Are you interested in a new Pandora LCD cable?

[Hardcoreufo]

The work looks great! Thanks for all the effort hns
!

 

[hns]

Good news: I found time to look into improving the production process for the cable samples and it is almost ok now:

These cables will go for testing now If they work inside a Pandora and do not break within some days, the project is ready to order the flex PCBs for the cables and then production can be done.
There is only one remaining problem with soldering the connector for the Pandora display panel. For these 5 cables, I had to check each connection pin under the microscope and repair some with a fine soldering iron. Since I do not want to do that for more than these 5, I have to think how to improve soldering the connector and find out where the problem comes from (e.g. warpage when heating of the flex PCB or the connector).
A final calculation shows that they can be offered at 34€ (incl. German sales tax), if I can produce (and sell) 100 units.

 

[Linux-SWAT]

Are they reinforced like the black ones ?

 

[hns]

They look quite similar to https://pyra-handheld.com/boards/th...in-a-new-pandora-lcd-cable.98970/post-1668847 and are based on the Gerber files "Rev 5".

 

[Linux-SWAT]

I mean is there special coating on it preventing teardown ?

 

[hns]

I mean is there special coating on it preventing teardown ?

No. I have no information how to make this and where to get it from and how to test if it is worth to be applied. But when being able to make these cables in small quantities it may be easier and cheaper to replace one again after let's say 3 years than trying to add coating and have 6 years.

 

[Linux-SWAT]

Yes, but only a few people is able to actually change cables on the Pandora, so that also means more back and forth.
Anyway, this is better than no cable at all :^) .

 

[hns]

Anyway, this is better than no cable at all :^) .

Yes, that is my idea to have at least a good one since we can't get the best one.

 

[FBnil]

@hns Ok, so do we have to buy at least 2 to get to 100 and keep the price low? Because I count around 50 people interested... and at double the price, I guess many will say now.
Maybe E.D. could want a few for his shop too, or was that already calculated into your calculations?

 

[TrashyMG]

I voted Yes, I like a spare... that said I have three Pandori

 

[FBnil]

@hns Hi HNS, were you able to work on this project? Any progress on the testing of the prototype cables you made and were they durable?

I meanwhile started looking on how to remove the screen... I think I tried three times over the last months. Each time I'm close, but unable to loosen the top (without resorting to damaging my screen even more). So I'll enjoy my purple screen a little more (instead of breaking the ribbon by opening it.

 

[hns]

@hns Hi HNS, were you able to work on this project? Any progress on the testing of the prototype cables you made and were they durable?

Thanks for asking! We were not yet able to test durability... But I had an idea. This needs a hell-machine... It is not yet completely finished, but already looks like this:

It will wobble the display lid of a Pandora. It should be able to handle roughly 80000 open-close-cycles per day so in 12 days it is a million... So we can expect real stress And ED can make a video. Who knows, maybe he can do a setup with a webcam showing the scene, the pandora display (initially working an broken at some point) and the counter?

When will it be finished? Well, in 2 months...

Regarding the question how to open the display assembly, others can likely answer better than me...

 

[FBnil]

You monster! You will subject your poor Pandora to torture? The heavens weep!
The contraption looks impressive. Be careful with the lcd screen, put something cushioning, or better, build a cage around the display. And remove the pin hinge, without it, the display goes floppy, but replace it with something that holds it in place.


Yes, a video! Like those Life stream videos on youtube? I think a few people, me included would watch that.


Thank you for the update. I'll ask again in 2 months....

 

[hns]

Once again almost 2 months later:


So the machine is finished and I have to think about safe transportation to ED since he will be running the tests.
What has not yet been completely elaborated is how to precisely detect that the LCD cable did break and after how many movements.
Well, the counter sows it but someone should sit near the LCD and watch some test image and note the number of cycles.
Or ED makes a video which is taping both the Pandora display and the counter and then the machine can run alone and one looks after it only every 8 hours or so.
If then the cable shows problems, the video can be rewound to the precise location where it started to fail.
Anyways the machine makes ca. 1 movement per second which translates into 60 per minute, 3600 per hour and 86400 per day.
So if we do not need the highest precision it could be run for one hour and then be checked. And then the next hour.
And if there is nobody who can come back after one hour to check the status, the machine is paused and resumed later.

 

[PowerGod]

Wild PANDORA appeared !!

"I choose you Death Machine !!"

Death Machine used DISPLAY FAPPING !!

IT'S SUPER EFFECTIVE !!

 

[FBnil]

That counter is really a finishing touch. Niiiiiice.

 

[stevenc99]

Generate QR codes on the screen, read them with a webcam/smartphone which saves them to a file (or even stops/pauses the test when no code has been read for X seconds)?
* include a counter or timer in the QR code
* reduce the number of error correction blocks when generating the QR code
* include a long string in the QR code so that it requires more, smaller barcode cells to display it (and thus be more sensitive to smaller parts of the image going missing)
* cycle through different foreground/background colors according to some fixed sequence (to detect if only one color channel stopped working)

 

[FBnil]

Generate QR codes on the screen, read them with a webcam/smartphone which saves them to a file (or even stops/pauses the test when no code has been read for X seconds)?
* include a counter or timer in the QR code
* reduce the number of error correction blocks when generating the QR code
* include a long string in the QR code so that it requires more, smaller barcode cells to display it (and thus be more sensitive to smaller parts of the image going missing)
* cycle through different foreground/background colors according to some fixed sequence (to detect if only one color channel stopped working)

sudo apt-get install qrencode

But are you sure QR is the way to go? If we are going to add a web camera that compares each image to the base image and stores it as images then a human looks at all the pictures (or an algorithm that calculates the difference between 2 pictures ImageMagickCompare) then the timedate of the picture (which could be in the name), and the number of the picture sequence is already enough to know when it died and how many folds it took.

edit: I would also only use the tv test pattern maybe a simplified one with only colors and not the black/white patterns, because those do not compress well and always generate a blurry. Just one image, no color rotation (or maybe after detecting that there are suddenly more differences than a certain baseline (for example, when turning on the light, or opening the curtains.

edit2: To "take a picture" (the camera is always-on, we only decide when we fetch a jpg from it), we need to put a mouse at the back of the arm, and let it click to "take a picture" (could also be a modified keyboard with a brokenout microswitch)

In the end, using a stopwatch to let it do 500 folds and manually boot and test it, put it back, let it fold some more, repeat can be much faster.

 

[hns]

edit2: To "take a picture" (the camera is always-on, we only decide when we fetch a jpg from it), we need to put a mouse at the back of the arm, and let it click to "take a picture" (could also be a modified keyboard with a brokenout microswitch)

Well, such a microswitch is already inside the windshield wiper motor (originally used to move it to the stow-away position) and it simply triggers the mechanical counter.
It should be possible to feed that signal (with proper voltage divider and overvoltage protection) to the device controlling the USB camera. Maybe I have some spare parts to add to the machine incl. some connector.

 

[comradekingu]

I think the paperclip is going to fall off.
Do a livestream