EMRFD Message Archive 629

Message Date From Subject
629 2007-04-12 04:07:07 neomag_magneo SSB signal distortion in homwbrew receiver
Hello group !

I am finalizing my homebrew receiver project (started some 3 yrs ago 1)
and what is puzzling me is the noticeable distortion in the received
SSB signals. The most obvious source (in my opinion ) might be found
in the audio chain which consists of a three stage opamp lowpass
filter and a LM386, more than sufficient for headphones. Another
source, of which I have no previous experience, is the 12pole ladder
crystal filter in the first if of 8 MHz. The measured BW is 2.0 kHz
with quite steep slopes.

Question: is this BW narrow enough for any noticeable audio
distortion. If so, I will stop worrying and concentrate on other
things, if not, I will need to continue my troubleshooting with all
the stages.

The 2 kHz BW is my intentional choise since I mostly run digital
modes, RTTY in particular and for that, it can be cascaded with
another 8pole
800 Hz filter. It is, however, more difficult to hear the distortion
in digitally modulated signals. I have concluded that if I am
satisfied with the audio quality in the phone mode, the others should
be OK too !


BR

Heikki (OH2LZI)
630 2007-04-12 04:33:59 Leon Re: SSB signal distortion in homwbrew receiver
----- Original Message -----
631 2007-04-12 08:31:17 exwb6lxa Re: SSB signal distortion in homwbrew receiver
--Hi group, I am a new member to the group and have not received a
copy of the EMRFD book. I have some experiance building receivers
though over the past 25+ years and thought I would throw my 2 bits
in, still building gear from the 1977 "Solid State Design for the
radio amateur" book, hi.Actually in my third copy from hard use. The
2.0 kc filter with steep skirts is a dandy filter for DX work on
SSB. It will cut much of the high frequecy QRM off as well as hiss.
It will make the audio on SSB sound rather "bassy" if you have good
low freq response and losse a little definition on voices, but not
bad.You should do a distortion test and freq response run on your
amplifier chain if you have a scope, or test the signal out of the
detector with a different amp to determine which is causing the
audiable distortion. One member already warned you about the LM386
amplifier. I have actually had good luck with them properly applied
for low current drain rcvrs for CW and field use. I do usually run
them in the low gain mode without any components on pin 1 and 8,
choosing to provide the extra gain in a well designed preamp, or
OpAmp stages proceeding it.All the while carefully adjusting the
gain distribution, not overdriving any one stage or the entire
strip. Like he said there are better choices for a high performance
receiver but they are very available, cheap and are very stingy on
current drain if that is a consideration. Before I just yanked it
out and replaced it I would find the real cause of the distortion
because you almost make it sound like it is pretty noticeable and
it might not be just the somewhat scratchy sound of a LM386. I am
not a audiofile in my ham band receivers and with a distortion of a
few percent or up to 10% they can sound plenty pleasent to me while
listening to SSB or CW, my primary mode of operation. Good luck on
your project, it sounds like you have been working hard on it for
quite some time, good work. I am sure you are almost there, a little
more testing and tweeking will get it operating fine. After you
change the amplifier out with a more refined device then you will be
able to honestly decide the pros and cons on using them. I
personally think for a headphone amp
632 2007-04-12 09:03:25 Kenneth Stringham Re: SSB signal distortion in homwbrew receiver
I have had no problems with the LM-386. The 2KHz.
bandwidth is sufficient for receiving SSB
transmissions, but there will be distortion. SSB
stations are sending signals with a pass band of
2.8KHz. You are rejectiong the upper 800Hz or
information in the 2KHz. to 2.8KHz. range.

Ken



Don't be flakey. Get Yahoo! Mail for Mobile and always stay connected to friends. http://mobile.yahoo.com/mail
633 2007-04-12 09:16:19 Joe Rocci Re: SSB signal distortion in homwbrew receiver
Remember that SSB is a linear amplitude modulation, and so any clipping of the signal anywhere in the signal path will produce distortion.  You mentioned your 12-pole crystal filter as a suspect element. I'd suspect a 12-pole filter too. Very sharp cutoff filters produce a "group delay", which means that some audio frequencies are delayed more than others as the signal passes through the filter. This 'distortion' of the phase relationships between the various components of the audio signal can make the audio sound unnatural, but it's hard to tell from your description whether it's the cause of your distortion. A good look at the signal with a scope is the best way to figure out where the problem is.
 
Joe
W3JDR
 
 
 
----- Original Message -----
634 2007-04-12 09:18:26 DSNman@comcast.ne... Re: SSB signal distortion in homwbrew receiver
Attachments :
    Heikki,

    Is the audio actually distorted? Could it possibly be that your BFO is not quite on a frequency that will cause the signal to be properly within the passband of the crystal filter? If it is distortion it may be possible to find it with a scope using a carrier that produces a tone of 800 hz.

    I

    --
    Loren Moline
    WA7SKT

    -------------- Original message ----------------------
    From: "neomag_magneo" <oh2lzi@sral.fi>
    > Hello group !
    >
    > I am finalizing my homebrew receiver project (started some 3 yrs ago 1)
    > and what is puzzling me is the noticeable distortion in the received
    > SSB signals. The most obvious source (in my opinion ) might be found
    > in the audio chain which consists of a three stage opamp lowpass
    > filter and a LM386, more than sufficient for headphones. Another
    > source, of which I have no previous experience, is the 12pole ladder
    > crystal filter in the first if of 8 MHz. The measured BW is 2.0 kHz
    > with quite steep slopes.
    >
    > Question: is this BW narrow enough for any noticeable audio
    > distortion. If so, I will stop worrying and concentrate on other
    > things, if not, I will need to continue my troubleshooting with all
    > the stages.
    >
    > The 2 kHz BW is my intentional choise since I mostly run digital
    > modes, RTTY in particular and for that, it can be cascaded with
    > another 8pole
    > 800 Hz filter. It is, however, more difficult to hear the distortion
    > in digitally modulated signals. I have concluded that if I am
    > satisfied with the audio quality in the phone mode, the others should
    > be OK too !
    >
    >
    > BR
    >
    > Heikki (OH2LZI)
    >
    >
    >
    >
    635 2007-04-12 09:53:49 Luiz Amaral Re: SSB signal distortion in homwbrew receiver
    In my opinion, it depends on the type of distortion you are detecting.
    Specially for digital communications, the BW is not the only filter
    parameter to worry about, but also its phase response. If the latter is not
    linear, severe distortions may be introduced in the sent information, making
    it sometimes unrecoverable.
    Speech analog audio signals, however, are insensitive to components phases
    and, therefore, easier to receive without distortion.

    Luiz - PY1LL

    ----- Original Message -----
    636 2007-04-12 10:54:57 Andy Re: SSB signal distortion in homwbrew receiver
    > and what is puzzling me is the noticeable distortion in the received
    > SSB signals.

    Exactly what kind of distortion?

    When most people say distortion to me, I think nonlinear, as in harmonic,
    distortion. But technically, any filtering, like sharp low-pass linear
    filters, is distortion too. But it sounds totally different than
    nonlinear/harmonic distortion.

    So, what do you mean?

    Too low a bandwidth doesn't cause any nonlinear (harmonic or IM) distortion.
    It just makes it muffled, and maybe adds a bit of "ring" if the Q is too
    high.

    Can you pick off your signal before the audio amps and see if it is
    distorted? Or inject a signal there?

    If I remember right, I've used the LM386 several years ago (someone else's
    design) and found them OK, though somewhat noisy.

    Andy
    637 2007-04-12 23:41:12 Kevin Purcell SSB bandwidth and intelligibility
    This raises another issue I've been thinking about over the past week
    or so: how does intelligibility vary with SSB bandwidth.

    I think everyone starts with the assumption that voice bandwidth fits
    nicely in 3.3kHz (300 to 3300kHz) or is it 2.7kHz (300 to 3000kHz)
    or, the number that appears in most TXs, 2.4Khz (300 to 2100kHz).

    So then why do receivers supply 2KHz or 1.8kHz narrow SSB filters in
    the receiver. Sure, a narrower bandwidth reduces noise bandwidth but
    if it also reduces intelligibility at the same rate then it's a poor
    trade off. Yet, these filters are common and people seem to use them.

    So if narrow bandwidths are good for SSB reception then shouldn't you
    transmit a narrow SSB signal. After all the receiver is going to
    ignore some of it you can put that transmitted power into parts the
    receiver will use?

    The Icom 756 Pro III has the ability to adjust the bandwidth of your
    transmitted SSB signal by changing the lower bound of the filter
    (100, 300 and 500Hz) and the upper bound (2500, 2700 and 2900Hz).

    <http://www.icomamerica.com/products/amateur/756proIII/>

    Perhaps it's more difficult to listen to narrower SSB for and
    extended period so higher fidelity SSB is more pleasant for rag-
    chewing but less effective for communications closer to the noise floor?

    The Extended/enhanced SSB folks make similar arguments though "Audio
    fidelity" and "intelligibility" are two different measures of a signal.

    <http://www.nu9n.com/essb.html>

    references a white paper (not a review -- it's a pitch) from Polycomm
    who argue extending the telephone bandwidth from 3.5kHz to 7kHz would
    help. But this is ordinary people using a high S/N communications
    system with little QRM. But this is heading in the opposite direction.

    <http://www.nu9n.com/images/Sound.pdf>

    So the question is "What are the trade-offs?" or more precisely:

    What work was done to define an "acceptable" bandwidth?

    How does intelligibility (articulation score/index?) vary with SSB
    bandwidth?

    How does intelligibility (articulation score/index?) vary with
    position of the filter for a given bandwidth SSB bandwidth? For
    example, if you only have 1.8kHz of bandwidth where is it best to
    place it? 300Hz to 2100Hz or 500Hz to 2300Hz?

    How does intelligibility (articulation score/index?) vary with SSB
    bandwidth at a given signal to noise ratio?

    Given "SSB" was used in FDM telephone systems I suspect AT&T Bell
    Labs probably did the early work in the 1940s on acceptable
    bandwidths and I presume arrived at the 300 to 3300kHz for telephone
    systems. but who did the work for radio communications systems that
    have a much lower signal to noise ration and other "interesting
    features".

    Anyone have any insights or references?

    Or another list where this might be discussed?

    Any comments welcome!

    638 2007-04-12 23:58:08 Ed Almos Re: SSB signal distortion in homwbrew receiver
    Try tracing the distortion by working backwards. Inject a low-level
    music* signal into the LM386 and see if it is distorted. If there is
    no distortion then inject your signal into the audio low pass filter.

    Edward Almos HA6SST

    * I menti
    639 2007-04-13 00:05:42 neomag_magneo SSB signal distortion in homwbrew receiver
    Thanks for all the response! Based on the answers, I do not have any
    major suspects on the 2 kHz filter and also the LM386 seems to be
    quite a decent amplifier, after all. In spite of that, I will try
    another one LM1877 (2 x 2 W), which according to the specs has 10
    times lower harmonic distortion. Have to check the correct BFO setting
    also, since that could be very critical with a filter with
    considerably less BW than the received signal.

    What was not mentioned in my earlier message is the testing
    (listening) conditions. Due to my daily schedule, I have done most of
    that in the evening in the 80 mb, which here in the OH country is
    quite congested due to the extremely active UA/RA stations just across
    the border. Since I have noticed better audio quality in the daytime
    and late afternoon, there is still the front stage, particularly the
    mixer, which could be overloading. This is my first trial of a H-mode
    mixer with 3125 gates (as shown in the EMRFD, and there could be
    something wrong. The frequency conversion as such is OK, but the
    strong-signal performance may be compromised, have to set up a
    standard DBM circuit and make the comparison.

    It was the H-mode mixer and its superior performance reported, which
    gave me the original idea to start this project. It would be quite
    ironic (but strictly according the Murphy's law)to find out that the
    rest of the circuitry is just fine and the very first borad having a
    serious deficiency, hihi.

    Thanks again!

    73 de Heikki (OH2LZI)
    640 2007-04-13 10:37:36 Ed Almos Re: SSB signal distortion in homwbrew receiver
    You may have found your problem.

    The input to an h-mode mixer using semiconductor switches must never
    be more than the supply voltage or the P-P voltage of the switching
    waveform. If this happens then there will be distortion.

    Try an attenuator on your receiver input, say 20dB, and see if the
    distorti
    641 2007-04-13 12:08:10 arv evans Re: SSB signal distortion in homwbrew receiver
    Heikki (OH2LZI)

    Placement of the BFO (receive) or Carrier Oscillator (transmit) with respect to your filter passband is interesting to play with.  Usually it will be OK to place the oscillator 12 to 16 db down the filter edge.  Which side depends on whether you are running LSB or USB.  This way the edge of your crystal filter will attenuate the lower audio frequencies (0 to 300 Hz or so) as well as the upper audio frequencies (2400 Hz and upward) that are attenuated by the opposite side of the filter.  Once you have a graph of the filter characteristics you can determine how much AF attenuation at any given frequency will be caused by the filter passband shape.

    In building several BITX20 transceivers <www.phonestack.com/farhan> (and BITX designs for other bands) it has been interesting to play with the BFO/Carrier_Oscillator frequency with respect to the filter passband.  By adjusting the BFO frequency to optimize for the other person's speech characteristics it is possible to make surprising changes in the audio quality and quantity.  The same thing happens when changing the Carrier Oscillator frequency to optimize for my own voice on transmit.

    In building homebrew crystal ladder filters for the BITX series it has been possible to make filters with -6db points of 1.4 to 3.5 KHz.  Narrower filters still work for SSB reception but the speech does not sound "natural" because much of the normally occurring harmonic content is missing.  With decreasing filter bandwidth the amount of audio does decrease, along with decreases in background noise.  It is possible to introduce a small amount of distortion in the speech by making the filter passband highly rippled (by mismatched filter input and/or output impedances) instead of being flat across the top.

    For SSB receiver tests it is possible to build a very low powered DSB generator with just an AF signal generator, an RF signal generator and a toroid and diodes (or a TUF-1) balanced mixer.  This lets you insert any AF frequency you want into the sideband signal to see just how it affects your receiver.

    By connecting receiver output to the sound card in your PC and running an AF Spectrum Analysis program you can see your receiver's audio passband and the presence of any carriers within that passband.

    Arv - K7HKL
    _._


    Ed Almos wrote:

    Try tracing the distortion by working backwards. Inject a low-level
    music* signal into the LM386 and see if it is distorted. If there is
    no distortion then inject your signal into the audio low pass filter.

    Edward Almos HA6SST

    * I menti

    642 2007-04-13 16:34:33 Tim Gorman Re: SSB bandwidth and intelligibility
    Kevin,

    If the only things you are receiving are the SSB signal and something like
    white Gaussian noise (i.e. "hiss"), a wider bandwidth will give you
    significantly increased intelligibility. If you are also receiving other
    masking noise (i.e. QRM) from a different, closely situated signal you can
    many times benefit from removing that masking noise with a narrower filter.
    The intelligibility gain from removing the masking noise goes up faster than
    the intelligibility loss from losing bandwidth.

    tim ab0wr


    On Friday 13 April 2007 01:38, Kevin Purcell wrote:
    > This raises another issue I've been thinking about over the past week
    > or so: how does intelligibility vary with SSB bandwidth.
    >
    > I think everyone starts with the assumption that voice bandwidth fits
    > nicely in 3.3kHz (300 to 3300kHz) or is it 2.7kHz (300 to 3000kHz)
    > or, the number that appears in most TXs, 2.4Khz (300 to 2100kHz).
    >
    > So then why do receivers supply 2KHz or 1.8kHz narrow SSB filters in
    > the receiver. Sure, a narrower bandwidth reduces noise bandwidth but
    > if it also reduces intelligibility at the same rate then it's a poor
    > trade off. Yet, these filters are common and people seem to use them.
    >
    > So if narrow bandwidths are good for SSB reception then shouldn't you
    > transmit a narrow SSB signal. After all the receiver is going to
    > ignore some of it you can put that transmitted power into parts the
    > receiver will use?
    >
    > The Icom 756 Pro III has the ability to adjust the bandwidth of your
    > transmitted SSB signal by changing the lower bound of the filter
    > (100, 300 and 500Hz) and the upper bound (2500, 2700 and 2900Hz).
    >
    > <http://www.icomamerica.com/products/amateur/756proIII/>
    >
    > Perhaps it's more difficult to listen to narrower SSB for and
    > extended period so higher fidelity SSB is more pleasant for rag-
    > chewing but less effective for communications closer to the noise floor?
    >
    > The Extended/enhanced SSB folks make similar arguments though "Audio
    > fidelity" and "intelligibility" are two different measures of a signal.
    >
    > <http://www.nu9n.com/essb.html>
    >
    > references a white paper (not a review -- it's a pitch) from Polycomm
    > who argue extending the telephone bandwidth from 3.5kHz to 7kHz would
    > help. But this is ordinary people using a high S/N communications
    > system with little QRM. But this is heading in the opposite direction.
    >
    > <http://www.nu9n.com/images/Sound.pdf>
    >
    > So the question is "What are the trade-offs?" or more precisely:
    >
    > What work was done to define an "acceptable" bandwidth?
    >
    > How does intelligibility (articulation score/index?) vary with SSB
    > bandwidth?
    >
    > How does intelligibility (articulation score/index?) vary with
    > position of the filter for a given bandwidth SSB bandwidth? For
    > example, if you only have 1.8kHz of bandwidth where is it best to
    > place it? 300Hz to 2100Hz or 500Hz to 2300Hz?
    >
    > How does intelligibility (articulation score/index?) vary with SSB
    > bandwidth at a given signal to noise ratio?
    >
    > Given "SSB" was used in FDM telephone systems I suspect AT&T Bell
    > Labs probably did the early work in the 1940s on acceptable
    > bandwidths and I presume arrived at the 300 to 3300kHz for telephone
    > systems. but who did the work for radio communications systems that
    > have a much lower signal to noise ration and other "interesting
    > features".
    >
    > Anyone have any insights or references?
    >
    > Or another list where this might be discussed?
    >
    > Any comments welcome!
    >
    >
    643 2007-04-13 17:03:07 Allison Parent Re: SSB signal distortion in homwbrew receiver
    644 2007-04-14 04:29:08 w4zcb77 Re: SSB signal distortion in homwbrew receiver
    645 2007-04-14 07:31:16 Giancarlo Re: SSB signal distortion in homwbrew receiver
    Hi Heikki,

    Edward gave you some data on max input signal to the H-Mode Mixer.

    If the distortion is associated to the Mixer itseld then you may have
    problems due to one of the followings:

    1) One or more switches not switching,
    2) Faulty FST3125
    3) Maybe inverted secondaries or primaries of the two transformers
    that are in parallel (I guess you assembled the original G3SBI
    version using 3 transformers) This will also give a high insertion
    loss
    4) Switches bias too low or too high (???)and having a very strong
    input signal, may generate distorsion
    5) If you use homebrew transformers using toroids then you may have
    unbalance of the two in parallel.
    6) Possible fault associated to the divider, uno of the waveforms not
    having the proper amplitude (faulty FF).
    7) Following stages.

    I do not know if you have checked waveforms and signals.

    If you are interested in homemade transformers then I suggest you to
    test my 2T version using fundamental Local oscillator.
    I do not recall if I ever uploaded the info in this newsgroup. I will
    check and if possible I will load a note in english.
    If you have friends members of RSGB you may look for additional
    infos on the H-Mode Mixer using FST3125:
    "I7SWX two-transformer H-Mode Mixer" – Technical Topics – RadCom
    (RSGB – UK) April 2003
    "High Level Mixers" - Technical Topics – RadCom - July 1998
    "H-Mode Mixer with FST3125" - Technical Topics – RadCom -Aug. 1998

    The H-Mode Mixer is used in the projects: CDG2000 and Pic-A-Star.

    73

    Gian
    I7SWX

    647 2007-04-15 23:38:47 neomag_magneo Re: SSB signal distortion in homwbrew receiver
    Hello group!

    It seems that my problems has raised a lot of discussion and many
    helpful items have been brought up. During the weekend despite of the
    20C outdoors temps I was able to do further testing and it seems that
    the 2.0 KHz filter and the excact bfo frequency setting do affect the
    audio quality a great deal. I have been kinda sloppy when adjusting
    the bfo, the one I have is tunable in the second if of 468 kHz.

    Thanks for Giancarlo for the further advice with H-mode mixer, we
    already discussed the problems with toroidal transformers I have used.
    I will certainly try the 2-tranformer design, but it may have to wait
    a little bit longer.

    Once I am done with the modifications, I will report my findings.

    73 de

    Heikki (OH2LZI)
    653 2007-04-17 00:26:26 Greg W:-) Re: SSB signal distortion in homwbrew receiver
    Hi Heikki

    I hope you managed to go outside and get some of the nice spring time
    sun , it was great. :-)

    gregW:-) OH2FFY


    http://www.swdxer.co.nr/
    =======================================================================

    654 2007-04-17 07:40:24 Giancarlo Re: SSB signal distortion in homwbrew receiver
    Just a similar temperature we had in I7-land !!!!

    When I was visiting for business OH3, this time in 2000, the sea was
    still half frozen...hi

    73

    Gian

    I7SWX


    665 2007-04-20 07:37:28 Bruno Re: SSB signal distortion in homwbrew receiver