Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1] 2 Next   Go Down
  Print  
Author Topic: SIGNAL SATURATION WITH FLDIGI  (Read 5654 times)
W6PJJ
Member

Posts: 55




Ignore
« on: August 12, 2012, 04:42:34 PM »


I have noticed an interesting behavior with FLDIGI. When a strong signal appears in the Bandpass of the audio spectrum it appears as though it saturates and causes de-sense in the demodulator and sometimes interferes with the received signal of the station being worked. I have searched the help files and several locations online but, cannot locate any discussions on this topic. The System is a iMac Computer running FLDIGI. Anyone else experience this during a QSO?

Don W6PJJ

 Cool
Logged
STAYVERTICAL
Member

Posts: 854




Ignore
« Reply #1 on: August 12, 2012, 05:22:54 PM »

Hi Don.

You have encountered one of the most common complaints about running digimodes, especially PSK31.

If you were using a narrow filter, the other signal would be outside the passband, and would not affect operation very much.
But, in digimodes, we commonly have the passband wide open to see as many signals as we can, typically 2.4Khz for a typical SSB filter.

The receivers AGC does not care whether we are looking at signal A or B or C - it just sees an aggregate strength and employs AGC appropriately.
So if a very strong signal appears anywhere in the passband - the receiver sensitivity is dropped and say goodbye to the weak signals.

There is nothing wrong with your transceiver or FLdigi - it is just what is supposed to happen.
Unfortunately, we cannot expect a transceiver to handle this - since this would be the same as expecting it to respond only to 1.3Khz in an SSB transmission.

The solution is to do what we do in CW work.
Use filters, I.F. shift, notch filters, adjust the AGC - and whatever else we can to remove the strong signal from the passband.

I know it is very attractive to see all those signals in the passband, but we need to think like a CW operator, since the situation is similar.
Many low-bandwidth signals squeezed together having separate qso's - so think what works in CW and you will have solved the problem.

Hope this helps, 73 - Rob

Logged
W6PJJ
Member

Posts: 55




Ignore
« Reply #2 on: August 12, 2012, 05:53:39 PM »


Hello Rob:

Thanks for your reply and remarks. Well... I had the distinct feeling that the Bandpass was far too wide but I cannot shift it sufficiently with the Paragon. I was hoping that maybe there was some upgrade for FLDIGI that would allow more parametric control and narrowing of the channel area or mask. A Computer is very capable of doing this especially if the Computer and the Transceiver were integrated and allowed control of the Passband. HI HI. I have worked with the Paragons PBT and BP it works a bit but not enough to eliminate adjacent channel interference. A receiver controlled by a Computer would be the Cats Meow wouldn't it?...... we could tighten it up or open it up or even disallow those offending signals but that would take some real power and I may have to use the Mac Pro if we get to that point someday. Thanks for responding... you have validated my suspicion. Maybe we will have to try other modes to see if this can be dealt with.

73's Don
W6PJJ

 Cool
Logged
N4CR
Member

Posts: 1652




Ignore
« Reply #3 on: August 12, 2012, 05:56:36 PM »

The problem can't be fixed after the stage in the radio that supplies AGC voltages.

This is the number one reason why I ditched my FT-857D and got the IC-746Pro. The IF DSP in more modern rigs occurs before the AGC is determined and thus you can narrow your IF bandwidth and cut out the station that is pumping your AGC.

It makes a world of difference.
Logged

73 de N4CR, Phil

We are Coulomb of Borg. Resistance is futile. Voltage, on the other hand, has potential.
STAYVERTICAL
Member

Posts: 854




Ignore
« Reply #4 on: August 12, 2012, 06:37:52 PM »

Hi again Don,

Yes, it is a common situation, but there are ways of mitigating the effect sometimes.

I am not familiar with the Paragon, but have had a quick look on the net and notice it has some optional filters available.
If its worth doing, and the cost is reasonable, a 500hz I.F. filter would be very worthwhile.
I suggest 500hz over 250hz because a lot of digimodes are 500hz and 250hz would be too narrow, although better with CW.

But without any money outlay there are also some things you can do:

When you send on digimodes, your transmit frequency is the sum of your R.F. (dial) and audio frequency (assuming USB).
So, if you adjust your dial up or down so that the offending signal moves out of the passband, you can just adjust the audio frequency to compensate.

For example:
Lets say you have the dial set to 14.070Mhz, and have a signal at 1000hz you are interested in working.
But you have the black hat QRO PSK31 at 2000hz on your waterfall display.
By adjusting your dial to 14.071Mhz the desired signal will be at 2000hz on the waterfall, and the black hat will be at 3000hz(outside your passband).

You get the idea - it can be done up or down and it is very easy with digimode programs, since you just click on where you want to transmit.
No math required - just move the dial to eject blackhat from your passband.

This stops the AGC being pumped by black hat as N4CR mentioned.

Just as an interesting aside, there is some benefit to audio processing, mainly since many soundcards have limited dynamic range.
So a very strong audio signal can cause problems as well.
I have had success using audio notch filters and even computer audio processing in helping reception, although it is obviously not as good as I.F. methods.

One technique I found helpful was to pass the audio from my USB signalink external soundcard interface through the internal P.C. soundcard for audio DSP processing.
It sounds complicated, but it just needs a jumper between earphone and microphone on the notebook in my case.
Having two soundcards, one is used to massage the audio signal before passing it to the digimode program.
It is possible to have brick wall filters and notches with this system - but it needs two soundcards and does nothing to stop AGC pumping.
Interesting to play with however.

Good luck - it is a hobby after all - a form of play - so enjoy.

73 - Rob
« Last Edit: August 12, 2012, 06:40:50 PM by STAYVERTICAL » Logged
N4CR
Member

Posts: 1652




Ignore
« Reply #5 on: August 12, 2012, 07:20:34 PM »

Another thing you can do is turn off your preamp(s) and dial in an attenuator. Also turn down the RF Gain. Perhaps turn off AGC entirely. I resorted to all of those things when I was using a radio without IF DSP or good filtering. Eventually I got a 250hz filter but it's not good for a lot of digital modes. It was great for PSK-31 though.

There are many times with my IC-746Pro that I'm filtering a PSK-31 signal through a 50 hz slot.
Logged

73 de N4CR, Phil

We are Coulomb of Borg. Resistance is futile. Voltage, on the other hand, has potential.
AA4PB
Member

Posts: 12679




Ignore
« Reply #6 on: August 13, 2012, 08:49:36 AM »

I'm not real familiar with FLDIGI but in MixW2 you are able to create your own rig control macros. I have written a narrow macro that narrows the DSP IF filter on my IC756PRO down to 50Hz and retunes the radio to center the currently selected signal in the filter - after first memorizing the original settings. Then a wide macro can return to the full waterfall. It all happens so fast that I can switch from wide to narrow without loosing more than two or three characters of a transmission.

The macros are availabe on the web site at www.ham-kits.com
Logged
KK4CPH
Member

Posts: 154




Ignore
« Reply #7 on: August 13, 2012, 09:16:24 AM »


. Maybe we will have to try other modes to see if this can be dealt with.

 Cool

Try Olivia.  Seems to work great even with QRM.   Cool

Eric
Logged
KG4RUL
Member

Posts: 2686


WWW

Ignore
« Reply #8 on: August 13, 2012, 09:34:10 AM »


I have noticed an interesting behavior with FLDIGI. When a strong signal appears in the Bandpass of the audio spectrum it appears as though it saturates and causes de-sense in the demodulator and sometimes interferes with the received signal of the station being worked. I have searched the help files and several locations online but, cannot locate any discussions on this topic. The System is a iMac Computer running FLDIGI. Anyone else experience this during a QSO?

Don W6PJJ

 Cool

You will see the same behavior with MixW.
Logged
AA4PB
Member

Posts: 12679




Ignore
« Reply #9 on: August 13, 2012, 10:30:56 AM »

It's not a matter of the software being saturated. It's a matter of the receiver AGC being driven by the strong signal in the passband and reducing the gain of the receiver to compensate (as it was designed to do). The only real solution is to narrow the bandwidth of the IF filter in the receiver. Since the IF filter is generally centered at a fixed frequency you usually have to retune the VFO dial frequency in order to move the desired PSK signal into the filter passband. That hopefully leaves the strong signal outside the filter passband where it will be greatly reduced in strength before hitting the AGC circuit.

As I said, with MixW2 and a receiver like the 756PRO you can make these adjustments automatically with rig control macros. I don't know about FLDIGI.

Logged
G0GQK
Member

Posts: 634




Ignore
« Reply #10 on: August 13, 2012, 01:17:05 PM »

Try using Airlink Express, there are four filters available which will solve some of the problems

Mel G0GQK
Logged
N4CR
Member

Posts: 1652




Ignore
« Reply #11 on: August 13, 2012, 03:51:09 PM »

There is NO filter that happens after the AGC is picked off that can fix a radio that's being pumped by a strong signal. You have to fix it before the AGC is sampled.

You can't fix it in software, you can't fix it in audio DSP or filtering.

The signal you were listening to was already lost way back when the AGC pumped and the signal you wanted to hear was attenuated by the AGC.

It's the way radios are designed. You can't fix this in a later stage and it's the reason that modern IF DSP rigs are selling like hotcakes for digital work. They really do solve the problem in a fundamental and correct way.
Logged

73 de N4CR, Phil

We are Coulomb of Borg. Resistance is futile. Voltage, on the other hand, has potential.
VA7CPC
Member

Posts: 2357




Ignore
« Reply #12 on: August 13, 2012, 05:06:05 PM »

+1 to what N4CR says.

Try turning off the AGC (if you can do that on a Paragon), and "riding gain" with the RF Gain and/or IF Gain controls. 

That's the best you can do, unless you add a narrow filter at the IF frequency.
 
               Charles
Logged
W6PJJ
Member

Posts: 55




Ignore
« Reply #13 on: August 14, 2012, 09:10:19 PM »



So many replies and so much information. I use the filters on the Paragon and sometimes turn off the AGC but, with local stations the Computerized demodulator gets swamped. A form of Software Defined Receiver would be real handy in this situation, we have such a very narrow channel mask with this mode that Computer defined filtering would be the Cats Meow... fortunately only the very strong stations swamp the Demodulator.Thank all of you for your remarks and recommendations.

73's-Don
W6PJJ

 Cool
Logged
DIDJATUL
Member

Posts: 9




Ignore
« Reply #14 on: August 25, 2012, 12:41:10 PM »

When the automatic gain control circuit reduces audio gain, tweaking the fast fourier trasform processing could greatly improve your signal processing in a lot of receive situations.

I always have the fldigi confirguraton window open to make real-time adjustments to the FFT processing as receive conditions vary. The "Blackman" fft is the default fast fourier transfrom prefilter window function in fldigi, but I tend to use "Rectangular", and then maybe "Hanning" for my fft window, but I have used all of the fft prefilters windows to tweak my signal processing based on receive conditions.

The audio sample rate, and converter type is important too. Use the Linux system monitor, or Windows task manger "performance tab" to determine CPU load. The "best sinc interpolator" can cause 100% CPU load in some old cpu processor types. Even my modern day intel atom processor can see very high load with the selection of the wrong interpolator type. I use the "fastest", with a high sample rate (96000 bps) via portaudio in linux.

I use the FFT signal line display in my waterfall window to reduce cpu load, and use the signal browser to select signals. You can get a "feel" for your fft latency rate, and gain a better understanding of your computers performance as it applies to the fft algorithm processing power.

Determine if your FFT Latency Adjustment is correct. Again, "4" is default, but I tend to use "2". I have many times used 6 or 7 based on fft processing performance during varying signal conditions, and the fft prefilter window I have selected.

The fast fourier transform algorithm is very powerful, and can be adjusted to meet the varying signal conditions that a digital mode amateur radio operator encounters, and used as a tool to overcome some of the limitations of the transceiver. FFT is the most important feature of fldigi, and with perfect adjustment, can help greatly if you do not have the best digital mode transceiver. Learning FFT processing adjustments, and optimizing FFT window function can help decoding significantly.

In PSK the computer, and software package (FFT) are just as important as the transceiver, and antenna. A powerful cpu, and a well adjusted fft window can decode a weak signal when it is reduced by AGC.

Practice adjusting fft with no filter, just the audio input, and then apply passband techniques when you have mastered FFT processing.

I have a Icom 718. It has no agc control at all. I have observed that perfect adjustments to Fast Fourier Transform have helped a lot in using my transceiver in PSK.
Logged
Pages: [1] 2 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!