View Full Version : PE e-ink side board


jackr
05-31-2011, 09:32 PM
I took a picture of both sides of the e-ink side's board of the pocket edge. I labeled most of the chips as best as I could. The image is in a zip due to the 1mb limit on png files :D.

Last_of_the_PEs
05-31-2011, 10:20 PM
Very nice. I hope you didn't have yours DIE before the image!
How can we benefit from the picture? Does it make it easier to program it? I know Emusan reports having done so already.

emusan
05-31-2011, 10:23 PM
Very nice. I hope you didn't have yours DIE before the image!
How can we benefit from the picture? Does it make it easier to program it? I know Emusan reports having done so already.

you must be mistaken... I have not yet BEGUN programming for the e-ink, I've just done some work identifying how to, lol I'm not that good(yet). e-ink programming will probably take (at least) another week to two before I get started, though others may beat me to it.

EDIT: didn't mean to snap, if you took it like that, I just don't want people to think I'm nearly that good until I am actually that good lol.

And nice pics by the way, they should be helpful, you may want to also post them in the Hardware technical thread(if they haven't been already)...

yugami
05-31-2011, 10:49 PM
Heres a breakdown. Most of it doesn't matter for an Android point of view. The main chips are the Marvell Armada 168 and the Epson EInk S1D13522

Weirdest thing is for the life of me I swear it has 1 Gig of ram installed. Where as my device only reports out 512meg.

chips:
Wacom W8001 - Pen enabled driver for eink screen.
EPSON D135220B1 - EInk driver chip (S1D13522)
SAMSUNG K9LBG08U0D - 4 Gig
MAXIM 17020E - pwm regulator (dual)
88AP1-BJD2 88AP168 - Marvell Armada 168
SEC 037 HCF7 K4T1G0840E - Samsung Ram chip 1 Gig
MXIC B103855 (MX25L6445EMI-10G) - 64MB Serial Flash
SMSC USB2514B D1033-A2P10 - USB Controller

emusan
05-31-2011, 11:00 PM
very strange about the RAM, it looks like the RAM chip is broken into smaller sections of memory, maybe entourage only wanted to use 512 so as to reduce power? Though I can't imagine you would save much, although for that matter you don't really need a whole gig...

jackr
06-02-2011, 03:33 AM
Sadly, I did come across some misfortunes w/ it :<. When I first put it back together, I broke the power switch, so I had to order a replacement [http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=P11157SCT-ND]. The switch from DigiKey has a stronger spring, but overall it works fine. So after I fixed the power switch, I turned it on to test, and it booted just fine. Thus I screwed on the shell only to find that I somehow killed the lcd D:.

emusan
06-02-2011, 09:30 AM
Oh no! Could it just be the ribbon cable and therefore easier to fix(I know others have had problems with it in the past), or is the whole thing smashed?

jackr
06-02-2011, 04:49 PM
I hope its the ribbon, right now I have the fabric removed up to the point where I suspect it could have been crimped, I will study it later. One thing that might have contributed is the battery. I plugged in the speakers and buttons/ir mouse cables before I plugged in the battery cable; when I plugged in the battery cable, a spark occurred around the 3rd pin from the edge of one side of the cable. This seems to occur when something is plugged into the board (if the battery cable is the only thing being plugged in, the spark doesn't happen). So it might be that a transistor fried or something, or it could very well be the ribbon, as another thing that occurred was the lcd screen flashing white when I lifted the board to re-seat it (this white flash happened after the battery plug spark, but before I managed to power it successfully for the last time.) The last possibility I am thinking of at the moment is that whatever switch/sensor that tells the edge that its closed is broken so that it thinks its always closed, and thus the lcd screen stays off, although if the lcd lights up during boot even if its closed, then that possibility is shot. At the moment I will take it easy and hope for one of those godsend $100 deals for another edge, and keep this one for parts :).

jackr
06-03-2011, 06:52 AM
I've attached a picture of the inside of the hinge. Lack of switches suggests to me that the pocket edge uses a different method to detect when the lid is closed. On the lcd side of the unit is two magnets. One is beneath where the pen sits, so I suspect it is for charging it. But the other magnet sits on the opposite side of the lcd from the pen, and seems to have no purpose, unless the wacom sensor is able to sense it when the lid is closed.

I haven't been able to fix the lcd yet, though it seems that the touch sensor isn't working either, which gives support to it possibly being the lcd cable thats gone bad.

jackr
06-03-2011, 07:54 AM
Took a leap of faith and removed the backplate to the lcd. The backlight consists of 27 leds on one side of the screen (the long side near the hinge). The lcd unit itself is a Innolux AT070TN92 V.2 AA0700013211 display. The mystery gets deeper as I found by holding the unit to light w/ the backplate and other pieces removed, the display part works fine. So what doesn't work so far is the led backlight and the touch sensor part.

I also counted the wires on the lcd cable, and it has 50 pins, which is the same as the ribbon connecting the lcd unit to the pcb that has the buttons and ir mouse. At the time when I had the backplate removed, I hadn't connected the pcb's connector (the black one with 12 pins) to the main board, so its likely that the touch sensor works.

ivanjt
06-03-2011, 08:53 AM
Oh to get my hands on a unit that I could take apart and map - started building, and taking apart, electronics units in the days of valves, now I build and repair computers and repair notebooks.

Try and trace the LED power line and somewhere you should encounter an electronic switch - either a single IC or an IC and transistor (IC will most probably have 8 pins). The photos that have been posted aren't good enough (clean well focused and evenly lit) for me to use to try and trace and give advice as to where to look.

jackr
06-03-2011, 08:45 PM
Oh to get my hands on a unit that I could take apart and map - started building, and taking apart, electronics units in the days of valves, now I build and repair computers and repair notebooks.

Try and trace the LED power line and somewhere you should encounter an electronic switch - either a single IC or an IC and transistor (IC will most probably have 8 pins). The photos that have been posted aren't good enough (clean well focused and evenly lit) for me to use to try and trace and give advice as to where to look.

My camera isn't in the best of conditions (same with my shooting skills as well) :<, but I will try to get a better shot of the traces. I was worried it would be an issue like a transistor or something (as the lcd had flashed white while it was off when I was moving around the main board.

The only chip on the pcb for the lcd side is a 06A3*7*41 (the 7 could be a 1, it's a really small chip) TSC20071. It appears to be the controller chip for the touch screen, so the ir mouse might be controlled on the mainboard instead.

http://pdf1.alldatasheet.com/datasheet-pdf/view/180915/BURR-BROWN/TSC2007.html


I tested the voltage and amperage of the backlight while the unit was powered on. I was reading 4.88v @ 0.78a. I had tested the leds with a power source of 7v and they had lit up at about the same brightness when the unit boots up.

jackr
06-03-2011, 10:28 PM
Attached is a picture of the pcb on the lcd side of the pe. The 4 pin connector is for the touchscreen.

I tested the backlight, and it still lights up, so its a chip/wire/trace somewhere on the board(s).

ivanjt
06-04-2011, 08:28 AM
Not too bad a picture ;-) to make really good ones it takes macro lenses and a stable setup.

Anyway, back to the problem.

I don't think that board has anything to do with the backlight other than being a passthrough. If U49 is the chip you indicate then you will have to look elsewhere. The board appears to be multi layer so we can't trace the tracks but I am thinking the controller is most likely on the main board.

Next test would be to find out if the + line for the LEDs is connected to any other socket, the - will most probably be a common ground and unswitched. If it is, then you will have to trace where that goes.

jackr
06-04-2011, 05:01 PM
I agree that u49 is most likely solely for the touch and maybe ir mouse and that the board is just a pass for the lcd/backlight unit. I tested and got the same voltage at both the lcd ribbon (where the backlight ribbon is soldered to the lcd's ribbon) and at the connector on the mainboard.

Attached is a shot of the connector on the mainboard. I labeled where the ground and positive pins are on the board for backlight. I tested the ground while the unit was unplugged from battery, and still got continuity.

ivanjt
06-04-2011, 08:24 PM
The area of interest is D18, U28, FB23 and U29. I can't tell if there is any damage because you are using a single light source which leaves shadows. Q1 may be a switch and I think U28 may be a set of 3 diodes. I can't read what U29 is but if it connects in any way to Q1 and D18 then it may be the LCD control chip.

If it is you will need a data sheet to tell which pins are which, somewhere there should be an inhibit, or switch, line that is the one you are looking for to find out its state.

I would like it to be something simple like D18 being open bit it most probably isn't.

One thought, if you sort this out you might have the answer to the WSOD that is bugging some people.

emusan
06-04-2011, 09:29 PM
One thought, if you sort this out you might have the answer to the WSOD that is bugging some people.

One user in another thread had the same thought, that the WSOD was due to the cable in the hinge, and was going to bring it to their local electronic repairman in order to see if they could get it fixed, though I'm not sure what became of it. It would make sense, as the hinge must put a lot of strain on that thin little cable...

jackr
06-05-2011, 01:42 AM
I tested d18, and before the diode, I got a reading of 4.96v @ 0.25a and after the diode a reading of 4.88v @ 0.20a. I will also try to get a better lighted shot of the area tomorrow.

ivanjt
06-05-2011, 06:37 AM
Something doesn't look quite right with those figures. Normal voltage drop across a standard diode, as this one appears to be, is 0.6v and you are showing 0.08v which could be almost correct if the diode was an avalanche type.

One other thing, how are you measuring the current?

jackr
06-05-2011, 06:45 AM
Something doesn't look quite right with those figures. Normal voltage drop across a standard diode, as this one appears to be, is 0.6v and you are showing 0.08v which could be almost correct if the diode was an avalanche type.

One other thing, how are you measuring the current?

Got me in one :). I don't have a lot of experience, so I used a yellow led that I had on hand and took a quick measurement.

ivanjt
06-05-2011, 12:02 PM
Thought it must be something like that - forget current measurements at the moment unless you have access to some very expensive equipment that gives you the figures after you plug in track width and thickness. The usual way is to unsolder one end of the component and put the meter between the solder pad and the unsoldered end of the component - not an easy task for the untrained (I some times make a mistake and break a component and I've been doing this type of work for over 30 years). When measuring voltage use the 'gnd' test points for the negative probe of your meter and be careful when applying the positive probe that you don't short anything out - could be very nasty.

For someone with little experience you are doing quite well so far. Just take care and we should get there.

jackr
06-05-2011, 08:02 PM
Attached is an edited shot of the u29 chip. It appears to be A84A 0116. The other shot shows a brown bead on L5. I couldn't yet find a good light source.

jackr
06-06-2011, 02:58 AM
I took readings on both sides of fb23 as well as the u29 side of c265 and each read @ 4.96v.

ivanjt
06-06-2011, 04:06 PM
If I can get the picture up on here I'm looking for the voltages at the numbered points with respect to TP3 GND.

Another thing, My replies to anything you post will be a little sporadic for the next couple of weeks - I have a lot of work on a new contract that should keep us going unlike Entourage.

jackr
06-06-2011, 05:25 PM
If I can get the picture up on here I'm looking for the voltages at the numbered points with respect to TP3 GND.

Another thing, My replies to anything you post will be a little sporadic for the next couple of weeks - I have a lot of work on a new contract that should keep us going unlike Entourage.

Hopefully it will be solved quickly then :). I took readings @ the points you want. The tilde ones are the ones where the reading fluctuated.

p1: 4.88v

p2: 4.96~4.97v

p3: 0v

p4: -0.1mV

p5: 3.25~3.26v

p6: 4.96~4.97v

p7: 3.23v

ivanjt
06-07-2011, 08:37 AM
OK, I have a little time while the server racks are installed.

Something doesn't seem right here, I'm not saying your measurements are wrong. just that the backlight should work since you are getting the nominal 5v at point 1.

Can you check that you get the same voltage (4.88v) at the + terminal of the LED array.

There may be two other explanations, Pulsing the supply to a higher voltage or they are switching the negative supply line. The first is possible but there is no logical reason for the second.

jackr
06-07-2011, 08:21 PM
OK, I have a little time while the server racks are installed.

Something doesn't seem right here, I'm not saying your measurements are wrong. just that the backlight should work since you are getting the nominal 5v at point 1.

Can you check that you get the same voltage (4.88v) at the + terminal of the LED array.

There may be two other explanations, Pulsing the supply to a higher voltage or they are switching the negative supply line. The first is possible but there is no logical reason for the second.

The closest I can get at the moment to the led's is where the led array ribbon is soldered to the lcd ribbon on the display. I have run 7.79v through it before and got the leds to light up; I also tried 6.47v but the leds didn't light up. 4.88v is present @ where the ribbon is soldered.

ivanjt
06-08-2011, 08:10 AM
Umm, scratches head and changes mental gears from server setup and testing to edge repair mode.

So the LEDs are in series hence the need for higher voltage. I am assuming the PE battery is a nominal 7v which doesn't give much leeway for discharge and running anything else - WiFi etc.

I must try and track down that A84A - or is it an AB4A? and see if it is a pulse switch or something like that. If it is then we need to know what switches it on since it can be controlled to shut off the screen. I'll see if someone in the office can track it down - hopefully it isn't a generic number.

Now back to keeping Co in business.

jackr
06-08-2011, 07:57 PM
I started a new thread with a more appropriate name of the problem over here: http://www.mobileread.com/forums/showthread.php?t=138257.

Hopefully someone else will try their hand @ testing voltage points on a working unit.

ivanjt
06-09-2011, 08:11 AM
Good idea jackr - if we get some comparisons we can move forward.

So far we've not been able to turn up anything on the A84A chip and some of my people think it is an OEM number but I'll keep trying.

Is it possible for you to get some high resolution pics of the board - by that I mean pics with enough resolution to be able to follow tracks.

End of lunch time so back to work - even the boss has to keep strict hours ;)

jackr
06-10-2011, 11:56 PM
Good idea jackr - if we get some comparisons we can move forward.

So far we've not been able to turn up anything on the A84A chip and some of my people think it is an OEM number but I'll keep trying.

Is it possible for you to get some high resolution pics of the board - by that I mean pics with enough resolution to be able to follow tracks.

End of lunch time so back to work - even the boss has to keep strict hours ;)

I went ahead and got a used PE on eBay, so I will test that when it arrives in a week or so. I will place my findings on this thread and the other thread as well. My big goal was to try and see if I can get the pcie working on the thing, but I hit a dead end as far as connectors go. Not to mention the board over there is unpopulated, so who knows how much work is needed :(. If I can get the backlight working again (or even if I can't-just unsolder the led ribbon and tie it to a 9v or something) I will have myself a tester board :D.

n6pfk
08-05-2011, 09:30 AM
How are the cables released. I see the hold downs but how do you activate them? Also does anyone have a digikey part number for the power connector? The center pin is missing in mine.

ivanjt
08-05-2011, 07:33 PM
How are the cables released. I see the hold downs but how do you activate them? Also does anyone have a digikey part number for the power connector? The center pin is missing in mine.
The cable loch is usually released by moving the plastic mechanism up along the cable - take care and don't break it, also use a plastic tool.

Don't know about digikey but the RS part number is 448-386 http://uk.rs-online.com/web/p/dc-power-plugs-sockets-leads/0448386/ which will give yoy the specs you are looking for.

n6pfk
08-06-2011, 12:34 AM
Thanks for the power connector info. The cables in the edge have some sort of metal clip and I want to really understand how they work before I damage any cables.

mokel22
08-06-2011, 02:21 AM
Thanks for the power connector info. The cables in the edge have some sort of metal clip and I want to really understand how they work before I damage any cables.

Just throwing in my two cents, but taking pictures and uploading them w/ your posts would probably help answer some of your questions with more confidence. Plus it would help future viewers with the same problem. Good luck on getting your edge fixed.

ivanjt
08-06-2011, 07:13 AM
Just throwing in my two cents, but taking pictures and uploading them w/ your posts would probably help answer some of your questions with more confidence. Plus it would help future viewers with the same problem. Good luck on getting your edge fixed.
That is one action I endorse completely. It is almost impossible to give hardware advice without knowing exactly what is being referred to, even if we do work with hardware every day.