OMAP4440 Dual Core

[fanoush]

We won't change the CPU right now as it would be unfair for all the guys who have been waiting.

That's the wrong reason. I would accept that you are still busy or that it is hard to source those parts or that it needs extra money or you already have older omaps ordered or whatever but not this one.


Maybe you can offer board upgrades to those really unhappy and make additional money on that? How much could be just the motherboard? I would get newer board with more RAM (512 or 1 GB) any minute. Faster chip is nice bonus but it is mainly for RAM.


Also some people want just emulators so those are happy happy with 256 megs, some don't need extra cpu power so I wouldn't be so sure it will piss off so many people. OTOH some people want real mobile computer with firefox and other stuff running concurrently. Some people today simply want to buy today's technology, not 2 years old so in fact you are already losing new customers by not offering upgrades.


Look at beagleboard or touchbook, once it was possible to upgrade it, they did it.

 

[Natro]

We won't change the CPU right now as it would be unfair for all the guys who have been waiting.

That's the wrong reason. I would accept that you are still busy or that it is hard to source those parts or that it needs extra money or you already have older omaps ordered or whatever but not this one.


Maybe you can offer board upgrades to those really unhappy and make additional money on that? How much could be just the motherboard? I would get newer board with more RAM (512 or 1 GB) any minute. Faster chip is nice bonus but it is mainly for RAM.


Also some people want just emulators so those are happy happy with 256 megs, some don't need extra cpu power so I wouldn't be so sure it will piss off so many people. OTOH some people want real mobile computer with firefox and other stuff running concurrently. Some people today simply want to buy today's technology, not 2 years old so in fact you are already losing new customers by not offering upgrades.


Look at beagleboard or touchbook, once it was possible to upgrade it, they did it.

This is just silly... its kind of like asking Sony to sell you a faster PSP than they currently produce... just because the technology is not state-of-the-art (SOTA) does not mean it is inadequate. Could it be better? sure... but the process and design changes along with the costs and logistics of doing it are simply not feasible. It took them years just to get the current version out the door... in a few years from now, today's SOTA will be obsolete and thus the cycle continues... this is the nature of the beast for electronic devices... there will and always will be something better, faster, smaller, and more capable on the horizon.

 

[Granitehead]

This is just silly... its kind of like asking Sony to sell you a faster PSP than they currently produce... just because the technology is not state-of-the-art (SOTA) does not mean it is inadequate. Could it be better? sure... but the process and design changes along with the costs and logistics of doing it are simply not feasible. It took them years just to get the current version out the door... in a few years from now, today's SOTA will be obsolete and thus the cycle continues... this is the nature of the beast for electronic devices... there will and always will be something better, faster, smaller, and more capable on the horizon.

For Sony you can be sure that there is no better /and/ cheaper (can that be? It's said somewhere on the previous pages...) processor which could be integrated by just switching the part without any other changes which they don't use.


This doesn't mean I want them to upgrade now, I'm not sure if I would like this, I just wanted to note this as it seems you haven't read this.

 

[Batou456]

Nuclear is old and busted. The new hotness is fuel cell technology.

If you're willing to settle for ethanol/methanol Fuel Cells the technology is basically available now actually. Airport security won't be happy with you, though.

Using the faster OMAP and more RAM is already possible with the latest board revision.
But as I already said a while ago: We won't change the CPU right now as it would be unfair for all the guys who have been waiting.


We changed it in case that sometime in the future, TI will no longer sell the slower version. You never know. And in case that happens, the faster SoC can be used without a board change,

there will and always will be something better, faster, smaller, and more capable on the horizon.

That was in response to a BeagleBoard revision that's been out and was pointed to as something OPL might do. The BeagleBoard's "SOTA", as you put it, competitor being the PandaBoard, which has also been out awhile making neither "SOTA" by your criteria, and thus your entire rationale rather irrelevant.

The relevant questions would be cost, need, and the associated analysis. Although from what I've gathered from Craigix the Batch 2 SoCs have already been ordered at this point and hence it's a rather moot point until such time as Batch 1 & 2 actually have enough orders to clear them out given how many units OPL has to pay for that currently are without a buyer, particularly as related to Batch 2.

 

[laurencevde]

Craig could sell those to the BeagleBoard, if at any time they'd need an extra batch. As the BB is also produced by CircuitCo, they won't even need to be shipped... (assuming both projects use the same chip-packaging...)

 

[Natro]

I work in an industry that is extremely sensitive to changes in our products primarily due to safety issues. The concept of strict configuration control is very important due to the costs (time, money, retraining, retesting, etc) involved. Even a change which at the surface seems completely transparent turns out not to be in many cases...


There is a very common misconception that something with the same form, fit, and function will also perform the same. Often times, the changes that were made to make the whatever-it-is better often impact how it performs in some way. Even if all the changes are internal to the replacing component, there may still be external effects that may not have been considered (such as changes in overall heat production and power consumption). Suddenly after the upgrade you notice excessive heat, a lower battery life, or some programs no longer run properly or randomly lock up now. My point is, theres almost ALWAYS a problem that manifests with any change that requires some other resolution to make it work. The more complex the system, the more likely any change will cause a problem.


The reward for making the change needs to justify the risk in making the change in the first place. Lets examine this...


REWARDS:


An updated SoC would provide more capability in the form of raw processing power, RAM, and possibly new features. In the case of the OMAP4 family, this would be roughly 150-200% increase in performance. In the case of the 720MHz version of the OMAP3530 this is only a 20% increase... but this is more than a mere numbers game... lets look at what this really means to the end user... In other words, what will the Pandora be capable of doing after the upgrade that it cannot currently do? The answer to this question is another debate and theoretical but we do actually have a basis for comparison... What can an 1GHz overclocked Pandora do that a 500MHz Pandora cannot? I think you could quickly come to the conclusion that the answer is not the 2x or even 1.2x you were expecting...


RISKS:


For the most part the exact issues will be unknown and varied, from OS issues, driver problems, and so on. Ultimately it will come down to time and money to fix all that. How much time and money? I would personally like to wait to find this out sometime AFTER the pre-sold Pandora's are delivered and the OPT can actually start to turn a profit. The cost of a change is much more than a simple difference in cost between components... I am not sure what type of FCC (or equivalent) testing/certification is necessary for changes, but there probably is something. Not to mention the costs for retooling the manufacturing equipment, developing new production tests, setting up the supply line, and other basic operations overhead associated with manufacturing electronics. That overhead costs alone can easily dwarf the difference in savings between the current and upgraded chip.


Looking at what the OPT has achieved to get the Pandora from concept to product is quite impressive when you consider they are just a "couple of people in a garage" so-to-speak. They are not a mega-corporation that can mobilize an army of engineers and lawyers, and leverage millions of dollars to entice lots of competitive bids from parts manufactures who are drooling for their piece of the cash pie.

 

[Batou456]

REWARDS:
An updated SoC would provide more capability in the form of raw processing power, RAM, and possibly new features.

I'm pretty sure OMAP4 is not what EvilDragon was alluding to instead of a revised OMAP 3530 or potentially the xM's AM/DM3730. Incidentally the OMAP DM3730 was actually was brought up in 2008 and shouted down under concerns of delays and otherwise after initial design finalization.

The AM/DM3730 is intended by TI as an upgrade of the 3530 line, and benefits from having a process advantage of 45nm over 65nm. This according to the TI charts results in twice the GPU power, native throttling at 1GHz, and should tend to run more efficiently.


With Batch 2 only roughly half sold we need to kick in the fundamental principle that a product people do not want is waste. A lot of people I've shown it to recently worry over the specs, and even the Archos types with a 3730 worry over being limited to 256MB of RAM. If viable a small upgrade of Batch 2 could make the overall product more attractive as a result, but at the same time raise the specter of involving further funds multiplied across 4000 units, particularly if the SoCs are already purchased and could not be resold for full price.


It should be noted that while nothing expect the NGP in terms of dedicated handhelds will compete with it in the forseeable future, but one does need to be mindful of the Xperia Play which already exceeds the Pandora's specs thanks to a newer process, etc.

RISKS:
For the most part the exact issues will be unknown and varied, from OS issues, driver problems, and so on.

The 3730, which is the farthest I see any legitimate speculation on, is designed by manufacturer for ease of switchover and literally mainly represents the benefit of a process shrink. There is no serious discussion of any major design beyond this which the mainboard has been revised to accommodate, and will be further revised anyway due to CircuitCo wanting board revisions to get failure rates down.

Hence there should be no changeover in tooling, or much of anything else as it's a socket compatible part with a already revised mainboard ready to accept it, and a fab company that already handles those parts for BeagleBoard. Further it will have to go through what teething problems may exist with any further revision anyway thanks to revisions intended to get board failure rate down.


Ergo you've demonstrated no increased risk, but failed to recognize the risk of not doing a revision due to a large portion of Batch 2 not being spoken for and hence has no guarantee it will sell.


Conclusion:


Personally I'm fine with it being left alone, but if it can be reasonably done I see no reason it should not be. The more attractive the now excess inventory is in terms of specs the easier it will sell. The less attractive it is along those lines, particularly if the 500USD pricetag is to remain, the harder it will be to sell.


OPL does not have Apple's marketing might, and 500USD is well into the realm where people normally will be quite cautious about committing. Do not ignore the psychology around buying involved here.


I also don't really see there being a major divide. The current Pandora design has led to people OCing as a matter of course with some emulation, with the potential to exceed the 3730's normal frequency. Hence there's only a divide in processing power if 3730 equipped units are considered OCed themselves as a matter of course. RAM is used for buffering and otherwise holding things in memory so shouldn't be a major dividing factor, although many Batch 1 people seem to comment about being anxious about being limited to 256MB and hence should be tend to be open to the potential of a near term upgrade to lay such anxieties to rest.


Perception management however is important if we're talking something in this price range, and if revisions aren't made it's probably advisable the first thing people see when they open this site isn't the spec sheet anymore. Instead clear demonstrations of what it does and offers along the lines of Evil Dragon's [once] Daily Videos would seem more appropriate. Particularly as there are noted issues with people simply not getting what it's about and then turning to whatever vaguely similar device is at hand and familiar and holding it up as better.

 

[DavidBowman]

Whatever solid reasons there might be to updating the hardware - and there are many - it all comes down to money. There simply isn't enough to make this happen. Not to mention all the current owners of a Pandora that would feel slighted in some way.


I'd much prefer everyone stay on the same hardware set. You know exactly what you're getting and the software should perform exactly the same from one unit to the next. This is still very much a niche project and there isn't any other device close enough to it to really take away its true market share. Most of the reasons offered for updating the hardware have the underlying reason of making it competitive with other devices, such as newer smartphones. And while the Pandora is not still "the most powerful handheld there is," even with the progress of technology it still fills it's original purpose very well.

 

[Dead1nside]

Let's just finish the Pandora 1 and then we can write out the dream list for Pandora 2, and hopefully get some of it Plus, hopefully time to market will be decreased.

 

[XxionxX]

I think that the OPT has learned a lot about the process of making a handheld, and they have also met a lot of people along the way. This is a learning process, and it can only make them stronger. That and the next version of the Pandora will be made by nanites.