I took the liberty of copying your post to which I am responding at the RB Club board. It sparked an interesting discussion between compewter doctor [who is someone that just might be a regular writer here, wink, wink, nod, nod] and sfx2000.
What does this have to do with Nokia? Not much I'm afraid. Might give Puck some pause with respect to his 1xRTT/GPRS predictions. For the benefit of all who are interested in the issue, I am again taking the liberty of copying the discussion between sfx2000 and computerMD, as follows:
____________________________________________________
Sfx, Rich, Vhuerta: I am convinced that the msm5000 chip did not work properly and that it required a new chip design (msm5105) for working handsets. It is clear that 1X is late, by anywhere from 6 months to a year. Qualcomm had hoped to produce significant numbers of msm5000 in the 4th quarter of last year. Interestingly, to date there seems to have been almost as many csm5000s sold as handset 1x chips. Sulpezio said 300,000 CSMs were sold last quarter (vs. perhaps less than 500,000 1X MSMs [anyone have a number here?]).
Qualcomm is late on this one. Qualcomm is a great company and very reliable. However, as a colleague of mine said "if you don't miss at least one plane a year you're spending too much time in airports". I would be disappointed if Qualcomm never missed a milestone date: The goal is not to make milestone date but to increase your competitve advantage (run faster in a race, not run at a given pace).
But it is very clear that there was a sea change at Qualcomm (and its partners) around February of this year, culminated with announcements from Verizon, Sprint, Nextel, and others with launch dates this year. That signaled that all the showstoppers had been solved, and there was left only the more mundane and predictable tasks of building it. It is no accident that this point came only after early availability of the msm5105. This was the key issue. [This insight is significant and important. Worth many, many times the price of a yearly subscription to SI. Means that the stock price will rocket later this year. CDMA 1x/ 1xRTT is the real deal and it's going to go gangbusters. Puck, take heed. C2]
The next milepost we will see is the launch, in quantity of 1x handsets by operators in advance of 1x infrastructure. This signals not only the availability of working handsets from multiple vendors, but also a strong belief by the operators that there are no problems with the handsets. Qualcomm only expects to ship 6 million 1x chips this year, so the earliest qualcomm will be shipping a majority of 1x chips is 1st quarter of next year (I am talking always about fiscal years here). That means the big rollout of 1x phones is slated for 2nd quarter (winter), since it takes about 10 weeks to manufacture a phone.
So, the reports from Korea may have been overstated. And there may have been non-handset problems (as I stated in my previous post). But, it is not a network build out problem since 1x phones are downward compatible so should work just fine on legacy networks. In other words, any problems they were having is with the technology.
--the doctor
____________________________________________________________
doc...
Sfx, Rich, Vhuerta: I am convinced that the msm5000 chip did not work properly and that it required a new chip design (msm5105) for working handsets.
I disagree. The MSM5000, according to QCOM's docs, was based on the older MSM3000 architecture, and it has been said by many sources that the 5000 was meant as a test platform so that vendors could move forward. QCOM's test tools group is still offering test phones that I suspect are based on MSM5000 for sale, at least according to their WWW site.
qualcomm.com
How much are you willing to bet that this is the 1x reference phone to which all are compared for interoperability with various network configurations (LU, NT, MOT, Samsung, others), as well as the benchmark for all handsets to be compared.
What is important about this link is that it shows the age of the chip. See the form factor of the phone? Same as the QCP-860 handsets, which given the subscriber sale to Kyocera, would mean that QCOM had these some time before February of 2000. Off topic, I've seen one of these phones at a tradeshow, and made a quick call on it (with a QCOM booth rep present of course). I also saw it demo'ed with data speeds much higher than my current QCP-860 could ever think of doing. Which means that perhaps those CSM5000 base station chipsets are already rarin' to go, just waiting for the standards support (which finalized sometime earlier this year).
It is clear that 1X is late, by anywhere from 6 months to a year. Qualcomm had hoped to produce significant numbers of msm5000 in the 4th quarter of last year.
The delay was not due to QCOM, but to the standards process, which took longer than many suspected (re - holy wars with ERICY regarding Chip Rate, last minute ideas by other members of CDG, etc, etc, etc). It should be noted that IS2000 was derived in large part from IS95C, which was around for some time.
Interestingly, to date there seems to have been almost as many csm5000s sold as handset 1x chips. Sulpezio said 300,000 CSMs were sold last quarter (vs. perhaps less than 500,000 1X MSMs [anyone have a number here?]).
I suspect that the CSM's are being put in place for network upgrades and replacements. Do note that PrimeCo replaced Motorola networks with Lucent, and Verizon seems to be moving away from Moto and towards Nortel and Lucent as well. Chips for those base stations have to come from somewhere. I suspect that that is where some of the chips went.
Qualcomm is late on this one. Qualcomm is a great company and very reliable. However, as a colleague of mine said "if you don't miss at least one plane a year you're spending too much time in airports".
I don't quite understand this comment, considering where the other prospective 3G and transitional 2G+ technologies are, I would suggest that QCOM is running on time and on schedule with their announced intentions.
I would be disappointed if Qualcomm never missed a milestone date: The goal is not to make milestone date but to increase your competitve advantage (run faster in a race, not run at a given pace).
See above... QCOM is pulling the nVidia trick, which is to keep elevating the level of capabilities faster than the competition can. See SnapTrak, RadioOne, BREW... they're doing fine...
But it is very clear that there was a sea change at Qualcomm (and its partners) around February of this year, culminated with announcements from Verizon, Sprint, Nextel, and others with launch dates this year.
Again, much of this has to do with IS2000 finally getting through ballot and everybody agreeing on the best path moving forward. Once in place, then it is just a matter of building the equipment, getting the software in place, moving forward with internal trials, etc...
That signaled that all the showstoppers had been solved, and there was left only the more mundane and predictable tasks of building it. It is no accident that this point came only after early availability of the msm5105. This was the key issue.
I think MSM5105 is a short term chipset, probably only until the 6xxx series is ready. Looking at QCOM's CDMA Technologies WWW site, it is clear that the 5105 drops in place of the 3100 chipset (from which it is derived I would suspect), this allows handset vendors to focus on product development and leverage their existing designs. Product development meaning more features and "killer apps"...
The next milepost we will see is the launch, in quantity of 1x handsets by operators in advance of 1x infrastructure.
Who knows, perhaps the phone you are using already has a 1x chip inside, just waiting for the moment.
This signals not only the availability of working handsets from multiple vendors, but also a strong belief by the operators that there are no problems with the handsets.
I think the operators have been quietly testing 1x for several months now, and both QCOM and the vendors have been refining the products. It would explain to some extend Dr. Jacobs' comments at the GSM World Congress about wideband being very late compared to others. QCOM hasn't been known for stirring up things unless they can back it up.
Qualcomm only expects to ship 6 million 1x chips this year, so the earliest qualcomm will be shipping a majority of 1x chips is 1st quarter of next year (I am talking always about fiscal years here).
That is a lot of chips, and they have to be going somewhere...
That means the big rollout of 1x phones is slated for 2nd quarter (winter), since it takes about 10 weeks to manufacture a phone.
Maybe you know more about this than I do, but I would expect that it would take more than 2 1/2 months. Based on when the MSM3100 was announced and when products that shipped, I would expect that it takes about 9 months or so from the availability of the reference platform from QCOM to when the final handsets start to ship.
So, the reports from Korea may have been overstated. And there may have been non-handset problems (as I stated in my previous post). But, it is not a network build out problem since 1x phones are downward compatible so should work just fine on legacy networks. In other words, any problems they were having is with the technology.
This is why I believe that this would not be an accurate report from Korea. There's too many uncertainties about it. A quick response to the issues and delays noted by WCDMA heavies (BT and NTT).
--the doctor
I'm starting to feel like Siskel and Ebert here...
I give the drama before us the "thumb's up".
;^)
sfx |
