Of cases and batteries

[bzar]

I think unmirrored would be even better, in case the cable gets eaten from outside in -- for a mirrored configuration, that would kill your first trace and its redundancy, right?

You're thinking of mirroring it along a different plane than I am

If you want to view it from the "I don't care if I don't see it" point of view, then I guess you are fine. But if there is a potential problem I am more inclined to learn about it rather sooner then later. Or are you happy with a product that has a higher potential of failure in it, as long as you don't notice it?

All units contain the possibility of failure in them. Sooner or later, all of them will fail. This is just a method of making the average time longer. It's not a matter of "I don't care, if I don't see it", but "I don't care, if it works as long as I need it to".

 

[asimov-solensan]

Pretty happy to hear that things are advancing. I hope everything continues to go this well.

 

[ible]

I think unmirrored would be even better, in case the cable gets eaten from outside in -- for a mirrored configuration, that would kill your first trace and its redundancy, right?

You're thinking of mirroring it along a different plane than I am

ok, so here's my very naive picture of what's going on:

SCREEN LID

ABCABC

ABCABC

MAIN BODY

where a line of A's indicate one trace running from the screen to the main body (and similar for B and C), and there is an unmirrored redundancy (as I suggested).

Wouldn't your mirroring look like this?

SCREEN LID

ABCCBA

ABCCBA

MAIN BODY

As I mentioned, this protects trace C the most, if we expect the cable to get eaten outside-in; in this case, A would die right away.

Perhaps I have the wrong picture in my head.

 

[bzar]

My bad, I recalled the cable being double-sided, where each trace would be mirrored on two planes:

side 1: ABCDEF
side 2: FEDCBA

But in any case you're right, it's better to use shifting instead of mirroring to get a good average protection for all traces. For the two-sided cable I recalled this would be something like:

side 1: ABCDEF
side 2: DEFABC

And for the one-sided the one you suggested:

side 1: ABCDEFABCDEF

It mostly comes down to optimizing for the largest sum of minimum distances to the edge of any trace.

I was wrong in my suggestion

 

[thatgui]

I understand what he's saying: the purple tint of death does get worse over time, losing other colours, and even failing completely in some cases. The concern thatgui seems to be having is that everything can be fine for a year, unknown that the cable is slowly failing, until one day you wake up and you have YELLOW tint of death.

I disagree with his concern. Technically it is true, but the odds of failure being small and not progressing is greater than zero. If out of 5000 cables 99 of them fail and one would have failed if not for redundancy then the redundancy has proved useful. And those other 99 managed to survive much longer than without the redundancy so really everyone wins something. Yeah, they still have to replace the cable eventually but they'd have had to do that anyway.

He's also talking about feedback but it's not like that would make a difference. If manufacturing goes similar to the Pandora they're made in large batches, and the cable problem is something that happens months, sometimes over a year later: getting early feedback isn't going to make that much of a difference: if a batch was installed incorrectly they're going to have to trust in the redundancy and hope that it doesn't progress, and if it does they're no worse off if it fails 18 months later than if it had after 12 months....

Exactly (except the difference in opinion about wether this is something to care about or not)

All units contain the possibility of failure in them. Sooner or later, all of them will fail. This is just a method of making the average time longer. It's not a matter of "I don't care, if I don't see it", but "I don't care, if it works as long as I need it to".

I disagree here. Of course everything breaks at some point in the future. But from my point of view, there is a big difference, between building something were you already expect it to fail due to a possible design flaw or some part simply reaching the predicted end of its life.

 

[WizardStan]

I disagree here. Of course everything breaks at some point in the future. But from my point of view, there is a big difference, between building something were you already expect it to fail due to a possible design flaw or some part simply reaching the predicted end of its life.

So picture this situation then: you are the proud owner of a Pyra. 12 months later the cable fails for whatever reason. A single trace has broken rendering the entire thing inoperable. You need to replace the cable.Alternate situation: you are the proud owner of a Pyra. 18 months later the cable fails for whatever reason. Redundant traces have broken rendering the entire thing inoperable and you need to replace the cable.

How is situation 2 worse than situation 1? And situation 2 is basically a worst case scenario, at least some cables will never progress far enough to break the redundancy. So as near as I can tell, outside of a few extra cents for wider cables, the redundancy is strictly better: those that would have broken either get more usage or never notice that anything is wrong.

 

[pmprog]

How is situation 2 worse than situation 1?

The only reason I can see is if there's a "end of life" date on the part (I guess not really with cables, but...) breaking after the end of life could mean not being able to fix at all.

 

[comradekingu]

Situation 2 is better than situation 1, because it breaks down later, since its engineered to do so. I wouldnt be suprised if it took 15 years to wear it down. It took a long while when the cable was

1. was more complex

2. looped around with no protection or padding

3. had no redundancy

4. was wider?

The only caveat is if the "extended" lifetime puts you just outside of warranty.

But then you have situation 0, which is ED is a really nice guy with the parts and fixing. Between choices 1 and 2, where 1 is a shorter no-trouble lifespan, ill always take nr2

 

[TrashyMG]

Shouldn't the real question be is why the LCD cable wearing out in the first place?

 

[comradekingu]

Yes, but that's why you are doing it this way, if you are making a cable, make it as sturdy as can be.

That isnt a reason to skimp on what makes it robust in other places, so that has to be improved too.

If it at no point catches against the casing at the part that bends, it should be fine. If the parts that it potentially could catch against are smooth, then thats somewhat of a safety-measure.

 

[Daniel]

Is there a webpage which counts interested buyers?

All peoples who are interested in buying a Pyra could register.
I think it would be helpfull if Evildragon could see how many pieces he will sell.

 

[Silent-Hunter]

I want to like this post, but there's no button.

 

[Hồng Thất Công]

I think there was a poll for this somewhere iirc.

 

[Exophase]

Shouldn't the real question be is why the LCD cable wearing out in the first place?

Does purple tint of death even happen with Rebirth and 1GHz Pandoras? I always associated this problem with CC units and thought some change was made to mitigate it. I don't think mine ever actually developed this problem, but I think it's because I never opened/closed the screen that much compared to more regular users.

Even if it is a fundamental issue with the Pandora's design we can't say that it's something that'll carry over to Pyra's much different cable, with or without redundant traces. Trying to do failure analysis on the Pandora's cable would be pretty time consuming and maybe expensive, although that doesn't mean it's not still important.

Is it possible that the traces on Pyra's cable can also be made thicker? Does that make any sense as an additional mitigation tactic?

 

[ible]

While such a registration might show general interest, the real information on how many willing buyers are out there is how many people are willing to put money down on a pre-order.

I don't know if it's possible to do this, but perhaps registration on this forum would be helpful for the general interest stuff: you have people sign up for an account here on the boards, and then e-mail everyone major updates from "taking pre-orders" to "shipping units"; these updates could also be opted out of, naturally, by setting e-mail preferences in your profile.

This might also encourage more people to post with issues/concerns/feedback/suggestions, when the Pyra does eventually come out, because they'll already have a forum account.

 

[Silent-Hunter]

I had it on my Ghz unit.

 

[TrashyMG]

Does purple tint of death even happen with Rebirth and 1GHz Pandoras? I always associated this problem with CC units and thought some change was made to mitigate it.

Just looking through the support section indicates there is at least a small amount of Rebirth and 1Ghz units that have the issue.

Even if it is a fundamental issue with the Pandora's design we can't say that it's something that'll carry over to Pyra's much different cable, with or without redundant traces. Trying to do failure analysis on the Pandora's cable would be pretty time consuming and maybe expensive, although that doesn't mean it's not still important.

Agreed, It's hard to say it would. Redundant traces are a good idea, however I would think putting some effort to trying to minimize LCD cable wear in the case design would also be something to strongly consider. Which may already be something that EvilDragon has in his plans.

 

[Kev2442]

I myself am quite interested in buying a prototype. (see here : http://boards.openpandora.org/topic/17697-of-boards-and-displays/page-2#entry372118/URL])

But ED, what are the actual differences between the proto and the final units ?

The only thing I know of is that it will be 3D printed instead of molded. Does that mean it won't be as tough ?

Could there be a reduced lifespan or cosmetic changes ?

I'm totally ready for a prototype pre-order, but only if I know what I'm paying for, you see ? :lol:

If there's nothing more than paying for the on-demand case printing, then my wallet is all yours !

Edit : also, while we wait for official Pyra ports of emulators and stuff, could we use the complete Pandora software lineup ?

 

[Binky]

Also, having diagnostic traces on the cables sounds like a very good idea.

Run a trace up one edge and back down the other, and then test the resistance of the loop. If its high, your cable is failing, if its off-the-scale high, you've actually broken one of the test traces.

 

[WizardStan]

You could even get real fancy and put multiple traces of varying resistances through it so that you can measure it and know exactly which traces have been broken, but that sounds like a lot more work than its probably worth. I honestly do believe we're making a mountain out of a mole hill here.