said by TheWiseGuy:
I believe that they both can bond 8 channels. The max download rates are relatively the same. I believe the 96MHz refers to where in the spectrum the channels can be located. If one channel were at 6xx MHz the other bonded channels for that radio would need to be within the 96MHz range.
New modems are reaching product cycles like new wireless APs/routers at Best Buy: there's a new Cisco Linksys model in the shelf every 3 months
In reality, we seem to be down to product cycles in the 18 month range now (what used to be more like 3-4 years), with new products constantly having reduced part count, lower power consumption, and lower cost: the cable modem landscape (chipsets) is evolving quickly these days.
Multiple tuners only add a little more plant flexibility for an operator, but the true use of that remains to be seen, as backwards compatibility with older equipment is tantamount.
And no, Mayuyu: silly - there's no 12 MHz-wide TV channels anywhere in the world. It's only 6 or 8 MHz (EURO-DOCSIS for the latter, in EU and JP). Just a google search away, you know!
First-gen D3 hardware (4x4 bonding) had single tuners with 60-96Mhz capture bands. That's edge-to-edge, btw., and was quite enough, as operators just had to ensure that four 6-MHz channels (24MHz) fit somewhere into the capture band.
With 8 channel bonding, we're already at 64 MHz required capture band, and that's with all channels stacked back-to-back: not realistic in any systems I've ever seen.
Take OOL for example: 603, 609, 615, 621, 627 and 687 MHz are being used: they do not use 603 in areas where 687 MHz is used for OOL (Brooklyn, according to posts here).
That creates an edge-to-edge capture requirement from 606 (3 MHz below 609, which occupies 606.0 to 612.0 Mhz) to 690 (687 + 3 MHz) MHz - 84 MHz total.
That's too far apart for some first-gen D3 single-tuner designs, but perfectly fine for dual 48 MHz tuners.
I don't think anyone here has posted a bonded channel screenshot showing 687 MHz as part of their bonding group - and the above is likely the reason why.
This under-the-hood improvement in hardware design does not affect performance to the customer, it just enables better plant/network management.
The vastly reduced power consumption on the other hand, should make a very noticeable impact on battery hold time, and that surely does impact the consumer: 8 hrs of battery life should be within reach, given those numbers.
8 vs 4-channel bonding:
Whether you get 5 DS channels or 4 or 3, it's more about managing area-specific demand at this point - and even 4x4 modems will benefit when there's more than 4 channels available: they'll get different *combinations* of the 5 (or more future) DS channels available, and hence spread out with different overlap, just like DOCSIS2.0 modems are now spread out across various (all?) DS channels in some upgraded areas.