Topics

One question only...


Alan de G1FXB
 

Tom, Thanks for the SDR link.

I really wish I had not re/discovered the RP project again.
https://www.redpitaya.com/
It's to RF processing what the RPi is to digital,  a solution awaiting the problem.
It even has the same initials....

Alan

On 10/08/2018 10:39, Tom, wb6b wrote:
I was reading and found this data sheet for the RTL-SDR dongle. https://www.rtl-sdr.com/wp-content/uploads/2018/02/RTL-SDR-Blog-V3-Datasheet.pdf
The information explains in the direct conversion mode there can be images around the 14.4 Mhz point because the sample rate in the direct mode is 28.8 Mhz. Maybe that was the image issue that was sticking around in the back of my mind about the direct mode.

The Red Pitaya info is really interesting. In a distant past life I designed with Xilinx programable Logic Cell Arrays and designed gate array ASICs. It would be nice to have a development board on my desk to play with. Been messing some with the AWS FPGA cloud instances. But, something affordable that I can hook my own wires onto is great.

Tom, wb6b


Tom, wb6b
 

I was reading and found this data sheet for the RTL-SDR dongle. https://www.rtl-sdr.com/wp-content/uploads/2018/02/RTL-SDR-Blog-V3-Datasheet.pdf
The information explains in the direct conversion mode there can be images around the 14.4 Mhz point because the sample rate in the direct mode is 28.8 Mhz. Maybe that was the image issue that was sticking around in the back of my mind about the direct mode.

The Red Pitaya info is really interesting. In a distant past life I designed with Xilinx programable Logic Cell Arrays and designed gate array ASICs. It would be nice to have a development board on my desk to play with. Been messing some with the AWS FPGA cloud instances. But, something affordable that I can hook my own wires onto is great.

Tom, wb6b


Alan de G1FXB
 

Thanks Tom,

I didn't have a problem reading any of your text,
more reading my text back and although I may have thought it, it didn't make it to the written stage.....
  On the subject of direct IQ and no opposite rejection:
"As a virtual SA we are not interested resolving either USB / LSB, (bothsides, if present) of a carrier"
As a virtual SA we are not interested resolving either USB / LSB, (bothsides, if present) we are not interested in demodulating the carrier, merely displaying it's presence and amplitude.



Gordon, and thank you from me, as well for the HF Explorer link.
That in turn reminded me to revisit Red Pitaya Dev Boards after +2 years, it was a greater cost for the sake of only a CW Skimmer project at the time.
However others have taken on brewing  so much around that project, one of it's original apps was a basic SA card......
So that's a fourth solution.


Alan



On 10/08/2018 04:31, Tom, wb6b wrote:
Hi Alan, welcome,

Yes, if you google around you will find instructions on doing the direct sampling mod on dongles that don't do direct sampling, and where to solder the antenna wire. The RTL-SDR dongle from here https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/ has the mod built in, with the advantage that they added an RF switch so the your SDR radio software can switch modes. 

I have both the RTL-SDR dongle and a NooElec dongle http://www.nooelec.com/store/sdr/nesdr-smartee-sdr.html that I use with the Ham-It-Up converter. I like both of these dongles. I'm amazed they work as well as they do. Fortunately, I do not have many strong broadcast stations and such nearby; that may help.

Also, in my last post, I switched from referring to the demodulator chit to calling it the decoder chip. I hope that did not cause any undue confusion.

Hi Gordon, the RF Explorer link you posted about looks interesting. I'll look at it further.

Tom, wb6b



Tom, wb6b
 

Hi Alan,

Just noticed your point about not needing to resolve which side of the carrier you are on, in the case of using a dongle as a SA. That is a good point. Even if you could see a difference in the energy within the bandwidth the scan is sweeping across the spectrum scanned, it does seem it would not matter and could be compensated for if it did. 

Tom, wb6b


Tom, wb6b
 

Hi Alan, welcome,

Yes, if you google around you will find instructions on doing the direct sampling mod on dongles that don't do direct sampling, and where to solder the antenna wire. The RTL-SDR dongle from here https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/ has the mod built in, with the advantage that they added an RF switch so the your SDR radio software can switch modes. 

I have both the RTL-SDR dongle and a NooElec dongle http://www.nooelec.com/store/sdr/nesdr-smartee-sdr.html that I use with the Ham-It-Up converter. I like both of these dongles. I'm amazed they work as well as they do. Fortunately, I do not have many strong broadcast stations and such nearby; that may help.

Also, in my last post, I switched from referring to the demodulator chit to calling it the decoder chip. I hope that did not cause any undue confusion.

Hi Gordon, the RF Explorer link you posted about looks interesting. I'll look at it further.

Tom, wb6b


Gordon Gibby <ggibby@...>
 

I just found a spectrum analyzer  out there for $165




On Aug 9, 2018, at 19:10, Alan de G1FXB via Groups.Io <g1fxb@...> wrote:

Thanks Jerry / Alison.

See below:-


On 09/08/2018 22:53, Jerry Gaffke via Groups.Io wrote:
Go for it, replicate the specan.
It will be very useful.

But that's a much more ambitious project than Arv's design.

That's scary because I thought the Specan looks more accessible, as it maybe possible to get some commercially built building blocks (Modules) for the important bits from ex-equipment.
The most critical looks to be the bandpass, Farhan used 112MHz and old TEK Analysers seem to use a 110MHz IF BPF in theirs. (But I would need to see how B/W compares..... )
And massage the sketch to suit.

So there is a choice of 3 "affordable" relative level Analysers
1/.    RTL SDR & HF convertor in the front end
2/.    Arv's Design
3/.    Farhan's Design

Maybe "Signal snooping" isn't so hard.
Thinking............

Alan

Simple is good if it's good enough for the task.

Jerry


On Thu, Aug 9, 2018 at 02:14 PM, Alan de G1FXB wrote:

Not meaning to be inflammatory to the author,

 but at the same time if there if there is no need to reinvent the wheel if it will give reproducible data comparable to other solutions:-

http://hfsignals.blogspot.com/p/specan-reboot-of-w7zoi.html

 



Alan de G1FXB
 

    Thanks Tom,
That sound like what I had poorly remembered, I thought there was some soldering somewhere.......
Antenna straight into either I or Q input but no opposite sideband rejection.
As a virtual SA we are not interested resolving either USB / LSB, (bothsides, if present) of a carrier

Alan

On 10/08/2018 00:14, Tom, wb6b wrote:
There is a "Direct Conversion" mod. The dongles have two chips, one is the tuner and the other the demodulator that was used to demodulation TV signals. Hackers reversed engineered the thing and discovered they could get the demodulator chip to pass the I/Q signals out the USB port.

The original operation was that the tuner mixed down to DC for the demodulator chip to do its decoding.

However, the decoder chip maker wanted to hedge their bets and make other possible configurations possible. So, they build another tuner into the decoder chip that could work up to about 30mhz. The hackers discovered that and by bypassing the original tuner chip and hooking the antenna to ether the decoder's I or Q input, the decoder chip will tune from around 100Khz to 30Mhz. However, in the direct conversion mode, the SDR software no longer has both I and Q so it can't determine the direction of the rotation of the signal vs the LO so images could be a bigger problem.  

Tom, wb6b


Tom, wb6b
 

There is a "Direct Conversion" mod. The dongles have two chips, one is the tuner and the other the demodulator that was used to demodulation TV signals. Hackers reversed engineered the thing and discovered they could get the demodulator chip to pass the I/Q signals out the USB port.

The original operation was that the tuner mixed down to DC for the demodulator chip to do its decoding.

However, the decoder chip maker wanted to hedge their bets and make other possible configurations possible. So, they build another tuner into the decoder chip that could work up to about 30mhz. The hackers discovered that and by bypassing the original tuner chip and hooking the antenna to ether the decoder's I or Q input, the decoder chip will tune from around 100Khz to 30Mhz. However, in the direct conversion mode, the SDR software no longer has both I and Q so it can't determine the direction of the rotation of the signal vs the LO so images could be a bigger problem.  

Tom, wb6b


Arv Evans
 

Jerry

In all fairness to Farhan VU2ESE and Wes W7ZOI, their spectrum analyzers parentage is
from even earlier ARRL Labs and other individual spectrum scanner designs.  Their
units work very well and are much closer to the push-button design of commercial
analyzers.  The little sweepy thing that I did is bare-bones just enough to be functional
and in keeping with the BITX philosophy it allows modification to make it do whatever
specific things you might desire.

This is a great time to be a ham radio operator/technician because component
prices are reasonable, components are readily available, and there are more good
ideas floating around than any normal ham can keep up with.  

Arv
_._


On Thu, Aug 9, 2018 at 3:53 PM Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io> wrote:
Go for it, replicate the specan.
It will be very useful.
But that's a much more ambitious project than Arv's design.
Simple is good if it's good enough for the task.

Jerry


On Thu, Aug 9, 2018 at 02:14 PM, Alan de G1FXB wrote:

Not meaning to be inflammatory to the author,

 but at the same time if there if there is no need to reinvent the wheel if it will give reproducible data comparable to other solutions:-

http://hfsignals.blogspot.com/p/specan-reboot-of-w7zoi.html

 


Alan de G1FXB
 

Thankyou Alison,
 "finally got it".

R820 & one of the HF mixer PCB's

Alan

On 09/08/2018 23:22, ajparent1/KB1GMX wrote:
Alan,

Most of the DVB-t sticks have a lower limit of 25-50 mhz.  To use them for HF a Upconverter
translates near 0 to 30mhz to higher frequency mine take that and goes from 120MHZ (0hz)
to 150mhz (30mhz).  Its a simple thing and works well.

Allison



Alan de G1FXB
 

Thanks Jerry / Alison.

See below:-


On 09/08/2018 22:53, Jerry Gaffke via Groups.Io wrote:
Go for it, replicate the specan.
It will be very useful.

But that's a much more ambitious project than Arv's design.

That's scary because I thought the Specan looks more accessible, as it maybe possible to get some commercially built building blocks (Modules) for the important bits from ex-equipment.
The most critical looks to be the bandpass, Farhan used 112MHz and old TEK Analysers seem to use a 110MHz IF BPF in theirs. (But I would need to see how B/W compares..... )
And massage the sketch to suit.

So there is a choice of 3 "affordable" relative level Analysers
1/.    RTL SDR & HF convertor in the front end
2/.    Arv's Design
3/.    Farhan's Design

Maybe "Signal snooping" isn't so hard.
Thinking............

Alan

Simple is good if it's good enough for the task.

Jerry


On Thu, Aug 9, 2018 at 02:14 PM, Alan de G1FXB wrote:

Not meaning to be inflammatory to the author,

 but at the same time if there if there is no need to reinvent the wheel if it will give reproducible data comparable to other solutions:-

http://hfsignals.blogspot.com/p/specan-reboot-of-w7zoi.html

 



Gordon Gibby <ggibby@...>
 

Hey, if somebody will build that I’ll buy it!


On Aug 9, 2018, at 14:42, Jerry Gaffke via Groups.Io <jgaffke@...> wrote:

I'm liking the looks of Arv's minimalist spectrum analyzer more and more.
Simple, sufficient for evaluating all the uBitx spurs (including any harmonics).
Perhaps somebody could kit it up such that it mates to a $3 Nano,
reports to a PC via the serial monitor.  Or to an LCD display.
Might take as little as $10 in raw parts cost?

The series tuned trap at the input could be left out.
If I'm measuring the 2'd harmonic and happen to also get some of the 4'th and/or 6'th, 
that's just going to add a dB or so to the measurement.   We're worried about
the sum of all those harmonics anyway.

Calibrate it first. 
The uBitx transmitting in CW mode feeding a 50 ohm dummy load is a good signal source.
Measure across the dummy load using a diode RF probe, set RV1 for about 1W using the RF probe before each measurement.
At one Watt of RF, that's sqrt(1Watt*50ohms) = 7.07 Volts RMS, or 10 volts peak, or 20 volts peak-to-peak into the RF probe.
Shield the rig and dummy load well, and maybe the first 20dB of attenuation after the dummy load too.
Set Arv's step attenuator to give the same reading out of his 2n3904 at each frequency you are calibrating for,
and record that attenuator setting. 

When looking at some dirt, set the step attenuator to give the same reading from the 2n3904 as you calibrated for.
The difference between that attenuator setting and the attenuator setting you recorded during calibration
is how many dB down from 1 Watt the offending dirt is.

Could watch the audio AC output of that 2n3904 using a diode rectifier feeding your Harbor Freight voltmeter.
Slowly sweep the frequency of the si5351 around the vicinity of the spur for maximum reading.

If everything proves linear enough, could slowly sweep the si5351 from 3 to 30mhz
and read the diode rectifier using a Nano ADC pin.
The Nano's program report any spurs found to a host PC through the Nano's serial monitor.  
The Nano's 10bit ADC has a maximum range of around 60dB, should easily cover a 40dB range accurately.
More than sufficient for evaluating our spurs, assuming the step attenuator gets set up correctly
before kicking off the sweep.
 
Jerry, KE7ER



On Sun, Aug 5, 2018 at 12:54 PM, Arv Evans wrote:
I am not close to my hamshack to check for accuracy. 
The attached drawing shows the general idea from what I can remember.
Sweepy.png
Looking at this reminded me that with very little change it could be made
into a CW or swept frequency signal generator (Si5351a out through mixer
and attenuator, with audio into mixer to modulate carrier).  Every time I look
at this it seem to suggest yet another modification. 

 


ajparent1/KB1GMX <kb1gmx@...>
 

Alan,

Most of the DVB-t sticks have a lower limit of 25-50 mhz.  To use them for HF a Upconverter
translates near 0 to 30mhz to higher frequency mine take that and goes from 120MHZ (0hz)
to 150mhz (30mhz).  Its a simple thing and works well.

Allison


Alan de G1FXB
 

H'Mmm

I have one of the cheapest of the cheap R820's sticks.

What has always cautioned me against using them as any sort of analyser substitute is the minimum conventional tuning limit is 24(?)MHz on mine.

(There is a possible hardware mod to pull the IQ signals direct for HF only? or some software Apps force a state to the front end (in the absence of the correct term) and the strong HF signals find a path through whatever input filtering that advertise to have....)


So in actual fact my concerns are unfounded, and sweep of 80 -> 10M (and beyond) looking for 5 order harmonics etc give valid levels.

Or do I need to apply any extra steps either above / below the 24MHz threshold??


Alan


On 09/08/2018 22:37, Tom, wb6b wrote:
I like the idea of using the rt-sdr dongles as a lower cost way for the larger population of Hams to do their own measurements on their equipment. That would allow me to do my own measurements without spending more money than I can justify at the moment on a commercial SA.

I have two dongles, one can do HF using "direct conversion", the other I use with a "Ham-It-Up" converter to cover the HF frequencies. The Ham-It-Up converter has a noise source that may be useable for filter bandpass measurement, and with a directional coupler to measure antenna bandwidth/SWR.

The SDR software I use is CubicSDR. While I like the software for general listening, it does too many behind the scenes gain adjustments and changes to the "DB" scale to make it of any use to quantify anything. Everything is a moving target.

I've used OpenWebRX on one of my Raspberry Pis, under the covers it is using libcsdr https://github.com/simonyiszk/csdr a modular set of SDR radio components that follow the UNIX tools philosophy. Possibly with this toolkit I could nail down the receive side, actually being to nail down a known fix gain.

Even with stepped attenuators (I think Keys makes a low cost one) I still need to have a way to measure how flat my SDR system is across the wide frequency range needed for measurement. 

But adapting SDR dongles to the needs of Hams to expand what they can measure to keep up with more modern technology and performance standards is a great idea.

Tom, wb6b


Jerry Gaffke
 

Go for it, replicate the specan.
It will be very useful.
But that's a much more ambitious project than Arv's design.
Simple is good if it's good enough for the task.

Jerry


On Thu, Aug 9, 2018 at 02:14 PM, Alan de G1FXB wrote:

Not meaning to be inflammatory to the author,

 but at the same time if there if there is no need to reinvent the wheel if it will give reproducible data comparable to other solutions:-

http://hfsignals.blogspot.com/p/specan-reboot-of-w7zoi.html

 


Tom, wb6b
 

I like the idea of using the rt-sdr dongles as a lower cost way for the larger population of Hams to do their own measurements on their equipment. That would allow me to do my own measurements without spending more money than I can justify at the moment on a commercial SA.

I have two dongles, one can do HF using "direct conversion", the other I use with a "Ham-It-Up" converter to cover the HF frequencies. The Ham-It-Up converter has a noise source that may be useable for filter bandpass measurement, and with a directional coupler to measure antenna bandwidth/SWR.

The SDR software I use is CubicSDR. While I like the software for general listening, it does too many behind the scenes gain adjustments and changes to the "DB" scale to make it of any use to quantify anything. Everything is a moving target.

I've used OpenWebRX on one of my Raspberry Pis, under the covers it is using libcsdr https://github.com/simonyiszk/csdr a modular set of SDR radio components that follow the UNIX tools philosophy. Possibly with this toolkit I could nail down the receive side, actually being to nail down a known fix gain.

Even with stepped attenuators (I think Keys makes a low cost one) I still need to have a way to measure how flat my SDR system is across the wide frequency range needed for measurement. 

But adapting SDR dongles to the needs of Hams to expand what they can measure to keep up with more modern technology and performance standards is a great idea.

Tom, wb6b


Alan de G1FXB
 

Not meaning to be inflammatory to the author,

 but at the same time if there if there is no need to reinvent the wheel if it will give reproducible data comparable to other solutions:-

http://hfsignals.blogspot.com/p/specan-reboot-of-w7zoi.html


Alan


On 09/08/2018 21:11, Jerry Gaffke via Groups.Io wrote:
Arv's mixer+si5351 would be easier to understand and hack,
no download of some bloated SDR software package required. 

The DVB-T route can be good, especially if interested in VHF and beyond.
Perhaps Arv's circuit could also serve as an up-converter so a DVB-T could be used at HF?

To be "easy" for anybody who hasn't done it already
there needs to be a couple pages of really explicit instructions.
Be it for Arv's mixer+si5351  or the DVB-T thing.

Jerry



On Thu, Aug 9, 2018 at 12:15 PM, ajparent1/KB1GMX wrote:
Easy way for the less technical builder...

DVB-T dongle, Upconverter, PC or Andriod tablet with software pick one. Done.

Likely the PC or tablet is already around.  I got a cheap 4core android for 35$(in 2016) at Walmart.
The DVB-T about 20$.
The upconverter about 40$ Nooelec.  Or make it its a crystal source, DBM 
and two filters.  Schematics on the net.
Just a fast look at Amazon no doubt both can be found much cheaper.


Jerry Gaffke
 

Correction:  down-converter.
Unless you somehow got a DVB-T USB dongle that can transmit.


On Thu, Aug 9, 2018 at 01:11 PM, Jerry Gaffke wrote:
Perhaps Arv's circuit could also serve as an up-converter so a DVB-T could be used at HF?


Arv Evans
 

Jerry, and others...

My little system is just a direct conversion receiver with Si5351a LO for wide frequency
coverage.  There are lots of potential ways to improve on the idea.

Arv
_._


On Thu, Aug 9, 2018 at 2:12 PM Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io> wrote:
Arv's mixer+si5351 would be easier to understand and hack,
no download of some bloated SDR software package required. 

The DVB-T route can be good, especially if interested in VHF and beyond.
Perhaps Arv's circuit could also serve as an up-converter so a DVB-T could be used at HF?

To be "easy" for anybody who hasn't done it already
there needs to be a couple pages of really explicit instructions.
Be it for Arv's mixer+si5351  or the DVB-T thing.

Jerry



On Thu, Aug 9, 2018 at 12:15 PM, ajparent1/KB1GMX wrote:
Easy way for the less technical builder...

DVB-T dongle, Upconverter, PC or Andriod tablet with software pick one. Done.

Likely the PC or tablet is already around.  I got a cheap 4core android for 35$(in 2016) at Walmart.
The DVB-T about 20$.
The upconverter about 40$ Nooelec.  Or make it its a crystal source, DBM 
and two filters.  Schematics on the net.
Just a fast look at Amazon no doubt both can be found much cheaper.


Jerry Gaffke
 

Arv's mixer+si5351 would be easier to understand and hack,
no download of some bloated SDR software package required. 

The DVB-T route can be good, especially if interested in VHF and beyond.
Perhaps Arv's circuit could also serve as an up-converter so a DVB-T could be used at HF?

To be "easy" for anybody who hasn't done it already
there needs to be a couple pages of really explicit instructions.
Be it for Arv's mixer+si5351  or the DVB-T thing.

Jerry



On Thu, Aug 9, 2018 at 12:15 PM, ajparent1/KB1GMX wrote:
Easy way for the less technical builder...

DVB-T dongle, Upconverter, PC or Andriod tablet with software pick one. Done.

Likely the PC or tablet is already around.  I got a cheap 4core android for 35$(in 2016) at Walmart.
The DVB-T about 20$.
The upconverter about 40$ Nooelec.  Or make it its a crystal source, DBM 
and two filters.  Schematics on the net.
Just a fast look at Amazon no doubt both can be found much cheaper.