It is currently Fri Dec 13, 2019 12:20 am



Reply to topic  [ 2173 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6, 7, 8 ... 55  Next
 RifTUP! FullHD Oculus DK1 upgrade kit 
Author Message
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
Who has the entire spec sheet for a snapdragon 800?

It may be possible to use the chip, but use the video ONLY. IF..it has the latency under control and IF it can address very high pixel densities and frame rates. (which I seriously doubt)

Again, the shortcut is the programmable chip as a LVDS to MIPI bridge. (lots of chips going the other way, MIPI to LVDS, TI has more than one)

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Thu Aug 08, 2013 8:16 am
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
Has OzOnE2k10 and Rozsnyo not been looking into an HDMI/DVI>>MIPI Bridge?

Could we not take the open Chalk-Elec's HDMI>>LVDS Bridge schematic and make a HDMI>>MIPI bridge by just replacing the parallelRGB>LVDS chip with a Solomon SSD2828 and maybe having to add some things and change the connector? Or better yet having them make one for us?

If we can make a universal HDMI>>MIPI circuit with adapters for each panel(custom cable from header to connector) it shouldn't have to take a year, right?


Thu Aug 08, 2013 12:16 pm
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
Inscothen wrote:
Solomon SSD2828



"-Support up to 1920 pixels per display row in Video mode, up to 60hz refresh rate"

perhaps it can be 'pushed' to do 1200 rows, as the next spec says:

"Support up to 2560 pixels per display row in Video mode, up to 30hz refresh rate"

2560x1440x30 eagles a mhz pixel rate of.. 110.59mhz


1920x1080x60 eagles a mhz pixel rate of 124.4mhz.

Thus the 1080p is already pushing the chip, or is at least it's stated highest spec. It may not be able to do the 1200p at 60hz, as that is a pixel rate of....138.24mhz

That puts the chip into a position of being over driven by 11%, if that is possible to do, just to get to 60hz at 1200p (WUXGA). Don't know how the chip is designed, don't know what it is doing.

Just some very rough improperly applied numbers to get a baseline of sorts.

OK. found some specs. Solomon's website is skint on info, which is the new sucky norm for leading edge tech these days.

"The SSD2828 connects an application processor with a traditional parallel LCD interface, and also a LCD driver with MIPI slave interface. Supporting AMOLED, a-Si LCD, metal oxide TFT and LTPS LCD panel technologies, SSD2828 can convert 24bit RGB interface into 4-lane MIPI-DSI interface to drive extremely high resolution display modules of up to 1200 x 1920 or 2560 x 1600."

But no pixel density info, if such a number applies to the chip and how it is internally implemented.

For example, the Toshiba chip says '165mhz bandwidth', which means the Toshiba chip MIGHT be able to do 1200p (wuxga) at 72hz..natively. Then, it may be possible to push it harder, if it is properly handled, as these system are always rated in their safe ranges. maybe to go to 85hz.

Which takes me back to the story on the 90hz 1080p panel that valve is rumored to be doing, which is possibly the Toshiba chip. The numbers that are rumored fit the Toshiba chip's stated specs.

OK. so..the Solomon 2828 and the toshiba chip appear to be the only possibilties that are commonly KNOWN to exist.

Well, there is one other, but I'm not sure they are playing this game yet. They can do all of this at 720hz, but not in a portable form factor and power ratings/use.

IT IS possible to use TWO of the Solomon chips in a parallel bridge and do high speed flipping from one to the other, but that is quite an engineering task to even formulate it and then test it. Thus, it maybe fairly quick to get to 120hz or even 144hz, but it would take a bit of engineering and head scratching.

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Thu Aug 08, 2013 9:18 pm
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
which, of course, circles back to this board:

http://www.aliexpress.com/item/Ssd2828- ... 92255.html

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Thu Aug 08, 2013 9:53 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
What Toshiba chip are you referring to? I see no product by them that provides a _________ -> MIPI-DSI interface.


Thu Aug 08, 2013 10:07 pm
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
KBK- that link you posted says "SSD2828 adapter plate maximum support 1920 * 1366 resolution"

MSat- could it be TC358768?


Thu Aug 08, 2013 10:15 pm
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
Inscothen wrote:
KBK- that link you posted says "SSD2828 adapter plate maximum support 1920 * 1366 resolution"

MSat- could it be TC358768?



http://www.toshiba.com/taec/components/ ... dBrief.pdf

Maximum HDMI clock rate is 165mhz, don't know if this translates to throughput to MIPI. suspected yes, but don't know.


Back to the Solomon chip:
1920x1366x60hz, gives a pixel rate at about 160mhz, which is about the same as ...everyone else. Ie, Toshiba saying 165mhz on their chip. everyone is brick walling with the technology in use. Some aspect of it's implementation, perhaps.

The Solomon chip numbers supports the idea that Toshiba's bit rate throughput limit will be similar.

The Toshiba is an integrated mouth to ass solution (HDMI in to MIPI out) , so it may be the way to go, for anyone attempting HMD futzing.

Remembering now that the Toshiba says 1080p is the max, so it may not accept out of range signal, as this is not really..analog.

The Solomon chip is probably the solution, as it is known quantity that can address these resolutions (or more specifically, their parameters). it can possibly be stuck on the back end of a known LVDS. A bit cludgy but it is the best bet so far, for both known capacities and limits and probably for latency.

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Last edited by KBK on Thu Aug 08, 2013 10:50 pm, edited 1 time in total.



Thu Aug 08, 2013 10:29 pm
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
That Toshiba you posted is MIPI CSI. Is CSI swappable with DSI?


Thu Aug 08, 2013 10:43 pm
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
Inscothen wrote:
That Toshiba you posted is MIPI CSI. Is CSI swappable with DSI?



Not the slightest Idea. Most of this is new video technobabble to me.

I'm stuck in the dark ages, trying to get digital to work as well as analog.

The good part is that most of my bag of tricks to improve function, still apply.

I mean, check this out. We could do this in home video over 10 years ago, but we had no source, so we did scaling:

BarcoReality 909 9" LC 1200lumen, 30,000:1CR, 37-210hz, 180mhz(pixel rate) 3200x2560p max rez.

All parameters totally analog and malleable. ALL. Any pixel, any time, any place.

Then, get in there and change 400-500-600-800 parts and do much much more than just that. And just crap your pants with the incredible image fidelity that digital is just now beginning to get close to.

Anyway, back to the rift issue.

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Thu Aug 08, 2013 10:52 pm
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
This thread has pretty much come back to where the DIY thread was. It appears the Rift controlbox PCB doesn't have enough traces to the cable connector from the scaler to support 1920*1xx0 so we need a HDMI/DVI>>MIPI cicuit. I guess we should wait for Rozsnyo to make a circuit? Or Chalk-Elec?

Without more info on the Nexus display(datasheet/pinouts), we may not see a 7" solution for a little while.


Thu Aug 08, 2013 11:38 pm
Profile
One Eyed Hopeful

Joined: Wed Jan 02, 2008 9:45 am
Posts: 3
Reply with quote
If your looking for a DVI/HDMI -> Mipi board, this could be a interesting solution:
http://www.buykorea.org/s/GOODS_SN/3007660#nolink
http://odm.hybus.net/eng/index.asp
I just send them a mail to get the price and a more detail datasheet, but so for this looks great.


Fri Aug 09, 2013 1:43 am
Profile
Binocular Vision CONFIRMED!
User avatar

Joined: Sun Oct 03, 2010 2:56 pm
Posts: 295
Location: England
Reply with quote
I have taken a picture of the IC side of the ribbon connector of the Nexus 7 2 generation here:

http://www.wastedspace.co.uk/cms/2013/0 ... on-tablet/

The writing says:


ME571K_LCD_FPC_JDI&AUO E219454

REV. 1.1 ICA C <> 1 ichia IHT

PCB MADE IN CHINA 94V-0 1324


Seems to be 28 of the 34 pins used.


Searching around you can find that the ribbon connector E219454 is the SAME code as the spare part for the 1st generation Nexus, apparently!

http://rounded.com/google-nexus-7-flex- ... html?sl=EN

Further searching indicates the FCC details should be released 26th November

_________________
My HMD Blog: http://www.wastedspace.co.uk/cms/category/projects/vintage-head-mounted-display-upgrade/


Fri Aug 09, 2013 3:41 am
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
E219454 may be the manufacturer's allotted "maker's number" for manufacturers, similar to the JB codes in japan, for binocular manufacturers. (Us and Canada have incorporation numbers allotted to them)

I've found that number elsewhere and the only thing that is the same, is the maker of the cable: "Ichia"

The key spec is probably this part of the numbering: ME571K_LCD_FPC_JDI&AUO

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Fri Aug 09, 2013 6:49 am
Profile
Binocular Vision CONFIRMED!
User avatar

Joined: Sun Oct 03, 2010 2:56 pm
Posts: 295
Location: England
Reply with quote
KBK wrote:
E219454 may be the manufacturer's allotted "maker's number" for manufacturers, similar to the JB codes in japan, for binocular manufacturers. (Us and Canada have incorporation numbers allotted to them)

I've found that number elsewhere and the only thing that is the same, is the maker of the cable: "Ichia"

The key spec is probably this part of the numbering: ME571K_LCD_FPC_JDI&AUO


ME571K is the motherboard for the tablet ...

_________________
My HMD Blog: http://www.wastedspace.co.uk/cms/category/projects/vintage-head-mounted-display-upgrade/


Fri Aug 09, 2013 8:11 am
Profile
One Eyed Hopeful

Joined: Thu May 02, 2013 7:28 pm
Posts: 41
Reply with quote
Wow great work guys!!

So have we narrowed it down? are we working with LVDS or MIPI?

Zach you should find something to plug that cable into see if it fits in anything else.

maxOh wrote:
If your looking for a DVI/HDMI -> Mipi board, this could be a interesting solution:
http://www.buykorea.org/s/GOODS_SN/3007660#nolink
http://odm.hybus.net/eng/index.asp
I just send them a mail to get the price and a more detail datasheet, but so for this looks great.



Amazing find max I knew somthing like that existed! Looks like regardless of what connector it is.. Were one step closer boys!


Fri Aug 09, 2013 9:55 am
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
Great find maxoh, let us know what they say.

Thanks for taking pics of the nexus zacherynuk.

Does anyone have an idea what the smaller, 90 degree bent cable is? Or the ic chips on the pcb?


Fri Aug 09, 2013 11:57 am
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
@KBK

MIPI-CSI is something else, and most likely wouldn't work. However..

@Insconthen
That chip you posted is another option besides the Solomon Systech part. Wonder how new it is as I haven't came across it before.

@zacherynuk

Great pics! You can see the traces WAY better than those from the iFixit teardown. I can count out 8 differential pairs which makes me think that the panel may indeed be LVDS. I had some doubts at first since some of the older Qualcomm Snapdragon SoCs only had MIPI-DSI outputs and not LVDS. Well, I'm still not 100%, but looking at the specs for an SBC based around a newer SnapDragon SoC shows that it does do LVDS without a converter chip like the older parts required.

Ok, this SBC has the same Qualcomm APQ8064 SoC as the new Nexus7, and if you look at the block diagram it shows support for both LVDS, and MIPI-DSI

http://www.inforcelive.com/index.php?ro ... path=72_73

So now the question is, would MIPI-DSI @ 1200p and (presumably) 24bpp require 8 differential pairs?


Fri Aug 09, 2013 1:43 pm
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
MSat- the LG 5.5" is 1080p MIPI 4 Lane and the cable is 30 pin. Everything the panel needs is on the 30 pin. 1 for ID. 8 for 4 lane MIPI. 2 for MIPI clock. 1 Vsync. 1 Reset. 10 grounds. 4 for LED. 1 iovcc. 1 vsp and 1 vsn. Does this help?

I'm still curious on the 90° angle cable in zacherynuk's pic of the PCB and cable connector. I'm thinking the long flat cable might be an extension cable from nexus motherboard to a PCB that takes in MIPI or LVDS. If that smaller cable is just MIPI signal, then it could be double sided with, on the top, you see the MIPI 4 data lanes(P) and 1 MIPI clock(P). and the other side would have the 4 lanes(N) and clock(N).


Last edited by Inscothen on Fri Aug 09, 2013 2:08 pm, edited 1 time in total.



Fri Aug 09, 2013 2:02 pm
Profile
Binocular Vision CONFIRMED!
User avatar

Joined: Sun Oct 03, 2010 2:56 pm
Posts: 295
Location: England
Reply with quote
Weird Right angle connectors... They are fairly standard affair nowadays, Molex make some common ones - I encountered them in 2010 here: http://www.wastedspace.co.uk/cms/2010/1 ... ct-part-9/ They are Bastards.... ... But mean squat diddly, they are wonderful snappy things that allow production components to be snapped together easily and modularity to exist; and they do not reflect the underlying electrical protocol.

The Chips / IC's on the panel side are just power regulation by the look of it, I couldn't easily see alternative power input so I am fairly confident it's all down the cable.

It simply smells of LVDS to me, not that it matters without some idea of pinouts! Dunno if it's feasible to probe or not.

I have emailed a couple of random LVDS driver board suppliers to see what they say - but since it seems we can 6 month confidentiality until November it would take something special to prove one way or another....

For spares, known knowledge and reusability of Fabs. productions, part numbers and existing IC's I would have thought LVDS would be easier for asus. Also; the connecting cable is identical having (google imaged it) from the nexus 7 and 7.2 - perhaps they are just straight swaps ? ... Be interesting eh ? So, for the sake of $33 I have ordered a replacement nexus 7 screen will plug it into this thing and see what happens... Android is good at determining it's own running PPI and Aspect ratio for rendering purposes and I assume it also have some generic drivers for use during recovery mode (Power and volDn) so we'll see.

_________________
My HMD Blog: http://www.wastedspace.co.uk/cms/category/projects/vintage-head-mounted-display-upgrade/


Fri Aug 09, 2013 2:06 pm
Profile
One Eyed Hopeful

Joined: Thu May 02, 2013 7:28 pm
Posts: 41
Reply with quote
zacherynuk wrote:
For spares, known knowledge and reusability of Fabs. productions, part numbers and existing IC's I would have thought LVDS would be easier for asus. Also; the connecting cable is identical having (google imaged it) from the nexus 7 and 7.2 - perhaps they are just straight swaps ? ... Be interesting eh ? So, for the sake of $33 I have ordered a replacement nexus 7 screen will plug it into this thing and see what happens... Android is good at determining it's own running PPI and Aspect ratio for rendering purposes and I assume it also have some generic drivers for use during recovery mode (Power and volDn) so we'll see.




Awesome cant wait to see! Do you have a shipping estimate?


Fri Aug 09, 2013 2:32 pm
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
here's the galaxy s4 test cable, which makes me think there's a chance the large/long nexus cable might just be an extension cable and that smaller cable may be a MIPI signal(4-lanes + clock) that feeds into the PCB the long cable is connected to.

http://www.etradesupply.com/samsung-galaxy-s4-series-lcd-screen-test-flex-cable-ribbon.html

zachery- do you have any other photos or is it too hard to get to the rest of the display panel?


Fri Aug 09, 2013 2:40 pm
Profile
Binocular Vision CONFIRMED!
User avatar

Joined: Sun Oct 03, 2010 2:56 pm
Posts: 295
Location: England
Reply with quote
Inscothen wrote:
here's the galaxy s4 test cable, which makes me think the large/long nexus cable might just be an extension cable and that smaller cable may be a MIPI signal(4-lanes + clock) that feeds into the PCB the long cable is connected to.

http://www.etradesupply.com/samsung-galaxy-s4-series-lcd-screen-test-flex-cable-ribbon.html

zachery- do you have any other photos or is it too hard to get to the rest of the display panel?


The pics I took were from the only cable I could see going from the panel to the motherboard - basically the one I wanted to have a good look at from the ifixit page here: http://www.ifixit.com/Teardown/Nexus+7+ ... wn/16072/1

The display panel I had visible (under the daughter board which controls the touch) was all what looked like power regulation to me, I was hoping to get something a bit more solid before delving too deep (hair dryers n poop are the next step!)

Having said that, there was another sticky pad south of the current pictures which may hide some IC's of note (You can see a load of tracks running that way) however generally speaking there is f'all you can see on the LCD's Eg: Nexus 1 (OEM2) : http://img.inkfrog.com/click_enlarge1.p ... =677206395

Poop? Seriously ?

_________________
My HMD Blog: http://www.wastedspace.co.uk/cms/category/projects/vintage-head-mounted-display-upgrade/


Fri Aug 09, 2013 2:53 pm
Profile
Two Eyed Hopeful

Joined: Mon Nov 19, 2012 12:14 am
Posts: 92
Reply with quote
@KBK For MIPI-DSI based interfaces you have to look at the maximum frequency limits of the PLL. For the SD2828 this limit is 1000MHz (see the PLCR register, 0xBA). MIPI-DSI requires that the PLL be at least 6 times the video pixel clock frequency when all 4 data lanes are used with the PLL locked to the pixel clock and slightly more than 6 times if asynchronous to the pixel clock. Based on this the absolute maximum video pixel clock is 166.66MHz for the SD2828. Given that pixel clock you can get a ballpark estimate of the max rez supported at 60Hz with minimum blanking intervals.


Fri Aug 09, 2013 3:39 pm
Profile
Two Eyed Hopeful

Joined: Mon Nov 19, 2012 12:14 am
Posts: 92
Reply with quote
I want to caution anyone from buying a DVI or HDMI to MIPI-DSI board without being absolutely sure the panel is supported by the board in question. At minimum the master MIPI-DSI LP clock has to have the frequency and associated delays configured to correctly match the requirements of the LCD panel. The SD2828 has several configuration registers for this purpose. Secondly the panel may require a sequence of post reset commands to initialize it for video operation. For example some panels must be told if the master will operate in burst mode or non-burst mode.


Fri Aug 09, 2013 3:49 pm
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
On the MIPI panels I've seen there's really only been a driver ic and some capacitors. Would the Nexus' Snapdragon require anything more than a driving ic and some power components on the display panel since it outputs LVDS?

LaserEdge- Thanks for the info. If you have any more info, don't hesitate to post. I think we all would appreciate it.


Fri Aug 09, 2013 3:53 pm
Profile
Two Eyed Hopeful

Joined: Mon Nov 19, 2012 12:14 am
Posts: 92
Reply with quote
For MIPI those two caps are most likely for the backlight driving circuits. You want to put them just before going over the FPC cable to make sure the voltage level remains constant for the LED strings in the backlight. Because of MIPI's ECC and CRC implemented directly into the packet structure plus differential signalling it doesn't require any components to filter noise from the MIPI clock and data lines.


Fri Aug 09, 2013 4:09 pm
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
zacherynuk wrote:
Weird Right angle connectors... They are fairly standard affair nowadays, Molex make some common ones - I encountered them in 2010 here: http://www.wastedspace.co.uk/cms/2010/1 ... ct-part-9/ They are Bastards.... ... But mean squat diddly, they are wonderful snappy things that allow production components to be snapped together easily and modularity to exist; and they do not reflect the underlying electrical protocol.

The Chips / IC's on the panel side are just power regulation by the look of it, I couldn't easily see alternative power input so I am fairly confident it's all down the cable.

It simply smells of LVDS to me, not that it matters without some idea of pinouts! Dunno if it's feasible to probe or not.

I have emailed a couple of random LVDS driver board suppliers to see what they say - but since it seems we can 6 month confidentiality until November it would take something special to prove one way or another....

For spares, known knowledge and reusability of Fabs. productions, part numbers and existing IC's I would have thought LVDS would be easier for asus. Also; the connecting cable is identical having (google imaged it) from the nexus 7 and 7.2 - perhaps they are just straight swaps ? ... Be interesting eh ? So, for the sake of $33 I have ordered a replacement nexus 7 screen will plug it into this thing and see what happens... Android is good at determining it's own running PPI and Aspect ratio for rendering purposes and I assume it also have some generic drivers for use during recovery mode (Power and volDn) so we'll see.


I would personally do a unpowered then powered probe first, to see if the pins match up, before plugging that puppy in. the rest of the circuitry can probably take/handle a signal mismatch, but power/pin mismatch might spell disaster. I'm advising due caution, but I'm also thinking that you'll do this 'check' anyway. Just saying it out loud.

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Fri Aug 09, 2013 10:31 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
I've been trying to scrounge up examples of MIPI and LVDS interfaces and compare them to what is observable on the ribbon cable. So to reiterate, it looks like there's 8 differential pairs.

MIPI would be 4 data lanes (and maybe a clock lane?) which means 5 pairs at most. This is based strictly on the part overview for the SSD2828 and Toshiba ICs (can't download the Solmon datasheet posted elsewhere).

When it comes to LVDS, the situation isn't as straightforward as I thought. For instance, I found the THine THC63LVDM83E which is a 24-bit LVDS transmitter. What's interesting about this device is that it can do 1200p @ 24-bits over 4 data + 1 clock pairs. This is in contrast to the more common interface I've come across, which brings me to:

Dual channel LVDS. From my understanding, it's essentially two separate LVDS streams each with their own data and clock lines - one channel controls the even pixels, while the other controls the odds. There may be exceptions (probably), but commonly dual channel LVDS @ 18-bit comprises of a pair of clock lines and 3 data channels, for a total of 8 diff pairs (Hey! That's exactly the number on the ribbon cable!). That means each data line sends a serial stream of 6bpp for the total of 18-bits. 24-bit is essentially the same, except there's an additional data pair which transmits the last 2 LSBs (I believe - might be MSBs) from each color channel - therefore it also serially sends 6bpp.

So, unless each data channel transmits more than 6bpp, I think what we're looking at on the Nexus7.2 is dual channel 18-bit LVDS.

Breakdown of zacherynuk's ribbon cable pic:

Attachment:
Nexus-7-2nd-Generation-Panel-Closeup-01.jpg


Of course, the labels are assumptions, but I think it's approximately close, at least as far as which group is the clock pairs and which is the data pairs. I also pointed out the lone trace which I assume is some sort of data line. Not sure where the I2C/SPI traces for the touch screen controller are. Maybe the other side? Those should be able to be found easily enough with a multimeter.


You do not have the required permissions to view the files attached to this post.


Fri Aug 09, 2013 11:16 pm
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
Doesn't the touchscreen use an entirely different cable?


Sat Aug 10, 2013 12:32 am
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
Inscothen wrote:
Doesn't the touchscreen use an entirely different cable?


Yep. Forgot about that.

I still feel a bit uncomfortable about my assumptions. At least as far as it being an 18bpp panel. I would expect a modern high-end panel to be 24-bit. Even the dev kit has a 24bpp panel (at least according to a previous statement by Palmer).


Sat Aug 10, 2013 12:31 pm
Profile
Binocular Vision CONFIRMED!

Joined: Fri Aug 31, 2012 11:21 pm
Posts: 203
Reply with quote
It's cool MSat. I think it's better to try to try to find the answer then to sit on our butts and wait for someone else to do it. It's probably 24-bit, but maybe not.

Here's a git for the nexus 7 2013 if there's a way to find out anything from there:

https://android.googlesource.com/device/asus/flo

And here's the binaries although I don't know if you can find any info from these:

https://developers.google.com/android/nexus/drivers#razor


Sat Aug 10, 2013 1:42 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Fri Jun 08, 2012 8:18 pm
Posts: 1329
Reply with quote
Inscothen wrote:

Here's a git for the nexus 7 2013 if there's a way to find out anything from there:

https://android.googlesource.com/device/asus/flo

And here's the binaries although I don't know if you can find any info from these:

https://developers.google.com/android/nexus/drivers#razor



Ah, good idea! When I was doing my research yesterday, I came across several forum posts where people were asking for the appropriate settings in linux to get LVDS panels working on their SBCs... There might very well be something in here to shed some light.


Sat Aug 10, 2013 2:15 pm
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
Another thing, is that one of the spacer traces does not necessarily have to be ground, it can be DC Power. When it comes to interference blocking... clean DC can be as good as a ground.

Another thing is that the sanpdragon is capable of 24 bit, from my understanding. No need to short the image fidelity for the sake of a cable and bad spec panel at this level of 'game'.

They'd get shot with a ball of their own poop if they were caught going to lower image fidelity with a panel of that rez and size.

~~~~~~~
With the HDMI/DVI to MIPI board showing up, the worst case solution is that this panel, whether it be MIPI or dual LVDS, can be fitted into a Rift chassis and be made functional.

extra edit credit: There is at least one line, possibly two on the back of the cable. A global ground, and then two wide traces.

double course credit edit: Any panel any time any place, that's what this new MIPI converter board means. For example, with diminutive IPD, one can go for the Samsung 5.5" and gain super high PPI.

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Sat Aug 10, 2013 5:06 pm
Profile
One Eyed Hopeful

Joined: Wed Jan 02, 2008 9:45 am
Posts: 3
Reply with quote
I did get a feedback about the HDMI/DVI->MIPI board. Sounds quite promising. Here the facts:
Attachment:
HBTM5(MIPI Bridge)_1.pdf

- the board will cost 150 USD
- Minimum quantity seams to be 100(http://www.alibaba.com/product-free/144559107/HBTM5.html), on buykorea it was 10, so let's see how to get a board
- still not clear what the refresh rate is, there's not that much info in the datasheet, but i would be surprised if the board wouldn't drive 60hz
- this board is the Version01, they are already planing a new board with smaller dimensions
- each board will include one interface board for a specific lcd panel and here it gets interesting! Like KBK mentioned in the last post, this board could open the door for other full hd displays, which are already available. Check out the page 3 + 4 on the pdf doc i attached:
Attachment:
hdmiMipiDia.png

Attachment:
lcdPanels.png

The sharp display seams to be the LS050T1SX01(see the diagram):
http://www.panelook.com/LS050T1SX01_SHARP_5.0_LCM_overview_17692.html
The other display could be the one from the LG Optimus G Pro, i didn't found a type nr:
http://www.panelook.com/LH550WF1-SH01_LG%20Display_5.5_CELL_overview_19484.html
http://www.etradesupply.com/oem-lg-optimus-g-pro-e980-lcd-screen.html

The LG display looks quite nice and seams to be easy available + the interface board has a nicer formfactor.
Shall we start a new thread about this MIPI bridge ?


You do not have the required permissions to view the files attached to this post.


Sun Aug 11, 2013 12:40 am
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
'silicon image' and the Solomon chip.

As for separate thread, not quite yet. Not sure 'bout the nexus panel yet.

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Sun Aug 11, 2013 7:12 am
Profile
One Eyed Hopeful

Joined: Thu May 02, 2013 7:28 pm
Posts: 41
Reply with quote
KBK wrote:
'silicon image' and the Solomon chip.

As for separate thread, not quite yet. Not sure 'bout the nexus panel yet.


I agree lets see if we can indeed do this swap first! Then we can move on to smaller displays... This is our best option for a HD dev Rift at the moment. We have 2 options to use the nexus 7 LCD regardless of connector once we get it all said and done 5.5 inch is next. :P


Sun Aug 11, 2013 8:23 am
Profile
Two Eyed Hopeful

Joined: Thu Jul 19, 2012 1:58 pm
Posts: 58
Reply with quote
The 5.5 is what oculus is testing now, that looks like the best choice at this point . Plus from the pics we know it all works!


Sun Aug 11, 2013 8:45 am
Profile
Terrif-eying the Ladies!
User avatar

Joined: Tue Jan 29, 2013 2:05 am
Posts: 910
Reply with quote
Flim wrote:
The 5.5 is what oculus is testing now, that looks like the best choice at this point . Plus from the pics we know it all works!



The 5.5" does not fit my IPD of 68mm.

If you've got an IPD of 60mm or less, then knock yourself out. However, that is notably below the 'mean' (Average) of western males.

Which means it can be a nice personal project but is not to be pushed as a global solution for everyone to join in with.

That may change with the introduction of larger 1200p MIPI panels, and whether the nexus is a MIPI panel or not, is yet to be established.

My original speculation placed the panel as being slightly more likely to be LVDS, as 'go with what you know, don't fix it if it ain't broke' was the way a company in a rush would go, instead of introducing the potential for problems in an area they had solved and was working just fine. To add, no need to experiment as the unit will never go past the rubber stamp level in it's entire product run, so no need to worry about experimenting with MIPI in this particular 'rush to market' of the second gen tablet (Nexus Google 7 gen 2 tablet). Asustek has HUGE experience with LVDS, across multiple platforms, thus, even less of a need to experiment with such a lucrative contract as this google tablet which was built like a high speed marathon run from start to finish.

MSat seems to be seeing potential for it to be LVDS, and I admit I see it too, after he explained it so well, and Zack backed it up from his direct observations and experiences.

Laseredge advised caution without breaching, uhm, er, potential personal concerns..if you were listening closely enough. (and that the panel may require a form of initialization at turn on)

Thus, it may require only a new LVDS driver board (with dual LVDS out), a cable connectivity conversion, and a flash of the panel parameters to said board. Then a bypass (just physical adjacent run, is all) of the tracker USB adjacent to the LVDS proper, and a hackenrift would be up and working at 1200p. Never mind the other issues, like drivers/software/SDK, etc.

_________________
Intelligence... is not inherent - it is a point in understanding. Q: When does a fire become self sustaining?


Sun Aug 11, 2013 9:29 am
Profile
Binocular Vision CONFIRMED!
User avatar

Joined: Sun Oct 03, 2010 2:56 pm
Posts: 295
Location: England
Reply with quote
The primary issue we have, I think, is that regardless of the panel type, we would be treading water until we knew the pinouts and timings and could get somebody to program a driver board.

I think, even if the panel is a sub-company of j-display, until the specs go public from JDI (unlikely) or until ASUS open up the specs to their developers and repair partners we just don't know how to configure it.

Currently I can find no WUXGA 7" panels nor specs for them... I think this has to change first and foremost.

I have asked a handful of driver suppliers who do not know of any such panel either and cannot match the product version we know it by.
I haver asked asus and JDI for tech specs; but I believe we are under the 6 month secrecy agreement still; active until November 26th.

Short of a miracle by an oscilloscope god (or his nominated messenger), I can't see progress being made too quickly from here...

_________________
My HMD Blog: http://www.wastedspace.co.uk/cms/category/projects/vintage-head-mounted-display-upgrade/


Sun Aug 11, 2013 11:25 am
Profile
Two Eyed Hopeful

Joined: Thu Jul 19, 2012 1:58 pm
Posts: 58
Reply with quote
Best bet is to get the 5.5 and mipi combination working... We can do it now.


You can adjust Ipd all day long, and since oculus is using the same setup without any complaints, it looks It will be a good hold over until commercial release.


Sun Aug 11, 2013 11:54 am
Profile
Display posts from previous:  Sort by  
Reply to topic   [ 2173 posts ]  Go to page Previous  1, 2, 3, 4, 5, 6, 7, 8 ... 55  Next

Who is online

Users browsing this forum: Bing [Bot] and 4 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
Powered by phpBB® Forum Software © phpBB Group
Designed by STSoftware.