Jump to content


Welcome to DBSTalk


Sign In 

Create Account
Welcome to DBSTalk. Our community covers all aspects of video delivery solutions including: Direct Broadcast Satellite (DBS), Cable Television, and Internet Protocol Television (IPTV). We also have forums to discuss popular television programs, home theater equipment, and internet streaming service providers. Members of our community include experts who can help you solve technical problems, industry professionals, company representatives, and novices who are here to learn.

Like most online communities you must register to view or post in our community. Sign-up is a free and simple process that requires minimal information. Be a part of our community by signing in or creating an account. The Digital Bit Stream starts here!
  • Reply to existing topics or start a discussion of your own
  • Subscribe to topics and forums and get email updates
  • Send private personal messages (PM) to other forum members
  • Customize your profile page and make new friends
 
Guest Message by DevFuse

Photo
- - - - -

Digital SWM theory and speculation


  • Please log in to reply
236 replies to this topic

#1 OFFLINE   Stuart Sweet

Stuart Sweet

    The Shadow Knows!

  • Super Moderators
  • 36,908 posts
Joined: Jun 18, 2006

Posted 23 December 2013 - 04:21 PM

This thread is devoted to the theory of Digital SWM technology. All participants need to be aware that any technologies represented here may or may not ever appear in any DIRECTV product and are, at the moment, pure speculation. As always, it's fun to imagine, but for real-world applications of Digital SWM technology, visit this thread: http://www.dbstalk.c...and-discussion/

Further, because this thread contains total and complete speculation, it's not off-limits to imagine anything happening with Digital SWM technology, so long as no person represents him or herself as being able to know the future.
Opinions expressed by me are my own and do not necessarily reflect
those of DBSTalk.com, DIRECTV, DISH, The Signal Group, or any other company.

...Ads Help To Support This Site...

#2 OFFLINE   slice1900

slice1900

    AllStar

  • Registered
  • 2,980 posts
  • LocationIowa
Joined: Feb 14, 2013

Posted 23 December 2013 - 04:23 PM

DSWM (Digital Single Wire Multiswitch) is Directv's implementation of what is generally referred to in the satellite industry as dCSS (digital Channel Stacking Switch) At least three dCSS implementations exist as products, which are all capable of DSWM functionality. The industry appears to have settled on 24 output channels as a standard. Subtracting the one Directv dedicates to the guide, this would permit as many as 23 channels per output, unless Directv chooses to limit the number of channels below that.

 

Directv listed some goals for DSWM in both their patent and the paper describing the implementation of that patent. Cost reduction through elimination of expensive analog components, particularly SAWs. Increased flexibility to handle "many more satellites", which likely refers to a single product for US and non-US markets. More output channels. Cost savings by eliminating parts from IRDs (i.e. some parts of up to the entire tuner) Enabling "future products", for "Home Gateway or MDU architectures".

 

Directv's current DSWM products, the DSWM13 switch and SWM 13 LNB utilize the same starting frequency of 974 MHz, with channel spacing exactly half that of the analog SWM, 51.03 MHz versus 102.06 MHz. This was a deliberate choice, and presumably must have been done to permit some forwards and/or backwards compatibility between ASWM and DSWM, but the full extent of such compatibility is not yet known.

 

 

Current DSWM products:

 

DSWM13 (Directv multiswitch) http://www.dbstalk.c...and-discussion/

This is a specialized product targeted at a specific market for hotels, and cannot be used outside of that market. It is limited to 13 channels on a single high powered output for reasons specific to its application. The DSWM13 uses Entropic's EN5400 chip. The DSWM13 is manufactured by WNC.

 

SWM LNB 13 (Directv LNB) http://www.dbstalk.c...ook-swm-13-lnb/

This is a general purpose product targeted at residential and commercial installs with a requirement for up to 13 tuners. It isn't known for sure why it was limited to 13 tuners, but may have to do with difficulties maintaining signal strength when split a number of ways (particularly when DECA is involved) as well as the likelihood that a comparatively tiny number of Directv residential customers have more than 13 tuners. It isn't yet known what chip this uses, but it is likely the EN5500. If that is the case, then this LNB very likely supports RDBS reception.

 

 

Current implementations of dCSS:

 

EN5400 (Entropic Communications) http://www.entropic....oor-unit/en5400

14 500MHz input bands, 20 channels. The existence of at least one competitor was mentioned but not named during Entropic's Q4 2013 earnings call. The EN5400's technology was originally developed by Teranetics under contract by Directv, it was acquired and completed by PLX Technology, then acquired by Entropic, along with the engineers who designed it, in 2012. This is the only dCSS known to be used in current Directv products.

 

EN55x0 (Entropic Communications) http://www.entropic....oor-unit/en5500

The EN5500/EN5510/EN5520 have six/seven/four 250-2350 MHz inputs, with one, three and two 24 channel outputs, respectively. Announced and sampling in April 2014. The EN5500 family is designed for worldwide use, which would account for the slightly higher top end frequency. Entropic claims it is the lowest power dCSS chip available, which allows it to be powered by receivers instead of external power inserter, but such a capability may not be utilized by Directv. The EN5500 was designed by Mobius Semiconductor, and it and the engineers who designed it were acquired by Entropic in 2013.

 

BCM4551 (Broadcom) https://www.broadcom....php?id=s817469

8 250-2350 MHz inputs, 24 channels, built in 28nm CMOS. Designed for worldwide use, allowing both DiSEqC and FSK control.

 

MxL86x (MaxLinear) http://www.maxlinear...m-capture-socs/

MxL865 has 5 250-2350 MHz inputs, 24 channels. MxL868 has 8 950-2150(?) MHz inputs, 3 outputs capable of 24 channels each. Designed for worldwide use, allowing both DiSEqC and FSK control. Maxlinear has announced design wins with Zinwell and Eagle Aspen to use the MxL865, with plans to deliver the chips in Q4 2014, but there is no way to know whether either has anything to do with Directv.

 

unknown name (NXP Semiconductor) http://dx.doi.org/10...CC.2013.6487719

14 500 MHz input bands, 24 channels. NXP was contracted by Directv to design this ASIC, presumably as proof of concept or for internal testing, as there is nothing to suggest it will ever be used in a product. Figures of merit for the ADCs such as ENOB and SQR exactly match those of the second embodiment in Directv's DSWM patent, demonstrating the patent was written based on the design/implementation of this ASIC. It was probably a proof of concept, and may actually be the same chip Directv designed with Teranetics that became the EN5400.

 

 

Links of interest:

 

Analog SWM (US 20060277578 A1)  http://www.google.co...s/US20060277578

RF Magic's patent application that describes Directv's current SWM. RF Magic was later acquired by Entropic Communications, who produces the RF5201 ASWM chip used in all current SWM products.

 

Digital SWM (US8238813)    http://www.google.co...tents/US8238813

Directv patent which describes digital SWM technology from their DSWM ASIC. Figure 6 shows the second embodiment, which exactly matches the implementation described in the paper. Rather than splitting up the signal to individual choose and frequency shift transponders of known size/spacing one at a time as the ASWM does, higher order digital functions allow operating on groups of transponders. The transponders need not have uniform width or spacing, so it can handle Directv's US and non-US satellites equally well.

 

Digital SWM LNB frequency drift estimation/correction (US8509722)    http://www.google.co...tents/US8509722

Directv patent that describes using the DSWM ASIC's DSP to estimate and correct for LO frequency drift (due to temperature or age) in a DSWM LNB, sensing deviations and correcting for it by adjusting the frequency of the DRO (i.e. LO) Additional diagnostic outputs can be generated which aid installation, troubleshooting, etc.

 

Digital Single Wire Multiswitch ASIC (paper from Feb 2013 ISSCC)    http://dx.doi.org/10...CC.2013.6487719

Describes a 45nm 7 metal layer + RDL CMOS implementation of the DSWM ASIC. The die area is not stated in the paper, but based on a die micrograph would estimate it at around 30 sq mm. The paper and associated materials detail 14 input bands in the 250-760 MHz range, 24 output channels, 9.4 watts maximum dissipation.

 

Maxlinear's dCSS (US20130205349 A1)  http://www.google.co...s/US20130205349

Maxlinear's patent application. Probably describes or is similar to the MxL86x family.

 

Mobius Semiconductor's dCSS (US20120189084 A1)  http://www.google.co...s/US20120189084

Mobius Semiconductor's patent application. Mobius Semiconductor was acquired by Entropic Communications in 2013, this likely describes or is similar to the EN5500. See the writeup for Figure 10 for the description of the embodiment related to Directv.


Edited by slice1900, 14 September 2014 - 05:46 PM.

SL5, PI-6S, SA-6AL 3xSWM16, 21 H20-100, 1 H20-600, 7 H24-700/AM21


#3 OFFLINE   P Smith

P Smith

    Mr. FixAnything

  • Registered
  • 20,061 posts
  • LocationMediterranean Sea
Joined: Jul 25, 2002

Posted 23 December 2013 - 04:25 PM

I would ask Mods to move many posts from original DSWiM13 thread to here; all of them contain speculations, theories, etc what would be make the thread is full of relevant posts



#4 OFFLINE   veryoldschool

veryoldschool

    Lifetime Achiever

  • Moderators
  • 41,991 posts
Joined: Dec 09, 2006

Posted 23 December 2013 - 04:27 PM

This thread is devoted to the theory of Digital SWM technology.

In which the basics are a change from analog filtering to digital filtering. This allows for closer spacing of the output channels.


  • FHSPSU67 likes this
A.K.A VOS

#5 OFFLINE   Stuart Sweet

Stuart Sweet

    The Shadow Knows!

  • Topic Starter
  • Super Moderators
  • 36,908 posts
Joined: Jun 18, 2006

Posted 23 December 2013 - 04:29 PM

My apologies, no posts will be moved to this thread as we don't have the ability to change their post dates and there are known issues with moving posts older than the starting time of the thread. Feel free to copy and paste.
Opinions expressed by me are my own and do not necessarily reflect
those of DBSTalk.com, DIRECTV, DISH, The Signal Group, or any other company.

#6 OFFLINE   slice1900

slice1900

    AllStar

  • Registered
  • 2,980 posts
  • LocationIowa
Joined: Feb 14, 2013

Posted 23 December 2013 - 04:30 PM

Replying to inkahauts' post from the DSWM13 thread.

 

But see that's the spot i tend to disagree looking forward three years from now. By then I expect HR24 to become no recover like the HR20 etc. when that happens we will be left with just genies. I have to believe by then they will allow Multiple genies if not sooner anyway. They really could use a lower cost solution for more than 10 channels than a swim16 once that happens IMHO. That gives them three years before it could become a Mainstream issue by my guestimate. While the number of people wanting two genies now likely is small the number if people wanting two DVRs in general is probably higher IMHO significantly. Not a exorbananr amount either way but enough that could Make them want to move to a larger swim. Also porting this to mdu makes sense as well at some point if the cost savings is there. And then my question becomes relevant again. You don't need that kind of power for mdu either correct? Again my question was simply does Dswim actually require higher output as the nature of the beast or is it pumped up do to its actual needs for its specific use in this specific product. Granted I doubt we know for sure. Hopefully we will know latter next month.

 

Everything will become non-recoverable, eventually. But what makes you think they won't keep making DVRs with two tuners? Or that they won't make a new Genie with even more than 5 tuners? They haven't released a new dual tuner DVR since the HR24, but that doesn't mean they never will.

 

Directv would clearly benefit from having a single wire LNB that supports more than 8 tuners, since they hit that limit more often, and that trend seems likely to continue. I doubt anyone would dispute that. However, a DSWM LNB doesn't make sense until it costs them less to install than a legacy LNB, four coax, and SWM16. We can make some pretty good estimate of what the SWM16 solution costs, but can only guess at the cost of a DSWM LNB, or if there are some issues that need to be worked out before that's something they can offer at all, or ever.

 

There's nothing inherent in the DSWM technology itself that requires a higher output. It is doing A to D and then D to A conversion, and the D to A output doesn't require high levels, and even if it did there's no reason they couldn't be attenuated as necessary. Likewise, there's nothing inherent in the analog SWM technology that requires the specific output level it uses. That's determined by the AGC amp at its output. If they had wanted the SWM8 to output at -20 dbm or -40 dbm, they would have chosen the AGC amp and its settings to produce that output.

 

It isn't a matter of "porting" DSWM13 to MDU or the home market. They won't so much use a DSWM13 as a starting point as they'll use the DSWM ASIC as a starting point. Undoubtedly it would share a good portion of its design with that of the DSWM13, but certain things would be different, such as the output level and the number of channels. Once the exact frequencies being used in the DSWM13 are made public, we should have a better idea of what the maximum number of channels the DSWM ASIC inside of it is capable of. From the information I have so far it looks like that ASIC is capable of a maximum of either 15 or 22 channels. A different implementation of the DSWM (i.e. a new ASIC built in a different process) could have different channel spacing from the ASIC used in the DSWM13, of course.


SL5, PI-6S, SA-6AL 3xSWM16, 21 H20-100, 1 H20-600, 7 H24-700/AM21


#7 OFFLINE   slice1900

slice1900

    AllStar

  • Registered
  • 2,980 posts
  • LocationIowa
Joined: Feb 14, 2013

Posted 24 December 2013 - 12:51 AM

I updated my post #2 with links to the patents & paper, along with a synopsis of what the patent describes as I understand it. Corrections for typos, additions, or something you think I'm misreading/misunderstanding appreciated!


SL5, PI-6S, SA-6AL 3xSWM16, 21 H20-100, 1 H20-600, 7 H24-700/AM21


#8 OFFLINE   P Smith

P Smith

    Mr. FixAnything

  • Registered
  • 20,061 posts
  • LocationMediterranean Sea
Joined: Jul 25, 2002

Posted 24 December 2013 - 01:39 AM

..

Thus there are 10 such bands for a SL3 LNB and 12 for a SL5 LNB. .

Please explain details of the numbers by sat/pol/etc


Edited by P Smith, 24 December 2013 - 01:39 AM.


#9 OFFLINE   HoTat2

HoTat2

    Hall Of Fame

  • Registered
  • 5,273 posts
  • LocationLos Angeles, CA.
Joined: Nov 16, 2005

Posted 24 December 2013 - 03:17 AM

Please explain details of the numbers by sat/pol/etc

AIUI P. Smith, the DSWM as described by the patent breaks down the conventional 250-750 MHz, 950-1450 MHz, 1650-2150 MHz triple 500 MHz band LNBF converted transponder band stack into 12 separate 500 MHz bands, each with a beginning frequency selected somewhere in the 10-100 MHz range, thus;

 

For the SL-5;

 

99W

 

1) LNBF 1 - Ka-lo LHCP

2) LNBF 1 - Ka-hi LHCP

3) LNBF 2 - Ka-lo RHCP

4) LNBF 2 - Ka-hi RHCP

 

101W

 

5) LNBF 3 - Ku LHCP

6) LNBF 4 - Ku RHCP

 

103W

 

7) LNBF 5 - Ka-lo LHCP

8) LNBF 5 - Ka-hi LHCP

9) LNBF 6 - Ka-lo RHCP

10) LNBF 6 - Ka-hi RHCP

 

110/119W

 

11) LNBF 7 - Ku 110 + 119 LHCP

 

119W

 

12) LNBF 8 - Ku 119 RHCP

 

Note: The SL-3 is the same excluding LNBFs 7 and 8 of course.

 

These then create the 12 "paths" (or 10 for the SL-3) each fed to a separate 9 bit ADC with a sampling rate of around 1.35 GigaSamples/s


Edited by HoTat2, 24 December 2013 - 03:34 AM.

DIRECTV sub. since Sep. of '95


#10 OFFLINE   veryoldschool

veryoldschool

    Lifetime Achiever

  • Moderators
  • 41,991 posts
Joined: Dec 09, 2006

Posted 24 December 2013 - 08:44 AM

Please explain details of the numbers by sat/pol/etc

There seems to be a bit of smoke and mirrors here.

There are six inputs in both the current analog SWM & DSWM.

With the three bands 250-750, 950-1450, 1650-2150 there are 18 possible 500 MHz blocks that can be used.


A.K.A VOS

#11 OFFLINE   P Smith

P Smith

    Mr. FixAnything

  • Registered
  • 20,061 posts
  • LocationMediterranean Sea
Joined: Jul 25, 2002

Posted 24 December 2013 - 10:54 AM

There seems to be a bit of smoke and mirrors here.

There are six inputs in both the current analog SWM & DSWM.

With the three bands 250-750, 950-1450, 1650-2150 there are 18 possible 500 MHz blocks that can be used.

That's what was my initial impulse ...did decide to ask the 'math' though



#12 OFFLINE   slice1900

slice1900

    AllStar

  • Registered
  • 2,980 posts
  • LocationIowa
Joined: Feb 14, 2013

Posted 24 December 2013 - 01:23 PM

There seems to be a bit of smoke and mirrors here.

There are six inputs in both the current analog SWM & DSWM.

With the three bands 250-750, 950-1450, 1650-2150 there are 18 possible 500 MHz blocks that can be used.

 

The patent doesn't mention flex ports at all, similar to how it (and the base analog SWM patents) leaves them out when describing the analog SWM.

 

However, it does say "while eight inputs 312-326 are shown, a larger or smaller number of inputs are possible without departing from the scope of the present invention". If we do get a 6 output LNB to add RDBS/BSS, the inputs on a future DSWM module might be labeled something other than "flex 1" and "flex 2" anyway...


Edited by slice1900, 24 December 2013 - 01:26 PM.

SL5, PI-6S, SA-6AL 3xSWM16, 21 H20-100, 1 H20-600, 7 H24-700/AM21


#13 OFFLINE   veryoldschool

veryoldschool

    Lifetime Achiever

  • Moderators
  • 41,991 posts
Joined: Dec 09, 2006

Posted 24 December 2013 - 01:28 PM

The patent doesn't mention flex ports at all, similar to how it (and the base analog SWM patents) leaves them out when describing the analog SWM.

Yet this doesn't change "the fact" that they are there and can be used.

You need to relate your patent readings to what is currently in use along with provisions not yet used.


A.K.A VOS

#14 OFFLINE   slice1900

slice1900

    AllStar

  • Registered
  • 2,980 posts
  • LocationIowa
Joined: Feb 14, 2013

Posted 24 December 2013 - 01:33 PM

Yet this doesn't change "the fact" that they are there and can be used.

You need to relate your patent readings to what is currently in use along with provisions not yet used.

 

I was only attempting to relate what the patent states in a clearer manner than the patentese it is written in that most people can't understand well.

 

Since I'd written only about the SL3 and SL5 outputs, that's why I mentioned 10 and 12. I added a note about six additional inputs via the flex ports, or future LNBs with that may have six outputs.


Edited by slice1900, 24 December 2013 - 01:40 PM.

SL5, PI-6S, SA-6AL 3xSWM16, 21 H20-100, 1 H20-600, 7 H24-700/AM21


#15 OFFLINE   veryoldschool

veryoldschool

    Lifetime Achiever

  • Moderators
  • 41,991 posts
Joined: Dec 09, 2006

Posted 24 December 2013 - 01:58 PM

I was only attempting to relate what the patent states in a clearer manner than the patentese

Good luck with that :lol:

I seemed to remember "explaining" the digital filtering for you.


A.K.A VOS

#16 OFFLINE   slice1900

slice1900

    AllStar

  • Registered
  • 2,980 posts
  • LocationIowa
Joined: Feb 14, 2013

Posted 24 December 2013 - 02:58 PM

Good luck with that :lol:

I seemed to remember "explaining" the digital filtering for you.

 

While I understand the computational side of the process quite well, I've never claimed to be an expert on the RF side. You are. That's why I asked for additions & corrections so others can make up for the gaps in my understanding and the patent description in the second post can be expanded and improved by applying expertise from both ends.


SL5, PI-6S, SA-6AL 3xSWM16, 21 H20-100, 1 H20-600, 7 H24-700/AM21


#17 OFFLINE   veryoldschool

veryoldschool

    Lifetime Achiever

  • Moderators
  • 41,991 posts
Joined: Dec 09, 2006

Posted 24 December 2013 - 03:03 PM

While I understand the computational side of the process quite well, I've never claimed to be an expert on the RF side.

Digital filtering is "anything but" RF. :hair:


A.K.A VOS

#18 OFFLINE   slice1900

slice1900

    AllStar

  • Registered
  • 2,980 posts
  • LocationIowa
Joined: Feb 14, 2013

Posted 24 December 2013 - 03:44 PM

Digital filtering is "anything but" RF. :hair:

 

The digital filtering in the DSWM is subject to the limitations of RF when has to go back out as RF and be received by analog tuners, i.e. requirement for skirts / guard bands. That is after all exactly what you "explained" to me about digital filtering...


SL5, PI-6S, SA-6AL 3xSWM16, 21 H20-100, 1 H20-600, 7 H24-700/AM21


#19 OFFLINE   inkahauts

inkahauts

    Hall Of Fame

  • DBSTalk Club
  • 16,285 posts
Joined: Nov 13, 2006

Posted 24 December 2013 - 07:20 PM

Replying to inkahauts' post from the DSWM13 thread.


Everything will become non-recoverable, eventually. But what makes you think they won't keep making DVRs with two tuners? Or that they won't make a new Genie with even more than 5 tuners? They haven't released a new dual tuner DVR since the HR24, but that doesn't mean they never will.

Directv would clearly benefit from having a single wire LNB that supports more than 8 tuners, since they hit that limit more often, and that trend seems likely to continue. I doubt anyone would dispute that. However, a DSWM LNB doesn't make sense until it costs them less to install than a legacy LNB, four coax, and SWM16. We can make some pretty good estimate of what the SWM16 solution costs, but can only guess at the cost of a DSWM LNB, or if there are some issues that need to be worked out before that's something they can offer at all, or ever.

There's nothing inherent in the DSWM technology itself that requires a higher output. It is doing A to D and then D to A conversion, and the D to A output doesn't require high levels, and even if it did there's no reason they couldn't be attenuated as necessary. Likewise, there's nothing inherent in the analog SWM technology that requires the specific output level it uses. That's determined by the AGC amp at its output. If they had wanted the SWM8 to output at -20 dbm or -40 dbm, they would have chosen the AGC amp and its settings to produce that output.

It isn't a matter of "porting" DSWM13 to MDU or the home market. They won't so much use a DSWM13 as a starting point as they'll use the DSWM ASIC as a starting point. Undoubtedly it would share a good portion of its design with that of the DSWM13, but certain things would be different, such as the output level and the number of channels. Once the exact frequencies being used in the DSWM13 are made public, we should have a better idea of what the maximum number of channels the DSWM ASIC inside of it is capable of. From the information I have so far it looks like that ASIC is capable of a maximum of either 15 or 22 channels. A different implementation of the DSWM (i.e. a new ASIC built in a different process) could have different channel spacing from the ASIC used in the DSWM13, of course.


I don't see another two tuner DVR for two reason. Money and direction.

Maybe a three tuner unit utilizing one four tuner chip. And sure we could se a seven tuner unit with two four tuner chips which is rumored To be what the hr44 was built with.

But at this point I'm Not sure you can make a two tuner unit for cheaper than at least a three tuner hunt anymore and still be able to perform at the same ability level as a genie. And I have a feeling internal box performance isn't why the hr44 is five instead of seven tuners. I think it's the other stuff that's limiting it.

#20 OFFLINE   slice1900

slice1900

    AllStar

  • Registered
  • 2,980 posts
  • LocationIowa
Joined: Feb 14, 2013

Posted 24 December 2013 - 08:23 PM

I don't know what sort of tuner chips are out there now, or their cost. It obviously costs more to use more than one like the Genie does, but you may be right that there's not much cost savings to be had if there are newer chips that support more tuners so newer DVRs may have more than two.

 

Regardless of what future DVRs look like, Directv has some good reasons to desire a single wire LNB that supports more than 8 tuners if it can be produced as cost effectively as a legacy LNB + running 4 coax + SWM16 solution.

 

Directv will need to introduce a new LNB relatively soon if they intend to use RDBS/BSS for customer broadcast. Depending on how wide an interest that content has, and what form the legacy version of this new LNB might take (in particular, if it has six outputs) will determine, along with cost, the likelihood of seeing a DSWM LNB at that time.


SL5, PI-6S, SA-6AL 3xSWM16, 21 H20-100, 1 H20-600, 7 H24-700/AM21





Protected By... spam firewall...And...