Separate names with a comma.
Discussion in 'DIRECTV General Discussion' started by west99999, Oct 31, 2013.
Anybody got any info on the new SWM13 switch that is coming out soon?
Halloween reference of some kind ??
No it is a real SWM switch. I just don't have much info on it and wanted to see if anyone else did.
It is for DRE and D2
Click for large view - Uploaded with Skitch
Funny to see a '13' thing on Halloween, in an electronic device no less. Engineers being big on 2, 4, 8, 16 etc., LOL.
Wonder how they did it to make room for 5 more SWiM carriers?
Reduced the SWiM carrier spacing below the current ~102 MHz?
Went to higher carrier frequencies than 1790 MHz?
Some of both?
Yes, interesting ...
It is interesting they're calling it a "Digital SWM". That implies it is somehow different from a regular SWM, beyond having 13 channels. Currently a legacy LNB frequency shifts satellite signals from 10-20 GHz range to lower frequencies that regular coaxial cable can carry, and stacks them onto 4 cables to make them easier to deal with. The SWM is sort of a second stage LNB, in that it does more frequency shifting and stacks individual transponders onto a single cable, making it even easier to deal with. But it still working only with analog RF, which needs to be decoded by a tuner for the original digital data to be recovered.
I've posted a few times that I thought Directv would eventually take the tuners out of the receivers and put them in the SWM. If you tune them in the SWM and deliver decoded digital data - which wouldn't even need to travel via coax - then everything is a client of the SWM. Even DVRs.
Maybe that's what they've done here, and you have 13 (14, if you include the guide channel) tuners in the SWM and there are no SWM channels, as such, but instead 13 DECA channels that are accessed by attached receivers and DVRs via RVU. Hotels would love that, if they had RVU capable TVs they wouldn't even need receivers!
That's my hunch at least, we'll see if I'm right or wildly off the mark
The "why 13" would be that they probably have implemented 16 tuners on a single chip (or maybe 8 on two chips, I don't know how big they are) Tuning is a parallel task, which would be implemented with sort of a semi-custom DSP (digital signal processor) Such highly parallel functions are often designed with 'spares' to account for defects in the chip fabrication process.
Those who are familiar with the specs of higher end GPUs will be familar with this, the not quite high end models from ATI and Nvidia are the same chip as the high end models with a few defective areas (cores, clusters, whatever marketing chooses to call them) disabled, because otherwise they'd be tossing out a lot of almost good chips to get a small number of fully functional ones, and high end GPUs would cost 10x as much as they do.
So if you need 14 (13 + guide channel) tuners to work, you design for 16. Two can be bad in manufacturing, or go bad later, and get disabled and you still have fully functional hardware.
13 would seem to be a very good number for a later SWM LNB residential version of this, to avoid needing SWM16 upgrades and running a bunch of extra coax when customers add a Genie to their account.
One more thought, this might be what Directv was talking about in late May when they made that comment about coming up with better solutions for customers receiving OTA broadcasts. Tuning satellite and tuning OTA ATSC broadcasts isn't the same, but they have enough in common that adding ATSC demodulation capability to a satellite demodulator isn't going to double the complexity, or even increase it by 50%. Perhaps much less than that.
The really low down hard RF stuff would be on a separate chip from the DSP anyway, because analog RF and high speed digital logic don't 'play well' together. Likely there would be separate chips for satellite and ATSC for that part of the process, since I don't think they'd share much there (but I could be wrong, I know the digital side but not the analog side) They'd share the DSP on the back end, it would merely run a different program to deal with the rest of the job for ATSC demodulation versus satellite demodulation. It just depends how much of the process is the kind of ugly analog stuff that a DSP wouldn't want to share silicon with.
I'm pretty sure that ATSC licensing is done per product, not per tuner, so it wouldn't cost Directv all that much more to make the DSWM13 capable of tuning 13 ATSC channels at once versus just one or two. 13 is probably overkill, but they could do enough that hardly anyone would ever run into issues with lacking ATSC tuning capacity. One model of the DSWM13 could include the ATSC analog side chip(s) and have an OTA in port to allow it to tune ATSC channels. You'd receive those via DECA/RVU just as you can today using a Genie client where the Genie has an AM21. Hopefully it'll be capable of scanning for channels
The one hitch to a DSWM LNB with OTA capability would be people with indoor antennas - how to get the signal back to the DSWM's tuner? The only way I can see would be if the DECA channels the DSWM used were higher in frequency than the current range, to allow carrying the 50-700 MHz ATSC from the user's antenna to the DSWM.
Hopefully it is cooler.
I googled for "digital SWM" and found a couple patents regarding it (as well as some guy's resume, but the link is dead)
US8509722 Frequency drift estimation for low cost outdoor unit frequency conversions and system diagnostics
US8238813 Computationally efficient design for broadcast satellite single wire and/or direct demod interface
Here's the overview from each:
I only did a quick scan of these patents, there's a lot of technical stuff in there that's way beyond my meager RF knowledge. Interesting that it is dealing with entire 500 MHz blocks at a time! I suppose in the far future this would allow Directv to use their 500 MHz frequency bands as a single block, without any spacing between transponders as currently implemented. I think I read somewhere that Directv is using 250 MHz wide transponders for some uplinks, so the technology is there. Taking advantage of that currently wasted guard band space between transponders would be many years away however, as everything would have to be replaced by Digital SWM for that to happen. However, maybe this would be done for new bands that might need new equipment anyway, such as the RDBS bands on D14 and D15.
It sounds like what I was describing, but this says they're doing it entirely digitally! I wouldn't have thought that could be close to cost effective, but maybe that's why it is targeted at DRE first, these things may make a SWM32 look like a bargain by comparison.
I'm sure someone who knows way more about RF than I do, such as VOS, would get a lot more out of an in-depth reading of these patents than I would. It does include an interesting overview of how SWM currently works, as well as how digital SWM is done differently.
I also found an asset purchase agreement where Entropic Communications (who invented the technology behind SWM) purchased Digital SWM ASIC IP from PLX last year.
I think they tightened up the channels and fit more in. I don't think its anything beyond that by I guess time will tell. And I suspect if it goes
Well we could see them in lnbs at some point.
Unless those patents aren't what the DSWM13 is, they did WAY more than tightening up the channels.
Looking at the DRE certification training ad slick from a couple of years ago, it recommends coursework/experience in Linux.
Thinking back, it seems like the theoretical practical limit of SWM was 11 channels.
What are the implications for existing equipment in the field?
If it's simply tightening up the current SWM channels or cramming more in there, what's the possibility that existing tuners can be modified by software (either changing their baseband or the software higher up the stack in the receiver) to be compatible?
If it is a wholesale re-do of SWM, including moving to digital processing of the signals, I wonder how flexible the current tuners are to use this as well?
I do not see existing equipment in the field being an issue. It was sated earlier that it was for DRE and D2 Advantage which are hotel and MFH technologies. Just like the old HR20i for MFH3, coming up with a new receiver for those markets would not be an issue.
Ok, I got it, I didn't pick up that those are MDU only. That brings up another question, though, I thought one reason/advantage for doing SWM in MDUs is that you can use the same equipment? Does this mean MDU users will slide back into the problems where a really old receiver is their only option?
The output of the DSWM is pure digital MPEG data, same as what a Genie's tuner decodes from the transponder embedded within the SWM channel which it sends to a client via DECA. So in theory any DECA capable receiver should work. There's no requirement for coax, in fact in hotels I suspect it is probably going to use TCP/IP over cat5 more often than coax and use wired C41 clients (or RVU capable TVs)
Based on a DECA "mesh test" maxing out at 250 Mb/sec and a full HD channel bandwidth of around 19 Mb, that divides out to 13 channels fitting within the current DECA point to point bandwidth capability (plus whatever is left over for the guide data stream) Maybe that's just a coincidence, since DECA as currently defined has only 3 50 MHz wide channels, but... More likely it increases the frequency range for DECA. 10 or 11 more 50 MHz DECA channels requires 500 or 550 MHz, and there's plenty of room in the sub 900 MHz range DECA inhabits. The implementation chosen between those two depends on the flexibility of the receiver's DECA hardware.
The patents indicate they have been working on this Digital SWM technology for some time, and likely waiting for the technology to mature to the point where the cost was reasonable. I'd say at the very worst it would be limited to the most recent SWM only equipment, or maybe they had enough forethought to build in capability long enough ago that all still supported receivers are included. Given that Directv knew they were going in this direction for a while, they'd be able make whatever accommodation to their hardware and software implementation of DECA to provide for DSWM and make as much of their current installed base DSWM ready as possible.
BTW, I found a few references in Linkedin to Digital SWM, which indicates that the ASIC was implemented in 45nm CMOS. That's one (and soon to be two) generations behind the state of the art, so there's ample room to drive improvements in cost/performance when they migrate this from DRE to residential since it is a fully digital design with no analog component implemented in Gallium arsenide to hold it back. The DSWM will be able to follow Moore's Law on down to higher performance and/or cooler and/or cheaper hardware every two years like other CE products. Just look at how much faster this year's iPhone is than the original six years ago.
The beautiful thing about this for hotels is that there is no reason that the DSWM's ASIC can't be programmed to do more than decode Directv. Decoding ATSC, QAM or even NTSC should be possible with the same hardware, so it'll be nice for adding in local OTA channels or modulated internal channels. Given that it handles a 500 MHz block at once, this thing should be capable of tuning the entire ATSC spectrum simultaneously...that's pretty cool! Since there's no additional hardware needed, it won't cost anything to provide this functionality, aside from the patent licenses needed for ATSC and QAM, and of course the developer effort. That's not likely to be a high priority for Directv, but I'll bet it is on someone's 'to do' or 'nice to have' list.
I found more info on this, which would tend to make me think it isn't necessarily using DECA - at least it would also have the option to use ordinary SWM. The ASIC was described in an ISSCC paper earlier this year, and the intro (which is all I saw, the rest is behind a paywall) states the ASWM [i.e. the SWM8/16/LNB of today] will "select and output signals from multiple LNB's to a fixed frequency slot for each set-top box on a single cable". That's the "frequency shift a transponder into a SWM channel" model we have now. In the DSWM "selected and reordered channels are digitally upconverted and are stacked into L-band through a 9b 5.46GS/s DAC" and "offers a full flexibility of channel selection, and supports many more satellites through a single L-band cable".
So as I previously outlined, the DSWM is digitizing the individual channels it wants, but this indicates it is then capable of digitally frequency shifting and converting them back to analog. ["DAC" means "digital to analog converter"] I had almost suggested this as a possibility in my last post, but I'm already writing such long posts I'm probably causing most people's eyes to glaze over already Analog L band - that would mean the DSWM could be utilized by any SWM capable receiver, even the venerable H20.
However, unlike current SWM, since it is working at the level of individual Directv channels, it could theoretically fit at least 40 unique HD channels in SWM this way (based on 5 HD channels per transponder) Maybe well over 100, depending on how receivers handle the SWM guard band. It would do this by placing channels intended for multiple receivers into shared SWM channels. Just like SWM does now, when two receivers happen to tune channels available from the same transponder, but it could cause as many receivers to share a single SWM channel as there is room for the channels they want. Not sure where the apparent limit of 13 in the current DSWM-13 would come from, unless it is simply limited by power/heat budget at this time. If so, that will improve with time.
One other thing stood out with these two quotes, first about DSWM: "supports many more satellites through a single L-band cable" and the second regarding the limitations of ASWM: "Although the ASWM solution solves the problem of multiple cables, it does not provide enough flexibility and capability to increase the number of inputs". That makes me wonder if they're planning ahead for adding even more capacity. Perhaps soon enough supporting RDBS on D14 and D15 would require two more coaxial cables, as has been discussed as a possibility in the D14 speculation thread. If you can support 40 unique channels in a DSWM (or even half that many) you don't have to worry about annoying residential (or small commercial) customers with six cables. Only larger MDUs would have to deal with that, everyone else would have a DSWM LNB and a single coax.
I didnt see any way they'd ever mess with the deca freq with the swim channels because then that'd screw up their Whole Home Service.
Well I have to wonder. Are they looking at the future? Maybe they will try as move deca to a offerer freq range up a bit higher and let over the air be diplexed in again? Also expand deca range to cover moca2 specs? So more nodes possible? Could that be the reason they are limiting it to 13 channels. Or maybe the equipment can't deal with more than 13 channels of swim. Don't know.
That would be perfect if they could run 40 channels of swim though which would equate to 40 channels on a lnbswim. Would Make everything easier for everyone. But even 20 channels would be plenty for most consumers.