Re: LCD evaluation and other preparations for FC handset
OK, so I picked up the package from HaoRan LCD company and examined their LCDs. Two observations, one expected, the other unexpected: The expected observation: the end of the FPC tail that is meant to go into a connector has straight edges, the corners are sharp 90 deg, no rounding. The unexpected observation: the contacts are on the "wrong" side. On both the historical GPM526A0 LCD in the Pirelli DP-L10 and the shiny new KD020C-2A from Startek the contacts appear to be on the top when the FPC tail is stretched out or on the bottom when the FPC tail is folded under the LCD; HaoRan's HT020K1QC36S has them on the opposite side. Looking at their datasheet more carefully I see that the drawing included therein does indicate the contact location, i.e., the delivered product does match the datasheet description - I simply overlooked it when I originally studied this DS, as I had naively assumed that the contacts would be in the "natural" location. Having the contacts on the wrong side fully invalidates my earlier idea of using a Hirose FH33 series connector - these connectors are not made in a "contacts on top" version, and it is difficult to fault Hirose for not making such connectors as the "contacts on top" arrangement is highly unnatural. In light of this discovery I have revised my plan of action: given that our current most immediate need is to compare the LCD picture quality between HaoRan and Startek and help guide the decision between these two vendors, in the interest of time expediency I am simplifying my HT020K1QC36S LCD test board idea: I will put this LCD on the test board with its tail stretched out (not folded under like it would need to be in the real handset), and use the known-good FH12-36S-0.5SH connector. This test board won't help with prototyping the mechanical arrangement for the handset, but it will hopefully allow us to solidify the Startek vs. HaoRan decision. Given that the two vendors' LCDs have their FPC contacts on relatively opposite sides, the mechanical arrangement would have to be quite different for one vs. the other, thus it would be easier to select the LCD vendor first based on the picture quality comparison, and then work out the mechanical arrangements for that chosen LCD. Now that I have the FPC connector picked out for the HaoRan LCD test board and the discussion with Hirose is being postponed until after the LCD picture quality comparison, there is no more obstacle and I can now proceed with the design of this very simple test PCB. I am hoping to get this PCB design done and sent out to fab some time this week. Hasta la Victoria, Siempre, Mychaela aka The Mother ___ Community mailing list Community@freecalypso.org https://www.freecalypso.org/mailman/listinfo/community
Re: LCD evaluation and other preparations for FC handset
> [...] this idea will NOT work for our FreeCalypso handset. Our PCB will > have 8 layers with lots of traces on every layer (intricate 2+4+2 HDI > structure with inner buried vias and two levels of staggered microvias > on each side), and the width of the PCB will extend only slightly past > the width of the LCD module. If we had to cut a slot the width of the > LCD's FPC tail in the middle of our PCB, it would cut through all of > our layers and present an obstacle for signal routing, completely > crippling our PCB layout. In short, just no - totally not suitable > for a PCB of our complexity. Then, another idea is rotating the LCD 180 degrees, then using some kind of LCD command that would allow to rotate the image accordingly - or rewrite the framebuffer output routines accordingly. I assume that, if you're not going through the center of the PCB, but making a slot on the top, the PCB routing shouldn't be as much of a pain. I understand that you've already picked this path, but I do think it's necessary to point out other options, since I know it might not be as obvious =) Cheers! Arsenijs ___ Community mailing list Community@freecalypso.org https://www.freecalypso.org/mailman/listinfo/community
Re: LCD evaluation and other preparations for FC handset
Hello FC community, An update on the FreeCalypso handset project: I am still in the process of evaluating candidate LCDs from different vendors. All of the candidate LCD manufacturers I am evaluating are based in Shenzhen, China - that city seems to be the world capital for fast-turn electronic design and manufacturing - and in the present case it is home to a large number of different LCD manufacturers. The only LCD I have successfully tested so far (as in got to the point of displaying pictures) is the one from Startek, model KD020C-2A - Startek was the first vendor to send me usable sample pieces, and they included a breakout board that made it much easier for me to assemble a working test rig. Here are some pictures of this test rig: https://www.freecalypso.org/members/falcon/pictures/lcd-test-rig/ The test rig consists of the LCD itself, the LCD vendor's breakout board, one of our FT2232D boards for interfacing and a rat's nest of wires connecting the LCD breakout board to the FT2232D board; the whole mess is mounted on a 6x12" sheet of acrylic material with double-sticky tape. The only outside connection to this whole arrangement is USB on the FT2232D adapter board, used for both power and programming: as soon as USB is plugged in, the LCD backlight lights up as it is powered directly from USB 5V through a resistor in series with each LED. I programmed the EEPROM on the FT2232D board with my own USB ID so the connected USB host won't treat it as a serial device (it is not a serial device - I am using the FT2232D chip in the so-called MCU host bus emulation mode, which is completely different from serial or SPI/JTAG), and in order to talk to the controller inside the LCD module and put pictures on the LCD, I run a special test program based on libftdi (published in the freecalypso-hwlab Hg repository on Bitbucket) that issues the necessary LCD controller commands through the FT2232D running in FTDI's MCU host bus emulation mode. The original KD020C-2A LCD had 12:00 viewing direction which we don't want (it is the opposite of the direction from which cellphone LCDs are meant to be viewed), but the sample pieces which Startek sent me are a modified version - it has been modified with some special film that somehow improves the viewing angle. I don't really understand it myself, but most of the test pictures I have tried displaying actually look better on this LCD from the 6:00 viewing direction (the one we want for the cellular handset application) than from 12:00. Overall the picture quality is very good, and this LCD from Startek would definitely make a good choice for our FreeCalypso handset product. It is definitely better than the LCD on the Pirelli DP-L10, and my S.O. says that the picture looks even better than on her Nokia phone. As a matter of due diligence though, I do need to evaluate one other LCD from a different vendor before we officially commit to using Startek LCDs. According to both the LCD datasheet and the vendor's communication with me, the other vendor's LCD is officially made for the 6:00 viewing direction, whereas Startek's LCD was originally 12:00, but then modified in some way which I don't really understand. It is possible that Startek's LCD with their special film is so good that it's even better than the other vendor's LCD made for 6:00, but I need to see both with my own eyes before I can commit to anything. The other vendor in question already sent me their sample pieces, and they are supposed to arrive on Monday California time. But they don't have a breakout board similar to Startek's for their different connector, so I will need to make my own test PCB. I plan on still using an external FT2232D breakout board rather than incorporating the FT2232D into my own test board - this way the test board will be simple enough for me to lay out myself. I'm going to give pcb-rnd a try on this LCD test board - I have previously used gEDA/PCB, but pcb-rnd has been making great progress and might be suitable for doing the actual handset motherboard layout way down the road, so I'm thinking of trying it out for the very simple board. I'm hoping to do this layout some time this coming week, and then it will probably be another 2-3 weeks to actually get the PCB made - but it is needed so we can properly evaluate both LCD vendors and make a fully informed decision. Stay tuned for further updates. Hasta la Victoria, Siempre, Mychaela aka The Mother ___ Community mailing list Community@freecalypso.org https://www.freecalypso.org/mailman/listinfo/community
Re: LCD evaluation and other preparations for FC handset
Hi Mychaela, You are unfortunately correct, that trace is on L6. Looking at L7 and bottom layers, there are many traces and I can't see any easy way that trace could be cut without a PhD in PCB microsurgery. :/ > > BTW is there an easter egg in the PCB layout? I can read "TimLee Allen > > Shawn Jean Shanny" above the SIM holder. > > These names/signatures made out of microvia patterns have been > preserved from Openmoko. It appears that back in the days of Openmoko > HDI PCB microvias were made by some different process compared to how > they do it nowadays, and on OM's original GTA02 boards these microvia > patterns (names/signatures) are actually visible with the naked eye on > the finished product boards, on the side with the display. When my > Iranian contacts made the FCDEV3B layout from OM's GTA02, they kept > OM's names/signatures (fairly so, as we are using the modem layout > which is originally OM's), but they were moved to a slightly different > location, which is where you see them now. But they are no longer > visible on the finished board (except perhaps via X-ray) because > today's PCB fabs make these microvias in some different way that > results in them being invisible. Cool! Thanks for the explanation. For the curious: https://imgur.com/a/3Jife If you look at http://wiki.openmoko.org/wiki/Who_is_Who apparently Tim Lee was the hardware manager, Allen may have been Allen Chang (hw engineer), Shawn was most likely Shawn Lin (RF Engineer), Jean and Shanny I'm not sure. --DS ___ Community mailing list Community@freecalypso.org https://www.freecalypso.org/mailman/listinfo/community
Re: LCD evaluation and other preparations for FC handset
Hi DS! > Looking at the project in Altium, it seems the signal trace from D5 to > U201 sits at the bottom layer. Do you have the ability to run Altium and open the PcbDoc natively? If so, highlight the net named FDP. I don't have the ability to run Altium myself (no Windows), but looking at the Gerber files in gerbv, I see that the trace from U301-D5 goes to a microvia right next to the ball, then on L7 it immediately goes to another microvia to L6, on L6 the trace goes to the area underneath U201, and then it goes through another pair of microvias back to the surface, coming out right next to U201 ball F4. Thus the long part of the trace is on an inner layer (L6), and the only parts that come out to the surface layer are the tiny runs on each end between the ball pad and the microvia, each completely under the respective IC package. Please double-check on your end - perhaps you were looking at a different signal? > I think it might be not too difficult to > do some surgery to cut the trace in the middle and pull it up elsewhere. I don't see how the trace cutting can done anywhere other than in the very short L8 (surface) run under U301 (requiring removal of that IC with a BGA rework station); cutting the trace at that point would be doable, but I don't see any way to pull up the U301-D5 side of the cut net. > BTW is there an easter egg in the PCB layout? I can read "TimLee Allen > Shawn Jean Shanny" above the SIM holder. These names/signatures made out of microvia patterns have been preserved from Openmoko. It appears that back in the days of Openmoko HDI PCB microvias were made by some different process compared to how they do it nowadays, and on OM's original GTA02 boards these microvia patterns (names/signatures) are actually visible with the naked eye on the finished product boards, on the side with the display. When my Iranian contacts made the FCDEV3B layout from OM's GTA02, they kept OM's names/signatures (fairly so, as we are using the modem layout which is originally OM's), but they were moved to a slightly different location, which is where you see them now. But they are no longer visible on the finished board (except perhaps via X-ray) because today's PCB fabs make these microvias in some different way that results in them being invisible. M~ ___ Community mailing list Community@freecalypso.org https://www.freecalypso.org/mailman/listinfo/community
Re: LCD evaluation and other preparations for FC handset
On Sat, Apr 07, 2018 at 06:21:15PM -0800, Mychaela Falconia wrote: > And we still need to fix and close the sleep mode bug. I have a high > confidence in my current hypothesis that FDP driving the flash chip's > reset line is the culprit, but it looks like we won't be able to test > it empirically until we build our next board, be it FCDEV3Bv2 or the > first version of the HSMBP. Hi Mychaela, Looking at the project in Altium, it seems the signal trace from D5 to U201 sits at the bottom layer. I think it might be not too difficult to do some surgery to cut the trace in the middle and pull it up elsewhere. I might try that on my FCDEV3B unless you want to try it on one of the boards you have. BTW is there an easter egg in the PCB layout? I can read "TimLee Allen Shawn Jean Shanny" above the SIM holder. --DS ___ Community mailing list Community@freecalypso.org https://www.freecalypso.org/mailman/listinfo/community