

I have to believe there is still something in the code that doesn't allow the AD9850 to work. When I get back from my camping trip there should be an AD9851 waiting for me and you should have received an AD9850 for testing.

I will be checking the forum once and awhile when I have Internet access.

Don
K9AQ

Reason: No reason

SteveAndrew
Posts: 213
Joined: Thu Nov 16, 2017 11:06 pm
Location: Gold Coast, QLD, Australia

Re: Feedback for version 0.9a here

Post by SteveAndrew » Tue Jun 26, 2018 12:46 am

K9AQ wrote:I made the changes to the DDS_tracking.h file and complied the sketch for an Arduino Mega and a AD9850
This is getting really weird. I'll wait for my AD9850 to arrive and then I should have the same setup as you have. We'll see where it goes from there. Out of interest, you may have a different variant of the DDS board that I'm using. It might just be worth checking that pins D0 and D1 are tied to +5V via a 10K resistor, this is to enable the IC in serial input mode. They probably are, as it works with your test code you put up earlier. In the meantime, if I do come up with anything, I'll post it here, you might get a chance to read it before you get back from your trip.

Reason: No reason

Kees_01
Posts: 24
Joined: Fri Jun 15, 2018 7:05 am

Re: Feedback for version 0.9a here

Post by Kees_01 » Tue Jun 26, 2018 7:25 am

Hi Steve,

A very basic question. I have a 4 MHz sinewave signal. I measure the power with SDR UNO using the RSP2, from an AD8307 (dB/mV) and a scope (read as Vtt . dBm is calculated). The results are -20,6 dBm, -20 dBm and -21 dBm respectively.
Then I quit SDR UNO and launch the SAS s.a. program. The 4 MHz signal is jumping in level very quickly with about 10 dB. The peak of the signal is read about -60 dBm. Then I switch off AGC and set the IF GAIN reduction manually to 59 dB (max). And the LNA state from 3 to 8. This reduces enormous the jump of the signal. Now the peak of the signal shows -40 to -35 dBm as a max.
I have read your releasnote to see if you have said something special for this issue. Only the right LNA state might be important. But doe snot succeed.
How can I adjust the settings in general to reach about -20 dBm as measured with the two other methodes? Or is the dBm trim setting a possibility?

Reason: No reason

SteveAndrew
Posts: 213
Joined: Thu Nov 16, 2017 11:06 pm
Location: Gold Coast, QLD, Australia

Re: Feedback for version 0.9a here

Post by SteveAndrew » Tue Jun 26, 2018 10:01 am

Kees_01 wrote:Then I quit SDR UNO and launch the SAS s.a. program. The 4 MHz signal is jumping in level very quickly with about 10 dB. The peak of the signal is read about -60 dBm. Then I switch off AGC and set the IF GAIN reduction manually to 59 dB (max). And the LNA state from 3 to 8. This reduces enormous the jump of the signal. Now the peak of the signal shows -40 to -35 dBm as a max.
Currently, the analyser runs in zero-IF mode. I think your signal is probably sitting on the DC spike. This will reduce the display level, and will make the displayed signal level very unstable. Try setting the DDS to 4.01MHZ, or setting the centre frequency of the analyser to 4.01MHz. Either setting will move your signal off the DC spike.

I'm running the analyser with a sweep width of 2MHz, centre frequency of 4MHz, and the DDS at 4.01MHZ. I'm seeing the DDS signal at -22dBm. Setting the DDS frequency to 4MHz gives me an unstable level of around -35dBM. You will see a similar effect with SDRuno if you run it in Zero-IF mode. I've just run some tests, and the problem only starts when the signal gets to within around 500Hz of the DC spike.The next release will fix this problem. I will be running all sweeps at a non-zero IF. Let me know if this helps.

By the way, you may see the ADC Overload message appear in the Input panel. Reducing the signal input level, or increasing the gain reduction does not always clear this message. It has now been fixed.

Reason: No reason

Kees_01
Posts: 24
Joined: Fri Jun 15, 2018 7:05 am

Re: Feedback for version 0.9a here

Post by Kees_01 » Tue Jun 26, 2018 8:29 pm

Hi Steve,

Thank you for your measurement. Yes, it helps when moving the DDS a bit from the DC spike. And then the carrier level is stable and ari=ound -22 dBm. And Zero-IF mode with SDR UNO gives also an unstable signal presentation.

Reason: No reason

nixiefreqq
Posts: 2
Joined: Thu Jul 05, 2018 2:30 pm

Re: Feedback for version 0.9a here

Post by nixiefreqq » Thu Jul 05, 2018 2:55 pm

attached files are a return loss sweep of a 10 element 1296 Mhz yagi antenna.

return loss bridge used was an amtronix sw 1200n. didn't want to mess with a tracking generator........so signal generator was set for Very slow sweeps to achieve a smooth peak hold trace. (8mhz sweep, 50ms delay per step, 1000 steps)

second picture is with the DUT removed to establish a reference line.

first picture is with the antenna attached to the DUT port of the RLB.

my wish list #1 would be a way to save and overlay the peak hold trace so that both the reference line and antenna plot could be displayed on one screen.

thanks

ps fantastic job on this software! first release is very very useful......and it will only get better!
Attachments
sweep 1296a.jpg
sweep 1296a.jpg (50.41 KiB) Viewed 24752 times
sweep 1296.jpg
sweep 1296.jpg (53.68 KiB) Viewed 24752 times
Last edited by nixiefreqq on Thu Jan 01, 1970 12:00 am, edited 0 times in total.
Reason: No reason

alantlk
Posts: 11
Joined: Tue Oct 31, 2017 1:07 pm

Re: Feedback for version 0.9a here

Post by alantlk » Sat Jul 07, 2018 11:20 am

I noted the 12dB difference in 2nd harmonic strength between the SAS and a calibrated SA on the forum, plus other comments in this thread. That prompted me to do some very simple testing. 5W of CW into a dummy load at 14.150MHz with the same SDR pickup arrangements (bit of wire fixed near the dummy load!) between various tests. Peak signal readings using the RSP2Pro:

1 - SDRUno -86.6dBm
2 - SAS -120dBm

Two other SDR software and hardware combinations had wildly different readings as well, -35dBm, and -68dBm. While some variation is to be expected with such a simple test setup I was surprised at the 85dBm range of values.

I queried my results with SDRPlay Support who replied to say there are calibration tables within the API that mean the SDRUno reported signal level is within 0.5dB of actual and that they are available to 3rd party software. While I note replies about zero IF, so I'll try again when I've got some more time, does SAS use those API calibration tables please?
Last edited by alantlk on Thu Jan 01, 1970 12:00 am, edited 0 times in total.
Reason: No reason

SteveAndrew
Posts: 213
Joined: Thu Nov 16, 2017 11:06 pm
Location: Gold Coast, QLD, Australia

Re: Feedback for version 0.9a here

Post by SteveAndrew » Sun Jul 08, 2018 12:47 am

nixiefreqq wrote:my wish list #1 would be a way to save and overlay the peak hold trace so that both the reference line and antenna plot could be displayed on one screen.
Several changes and additions are being made to the display system, including the ability to save one or more traces to a memory bank. These changes will be available in the next release, and will allow the user to recall selected traces for comparison and measurement purposes.

Reason: No reason

SteveAndrew
Posts: 213
Joined: Thu Nov 16, 2017 11:06 pm
Location: Gold Coast, QLD, Australia

Re: Feedback for version 0.9a here

Post by SteveAndrew » Sun Jul 08, 2018 1:35 am

alantlk wrote:I noted the 12dB difference in 2nd harmonic strength between the SAS and a calibrated SA on the forum
I cannot find any post comparing the levels between the SAS and a calibrated SA. Could you post a link to the post your are referring to ?
alantlk wrote:Two other SDR software and hardware combinations had wildly different readings as well, -35dBm, and -68dBm. While some variation is to be expected with such a simple test setup I was surprised at the 85dBm range of values.
Could you provide some details, such as which software you are using. Some software, such as SDR# does not display signals in actual dBm, and the reported signal level will vary with SDR gain settings.
alantlk wrote:I queried my results with SDRPlay Support who replied to say there are calibration tables within the API that mean the SDRUno reported signal level is within 0.5dB of actual and that they are available to 3rd party software. While I note replies about zero IF, so I'll try again when I've got some more time, does SAS use those API calibration tables please?
I've compared levels between SDRuno, a signal generator and the SAS. The difference in levels are usually within 1-2 dB. There is some slight loss with SAS due to (currently) uncompensated windowing loss. The values you are reporting seem to be wildly out compared to what I've seen so far. Can you provide some more detail about the tests you have made ?

The total system gain, and gain reduction figures are obtained from the API which uses the gain tables to calculate the returned values.

Reason: No reason

Post Reply