[TangerineSDR] PSWS System Specification preliminary Ver 0.1

Scotty Cowling scotty at tonks.com
Tue May 7 14:08:55 EDT 2019


Hi Tom,

Yes, I was dismayed to see that limitation carried over to protocol 2. 
My planned network has one server and multiple clients, something that 
neither HPSDR protocol can handle.

The LCC can live with the discovery limitation, can it not? Since it is 
intended to be Host<-->SDR communications and never be exposed to the 
outside world (for security reasons), it will only ever communicate on 
one subnet.

The RCC, however, should not have that limitation. I am not sure how 
individual stations will "hook in" to the network. Will they have to 
register with a central server? Will they then be polled? Or will they 
push data, once set up by the central server? How is authentication done?

The limitations of the IP stack in the FPGA certainly contributed to the 
limitations. Phil VK6PH implemented it as a state machine. I would not 
do it that way. I would use a soft-core CPU (NIOS II) and a steerable 
state machine to make it more easily programmable. Protocol changes 
would then be mostly (if not all) software changes. They would still be 
FPGA changes, but not low-level hardware changes to state machines.

My intent for the protocol is to provide for discovery, re-programming 
(for updates) and control using UDP with an ACK layer added. The 
"control" part would be for hardware specific control, as well as for 
setting up stream to/from the SDR and external IP addresses. By external 
addresses, I mean the SBC's address, other hosts' addresses on the local 
subnet as well as addresses anywhere else on the Internet. These streams 
are set up by the SBC, but once set up, stream directly between SDR and 
designated IP address via UDP.

We should define several different formats for streams, and leave it 
open to add formats as desired. For example, we could have a 32-bit I/Q 
format (16 bits I and 16 bits Q) a 16-bit raw sample format (16-bit 
samples, back-to-back), a 64-bit 192Ksps I/Q stream (32-bits I and 
32-bits Q), etc, etc. Each stream would be set up with sample rate, data 
width, bit/byte order, packet size, etc. Include a stream format ID (or 
maybe this is implied, since the stream was set up in advance), and we 
can add formats as needed. So we would have a Meta-tagged format that 
included a metadata header with GPS time stamp, timing mark, etc. 
followed by I/Q or raw samples (maybe multiple different formats as 
required).

One other comment on your PSWS spec v 0.2. I still think that the 
delineation between the host and SDR is dictating an SBC-centric 
architecture. It is not clear to me that the "host" and "DE" are 
processes and not hardware blocks.

73,
Scotty WA2DFI

On 2019-05-07 09:58, Tom McDermott wrote:
> HI Scotty - I did not know Protocol 2 had the same single endpoint IP 
> address issues.   The limitations of protocol 1
> made remote discovery impossible without a proxy (the IP address of 
> both devices have to be on the same subnet).
>
> Is it possible that the FPGA intellectual property (IP stack) was 
> responsible for this limitation?  If so, is something
> else without those limitations available for this project?
>
> LCC has a couple of key requirements (listed in the System Spec) 
> needed for PSWS.  First is the need to mark
> which sample is aligned with the timing mark, and identification of 
> GPS time for the frame data.  It's really fundamental
> to the project.  I mention it in section 6, but not in section 8.  
> Probably should update the document for that.
> The other requirements for Section 6 could be handled outside of the 
> data packets in any of several different ways.
>
> Nathaniel sent me the grant he wrote, and some ideas on Ham usage.  I 
> will try to wrap all that and the user cases into
> some kind of document.
>
> Nathaniel also said they are working on finalizing the Magnetometer 
> decision.
>
> -- Tom, N5EG
>
>
>
>
>
> On Mon, May 6, 2019 at 4:31 PM Scotty Cowling via TangerineSDR 
> <tangerinesdr at lists.tapr.org <mailto:tangerinesdr at lists.tapr.org>> wrote:
>
>     Hi Bill,
>
>     I think it is unworkable.
>
>     The HPSDR protocol is hardwired to talk from exactly one SDR
>     (referred to as "SDR hardware") to exactly one Host (referred to
>     as "PC").
>
>     No provision is made for multiple discovery requests, and you
>     cannot stream data to more than one IP address. So all 192K slice
>     receivers would have to stream to the same hardware (same IP
>     address). Command and control definitions are so married to the
>     HPSDR and ANAN hardware that we would probably not be compatible
>     with any off-the-shelf radios anyway.
>
>     While I think we can base the protocol loosely on the HPSDR
>     protocol as a starting point, I think it is too restricted to be
>     of general use. I am working on a more general-purpose protocol
>     for LCC that would eliminate most, if not all, hardware-specific
>     restrictions. We can use some SDR hardware that I have already
>     built to develop the protocol, or use existing TAPR or Apache Labs
>     Hermes boards or even ANAN radios. They all have FPGAs in them, so
>     we can program them to do the SDR end of the protocol for testing.
>     I would be willing to work on the FPGA end if someone wants to
>     work on the Host end so we can get a system up and running.
>
>     My goal is a hardware-agnostic, multiple client, multiple server
>     protocol that would handle streaming of I/Q data, raw sample data,
>     RX audio, TX audio, and control commands, as well as
>     re-programming of the hardware over Ethernet. HPSDR protocol 2 is
>     far from this goal.
>
>     Do you have a networking expert (maybe that would be you?) that we
>     can get in on the definition?
>
>     I don't want to interfere with your progress, but the above were
>     features that I wanted to be included in protocol 2 that were not.
>     I think that the use of different (changeable) ports for every RX
>     stream, but no changeable ports (or limited ports) for other types
>     of streams makes too many assumptions on the hardware that is on
>     each end.
>
>     So can we start with a clean slate? Pretty please? :-)
>
>     73,
>     Scotty WA2DFI
>
>     On 2019-05-06 14:29, Engelke, Bill wrote:
>>
>>     Hello Tom & Scotty:
>>
>>     Nathaniel has given me the go-ahead to start working on the specs
>>     for the software, which I understand to include:
>>
>>     -Functional Specifications for the Central Control system and
>>     related database
>>
>>     -Remote Command and Control Protocol
>>
>>     -Functional Specifications for the local control system (this is
>>     the SBC often referred to as the Host)
>>
>>     I guess a part of this would touch on the Local Command and
>>     Control Protocol (between the Host and Data Engine). Here’s my 2
>>     cents:
>>
>>     If I would suggest  to use openHPDSR protocol (or at least a
>>     subset/superset of it). Reasons:
>>
>>     ·We can stand on the shoulders of the existing, proven design.
>>
>>     ·If somebody has another SDR (e.g., Red Pitaya, Hermes, etc.)
>>     they could use these as the radio for the PSWS.  (Note that the
>>     central database will know what kind of radio every observation
>>     came from, and if you are concerned about the quality of data
>>     from a different type of radio, you can simply exclude from any
>>     analysis that is sensitive to that factor. WE WILL NOT SUPPORT
>>     any of these different radios except to conform to our selected
>>     ported of the spec).
>>
>>     ·We may be able to re-use some existing working code.
>>
>>     ·Maybe a user could use their PSWS as a radio (a receiver, at
>>     least) under PowerSDR if they wanted to.
>>
>>     ·We can get started with software development even before the new
>>     hardware is ready, because there are radios that conform to the spec.
>>
>>     ·I have found the openHPSDR-Protocol-2 online at
>>     https://github.com/TAPR/OpenHPSDR-Firmware/blob/master/Protocol%202/Documentation/openHPSDR%20Ethernet%20Protocol%20v3.8.pdf
>>
>>     If anyone thinks this is definitely not workable, please let me
>>     know, so that I can learn; I expect there are aspects of this
>>     that I am not aware of, so I am open to all feedback…  thanks….
>>
>>     -73- Bill, AB4EJ
>>
>>     *From:*Tom McDermott <tom.n5eg at gmail.com>
>>     <mailto:tom.n5eg at gmail.com>
>>     *Sent:* Sunday, May 5, 2019 8:36 PM
>>     *To:* Engelke, Bill <bill.engelke at ua.edu>
>>     <mailto:bill.engelke at ua.edu>
>>     *Cc:* TAPR TangerineSDR Modular Software Defined Radio
>>     <tangerinesdr at lists.tapr.org> <mailto:tangerinesdr at lists.tapr.org>
>>     *Subject:* Re: [TangerineSDR] PSWS System Specification
>>     preliminary Ver 0.1
>>
>>     Hi Bill - one of the problems when writing a new specification is
>>     that the author has
>>
>>     undocumented assumptions in their mind.  The review process helps
>>     get those discovered
>>
>>     and then written into the spec or some other document.
>>
>>     This thread is good because it has uncovered several of those
>>     that need writing out.
>>
>>     For this project, it is apparent that we need a use case
>>     document, covering the two cases,
>>
>>     and perhaps more later on.
>>
>>     -- Tom, N5EG
>>
>>     On Sun, May 5, 2019 at 1:16 PM Engelke, Bill <bill.engelke at ua.edu
>>     <mailto:bill.engelke at ua.edu>> wrote:
>>
>>         Tom – OK, I get it.  Thanks for being patient with me, I’m
>>         trying to understand as much as possible about this so that I
>>         can properly guide the software development. There are indeed
>>         people / organizations that have major local horsepower, and
>>         it is certainly advisable to ensure they can roll their own
>>         system if they wish to.  -73- Bill AB4EJ
>>
>>         *From:*Tom McDermott <tom.n5eg at gmail.com
>>         <mailto:tom.n5eg at gmail.com>>
>>         *Sent:* Sunday, May 5, 2019 10:23 AM
>>         *To:* Engelke, Bill <bill.engelke at ua.edu
>>         <mailto:bill.engelke at ua.edu>>
>>         *Cc:* TAPR TangerineSDR Modular Software Defined Radio
>>         <tangerinesdr at lists.tapr.org
>>         <mailto:tangerinesdr at lists.tapr.org>>
>>         *Subject:* Re: [TangerineSDR] PSWS System Specification
>>         preliminary Ver 0.1
>>
>>         Hi Bill - Your understanding is one particular
>>         implementation, it's the same implementation that Scotty is
>>         assuming.
>>
>>         Given today's silicon technology it may be the only practical
>>         implementation. Given tomorrow's technology maybe not.
>>
>>         There could be several use cases:  the one you are assuming,
>>         where the remote PSWS is accessed over the Internet. and
>>
>>         where remote-->server bandwidth is a significant limitation. 
>>         For that case Nathaniel laid out the need for the remote
>>         station to
>>
>>         stream to local non-volatile storage, then be triggered
>>         during an event of interest to send a small data subset back
>>         to a central
>>
>>         server.
>>
>>         But there is another use case where the PSWS is located on
>>         the same LAN as the server. Phil Erickson expressed interest
>>
>>         in this case.  For this case perhaps folks would want to
>>         stream continuously over the LAN to that server, providing
>>         greater
>>
>>         data capture capability.  For this use case an inexpensive
>>         Host function might not be able to keep up.  Essentially,
>>         it's replacing
>>
>>         the low-cost host function with one that's a lot more
>>         powerful, perhaps integrated with a local server.   With
>>         software modularity
>>
>>         perhaps the host software could be portable, or maybe that's
>>         phase 2 (or maybe phase never).
>>
>>         So the spec is drafted in an attempt to try not to limit the
>>         use case.
>>
>>         -- Tom, N5EG
>>
>>         On Sun, May 5, 2019 at 7:23 AM Engelke, Bill
>>         <bill.engelke at ua.edu <mailto:bill.engelke at ua.edu>> wrote:
>>
>>             Hi Tom – a couple of notes – My understanding is that the
>>             Host Computer (local control system, SBC) will be
>>             connected to the DE vu gigabit Ethernet.
>>
>>             Also I have a concern about the statement in System Spec
>>             rev 0.2 -
>>
>>             If the DE and Host are physically separate devices, then
>>             the DE may need to stream data directly to the central
>>             server if the Host cannot process fast enough.
>>
>>             There will not be sufficient capacity either in the
>>             central sever nor the user’s internet bandwidth to stream
>>             directly. Very high speed upload is the exception rather
>>             than the rule in the US, and we can’t assume we will have
>>             the funding for a central host able to receive this data
>>             rate from lots of PSWS units at once.  I think we have to
>>             *ensure that the local control system is able to process
>>             fast enough* by how we design the whole system. If I am
>>             missing something, please let me know.
>>
>>             -73- Bill AB4EJ
>>
>>             *From:*TangerineSDR <tangerinesdr-bounces at lists.tapr.org
>>             <mailto:tangerinesdr-bounces at lists.tapr.org>> *On Behalf
>>             Of *Tom McDermott via TangerineSDR
>>             *Sent:* Sunday, May 5, 2019 8:13 AM
>>             *To:* TAPR TangerineSDR Modular Software Defined Radio
>>             <tangerinesdr at lists.tapr.org
>>             <mailto:tangerinesdr at lists.tapr.org>>
>>             *Cc:* Tom McDermott <tom.n5eg at gmail.com
>>             <mailto:tom.n5eg at gmail.com>>
>>             *Subject:* Re: [TangerineSDR] PSWS System Specification
>>             preliminary Ver 0.1
>>
>>             Hi Scotty
>>
>>             The spec says that there is a DE function and a Host
>>             Computer function - it doesn't say how they are
>>
>>             physically connected or implemented.  If that is unclear
>>             then the spec is not written well
>>
>>             enough and other folks will likely make the same
>>             (mis)interpretation, so it should be reworded.
>>
>>             Thanks for idea about the synthesizer.  You are right,
>>             the SL spec does say that the output dividers are all
>>
>>             synchronously reset at power up and can also be reset by
>>             the programmer to re-sync. It also says they
>>
>>             will maintain phase-sync across outputs that are derived
>>             from the same synthesizer.  I will make some
>>
>>             measurements on the Eval Board across the 4 outputs to
>>             see how well they stay aligned.    I hope it's
>>
>>             not susceptible to glitches un-syncing them. That would
>>             be a dog to debug in the field.
>>
>>             -- Tom, N5EG
>>
>>             On Sat, May 4, 2019 at 6:27 PM Scott Cowling via
>>             TangerineSDR <tangerinesdr at lists.tapr.org
>>             <mailto:tangerinesdr at lists.tapr.org>> wrote:
>>
>>                 Hi Tom,
>>
>>                 OK, I was thinking that the four outputs of the
>>                 AD5344 would be our
>>                 Clock Module outputs. One to the FPGA, one to each RF
>>                 Module, one
>>                 external. If you wanted two or more of them
>>                 phase-coherent, just drive
>>                 them from the same AD5344 synthesizer. Then the
>>                 AD5344 output block
>>                 would perform our clock distribution, making an
>>                 additional chip
>>                 unnecessary. The variable frequency provided by the
>>                 AD5344 synthesizers
>>                 would be used to configure clock outputs for
>>                 different kinds of RFMs.
>>                 This would enable us to build a low-cost LowFER RFM
>>                 (100kHz-500KHz) that
>>                 samples at, say, 10MHz if we want to. NCOs are
>>                 totally under the control
>>                 of the DE, so they can be designed as one shared NCO
>>                 or multiple NCOs as
>>                 desired.
>>
>>                 Yes, the whole point it to have the hardware that is
>>                 best suited to each
>>                 task, do that task.
>>
>>                 Implementing a full TCP/IP stack or SSH in the FPGA
>>                 is an expensive use
>>                 of hardware (so let the SBC do it). Implementing
>>                 decimation and slice
>>                 filtering in the SBC puts an unreasonable burden on a
>>                 low-cost
>>                 general-purpose SBC (so let the FPGA do it).
>>
>>                 Tom, it just seemed like you were picking one way
>>                 over another in your
>>                 document by having a DE section 6 and a Host PC
>>                 section 7. I agree that
>>                 we need to specify the tasks, but not necessarily the
>>                 hardware that does
>>                 each one. Hopefully it will become obvious where each
>>                 task is best
>>                 handled when we document the needed steps. And I am
>>                 trying really hard
>>                 to keep the SBC out of the main flow, if possible.
>>                 (Maybe it is not.)
>>
>>                 It was never my intention to allow the user to pick
>>                 any old SBC for
>>                 PSWS. My intent was to be flexible enough that *we*
>>                 could pick from many
>>                 alternatives and pick the one that works best for us.
>>
>>                 All these options will get out of control very
>>                 quickly if we let users
>>                 pick whichever ones they want. It will turn out
>>                 exactly as you describe!
>>                 Just look at the early days of Flex, when users got
>>                 to pick their own
>>                 sound card to use with the SDR-1000. Total mayhem,
>>                 even after Flex TOLD
>>                 them to use "brand X, model Y" cards.
>>
>>                 The options are for US to more easily build something
>>                 that meets the
>>                 requirements that this specification is enumerating
>>                 (and others that we
>>                 decide to support). And if users want to play, that
>>                 is fine. They do so
>>                 at their own skill level. We support only those
>>                 configurations that we
>>                 have engineered to meet specific targets.
>>
>>                 OK, I'm off my soapbox now. :-)
>>
>>                 73,
>>                 Scotty WA2DFI
>>
>>
>>                 On 5/3/19 8:36 AM, Tom McDermott wrote:
>>                 > Hi Scotty  - yes, having one synthesizer output
>>                 feed two clock drivers
>>                 > (one per RFM) would be fine.
>>                 > It's not a problem to have a phase offset, provided
>>                 that offset does not
>>                 > change.
>>                 >
>>                 > I don't think it's a problem to have fixed
>>                 frequency on the ADC clocks.
>>                 > That clock is not used to tune the receiver,
>>                 > the NCO in the DE is used to tune the receiver. So
>>                 the DE can implement
>>                 > one NCO for the synchronous case,
>>                 > and two NCOs for the case where we want the two
>>                 receivers to tune to
>>                 > different frequencies.
>>                 >
>>                 > While the spec describes SBC and DE, those two
>>                 functions could
>>                 > theoretically co-exist on on module (for
>>                 > example the Red Pitaya) where the FPGA does both
>>                 the DE function, and
>>                 > via the CPU-block does the SBC
>>                 > function.  A concern with having the FPGA do both
>>                 is the lack of
>>                 > sufficient computational capability in the
>>                 > limited CPU performance that's implemented on the
>>                 FPGA die.   So the
>>                 > spec tries to separate the functionality
>>                 > without saying where it's physically implemented.
>>                 >
>>                 > I do hope we restrict ourselves to one
>>                 implementation. If folks can
>>                 > choose any SBC, any Receiver, etc. then there
>>                 > will be no end of "The software doesn't work on my
>>                 combination",  "My
>>                 > hardware receivers aren't coherent",
>>                 > "My hardware is overflowing buffers", "My hardware
>>                 doesn't support
>>                 > amplitude calibration"    ad nauseum.
>>                 > The data that is produced risks being largely garbage.
>>                 >
>>                 > -- Tom, N5EG
>>                 >
>>                 >
>>                 >
>>                 > On Fri, May 3, 2019 at 7:08 AM Scotty Cowling via
>>                 TangerineSDR
>>                 > <tangerinesdr at lists.tapr.org
>>                 <mailto:tangerinesdr at lists.tapr.org>
>>                 <mailto:tangerinesdr at lists.tapr.org
>>                 <mailto:tangerinesdr at lists.tapr.org>>> wrote:
>>                 >
>>                 >     Hi Tom,
>>                 >
>>                 >     Regarding the clocks, it seems that the AD5344
>>                 addresses this
>>                 >     problem with the cross-point switch and
>>                 synchronous dividers.
>>                 >
>>                 >     We can feed two outputs with one synthesizer if
>>                 we like, for a fixed
>>                 >     (and very small) phase offset. Routing one
>>                 clock output to two RF
>>                 >     Modules is problematic, especially with LVDS
>>                 outputs. Wouldn't this
>>                 >     work?
>>                 >
>>                 >     Alternatively, we could put a clock
>>                 distribution chip on the DE to
>>                 >     take one Clock Module output and distribute it
>>                 to FPGA, RFM1 and
>>                 >     RFM2, but then we lose the capability to clock
>>                 the RFMs at different
>>                 >     frequencies (for example, when one is a TX and
>>                 one is an RX).
>>                 >
>>                 >     One thing to keep in mind is that the FPGA has
>>                 multiple clock
>>                 >     inputs, so we can route one from each RFM, one
>>                 from the Clock Module
>>                 >     and one from a local oscillator and let the
>>                 FPGA code decide which
>>                 >     one to use. Of course, we would have to either
>>                 let the FPGA generate
>>                 >     the RFM clocks or route multiple clocks to each
>>                 RFM to clock the
>>                 >     ADCs directly. The reason I bring this up is
>>                 that I want to use an
>>                 >     on-board inexpensive oscillator (and no Clock
>>                 Module at all) as the
>>                 >     inexpensive, entry-level SDR. No, the
>>                 performance would not be as
>>                 >     good. But the cost would be much lower than
>>                 with any clock module,
>>                 >     and it would be selectable by simply loading
>>                 the FPGA with a
>>                 >     different image and unplugging the Clock Module.
>>                 >
>>                 >     The section 7 comment was not intended to
>>                 dictate a data flow path.
>>                 >     I just wanted to show that we need (and will
>>                 have) two GbE ports to
>>                 >     use that path in case we want to handle data in
>>                 this way (for
>>                 >     example, if the SBC cannot keep up). We can
>>                 word it in whatever way
>>                 >     makes it clear that we can do it either way.
>>                 Isn't your description
>>                 >     implementation specific also (i.e., data flows
>>                 through the SBC)?
>>                 >
>>                 >     Maybe for PSWS, the data has to flow through
>>                 the SBC? That is the
>>                 >     impression I got, and I am not sure we want to
>>                 restrict the
>>                 >     architecture that way, but we could if that is
>>                 our intent.
>>                 >
>>                 >     73,
>>                 >     Scotty WA2DFI
>>                 >
>>                 >
>>                 >     On 2019-05-03 04:09, Tom McDermott wrote:
>>                 >>     Thanks for the good comments, Scotty !
>>                 >>
>>                 >>     Section 3:  the ADC clocks MUST come from the
>>                 same one synthesizer
>>                 >>     output. If they
>>                 >>     come from two different outputs there is a big
>>                 problem...
>>                 >>
>>                 >>     The clocks to the two ADCs  must be (and
>>                 remain) phase coherent.
>>                 >>     If the ADC clocks come
>>                 >>     from different outputs of the synthesizer,
>>                 then each time the
>>                 >>     synthesizer starts, there could
>>                 >>     be any phase difference between the two.
>>                 Further they could drift
>>                 >>     back and forth relative to one another
>>                 >>     a little bit because of the way synthesizers
>>                 work. Thus the phase
>>                 >>     between them under
>>                 >>     the best case would be +/- 180 degrees.  At 30
>>                 MHz that would
>>                 >>     represent +/- 45 degrees.  This
>>                 >>     means that the baseband sampled signal at 30
>>                 MHz would also have
>>                 >>     an unknown phase
>>                 >>     difference of +/- 45 degrees.  That completely
>>                 wrecks the ability
>>                 >>     to discriminate polarization.
>>                 >>
>>                 >>     There can be a difference in phase between the
>>                 two ADC clocks, but
>>                 >>     it must remain constant over time.
>>                 >>     Any difference will be calibrated out when the
>>                 antennas and
>>                 >>     feedlines are calibrated for phase delay.
>>                 >>     But then the phase differences cannot change. 
>>                 The only way I see
>>                 >>     to do that is to have one
>>                 >>     synthesizer output that is distributed to the
>>                 two ADC clocks.  If
>>                 >>     the ADCs are on different modules
>>                 >>     then one clock signal has to be routed to the
>>                 two of them in some
>>                 >>     manner (parallel, daisy-chained, etc.)
>>                 >>     Maybe there is some other way but it's
>>                 difficult to see.
>>                 >>
>>                 >>     This is one of those things that makes phase
>>                 coherent receivers
>>                 >>     very difficult, and why off-the-shelf units
>>                 >>     have to be carefully evaluated, as virtually
>>                 none of the ones I've
>>                 >>     seen address this problem properly.
>>                 >>     It's the Radio Astronomy problem.
>>                 >>
>>                 >>     Section 7:  The spec attempts to be
>>                 implementation-non-specific.
>>                 >>     Forcing a particular method
>>                 >>     to distribute Ethernet data may eliminate all
>>                 other potential
>>                 >>     solutions.  The approach you outline
>>                 >>     is a really good and elegant solution, but the
>>                 spec should not
>>                 >>     mandate it (it should allow it).
>>                 >>
>>                 >>     Section 8:  great comment.  I will restructure
>>                 as you outline.
>>                 >>
>>                 >>     -- Tom, N5EG
>>                 >>
>>                 >>
>>                 >>
>>                 >>     On Thu, May 2, 2019 at 5:27 PM Scotty Cowling
>>                 via TangerineSDR
>>                 >>     <tangerinesdr at lists.tapr.org
>>                 <mailto:tangerinesdr at lists.tapr.org>
>>                 <mailto:tangerinesdr at lists.tapr.org
>>                 <mailto:tangerinesdr at lists.tapr.org>>>
>>                 >>     wrote:
>>                 >>
>>                 >>         Hi Tom,
>>                 >>
>>                 >>         Here are my comments on your excellent
>>                 document.
>>                 >>
>>                 >>         73,
>>                 >>         Scotty WA2DFI
>>                 >>
>>                 >>         Notes on PSWS Specification V 0.1 5 May 2019
>>                 >>
>>                 >>         Section 3.0
>>                 >>         Clock module will have 4 *programmable
>>                 clock* outputs:
>>                 >>         1. FPGA
>>                 >>         2. RF Module #1
>>                 >>         3. RF Module #2
>>                 >>         4. High-speed Reference Clock
>>                 >>
>>                 >>         In addition, two more outputs:
>>                 >>         5. 1 PPS timing
>>                 >>         6. 10MHz fixed reference
>>                 >>
>>                 >>         Clock outputs should be differential LVDS.
>>                 Single ended clocks
>>                 >>         will be
>>                 >>         too noisy, especially across a connector
>>                 boundary.
>>                 >>
>>                 >>         The FPGA should not provide the ADC
>>                 clocks, they should come
>>                 >>         directly
>>                 >>         from the clock module. The ADC may supply
>>                 a source-synchronous
>>                 >>         data
>>                 >>         clock to the FPGA.
>>                 >>
>>                 >>         Section 4.0
>>                 >>         Magnetometer interface can be I2C, SPI,
>>                 serial UART or RS-485.
>>                 >>
>>                 >>         The magnetometer will almost certainly
>>                 need to be remotely
>>                 >>         mounted. In
>>                 >>         this case, RS-485 is recommended. We can
>>                 specify two-wires for
>>                 >>         communictions and two wires for power
>>                 (typically 5V).
>>                 >>
>>                 >>         Section 5.0
>>                 >>         It is not immediately clear that *each* RF
>>                 Module has two
>>                 >>         synchronous
>>                 >>         channels.
>>                 >>
>>                 >>         Note that the *pluggable filter* can be
>>                 bypassed with a
>>                 >>         jumper, making
>>                 >>         it optional. Maybe call it "optional
>>                 pluggable filter".
>>                 >>
>>                 >>         Section 6.
>>                 >>         The DE shall also be capable of sourcing
>>                 or sinking UDP data
>>                 >>         streams
>>                 >>         to/from any IP address, under direction of
>>                 the host processor.
>>                 >>
>>                 >>         The DE will have a three-port GbE switch,
>>                 connecting the FPGA,
>>                 >>         host PC
>>                 >>         and external network. How do you want to
>>                 explain this?
>>                 >>
>>                 >>         Section 7.0
>>                 >>         Processing the stream from the DE is
>>                 optional, since it may
>>                 >>         not be able
>>                 >>         to process this much data. The metadata
>>                 tasks will fall to the
>>                 >>         DE in
>>                 >>         this case.
>>                 >>
>>                 >>         The host will not always transmit the data
>>                 to the central
>>                 >>         server. The
>>                 >>         host may direct the DE to stream data
>>                 directly to the central
>>                 >>         server if
>>                 >>         it cannot process the volume or rate of data.
>>                 >>
>>                 >>         Section 8.
>>                 >>         I have been using "Command and Control
>>                 Protocol" for the
>>                 >>         protocol used
>>                 >>         between the central server and the host
>>                 PC. I have been using
>>                 >>         "Local SDR
>>                 >>         Protocol" for the protocol between the
>>                 host PC and the DE.
>>                 >>
>>                 >>         We could use "Remote Command and Control",
>>                 or "RCC" for the
>>                 >>  host<-->central server communications, and "Local
>>                 Command and
>>                 >>         Control"
>>                 >>         or "LCC" for host<-->DE communications.
>>                 Whatever we use, we
>>                 >>         should
>>                 >>         define it.
>>                 >>
>>                 >>         The Remote Command and Control section
>>                 seems to be missing
>>                 >>         (although you
>>                 >>         refer to it once in section 10). Even
>>                 though we don't know
>>                 >>         what the
>>                 >>         protocol    is, I think it should be
>>                 mentioned (as defined an
>>                 >>         a separate
>>                 >>         document?)
>>                 >>
>>                 >>         I think we need to make a clear
>>                 distinction between Remote C&C
>>                 >>         and Local
>>                 >>         SDR C&C. We will need security and maybe
>>                 encryption on the
>>                 >>         Remote C&C,
>>                 >>         but not so much on the Local SDR protocol.
>>
>>
>>
>>                 -- 
>>                 TangerineSDR mailing list
>>                 TangerineSDR at lists.tapr.org
>>                 <mailto:TangerineSDR at lists.tapr.org>
>>                 http://lists.tapr.org/mailman/listinfo/tangerinesdr_lists.tapr.org
>>
>
>     -- 
>     TangerineSDR mailing list
>     TangerineSDR at lists.tapr.org <mailto:TangerineSDR at lists.tapr.org>
>     http://lists.tapr.org/mailman/listinfo/tangerinesdr_lists.tapr.org
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.tapr.org/pipermail/tangerinesdr_lists.tapr.org/attachments/20190507/b0df00cf/attachment-0001.html>


More information about the TangerineSDR mailing list