Author Topic: "NOSE SOURCE BG7TBL 2016-03-06"  (Read 13257 times)

0 Members and 1 Guest are viewing this topic.

Offline technogeeky

  • Frequent Contributor
  • **
  • Posts: 555
  • Country: us
  • Older New "New Player" Player Playa'
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #25 on: October 11, 2017, 05:08:52 am »
I love the screenshots.  How did you get the animation?

Thank you! I have spent some time getting good at making them, because I think they are a great way to share stuff here.

It's pretty simple, really:

  • I use Rigol Bildschirmkopie to capture screenshots at "as fast as possible" or every 1 second[1]
  • I import the screenshots as layers into a GIMP file, then Optimize for GIF, and save as GIF with "replace" turned on

[1] It's important to get the smooth fast images to use USB and not LAN. LAN can only capture about 1 picture every 3 seconds, whereas USB can capture 3 to 4 pictures a second.[2]

[2] If you use the fastest capture rate, some display elements (like the measurements at the bottom) will be slow to update and thus will be incorrect in the screenshots. Even worse, some display elements (like the FFT itself!) will never actually get computed because the screenshot requests interrupt them. In these cases, I have to slow down the acquisition to once per second.

If you have any more questions, feel free to ask here or in PM.
« Last Edit: October 11, 2017, 05:15:38 am by technogeeky »
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #26 on: October 11, 2017, 01:55:05 pm »
@technogeeky

VERY nice work!  Many thanks.  Based on your work, putting it in a proper metal enclosure with a high pass filter on the output and an LM317 trimmed to produce the best spectrum  will turn it into a useful tool.

 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #27 on: October 30, 2017, 12:44:04 am »
I've been playing around.  I built a -40dB attenuator so I could feed it to my SDRplay RSP2.  With 12.1V (as close as my old PS could be adjusted.  To be fixed) it looks pretty good overall.  There is an issue with spurs in the  RSP2 which is quite predictable from the scope trace.  I thought I'd show those.

First image is just the spur using 4096 point averaging. The noise generator is off.  Second is with the noise generator on and shows the low frequency noise convolved with the spur in the mixer.

It looks far worse than it is.  This is a zoom.  In practice one would shift the center frequency of the RSP2 to move the spur and associated artifacts away from the region of interest.
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #28 on: October 30, 2017, 01:47:12 pm »
I neglected to mention that the spur is  LO bleed thru.  I posted it because it was exactly what looking at the output on a scope predicts., lots of crud for 2 MHz around any significant mixer products.

The really interesting part is comparing spectra from different SDR programs.  They have been very different in some instances.  Not sure if software or user error yet. 

The -40dB attenuator I put together in a 2" x 3/4" x 1" Pomona box appears to have some flatness issues above 1 GHz.  Unfortunately, tough to analyze with 10 MHz chunks of RF.  I suspect it's due to standing waves in the box.  Anyone know of a readily available RF absorbing material?

With a hi pass filter to kill the crud below 1 MHz and a microstrip switched  attenuator it should prove very useful for examining SDR performance.
 

Offline technogeeky

  • Frequent Contributor
  • **
  • Posts: 555
  • Country: us
  • Older New "New Player" Player Playa'
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #29 on: October 31, 2017, 09:32:52 am »
I neglected to mention that the spur is  LO bleed thru.  I posted it because it was exactly what looking at the output on a scope predicts., lots of crud for 2 MHz around any significant mixer products.

The really interesting part is comparing spectra from different SDR programs.  They have been very different in some instances.  Not sure if software or user error yet. 

The -40dB attenuator I put together in a 2" x 3/4" x 1" Pomona box appears to have some flatness issues above 1 GHz.  Unfortunately, tough to analyze with 10 MHz chunks of RF.  I suspect it's due to standing waves in the box.  Anyone know of a readily available RF absorbing material?

With a hi pass filter to kill the crud below 1 MHz and a microstrip switched  attenuator it should prove very useful for examining SDR performance.

I just want to point out that you originally mentioned that you were going to put the noise source in a metal box. Just make sure that isn't the source of whatever problems you are having.
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #30 on: October 31, 2017, 11:18:29 pm »
I'll get to putting it all in a box.  But it doesn't change the point.  Without a high pass filter the low frequency crud gets replicated either side of the LO signal once it hits a mixer.

The really interesting part is how different the spectrum looks with the same radio and setup and 3 different SDR packages.

 

Offline medical-nerd

  • Regular Contributor
  • *
  • Posts: 198
  • Country: gb
  • What's that coming over the hill?
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #31 on: November 01, 2017, 12:04:36 am »
Hiya

I've only used these to look at a few filters in the 300Mhz-1500Mhz range and bought one after reading:

https://www.rtl-sdr.com/rtl-sdr-tutorial-measuring-filter-characteristics-and-antenna-vswr-with-an-rtl-sdr-and-noise-source/

I've found this and a cheap rtl dongle sufficient for my needs thus far.

They have also been used as a noise source in the input pathway of amateur radio telescope setups when cooled.

I agree that for $20 and a cheap dongle it gives hobbyists an instrument that would be beyond their normal reach.

Cheers 
'better to burn out than fade away'
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #32 on: November 01, 2017, 01:11:15 am »
Take a look at this:

https://www.rf-microwave.com/datasheets/6107_generic_NSG-10B_01.pdf

in particular, note the comments about the size of the enclosure and the need for RF absorbing material for larger enclosures.
 

Offline medical-nerd

  • Regular Contributor
  • *
  • Posts: 198
  • Country: gb
  • What's that coming over the hill?
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #33 on: November 01, 2017, 01:38:11 am »
Hiya

a suggestion for microwave absorber material:

https://www.rf-microwave.com/en/shop/0/377-microwave-absorbing-materials.html

Cheers
'better to burn out than fade away'
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #34 on: November 01, 2017, 08:48:11 am »
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #35 on: November 01, 2017, 09:13:21 pm »
I built a 1 MHz 3 element capacitor input LC filter in  another Pomona box using thruhole parts.  It's OK for VHF but will obviously need to be redone for UHF work using SMD and microstrip construction.

First image is a scope shot.  Top trace is filtered, bottom trace is the noise source output.  Somewhat cleaned up, but still needs improvement.

2nd image is from the RSP2 using SDRuno.  It shows the images of the low frequency components visible on the lower scope trace.

3rd image is with the filter inserted which suppresses the images by about 6 dB.

I'll try a 2nd iteration using a 5 or 7 element filter, but I doubt that I can achieve good isolation between the input and output using the parts I have on hand.  But worth a try.

Hopefully this explains my initial criticism.  It really needs an output high pass filter.   That should not cost more than a few cents using SMD parts.
 

Offline technogeeky

  • Frequent Contributor
  • **
  • Posts: 555
  • Country: us
  • Older New "New Player" Player Playa'
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #36 on: November 01, 2017, 11:05:29 pm »
I built a 1 MHz 3 element capacitor input LC filter in  another Pomona box using thruhole parts.  It's OK for VHF but will obviously need to be redone for UHF work using SMD and microstrip construction.

First image is a scope shot.  Top trace is filtered, bottom trace is the noise source output.  Somewhat cleaned up, but still needs improvement.

2nd image is from the RSP2 using SDRuno.  It shows the images of the low frequency components visible on the lower scope trace.

3rd image is with the filter inserted which suppresses the images by about 6 dB.

I'll try a 2nd iteration using a 5 or 7 element filter, but I doubt that I can achieve good isolation between the input and output using the parts I have on hand.  But worth a try.

Hopefully this explains my initial criticism.  It really needs an output high pass filter.   That should not cost more than a few cents using SMD parts.

Ah, ok. That also explains why I wasn't thinking of the same sort of problems. Because the devices I was thinking of using have excellent input filtering (like my spectrum analyzer) which can easily overcome this problem if you just avoid the band in question. If you have a SDR, the LNA's performance will be subject to whatever signal is the most powerful. So you can't get around the spikes. Right?

I wonder if there is space on the board to implement this stuff. Also, I wonder -- if there isn't space yet -- if you could remove the last amplifier and replace it with passive filtering.

This could be an opportunity to re-make this board/project so that it's compatible with SDRs in terms of performance without prefiltering.
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #37 on: November 02, 2017, 02:23:41 am »
I'm hoping that "BG7TBL" will add an output filter.  It really needs an attenuator also.  Bota goes into considerable detail on that subject in his description of the noise generators he built for the 18th EME conference in Florence.  Mine is just a hack.  To be redone later in proper fashion.

For my unit, I'll just add another board (or two) with a filter and an LM317 with some large capacitors so that it will run properly using a generic wall wart.  I scavenged some zinc coated steel from an HVAC shop and will bend up a box to fit once I know how big it is.  I was looking into SMD inductors earlier and it looks as if it will be $3-5 to make a filter.  So a bit more expensive than I thought.

Lacking a spectrum analyzer I had to rely on my knowledge of the Fourier transform and the scope.  Your traces gave me reason to keep going which is greatly appreciated.  All the cheap SDRS have rather marginal input bandpass filtering.   High order analog filters are hard to build because of part tolerances.  Problem is, you need 6dB for each bit of ADC resolution.  That's 72 dB for the RSP2.  This is why oversampling is so popular.  You can use a 2 pole analog filter followed by a digital filter to do the rest of the antialiasing.

My next problem is QC on an Instek MSO-2204EA :-) 

I bought a great load of through hole parts a couple of years back, mostly resistors and capacitors so I had a complete set of values.   I was able to build an attenuator and filter without being beat to death which was a real treat. Probably the first time I've been able to build something in a few minutes. Usually I get to play "hunt the wumpus".  Not much, 3 parts each, but I was able to calculate the values and then pick them out of the bins.  Very sweet.

Now I just need to figure out what the various spectrum displays from different SDR software packages are actually showing me.

 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #38 on: November 03, 2017, 06:24:27 pm »
To wrap this up.  I rebuilt the filter as a 5 element.  Except for some large spurs produced by the RSP2, the spectrum now looks clean.  Level changes a bit, especially in the HF region.  But pretty satisfactory. The filter was designed using Elsie and a couple of slug tuned coils.

Shots are:

no_filter_5.png - no filter. included to keep all settings the same
filter_5_int.png - just the RSP2 internal MW/FM filter
filter_5.png - external 5 element 1 MHz high pass filter only
filter_5_int.png - external 5 element 1 MHz high pass filter with internal MW/FM filter
 

Offline Nauris

  • Regular Contributor
  • *
  • Posts: 188
  • Country: fi
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #39 on: November 03, 2017, 08:17:48 pm »
The flyback converter that´s included on the PCB to generate the ~36V to bias the 24V "nose" zener (IIRC) is crap. I tried everything (different inductors, capacitors, feedback networks...) to keep the output voltage from oscillating without success. To me it seems the STM flyback controller (MC34063A) just isn´t able to provide a sufficiently clean output voltage at the low current load (loop oscilation). Otherwise, the noise performance of the little gadget doesn´t seem too bad.

Cheers,
Tom
MC34063 is on/off hysteretic design it is never going to give clean output. It is just not intended for anything needing clean power as no matter load, it is always running in pulse skipping mode.
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: "NOSE SOURCE BG7TBL 2016-03-06"
« Reply #40 on: April 28, 2020, 01:54:32 am »
For anyone who followed this, here is a comparison to a recent ebay diode noise source of rather better performance.  At least in what I expect from a broadband noise source.

First the BG7TBL unit:

https://www.eevblog.com/forum/rf-microwave/another-ebay-noise-source/msg3038324/#msg3038324

and then the new unit:

https://www.eevblog.com/forum/rf-microwave/another-ebay-noise-source/msg3036288/#msg3036288

except for the reference level, everything should be the same on both sets of sweeps.

Reg
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf