[Ground-station] Need advice on DVB-S2 modem

[Wally Ritchie]

  Several P4G members have played with the SR-1 from Ayecka. These are
based on the STV0910ADB (an enhanced version for GSE, etc.). They have a
couple of other models with different capabilities - some which seem to be
spec'd down to less than <200k. These are reasonably inexpensive for a
large ground station but probably too expensive for ordinary widespread
use. For P4G we focus only on the GSE or raw BBFRAME level.

The modem ASIC chips are generally characterized by the chip manufacturer.
The technical information and the test results are almost always
proprietary and under NDA. But they establish the tested and verified
performance of the ASIC, under some set of test conditions and often in
combination with other members of the chipset (like tuners). When you go
outside these boundaries you are in unverified territory so the
test/verification burden now shifts to the modem vendor that incorporates
the chips. You generally can't count on this for low cost modems. But for
military and commercial this data is generally available. Verification and
testing can be a large component of the engineering effort and that is
reflected in the price for such equipment.

For most DVB-S2 the analog system ends with IQ inputs to the ASIC from the
tuner. The tuner has filtered these to some bandwidth, typically
programmable from about 5 to 45MHz or so in the tuner. When there are
multiple signals within this IQ baseband, they have to be filtered
digitally in the ASIC. The performance question for many application is how
well it performs given adjacent signals that may be equal or stronger than
the desired signal.

The motivation for reduced symbol rates, as you know, is to improve the
C/N. It's being assumed that you are reducing the noise bandwidth in order
to realize the improvement. But the noise bandwidth can be reduced only by
filtering out the noise outside the bandwidth. And this may also include
interfering signals. So to verify performance, the modem manufacturer needs
to be able to generate both desired signals, undesired adjacent signals,
and noise. None of this is particularly hard to do. But it's all necessary
to verify that the system will perform as expected under the expected
conditions.

If you have no adjacent signals as you are the only thing in the 5MHz
passband, you still have to contend with the noise bandwidth. The noise in
5MHz passbnad is the same whether the signal is 100kHz or 5MHz. So the C/N
is the same. You can improve the C/N ONLY if you filter the noise to a
smaller bandwidth. Only then can you pick up that the 17dB improvement from
going from 5MHz to 100kHz.

With DVB-S2 there are a lot of tradeoffs between modulation and coding as
well as bandwidth. During verification you can measure that you get the
expected performance changes with changes in only modulation and coding and
with only changes in bandwidth. Not seeing the expected changes can
indicate that the filtering necessary to achieve the improvement possible
with reduced bandwidth is not present in the system (or overwhelmed by
other factors).

QO-100 is a fantastic testbed for this, at least for non-Doppler signals.
QO-100 wideband transponder has an 8MHz bandwidth, a strong 2MHz wide
DVB-S2 carrier and other experimental signals, many of them relatively
narrow. So some real world testing can be done through QO-100.

As to Doppler, the receiver should not have a problem tracking it. But to
verify the performance in the presence of Doppler, the modem has to be
verified both with or without any compensation that you are introducing.
Generally, it is best to provide the simulated Doppler in a separate mixer
stage such that it does not interfere with any tracking mechanism present
in the primary signal chain. The compensation, if any,  should be applied
in the same manner as it will be in the final system. As you indicate that
it is the plan to get Doppler correction from ephemeris data, the
verification will generally need to simulate the errors in that process as
well. It should be noted that the ASIC chips generally have a number of
internal parameters that affect the tracking loops of which there are
usually several (course, fine, symbol etc.). These are normally a set and
forget (or ignore) thing in normal DVB-S2 applications. But they do allow
adjustments that can affect the Doppler performance. There are
configurations where higher performance may be obtained with Doppler
correction and tighter tracking, especially in the presence of adjacent
signals.

There is some modem work underway on a low-cost STV0910ADB based modem with
additional IQ input baseband filtering. Some of this might become
applicable in your time-frame.

Your team will find lots of support and assistance on the Phase 4 Ground
slack channels. There is also a lot of historical discussion there that
might be of use to your team.

WU1Y (Wally)




On Sun, Dec 8, 2019 at 8:09 PM Mike Parker via Ground-Station
<ground-station at lists.openresearch.institute> wrote:

> A team headed by the University of Arizona is working on a 6-U satellite
> that will fly at 500 km altitude.  The satellite, named CatSat after the
> UofA wildcat mascot, was described in a paper presented at the Amsat Annual
> Meeting.  We are shooting for a launch that could be as early as one year
> from now.  The plan is to have a downlink using DVB-S2 modulation generated
> by an FPGA on the AstroSDR card provided by Rincon Research.  This
> modulation was chosen so that hams that have been working phase4 ground
> DVB-S2X might receive and demodulate the link.  We are planning to transmit
> a limited number of ModCod’s.  Modcod 7, QPSK 3/4, and modcod 17, 8-PSK
> 9/10 are likely choices.
>
> Two experiments plan to use DVB-S2 modulation on the downlink.  One
> involves transmitting high quality video, and the other will capture and
> retransmit narrowband pieces of the HF ham bands containing WSPR and FT8
> signals.  Closing the link to our 6.1 meter diameter ground station is no
> problem, and we are planning on a modulation bandwidth up to 20 MHz and bit
> rates over 50 Mbps when using the satellite’s inflatable directive antenna
> provided by FreeFall.  But closing the link to a ham 0.6 meter dish when
> operating with an onmi-directional transmit antenna poses a very different
> problem. We will likely require small downlink bandwidths on the order of
> 200 kHz.
>
> We are looking for a demodulator to assist in testing the satellite, and
> also a demodulator for use in the ground station.  So we are reaching out
> to you for suggestions.  We hope that an available commercially for an
> affordable price, or perhaps something designed by a member of this group.
> We have neither the desire or time to reinvent the wheel.
> Several things concern us we would appreciate advice on.
> 1) We read with interest a paper by Downey, Evans, and Tollis, “DVB-S2
> Experiment over NASA’s Space Network”.  It said “ typical commercial DVB-S2
> receivers are not designed for symbol rates below 300 kbaud”.  That is
> consistent with our observation that many commercial demodulators do not
> seem to have a lowest bandwidth specification.  Anyone know of one that
> goes lower in bandwidth while having a high bandwidth capability?
> 2)  We need to have a demodulator that will output raw DVB-S2 frames,
> bypassing any transport layer protocols which are normally used with DVB-s2
> such as Multi-protocol Encapsulation(MPE) or Geeric Stream Encapsulation
> (GSE).  Downey, Evans, and Tollis used a Newtec MDM6000 modem.   Is there a
> better or cheaper solution? (I haven’t priced one yet).
> 3) Doppler shifts are also a concern, especially at a low data rate, but
> we have a plan to solve that if necessary using a local oscillator in the
> ground station that is swept according to ephemeris predictions to
> de-Doppler the signal before demodulation.
>
> Oh yes, some good news.  The first of our 6.1 meter dishes has been
> reassembled in Rincon’s parking lot in Centennial, CO.  A picture with
> assembly in progress is attached.  I’m flying up tomorrow with a 10 GHz
> feed horn and LNA to see if we can hear signals!
> Mike Parker, KT7D
> [image: image001.jpg]
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openresearch.institute/pipermail/ground-station-openresearch.institute/attachments/20191209/5f2ee13e/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: image/jpeg
Size: 61792 bytes
Desc: not available
URL: <http://lists.openresearch.institute/pipermail/ground-station-openresearch.institute/attachments/20191209/5f2ee13e/attachment.jpg>