HF Digital Error correcting? Also, what's up with PSK31?

Discussion in 'General Technical Questions and Answers' started by N0NS, Oct 9, 2008.

Thread Status:
Not open for further replies.
ad: L-HROutlet
ad: l-rl
ad: Left-2
ad: Subscribe
ad: L-MFJ
ad: Left-3
ad: FBNews-1
  1. K0RGR

    K0RGR Premium Subscriber QRZ Page

    The first secret to PSK reception is to turn off the AGC. And, if you can find a way to use your filters, that will not hurt, either. Noise blankers also greatly limit your dynamic range, though I find DSP filters often help the reception.

    I am not aware of too many receivers that can handle an extremely strong signal 60 or 100 Hz. away from a very weak one, regardless of the filters in play, or the modes involved. The '31' in PSK31 refers to both the bandwidth and the baud rate of the mode. 31 is slow and narrow. 64 is faster and less narrow. 256 is wider yet, and but MUCH faster.

    Now, if you park your KW 2 Khz. away from me, I won't care that much. With the AGC off and the DSP filters on, my IC-7000 won't know you're there.

    It isn't socialism, it's civilization. You can't poop on the public commons, either. If you want to run a KW, go up the band a little and run RTTY. I run high power on RTTY, when I don't care about my power bill.

    Thanks to AG3Y - I was thinking about FLARQ in my previous post - couldn't remember the name or the guy who created it. I hope it catches on - I really miss AMTOR!
    Last edited: Oct 10, 2008
  2. KI4NGN

    KI4NGN Premium Subscriber QRZ Page

    Why? It's an audio signal, and a properly adjusted PSK transmission has the audio level set to just below triggering any ALC action. Just what is it that you're implying will be wrong with that signal?

    So? What's your point?

    Well, the IF DSP filters on my 756ProII are before the AGC processing. How much further 'upstream' do you need to be?

    I'm not sure what you are trying to say here, but I would rate weak signal PSK second only to CW in terms of being able to copy. What else matters?


    As already noted, these complainers need to know how to use their filters, or as noted by RGR, turn off their AGC.


    I haven't tried it, but in thinking about it a little more I've come to the conclusion that turning off the AGC is likely not a solution to deal with strong adjacent signals.

    The problem with the AGC being on is that the strong signal in the passband cause the gain to drop, also reducing the weaker signal.

    The problem with turning off the AGC is to remember why it reduces gain to begin with; to prevent distortion from overload. If it's off then the distortion being caused by the strong signal is going to effect the entire passband, including the weaker signal.

    Filters are the answer.
    Last edited: Oct 10, 2008
  3. N9DSJ

    N9DSJ Ham Member QRZ Page


    Not trying to take anything out of context but wanted to limit my comment to this statement alone. There are many modes that far exceed BPSK31 in terms of digging signals out of the noise; some even do better than CW. BPSK31 is very popular but not overly robust and does poorly in over the pole paths due to phase distortion. Probably the most sensitive are various WSJT modes such as JT65, WSPR, JT2 and JT4x but these are hardly "conversational" as they are aimed at conveying the minimal information needed for a "good" contact using EME or other weak paths. Modes such as PSKAM, Thor, DominoEx, Throb, Olivia, MFSK and others are less sensitive but still surpass BPSK31 by a large margin. CHIP and MT63 are less sensitive, but due very well in heavy QRN and QRM, albeit at the expense of bandwidth. Also some ARQ sound card modes are effective yet reasonably narrow, such as ALE400, PAX and various NBEMS with Flarq modes (although that is meant for file transfer). In addition some of these modes "play" very well with strong adjacent signals as high power and good antenna(s) are a necessity for tough paths. This is not meant to diminish operating QRP, but have never seen a complaint of someone with "too strong a signal" when using JT65, for example.


    Bill N9DSJ
  4. W8JI

    W8JI Ham Member QRZ Page

    While it is true that turning off AGC, turning back RF gain, and adding selectivity helps a PSK system you are still limited by the very close spaced dynamic range of the entire system.

    If a transmitter has harmonic distortion of -50dB in the audio chain, which would be a VERY good transmitter, and you transmit PSK you have a harmonic -50dB down unless you park the transmitter up near the upper cutoff of the filter. Even if you do that you have IMD that will fall within passband, and the very best transmitters are again around -50dB in that respect.

    A CW transmitter only has the problem of waveshape in the keying circuit and linearity after that point. The entire audio system is not at play.

    The receiver, as much as you like your DSP filters, has limited dynamic range when only the DSP filters are at work. It gets a whole lot better when the roofing filter kicks in, but that is several kHz wide.

    Look at this table:

    at 2 kHz spacing the IM DR is -75 dB for a Pro III, and that is just with two EQUAL level signals 2 kHz apart. It does not include audio system distortion including the detector and later stages. It gets much worse when the additional stages are tossed in, or when more signals are present.

    As a matter of fact very few competitive stations, if any, use the PRO III because the close spaced performance with multiple signals is so poor.

    All of this stuff works together to prevent PSK or other modes that use the SSB and entire audio systems on both ends from having the high dynamic range possible if it was a different modulation scheme, like direct DSP generated and decoded digitized signals with suitable roofing filters if needed.

    This is why PSK people go nuts about everyone running a tiny little antenna with low power. The system has poor dynamic range, so it becomes a socialist mode where everyone has to be a signal peasant.

    You may not like it, you may not agree with it, but it unfortunately is a fact. We can pack a clean 1500 watt that is 30 over 9 a few hundred Hz away from a noise floor CW signal with off-the-shelf good equipment, contesters do that all the time. If you tried that on digital modes that run soundcards into regular SSB radios and decode after passing through the entire receiver system you would be tarred and feathered by the masses regardless if they had AGC on or off or had a narrow filter selected.

    Now it is possible to have very good dynamic range, and it will probably come from things like a Flex radio that can get into the 90 dB or more dynamic range with detecting inside the computer, or with a DSP system that uses a very good DSP (better than the stuff we commonly see).

    Any system that uses a common SSB radio's system from mic jack to line output (or worse yet speaker output) to decode will have terrible dynamic range at very close spacing. There is just too much unnecessary junk in the signal path.

    73 Tom
  5. NN4RH

    NN4RH Premium Subscriber QRZ Page

    Never understood why they feel they must have their 1000 Watt CW station smack in the middle of the digital modes playground - they aren't even following their own "Considerate Operators Frequency Guide".
  6. KR2D

    KR2D Ham Member QRZ Page

    Honestly, W1AW was there first. I believe that Europeans came up with the suggested frequencies for the PSK watering holes. I believe the logic was to set the PSK area 10KHz below the RTTY area. That put it smack on 3580.

    When I visited W1AW, the operator (forget his name and call) said he was not aware that the PSK watering hole was surrounding 3581.5.

    W1AW is LOUD here on 80m, but I can mostly filter them out. A while ago, they had a serious problem with key clicks that killed anything +/- 500 Hz, but that is fixed.
  7. TI2AEB

    TI2AEB Ham Member QRZ Page

    The Mode ARQ is one in which there can not be errors, but is rather slow. Besides it requires fair signals to be effective.

    Try Olivia or Domino. Althought they do not have true error correction. As as a CW operator here, I can tell you that with Domino 4 or 8 or Olivia, you cant hear the signal, but the machine copy the message almost without any error.In fact hams likes to have QRP, some 2.5 watts plus a vertical antenna, nevertheless I can workout them 100%, thoughout the QRN. Under almost closed band (around 6pm in 14 MHz)

    OLivia and Domino can work quite well with a S/N ratio of only -13 dB. You can hear a signal with that S/N at all.
    Please download MULTIPSK, were you can find excellent explanations of the several modes, their speed and S/N. Also you can find there the QRG of the modes other than PSK31.

    I use DM780, and with PSK31 you can watch at the same time around 10 qso.
    It is possible to do so, because PSK31 takes so little BW. That is the advantage of this mode.
  8. AB0WR

    AB0WR Ham Member QRZ Page


    I think you are trying to blow some smoke at us here.

    YYou started off talking about IMD. You will get the same IMD performance out of the driver and amplifier in a SSB/CW transceiver whether you are using CW *or* psk31. Most transcievers don't change the bias settings in order to affect any significant difference for one mode over the other.

    Now in this mesage you jump to speaking about harmonic distortion and try to somehow tie that into increased IMD. You know darn well that harmonic distortion in the audio chain doesn't cause increased IMD in the amp chain. The harmonic distortion might cause a "doppleganger" signal to be sent but that is not the same thing as driver or amplifier IMD.

    A high speed CW signal with the same throughput rate as psk31 will probably take up *MORE* spectrum than a psk31 signal.

    On the receiver side, the dynamic range is going to be limited in the exact same way with either CW or SSB, i.e. rf amplifier, agc, etc. dynamic range limitations.

    The only difference in the typical receive chain would be the IF filter. If you don't have a narrow CW filter installed then you will be listening to CW with the exact same components you listen to psk31 with. It would be hard to claim any significant advantage for CW over psk31 in this case. You'll have exactly the same noise contribution, exactly the same dynamic range, and exactly the same adjacent channel interference results. If you use a narrower filter for *both* modes, you'll still have equal receiver performance.

    Face it folks. The *prime* draw that psk31 has is its ease of operation for appliance operators. They can use a computer and a SSB transceiver to sit down for a few minutes in the evening and make a "run" through the entire SSB passband making multiple contacts without having to do anything except click the mouse on the next signal.

    This operational convenience *FAR* outweighs any limitations, at least for most operators -- and many times that includes me. Rather than using my 250hz CW filter and trying to tune through the psk31 watering hole looking for contacts it is just so much simpler and quicker to just look at the waterfall, pick the next station, and click on it.

    Bottom line? Psk31 doesn't have significantly more problems with IMD than CW. -50db from a 30 watt psk31 signal would be what? .0003watts? I don't think this is going to be above the noise floor for most people. And the operational convenience is much greater than the inconvenience of a limited dynamic range in the speaker audio or the computer sound card, especially when exactly the same receive chain is used by many people for both CW and psk31.

    tim ab0wr
  9. W8JI

    W8JI Ham Member QRZ Page

    Not true.

    Systems using the entire audio section of conventional radios to process very narrow signals are loaded with problems. Collins, as good as they were, tried this with the 32S1. They learned a painful lesson, and the 32S1 is historically one of the cleanest transmitters every built. They had to abandon the audio generated CW in the 32S1 and change it to real CW.

    You have IM problems from all the audio stages, hum and noise from external and internal sources, carrier balance, sideband suppression, and half dozen other issues to deal with when people try to dump narrow modes into a mic jack on a regular radio.

    This is compounded by doing the decoding external to the radio after it passes through the audio system of the receiver.

    I'm sure a good wide dynamic range directly generated PSK system could be built using modern DSP technology, but the fact is a common system with even the best radios and installation is generally limited to the 50 db range in a bandwidth dozens of times the actual signal bandwidth. That is terrible compared to other modes.

    This is why PSK people are such constant whiners and complainers about signal level. They demand everyone be weak because it hides the dynamic range flaws, and the lack of allowed ERP severely limits the effective range of PSK.

    It is a "toy mode" as it stands today. It could be a good system if the equipment didn't use audio in and out.

    73 Tom
  10. WB3BEL

    WB3BEL Ham Member QRZ Page

    I have to agree with Tom W8JI on this one.

    He is being super generous when he uses an example of -50dBc IM products from a transmitter. The norm is more like -25 dBc. Sure there are a few transmitters that can run class A and might get into the 40+ dB range if perfectly adjusted.

    The problem is that once the transmitter has created these IM products no receiver filter is going to remove them. Well theoretically if you demodulate a strong signal and recreate the TX chain distortion perfectly you can cancel it in the receiver, but this is not the stuff you see in Ham radio equipment. And it would not work very well on HF anyway.

    His point that using a transmit architecture that was made for analog voiceband audio is sub-optimal for narrowband data is on target.

    Thats not to say that you cant have fun with lots of these easy to use modes, but you should recognize some of the limitations.

    73, Harry
Thread Status:
Not open for further replies.

Share This Page

ad: elecraft