#milkymist IRC log for Thursday, 2012-01-19

kristianpaulrecongiguration indeed MAIN feature for geek loving :)02:17
kristianpaulah the cpu combo. well, i tought about it because the usb part...02:18
kristianpaulbut i still prefer give a try to a usb host core from opencores than using an asic to this state of comitment ;-)02:18
wolfspraulcladamw: hi there, good morning07:24
wolfspraulwe were wondering yesterday whether you already selected a DVI-I connector07:24
cladamwhi good morning07:24
wolfspraulthen we quickly did a digikey search for molex parts and found this http://search.digikey.com/us/en/products/74320-1004/WM5600-ND/35606607:25
wolfspraul74320 series07:25
cladamwno yet, but just refer. to Molex one: 74320-100407:25
wolfspraulquite expensive, 3-3.50 USD even for 50pc, but I guess we can find others with very similar specs that are cheaper07:25
wolfspraulor we just take the molex part for now, 3.50 USD is not totally bad07:26
wpwrakah, everybody has the same question at the same time :)07:26
cladamwyes, quite expensive but still have FCI likely07:26
cladamwi can check if our previous vendor of VGA connector for that.07:27
wolfspraulwow the bom says we pay 20 cents for our current VGA connector :-)07:28
cladamwthat VGA is traditional though.07:28
cladamwtaiwan or china should can source more chipper one.07:29
wolfspraulok, we can start with the molex one and then you check a few other vendors maybe, but I'd say not too many07:29
wolfspraulmaybe we can find another big and quality one that is substantially cheaper?07:29
wolfspraulif it's a small shop or hard to reach etc. then rather not, I think07:29
cladamwbut you know the request time from taiwan here is not good and fast, so use Molex from digikey for drawing firstly. ;-)07:29
wolfspraulonly causes problems later on07:29
wolfspraulyes, and most likely others will follow or be very close to molex dimensions, true?07:30
wolfspraulI would only compare to others we can easily source at digikey/mouser/farnell07:30
wolfspraullet's say that's a requirement?07:30
cladamwclose to surely the pins dimension but not all equalent to the same plasic size i think.07:31
cladamwto replace DVI-I is not the time now though. but surely we need to find another before R4 run.07:32
cladamwand also all include J21 and Jxx etc.07:32
wolfspraulit would help Werner with the expansion system dimensions07:32
wolfspraulbut we have a starting point with the molex one07:33
wolfspraulwpwrak: should we make sourcability of such a part on digikey/farnell/mouser a design criteria?07:33
wpwrak`yeah, if the molex connector is roughly what adam will use, that's all the information i need07:33
wpwrakat least of a drop in replacement07:33
cladamwneed to reading werner's reply... reading...07:33
wpwrakif you can find some lower-cost substitute, great. but you'll want to have at least a well-known part to fall back to.07:34
wolfspraulstill the same question, let's be more specific07:36
wolfspraul"lower-cost substitutes" may well be on dk/mouser/farnell as well07:37
wpwrakoh, sure. but it doesn't matter.07:37
wolfspraulthe question is whether for that type of part, we make sourcability on a big online distributor a *requirement* for us, or not07:37
wpwraksourcability of the same or an equivalent part07:38
wolfspraulif not, we may source from a Taiwanese company without significant online presence, sales staff without good English, etc.07:38
wolfspraulyou won't get a 100% drop-in match probably07:38
wpwrakwell, same specs :)07:38
wpwrakyou'll notice if the tolerances are way off :) and if there's not much of a price difference, you may want to save yourself the hassle with the small shop07:39
wpwrakbut each part should have a properly documented equivalent at one or more major distributors07:40
wolfspraulso it is a requirement07:40
wpwrakyes, this is07:40
wpwrakotherwise, you have weird bits in the system you cannot second source07:40
wpwrak... and they also have a high disappearance risk07:41
wpwrakwhile it would be less likely that, say, xilinx would just disappear over night07:41
wpwrake.g., was fun to go over the gta02 BOM ~2 years after the design. (in gta02-core) for about half of the .tw components, we couldn't even find the company anymore07:43
cladamwwpwrak, exactly you're right !07:43
cladamw.tw manufactures are going to or already either close or cut down on stuff member.07:46
cladamwbigger chips/components manufactures become more bigger.07:47
wpwrakheh :)07:48
wpwrakthey feed on the small :)07:48
cladamw(Y1) surely we can source flatter one.07:48
cladamwlike (X1, 12 MHz) in jtag-serial board, just need to find 24.576MHz.07:50
wpwrakkewl. something in the 3.0-3.5 mm range would be great. that would leave a nice round 5 mm for the extension board. an easy number to remember :)07:50
wpwrakah, that's a tiny one :)07:51
cladamwwhich means in the other hand ..... cost rising. ;-)07:51
wpwrakyeah, the really small ones can be expensive07:52
wpwrakbut it doesn't have to be *that* tiny. just a little thinner :)07:53
wpwrakthis one seems very compatible: http://search.digikey.com/us/en/products/ABMM2-24.576MHZ-E2-T/535-9639-1-ND/123710608:01
wpwraksame tolerance, better drift, same capacitance, other parameters (ESR, temp.) similar08:01
wpwrakor, even cheaper and better: http://search.digikey.com/us/en/products/NX3225SA-24.576000MHZ/644-1053-1-ND/112892508:02
wpwrakone of them can't beat the 10 cents of Y1, though08:03
cladamw(newest R4, draft) http://downloads.qi-hardware.com/hardware/milkymist_one/sch/tmp/Milkymist%20One%20R4%20-%2020120119.pdf08:03
cladamwI'm writing for review helps.08:04
wpwrakheh, schematics diffs would be handy now :)08:13
wpwrakJ22 looks good08:13
wpwrakhmm, so all 3.3 V pins are occupied now ?08:15
cladamwbank0~2 are run out!08:15
wpwrakhmm ...08:15
cladamwexcepts bank3, but it's ddram part...so you can see firstly08:15
cladamwalso the LEDs I also added to be as LED_USB[A:F] totally 6pcs.08:16
cladamwalso be noted that some driven by 2.5V, others by 3.3V.08:17
wolfsprauli think we have to keep that LED idea under control08:17
wolfspraulwe have how many now? 18?08:17
wolfspraulis that too much?08:17
wolfsprauldo we need one wire per led?08:17
wolfspraulsince we are trying to have a good design overall etc. - maybe that's not a good way to control 18 leds with 18 wires?08:17
wolfspraulI don't want to ruin a good design with a wasteful bunch of lights ;-)08:18
cladamw20 LEDs now, it's easy to delete them later.08:18
wolfspraul20, wow. and they need 20 fpga wires?08:19
cladamwsee page 6 of Misc.SchDoc08:19
cladamwespecially I put 6 ports USB leds in bank3, if no needs few of them , it's all good to remove later. ;-)08:20
wpwrakyou need one wire per LED only if you need 100% duty cycle08:22
wolfspraulI feel the led proposal we have right now is out of hand08:23
wolfspraulit feels wasteful08:23
wolfspraulbut I'm not 100% sure what the latest proposal is :-)08:24
wpwrakyeah, it does look a bit excessive08:24
wolfspraul20 leds, really?08:24
wolfspraulwith 20 wires?08:24
wpwraki don't like the idea of filling up the FPGA like this08:24
wolfspraulwhere in other places we are thinking carefully how to reuse and overlay wires?08:24
Fallenouwhat are those leds for ?08:26
cladamwhey...hehe...i just followed r3 known issues, we can downsize them now though...08:26
wolfspraulFallenou: we want to say m1 is as cool as arduino, you can have a lot of blinking lights on it! :-)08:27
Fallenouone of r3 issue is "not enough led for user to be happy with blinking stuff" ?08:27
wolfspraulok seriously, from my side I do like more lights, but I also think it still has to be a good idea and good execution08:28
Fallenouwolfspraul: I/O helps to sell that's right (buttons / switches / leds) because it helps prototyping or just have fun08:28
Fallenoubut 20 seems a lot08:28
wolfspraulat some point werner proposed a light at each port to indicate what a patch wants08:28
Fallenou16 would allow for the user to output 2 bytes08:28
wolfsprauleven though practically that will get very hard with things like usb where you don't really know what is being attached, that was a strong idea I thought08:28
wolfspraulFallenou: yes yes, we know08:29
cladamw2*dmx, 2*midi, 3*video-in, power, 6*usb, ir, dvi-i, ethernet, memcard, rendering, on/off08:29
cladamwlet's determine what we need them now. ;-)08:29
wpwrakwe could do some multiplexing. e.g., a simple NxM matrix would need N single-load sources and M N-load drains, have a duty cycle of 1/M, and use N+M pins08:29
Fallenouoh ok led to show what external device the current patch is using ? nice idea :)08:30
cladamwno rendering? no On/Off? 6 USB -> 3 ?08:30
cladamw3*video-in -> 1?08:31
wpwrakFallenou: the idea is to have the following states: 0) unused (off), 1) ready for use (e.g., permanently on), 2) error (fast regular blink), 3) in use (activity blink)08:31
Fallenouwill you put plastic kind of optical fibers to conduct light from led all the way to the side of the case ?08:31
wolfspraulnot random reductions, let's think about what those lights mean08:31
Fallenouwpwrak: ok sounds nice :)08:31
wpwrakFallenou: no fibers. just LEDs on the PCB.08:32
wolfspraulwhy does 'rendering' have a light btw?08:32
wolfspraulright now we have 308:32
wolfspraulif it doesn't render, you will see right away on the vga08:32
Fallenouwpwrak: will it be easy to locate/read for the lambda user ? "this lead means this" ?08:32
wpwrakthat's the idea. otherwise, it wouldn't make all that much sense :)08:34
Fallenouhow would you make it easy ? locate the led near the device/connector ? print the meaning on the PCB ? (printing maybe not easy to read during a party)08:35
wolfspraullocate near connector08:36
wpwrakhttp://en.wikipedia.org/wiki/Classical_unities ;-)08:37
wolfsprauldo we need 3 leds on rc3 ?08:39
wolfspraulpower-on seems good, but how about the other two?08:39
wolfsprauljust trying to understand what people feel about them. I only need the other two to workaround our boot-hang bugs in earlier hardware revisions08:39
wpwrakdunno. we've kinda gotten used to them08:40
wpwrakone is just power. that could go to the DC connector08:43
wolfspraulalthough it cannot perform the "I want power" function unless we have a battery to light it :-)08:46
wolfspraulhow do people implement this 'breathing'/dimming feature of leds that Apple popularized to indicate a sleeping notebook?08:47
wpwrakduty cycle variations08:48
wolfspraulcan we do that with the 3 leds we have today, the way they are wired up?08:48
wolfspraulI mean on rc308:48
wpwrakwith two of them08:48
wpwrakthe 3rd one is fixed to power, without sw control08:49
wolfspraulah k08:49
wolfspraulso yeah, those other two I personally don't need08:49
wolfsprauland the power one could be near the DC jack08:49
wolfspraulit will be difficult to get complete consensus on the leds among a majority of people here08:49
wolfspraulsome may not care, and then everybody else will have small or large differences in what they propose08:50
wolfspraulhow do we make everybody happy? :-)08:50
wolfspraulwe agree that > 10 fpga pins for this is wasteful?08:52
wolfspraulif we even have that many fpga pins, I think Adam is basically just drawing up a wishlist now08:52
wolfspraulI think we agree that we don't want any additional control ics08:52
wpwrakhmm. need to find out how low a duty cycle would still be clearly visible08:52
wolfsprauldo we agree that all leds should have the same color, green?08:53
wpwrakwe could add a little AVR ;-))08:53
wolfspraulyeah exactly08:53
wpwrakdunno about the color yet08:53
wolfspraulwpwrak: ok, can you think about the led idea from a holistic point a bit more?08:54
wolfspraulif you did, you would take the lead and make other people's lifes much easier08:54
wolfspraulit's very hard to find compromises on this kind of proposal08:54
wolfspraulit either makes sense as a whole or not08:54
wpwraki don't see much of a need for compromises just yet. it's merely an implementation question.08:55
wolfspraulah then let's go through08:55
wolfspraulyou just said "don't know yet" to the color question08:55
wolfspraul> 10 pins is wasteful?08:55
wolfspraulthat seems to be consensus08:55
wpwrakthe color is not a compromise - it's a question of picking one that works :)08:56
wpwrakyeah, > 10 pins is evil08:56
wolfsprauldo we really need that many leds? for all usb ports? 3 for video-in ?08:56
wolfspraulhow about the 2 non-power ones we have on rc3?08:56
wolfspraulwe just leave them because they were there before, or we remove them because they are not needed in the new overall design?08:56
wolfsprauland the power-on one - move to DC jack?08:57
wpwraki'd move the power LED08:57
wpwrakthe others ... well, at least one shows us how the booting it progressing08:57
wpwrakthe other too, but to a lesser extent08:58
wpwrakalright .. let's switch to the leds ...08:59
wolfspraulif we really have those 14+ leds, there will be ample opportunity to watch the booting progress :-)08:59
GitHub45[milkymist] wpwrak pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/f2650b256ad0a654f465a5d2cc215c4999f1930a09:00
GitHub45[milkymist/master] libfpvm: rename op_not to op_bnot, to follow MilkDrop naming - Werner Almesberger09:00
GitHub77[flickernoise] wpwrak pushed 3 new commits to master: http://git.io/6Lo3dQ09:00
GitHub77[flickernoise/master] compiler: rename op_not to op_bnot, to follow MilkDrop naming - Werner Almesberger09:00
GitHub77[flickernoise/master] compiler: add bnot(x) as MilkDrop-style synonym for !x - Werner Almesberger09:00
GitHub77[flickernoise/master] compiler: boolean operations "and" and "or" (WIP, untested) - Werner Almesberger09:00
wpwraklekernel: where did you find the per-bank current rating ? i'm searching in "Spartan-6 FPGA Data Sheet: DC and Switching Characteristics" but don't see it :-(09:14
cladamwuse chip equipped with I2C control with output bits to save fpga pins? or AVR?09:39
wpwrakstep 1: see what multiplexing can do :)09:40
wpwrakmultiplexing can give you something in the order of 2*sqrt(N) or perhaps even 2*sqrt(N/2)09:40
wpwrakso for 20, that would be 4+5 = 9 or maybe even less09:41
wolfspraulcladamw: no other ic09:52
wolfspraulI want to focus M1 on the Milkymist technology09:52
wolfspraulnot pile more and more things on it09:52
wolfspraulthat's my thinking at least09:53
wpwrakso much about the clever plan to sneak in an ARM 1 GHz dual core, disguised as LED controller :)09:53
cladamw1. added 20 LEDs to issue unknow bank maximum currents 2. multiplexing chip makes it possible but cost up 3. if we do really need this feature? If not, do we want to change 9*2 to 10*2 for J21/J22? 4. or pull saved fpga pins with suitable connector and find somewhere for it?09:54
cladamwnow question is should we really need this feature?09:55
wolfspraulwe don't want to waste a lot of pins on leds09:55
wolfspraulwe don't want an i2c or so control ic09:55
wolfspraulif anything we reduce the number of leds or multiplex them without control ic if possible09:55
wolfspraulI think, waiting for Werner's thoughts...09:55
wpwrakfor 18 LEDs, with multiplexing, we'd need: 18 LEDs, 3 resistors, 6 I/O pins (3 must be capable of If(m1), the other 3 must be capable of 3*If(m1). If(m1) is what we choose, e.g., 6 mA.)09:57
wpwrakthat multiplexing approach has one disadvantage, though: the LEDs may never be entirely dark. but that could be considered a feature :)09:58
wolfspraulsounds scary10:02
wpwrakit would be a very faint glow. basically whatever If you get at Vf = 3.3 V/2. in theory, that should be zero. in practice, it's a little bit more. so there will be a photon every once in a while10:04
wpwraklet's see how much it is ...10:04
wpwrakabout 70 uA at 1.7 V with a C190KGKT10:08
wpwrak> 100 uA at 1.65 V with a C190KRKT10:08
wpwrak(all with 220 Ohm series)10:09
wpwrakactually, it would be a bit less in practice, since we'd share the voltage drop10:09
lekernel_mh? why can't you turn LEDs entirely off with multiplexing?10:13
larscwith a n to n**2 multiplexer one port will always be on, of course you can turn of the multiplexer itself10:18
wpwraklekernel_: in an N*M*2 multiplexer, each crossing has two LEDs (anti-parallel)10:26
wpwraklekernel_: so on each row that's Z, between columns that are H and L, there will be some leakage current10:27
wpwrakmaybe you could avoid it entirely by Z-ing (instead of 0/1-ing) the column, though10:29
wpwrakin any case, the effect should be minor, given that we already drop to a "normal" Vf on the series resistor. so there's not much left for leaking10:31
wpwrakwe should be good with a duty cycle down to ~10%10:32
wpwraktried green and red. both looks fine through the light blue acrylic.10:33
wpwrakgreen suffers very little dampening, red a bit more. but red is much brighter than green. at least with the LEDs, i used.10:33
wpwraklet's check that current ... not sure if the two Vf are really the same10:37
wpwrak8 mA both. hmm. a little high10:39
wpwrakah no, power supply exaggerated. only 6.3-6.4 mA, according to the multimeter. that's acceptable.10:42
wolfspraulthe power led is not software controlled, right?10:48
wolfspraulcould that be considered a problem if someone wanted to turn the board entirely dark?10:48
wpwrakwell, there's always black paint ... :)10:51
wolfspraulI'm thinking about the user10:51
wolfspraulwho has a closed case in front of him in a dark setting, and maybe is frustrated that there is this stupid green light that cannot be turned off at all? I can imagine that.10:52
wpwrakthe user of black paint will succeed. or the user of a soldering iron :)10:52
wpwrakor the oser with a piece of black cloth :)10:53
wolfspraulneither is practical or user friendly10:53
wolfspraulwould there be a way for software to override/disable the power led?10:53
wpwrakthen a way to control the LED may be appropriate :)10:53
wolfsprauljust a thought, I haven't heard that from actual users10:53
wpwraklekernel_: what's the default setting of the pins after reset but before configuration again ? pull-down ?10:54
lekernel_pull up10:56
wpwrakmaybe we should make a qi-hw/m1 blog and call it "the overengineer" :-)10:56
lekernel_feel free to post on the wordpress :p10:56
lekernel_and yes... with those leds... haha10:56
wpwrakeven easier. then a NPN or equivalent FET should do. by default (pull-up), it'll conduct10:56
lekernel_unrelated - is there anything special about http://www.lab126.com ? seems to me it's only the EE lab of amazon10:58
lekernel_someone recommended me I check it out when in the US10:59
lekernel_and since I have absolutely no intention to work for amazon, well... :)11:00
wolfspraulif you have an introduction, sure why not11:00
wolfspraulyou should be open minded to see how other people work and listen to them11:00
wolfsprauleven though Amazon is on a good path to match or even surpass Apple in building the perfect secrecy system :-)11:00
wolfspraulkinda exactly the opposite of what we do here :-)11:01
wolfspraulif you work for them, it's the last time we've heard from you here, for sure...11:01
wpwrakthis would be the multiplexing matrix: http://downloads.qi-hardware.com/people/werner/tmp/mpx.pdf11:02
wpwrakcladams: this would be the multiplexing matrix: http://downloads.qi-hardware.com/people/werner/tmp/mpx.pdf11:37
cladamwwpwrak, wow ~11:56
cladamwactive-low and high to get combinations, but how many LEDs in the end we want to have?11:59
wpwraki don't know yet :)12:00
wpwrakcan you place LEDs under the R/G/B video in connector ? (there's room there. just don't know if the fab would like that)12:01
cladamwexcelent on that combination. ;-)12:01
cladamwshouldn't have problem.12:02
wpwrakkewl :)12:04
cladamwthat's also cool idea !12:05
cladamwdo you want those half-reserved pins to be pulled out?12:07
wpwrakthe pins you now assigned to LEDs ? whatever is more convenient for you. in the end, we should only need ~6 pins12:14
wolfspraulmaybe we can switch the color of all leds to red?12:15
wolfspraulif I think about red vs. green, I can be a little more excited about red, I think12:15
wpwrakheh :)12:15
wolfspraulthe pcb is black, the case light-blue12:15
wolfspraulblack & red is good12:15
wpwrakred is often brighter12:16
wolfspraulthen they are just all red, nothing green12:16
cladamw18 LEDs? don't know which two you didn't want?12:19
Fallenoublue is often very bright too (or maybe it just depends on the led properties)14:14
FallenouI just noticed that on a few dev boards with several led colors, the blue is generally the brighter one IIRC, don't know if this means anything14:15
Fallenouwhen you chose a led you can chose the number of cd (light intensity) on farnell website :o14:17
wpwrakyes, blue is bright. also needs a high voltage, though14:44
wpwraklet's see aboutthe cost ..14:44
wolfspraulred works well with the black pcb14:47
wpwrakred (the one i regularly use): LTST-C190KRKT, 54 mcd @ 20 mA, 2.0 V, USD 62.10 (1000)14:48
wpwrakbest blue < 3.3 V: LNJ937W8CRA, 17 mcd @ 5 mA, 2.9 V, USD 133 (1000)14:49
wpwraklet's see if there's a better red now14:49
wpwrakand indeed, there is. e.g., this one is almost twice as bright: APT1608SRCPRV, 100 mcd @ 20 mA, 1.85 V, USD 58.00 (1000)14:52
wpwrakeven this one is cheaper: APT1608SURCK, 220 mcd @ 20 mA, 1.95 V, USD 62.00 (1000)14:53
wpwrakall these are upward-facing. not sure if we'd rather want sideways-facing14:54
wpwrakin any case, they have an angle of ~120-130 deg. so you can see them well from the side14:56
wpwraknow .. let's see if i can find all of adam's 20 LEDs :)15:05
wpwrakah, that's the old ones plus 17 new. got it.15:06
wpwrakof these, one becomes a special case (LED3, DC in). that is, if we want to add the possibility to turn it off completely15:08
wpwrakat least LED1 (?) should probably stay the way it was, because the use would be a simple boot status indication15:10
wpwrakwill we reduce the number of buttons to one ?15:16
wpwrakadams' LED_ETHERNET seems unnecessary. there are no less than two LEDs already, beautifully integrated in the ethernet connector. and they work today :)15:17
wpwrakon the other hand, adam has no LEDs for audio in/out/mic15:18
wpwrakso if we keep all the ones he suggested, plus audio, minus ethernet, that would then be +19 LEDs15:19
wpwrakthe USB LED idea is a tad problematic. first, it seems to be quite uncommon. but then, i could imagine that a status indication could be rahter useful15:22
wpwraksecond, with vertically stacked USB receptacles, how do we indicate which is which. hmm.15:22
wpwrakah, and perhaps we should make all the LEDs glow a little intentionally, so that one can see where the others are (that would be a sw problem)15:24
wolfspraulI think we can reduce the buttons to 115:42
wolfsprauland I think we can reduce the number of leds, but I would have a hard time telling you which one to cut15:43
wolfspraulif the pins are reduced now that's a first step though, maybe enough15:43
lekernel_19 LEDs? USB LEDs? huh...15:49
wolfspraulrelax, let's trust werner's aesthetic judgment15:50
wolfspraulhe is testing...15:50
wolfspraullekernel_: 19 leds is the ultimate arduino attack! :-)15:51
wolfspraulthat's why m1 is more expensive :-)15:51
wpwrakmaybe we can then save the DMX output ;-)15:54
wpwraki also tried blue. but red is brighter, even the old red i have15:55
wolfsprauli've heard several times that blue looks bad and cheap15:55
wpwrakand you can get twice the mcd for a minimum difference15:55
wolfspraullast time I think from roh15:55
wpwrakhehe ;-)15:55
wolfspraulbut you are testing, I think we should follow with what you *see*15:55
wolfspraulI'm not going to remote-fiddle with that15:56
wpwraknothing is as obsolete as yesterday's fashion :)15:56
wolfspraulthe blue leds I have seen do look strange, like not really blue or so, don't know15:56
wpwrakremote-fiddling would be nice. but i think it's hopeless to try to convey the impression of colors and led intensity in pictures :-(15:56
wolfspraulno, let's not15:56
wpwrak(well, convey, yes. but it would show what i want you to see :)15:57
wolfspraulyou have the transparent-blue case, you have everything needed. so you can make the aethetic judgment, and if you do I would be more than happy about that.15:57
wolfspraulmy #1 feedback before was "oh my, too many fpga pins wasted"15:57
wolfspraulthat seems to have been addressed15:57
wolfspraulmy #1 feedback would be, since we work on leds anyway - can we also disable *all of them* ?15:58
wolfspraulsorry that's the #2 feedback :-)15:58
wpwrakit's more a functional judgment. for the aesthetics, i don't feel anything that says "wow" or "aww"15:58
wolfspraulyes agreed, I don't believe this will drive sales directly, it's more what makes you feel good about the product, and then we can hopefully convey that feeling to our users15:58
wpwrakyes, we can disable all of them. needs one FPGA pin and a transistor15:58
wpwrakwhat i'm more worried about is finding good places, so that one can actually see what they refer to. and now just get drowned in a sea of lights.15:59
wolfspraulthen my #3 feedback would be that the 3 leds we have could be focused better.16:00
wolfspraulyes, agreed16:00
wolfspraulI think the 3 leds we have is just an initial design idea (and not a bad one), to have 'three' because that's a nice initial choice16:01
wolfspraullike the 3 buttons16:01
wolfspraulbut if we go and work on leds now, the meaning of those 3 should become part of the newer larger idea, and we don't just leave those 3 around there because there were there before16:02
wolfspraulif we add leds for ports, I think we can remove 2 of those 3, and just leave the power-on one16:02
wolfspraulwhat would people really do with those leds, I'm still wondering ;-)16:03
wolfspraulI just imagine my m1 here would have 15+ leds :-)16:03
wolfsprauland then what?16:03
wolfspraulit would blink like crazy all the time?16:03
wolfspraulthat would definitely make it look alive16:04
wolfspraulyeah why not, m1 illuminated...16:05
Fallenouhope it will not be an eye killer16:05
wolfspraulwpwrak: can you include the existing 3 leds in your thinking?16:12
wolfspraulhey, just because we are discussing - how about leds at the bottom side of the pcb?16:14
wolfspraulwith the semi-transparent acrylic we have that might look quite interesting16:14
wolfspraullike the m1 is floating16:14
wolfspraulit could cause emi trouble though and we also still have this whole issue with the bottom metal sheet and Joerg's proposal with 'X' aluminum wires...16:15
wolfspraulanyway, just had that idea :-)16:15
wpwrakFallenou: we can always dim things :)16:21
wpwrakwolfspraul: (bottom) hmm, would have to give this a try16:22
wpwrak(existing leds) agreed on moving LED1 (DC power)16:23
wpwrakfor LED2/LED3, i wouldn't mind getting rid of one of them. we'd lose a little bit of boot information, but not a lot16:23
wpwrakthe remaining one could then snuggle up to the button16:24
wpwraknow, waiting for my camera to charge ...16:50
wpwraklet's hope the camouflage painting on the led cable does the trick16:51
wpwrakany preference on where LEDs should go if we can't place them under/in front of the item ?17:09
wpwraki.e., always left, always right, maximize spread, ... ?17:10
wolfspraulhmm. which one for example?17:10
wpwraklet's say the audio-mic-vga line17:11
wolfspraulfor those 3 I'd say on the right side, towards the front17:12
wpwrakright side as seen from the front ?17:13
wolfspraulclose to the acrylic, but on the right side of the connectors17:13
wpwraksounds good. how about dmx and midi ?17:14
wpwrak4 leds there17:14
wolfsprauldmx looks like they could be under the connector, towards the front17:14
wolfspraulbut we need to ask adam about mechanical or assembly problems17:15
wolfspraulfor the midi connectors, there is no such space, so it can only be on the side again, towards the front imho17:15
wpwrak(dmx) hmm, partially. seems a bit too narrow in the middle. but a bit on the side should be fine17:18
GitHub81[migen] sbourdeauducq pushed 1 new commit to newnamer: https://github.com/milkymist/migen/commit/4eac60d181748d1b09583b63f00e0096bdeb8a1017:30
GitHub81[migen/newnamer] New naming system: second attempt - Sebastien Bourdeauducq17:30
GitHub126[migen] sbourdeauducq pushed 1 new commit to newnamer: https://github.com/milkymist/migen/commit/00d3eb7989c9d4d8ada2f3ae60ef442b285b261e17:47
GitHub126[migen/newnamer] Always include last step in names - Sebastien Bourdeauducq17:47
wpwrakboom. nice X crash18:17
wpwraknice :)19:10
wolfspraulyes! great!19:12
wolfspraullet's support these guys well and let's all believe in the future of Milkymist!19:12
wolfspraulthen more will join for sure...19:12
wolfspraulhe has a very positive attitude there, that's inspiring19:13
wolfspraulwpwrak: I'm thinking about the expansion board again23:05
wolfspraulwhy is the distance between the two headers 45mm ?23:06
wolfsprauland second, what do we say, in general, about stacking expansion boards?23:06
wolfspraulif someone stacks them, the 45mm area between the headers is the largest contiguous space, and that's quite small23:08
wolfspraulthat was all ;-) just thinking...23:08
wolfspraulha yeah, that's funny!23:14
wolfspraulMT is Montana?23:15
wolfspraulyes, it is23:17
--- Fri Jan 20 201200:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!