DSP lightning noise reduction for WEAK signals

[KF1Z]

You need a board for whatever frequency you'll be using....

If you use it as a stand-alone RX, then you want an 80meter, or 40meter etc...

An IF version.. (same thing...)  you want Whatever IF freq you're putting into it.

You'd have to use your 455khz, or 8.2mhz.

There is an article on the SR40 files section specific to the FT1000-D

[K1JJ]

A special receiver, that receives and separates both the upper and lower sidebands (an SDR with the appropriate DSP-based demodulator)... and which includes a 1 second delay... would be able to insert the audio recovered from the upper sideband or the lower sideband... depending upon which one is the least noisy (which one doesn't have a lightning crash on top of it).
Stu

Pretty cool idea, Stu.

Something like a voting system - you create the choices.   Taking it further... put a THIRD signal on a pilot carrier with a 2 second delay and give it a third choice ... :-)

EZ way: Crank up the transmit audio echo delay box and set it for 1 second.  Put on flak jacket. Caw Mawn.  

T

[K1JJ]

You need a board for whatever frequency you'll be using....

If you use it as a stand-alone RX, then you want an 80meter, or 40meter etc...

An IF version.. (same thing...)  you want Whatever IF freq you're putting into it.

You'd have to use your 455khz, or 8.2mhz.

There is an article on the SR40 files section specifice for the FT100D

Bruce,

OK, I understand now.... so it's either an 8.2 mhz or 455 khz.  I suppose the 8.2mhz BEFORE the filters might be the best choice to be able to create and control any new DSP filtering desired.  I assume the DSP filter software can create a better filter than the current INRAD $110 2.8khz types, right?

Do you have a link to the FT-1000D articles... on this site or elsewhere?  You DID mean 1000D not 100D, right?

T

[KF1Z]

Tom, I was going to email it... but my email server has been down for a couple days...

Yes I meant ft-1000d

The  8 mhz IF is the way to go... there may me a mod to do....

Here's the link

http://f1.grp.yahoofs.com/v1/sB4WSIMSvEaNb5LsGb2LvFO3xLzMZe07fCQnMQIY7fGztPiQs8AJRjHZsFuzT6KqY61ShRdCajE18rxk1hcG7S_gKuoZugT3vcKx-nLq/FT-1000WithSoftRock.pdf

[WA1GFZ]

Tom,
I sent you all this stuff a while ago but there may be some newer stuff.
The input transformer design is a PITA. I have a better one and can make you one in about 10 minutes rather than winding a mile or #30 on a low perm core. fc

[K1JJ]

Bingo.  That 1000D article was just what I needed, Bruce.  Tnx.

http://f1.grp.yahoofs.com/v1/sB4WSIMSvEaNb5LsGb2LvFO3xLzMZe07fCQnMQIY7fGztPiQs8AJRjHZsFuzT6KqY61ShRdCajE18rxk1hcG7S_gKuoZugT3vcKx-nLq/FT-1000WithSoftRock.pdf

It said:  "I can say conclusively that weak cw and ssb signals are more readable/more confortably readable listening through the digital filters of the softrock or Delat44 than through the crystal filtering of the FT-1000D when there is QRN or static crashes present....."

This is a good start in the right direction for static crash reduction. Better filters (DSP) for static is an area I never considered.

The author wrote it in Jan/2007 and said he wud add updates to the forum. Hopefully there's some newer progress. But it looks do-able and I'll try it once I can zero in on the easiest board available. The article said YIG may have some more 11 mhz crystals available for the 1000D interfacing. That's the info I need to dig out, as well as the input transformer you mentioned, Frank.

T

[WA1GFZ]

I did not observe the performance with static. The other features are cool though. You can't beat a good FFT spectrum display

[KF1Z]

Tom,
I sent you all this stuff a while ago but there may be some newer stuff.
The input transformer design is a PITA. I have a better one and can make you one in about 10 minutes rather than winding a mile or #30 on a low perm core. fc

It's actually quick and easy for the one he's going to need... (8.2 mhz)
Can't speak for the performance compared to something else though.....

Tom if you want it... I have one here as I said.... change the xtal and filters.... and you've got it....

[K1JJ]

Tom,
I sent you all this stuff a while ago but there may be some newer stuff.
The input transformer design is a PITA. I have a better one and can make you one in about 10 minutes rather than winding a mile or #30 on a low perm core. fc

It's actually quick and easy for the one he's going to need... (8.2 mhz)
Can't speak for the performance compared to something else though.....

Tom if you want it... I have one here as I said.... change the xtal and filters.... and you've got it....

Sounds, good, Bruce -

See Below - Here's the info YIG posted about converting a SoftRock board for the 8.215mhz  FT-1000D.  Is this what you mean by changing the "filters?" (C5, C6... L1, L2)

And I need an 11 mhz xtal and also the T1 is the transformers you meant, Frank?

Sounds easy enough.

Yes, Bruce, this is a V 6.1 or equiv board?  If compatible, send me your PayPal and total cost w/ shipping and we'll do it.

T

------------------


Re: [softrock40] Using at FT-1000D IF of 8215 KHz


Hi Jerry,

Here are a set of component values that I think could work properly for the FT-1000D IF application of the v6, (v6.1 circuit board), SoftRock.

SoftRock operated in 1/3 sub-harmonic sampling mode. Reverse audio cable connection to board with tip going to R on board and ring going to T on board. 

C5 180pF

C6 100pF

L1 8T of #30 on a T30-6 core (0.24 uH)

L2 30T of #30 on a T30-6 core (3.3 uH)

T1 8T of #30 on primary and 4T of #30 in each of the two secondary

windings "bifilar" over the top of the primary on a T30-6 core (0.24 uH on primary)

C21 = C23 = 1500 pF

C22 = 100 pF

Crystal X1 is 11.0 MHz

73,

Tony KB9YIG

----- Original Message -----
From: Tony Parks
To: softrock40@yahoogroups.com
Sent: Monday, August 28, 2006 9:40 AM
Subject: Re: [softrock40] Using at FT-1000D IF of 8215 KHz



Hi Jerry,
 
In July I built a v6 SoftRock for Steve Ireland, VK6VZ, to use with his FT1000 for IF signal processing.  Steve has the unit now and should be giving feedback pretty soon on how it is working out. 
 
The crystal I used for the FT1000 IF application was a 11.0 MHz crystal and 1/3 sub-harmonic sampling was used.  The center frequency was about 3* (11.0 -0.003 ) / 4 = 8.248 MHz, 33 kHz above the IF frequency.  If a soundcard is used that can sample at 96 kHz, the IF tuning with the SoftRock will be from 8.2 MHz to 8.296 which, I think, should give reasonable coverage of the FT1000 IF passband.
 
I think I still have a few more of the 11 MHz crystals and the associated components.
 
73,
Tony KB9YIG
 
 
----- Original Message -----
From: Jerry Flanders
To: softrock40@yahoogroups.com
Sent: Monday, August 28, 2006 8:39 AM
Subject: [softrock40] Using at FT-1000D IF of 8215 KHz


Need recommendations for using 80/40 softrock at 8215 KHz IF of FT-1000D.

Jerry W4UK

[KF1Z]

Tom, the info above is old, and not the same version....

But that's the right track.....

Schematic is here:

http://f1.grp.yahoofs.com/v1/0DoWSB7ahneCUZUMyMij8AbA3lMSSpn_RHEEQuwtXcupOq3lSyZRVRLpyaXm3qKOD4psMBAX5jJdqMtbbFLEQwiP9AG2db-xAc84UNcS/SoftRock%20v6.0%20docs/SoftRock%20Lite%20docs/v6.2%20SoftRock%20Lite%20%20schematic%2012_18.pdf


I can't locate the values you'll need,,,  might have to ask on the group, or email Tony...

[K1JJ]

Bruce,

OK - all set and ready to play....

If you have a link to the tech manual, let me know and I can figure it out the following questions.... :-)

I looked over the schematic and everything looks clear except for the "audio out."   What frequency is the output and does it simply plug into the computer sound card auxilary input with a standard coaxial cable?

I see they recommend a 98K sound card.  How do I determine my current sound card's ratings.... any tips in this area or recommended cards to use?  Anything else I need to prepare to get this up and running?


T

[KF1Z]

Bruce,

OK - all set and ready to play....

If you have a link to the tech manual, let me know and I can figure it out the following questions.... :-)

I looked over the schematic and everything looks clear except for the "audio out."   What frequency is the output and does it simply plug into the computer sound card auxilary input with a standard coaxial cable?

I see they recommend a 98K sound card.  How do I determine my current sound card's ratings.... any tips in this area or recommended cards to use?  Anything else I need to prepare to get this up and running?


T

"Audio-out"  is the I/Q data to the soundcard...
One left and one Right channel line inpoot to the card.

Use the soundcard you have.... you should be able to get the properties of the card under "device manager" or "sounds and sound devices".....

Don't worry about the bandwidth of the card for a while... it's not that important for your use...
(The 96k card gives you twice the spectrum display)

My "on-board" card worked ok-fine....
But, If you get into this idea, you may well find yourself upgrading.....


My email is STILL not working (thanks to Hughesnet, and their 24 hour upgrade, that has gone on for almost 72hours now..)

So send a PM with your mailing addy....

The "manual" is http://f1.grp.yahoofs.com/v1/8FYWSEh7eSz7QnNTGJ5Kr9lBjb8yNhGsm90d7iE8t6SBnNnXY8r_8CIqyNt70PgwNRTonsSnbfbWVRWpjMgIqWfBPj9U06E5Rv-7UuQ_/SoftRock%20v6.0%20docs/SoftRock%20Lite%20docs/v6.2%20SoftRock%20Lite%20Builder%27s%20notes%2011_17.pdf


That's actually "builders notes", but may be usefull...

You should go ahead and join that group, so you have access to all the info...
http://groups.yahoo.com/group/softrock40

[KF1Z]

I just got this responce from Tony...
These are the values you need....


""
Hi Bruce,
 
8.215 MHz SoftRock Lite IF use

SoftRock operated in 1/3 sub-harmonic sampling mode. Reverse audio cable connection to board with tip going to R on board and ring going to T on board.

C3 180pF

C4 100pF

L1 34T of #30 on a T25-2 core (4.0 uH)

T1 9T of #30 on primary and 5T of #30 in each of the two secondary

windings "bifilar" over the top of the primary on a T25-2 core (0.25 uH on primary)

C21 = 100 pF

C22 = 1500 pF

Crystal X1 is 11.0 MHz for a center frequency of about 8.249 MHz. Clock division set for x4.

73,

Tony KB9YIG

[Bacon, WA3WDR]

Looking at this today, an analog solution would probably work pretty well.  DSP would be a way to better recognize and predict the noise on a microsecond by microsecond basis, and a convenient way to apply controlled soft muting during microsecond to millisecond pulses.

An anti-jam technique that was aimed at strong-signal interference to weak signals was passing everything through an amplifier with serious crossover distortion, such that the desired signal did not come through it.  Then the strong jammer would be the only thing to come through, and then it would be inverted and mixed with the direct undistorted signal, as a pulse-cancelling signal.  This I think would greatly reduce the dynamic range that the receiver would confront in a lightning noise scenario.  I am thinking that receiver AGC can be applied to make dynamic adjustments to the size of the crossover notch.

As for muting during the noise bursts, I think it might be possible to mute the remaining clicks in sub-millisecond muting windows, and pass signal between these.  Depending on the actual quality of the noise, there may be quite a bit of improvement possible.  I imagine a peak detector looking at either raw signal, or maybe the 'jammer-cancelled' stuff, converted into a mute or no-mute square wave signal, followed by a low-pass filter (around 10 KHz to 100 KHz cutoff) controlling a multiplier that passes delayed 'jammer-cancelled' signal except when the peak detector says mute, for soft-edged muting.  Or this might be done digitally such that any mute is done by following a window curve such as Hanning or Hamming or raised-cosine, etc, from 1 to 0 and back to 1 again.  It is the soft curve from 1 to 0 to 1 again that is most important for minimum IM products, and a slight signal delay would align the 0 with the click.

DSP should help in responding to, if not predicting the noise snaps, and it should be able to help against the myriad of background clicks and ticks that are around the level of the signal being received.

[WA1GFZ]

The problem with lightning is the spectrum. It covers from audio to well into MHz.
This makes it hard to determine the difference between lightning and signal when the level is simular to the signal you want to listen to.

[K1JJ]

Bruce,

I signed up for the SoftRock group and have read a number of the FT-1000 related postings.  I should be able to figure things out enough to get it running.

BTW, I noticed the new values YIG just gave you today and the ones he posted from Aug/2006 (shown in a few posts above) have some values that are way different... like C21 and C22 switching positions from 100pf to 1500pf.  Hope there's not a mistake in the latest. But I'll use the latest ones YIG gave you today.

Yes, my adr is OK in QRZ.com. Thanks for getting me started.

Well, looks like the WX OM will be sunny for Nearfest this weekend. The forecash cleared right up! Hope to be there and see some of you mugs.


T 

[K1JJ]

Never Mind....

[K1JJ]

Here's a guy (PA0SIM) who is doing just what we are talking about.

He has two phased loops and uses two receivers into DSP processed software.  Even uses an FT-1000D for some tests.

Listen to some of his QRN and QRM reduction examples on 80M.  Very encouraging material...

http://www.pa0sim.nl/Results%20and%20Audio%20Samples.htm


Here's a diagram of the setup: (the second diagram of the K2 rigs using two SoftRocks is the better example)
http://www.pa0sim.nl/Hardware.htm

Check out his software choices and results: (MatLab, Simulink, Labview - now we're talking!)
http://www.pa0sim.nl/Software.htm


Here's how he gets full control of his signals in Labview - nice customized cancelling and phase and amplitude controls and graphics -  I gotta do it!
http://www.pa0sim.nl/screenshot.JPG


It appears the answer is not in a single box, but in a larger combined effort.


He says,

"Effective directivity by DSP -
Digital Signal Processing makes effective directivity possible similar to large directional antennas using an array of two antennas (two active small “magnetic” loop antennas) and two phase coherent receivers (FT1000D or two K2’s).

Both signals of the antennas/receivers are converted into the frequency domain.

For each frequency component the phase difference (actually Time Distance Of Arrival, TDOA) between the two signals is calculated.

Frequency components are then selected or rejected on their actual phase difference.

Phase selectivity equals directivity."


I need a SECOND SoftRock board....


I'll bet Big Country, you could pull this off with a second loop, caw mawn.

T

[KF1Z]

BTW, I noticed the new values YIG just gave you today and the ones he posted from Aug/2006 (shown in a few posts above) have some values that are way different... like C21 and C22 switching positions from 100pf to 1500pf.  Hope there's not a mistake in the latest. But I'll use the latest ones YIG gave you today.
.

There is a difference in the ver 6.1 and 6.2 boards...

The input filter in 6.1 had 2 inductors...
In 6.2 (the one you're getting) there is only 1 inductor.

So yes, use those I just gave you.....

[K1JJ]

Frank or Bruce,

I'm trying to piece together the last details of what I'll need to combine the two receivers into the soundcard.

I want to check to be sure before proceeding:

Click the link below and then look at the diagram on the bottom of the page.  He is using Tayloe QSD quadrature boards.  The SoftRock Lite 6.2 does the same basic function, right?   He was talking about preserving the phase of both signals for further nulling/ cancelling in the software. If I substitute the SoftRock Lites for these Tayloe's, it will work the same and preserve relative signal phase, correct?

I'm thinking of tapping the FT-1000D 8.215mhz AFTER the 2.8khz ssb filter to provide a roofing filter effect - at least that's what he did with the ElectroCraft K2's. (I wouldn't have the spectral display feature)  I would disable the AGC on both receivers.

I would need two Lite's to do the job.

http://www.pa0sim.nl/Hardware.htm

Does this look OK?



BTW, I read through the long 13 page thread you guys did here about SDR. Fasinating stuff and moving fast.  I got a kick out of your "buyers remorse", Bruce.... :-)     

Frank, did you get your Mercury yet and are you still excited about it?

T

[WA1GFZ]

Tom the Mercury is a couple months away. They are supposed to demonstrate one at Dayton. I read there was an artwork error under the A/D that slowed things down a bit. My 1/2 watt exciter should should show up in a couple weeks. Third order crud down 50 dB come on.
Softrock is a form of single balanced Tayloe. My HB unit is double balanced Tayloe. I would like to hear more about your dual set up. Sounds cool for the Cubic R3030. I bought a couple SR6s for it.
You don't want to put the I/Q after a tight filter because it will screw up the spectrum display width. The software filters will work as well as a single crystal filter.