whitequark changed the topic of #glasgow to: glasgow debug tool · code https://github.com/whitequark/Glasgow · logs https://freenode.irclog.whitequark.org/glasgow
_whitelogger has joined #glasgow
<_whitenotifier-9> [whitequark/Glasgow] whitequark pushed 1 commit to master [+0/-0/±1] https://git.io/fhbVB
<_whitenotifier-9> [whitequark/Glasgow] whitequark e54fb1a - protocol.vgm: fix off-by-1 in 0x7X opcodes.
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/497959732?utm_source=github_status&utm_medium=notification
<bgamari> how much RAM should the fx2 firmwre require?
<bgamari> My build seems to be about 10 kBytes
<bgamari> which is larger than the 8 kB size of the CY7C68013's RAM
<bgamari> whitequark, ^
<_whitenotifier-9> [Glasgow] bgamari opened issue #99: FX2 firmware too large? - https://git.io/fhbwW
_whitelogger has joined #glasgow
<_whitenotifier-9> [Glasgow] marcan commented on issue #99: FX2 firmware too large? - https://git.io/fhboG
ec0 has quit [Quit: ZNC - http://znc.in]
ec0 has joined #glasgow
emily has quit [Remote host closed the connection]
emily has joined #glasgow
_whitelogger has joined #glasgow
emily has quit [Remote host closed the connection]
emily has joined #glasgow
<_whitenotifier-9> [whitequark/Glasgow] whitequark pushed 1 commit to master [+0/-0/±1] https://git.io/fhb12
<_whitenotifier-9> [whitequark/Glasgow] whitequark 3cfc95f - applet.yamaha_opl: unbreak handling of VGMs for a different chip.
<whitequark> bgamari: CY7C68013 has 16 KB of RAM
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/498182860?utm_source=github_status&utm_medium=notification
<whitequark> can you please run `glasgow -vv factory`?
<_whitenotifier-9> [Glasgow] bgamari commented on issue #99: FX2 firmware too large? - https://git.io/fhbMn
<_whitenotifier-9> [Glasgow] whitequark commented on issue #99: FX2 firmware too large? - https://git.io/fhbMW
<_whitenotifier-9> [Glasgow] bgamari commented on issue #99: FX2 firmware too large? - https://git.io/fhbMu
<_whitenotifier-9> [Glasgow] bgamari commented on issue #99: FX2 firmware too large? - https://git.io/fhbMK
<_whitenotifier-9> [Glasgow] whitequark commented on issue #99: FX2 firmware too large? - https://git.io/fhbMM
<_whitenotifier-9> [Glasgow] bgamari commented on issue #99: FX2 firmware too large? - https://git.io/fhbMp
<_whitenotifier-9> [Glasgow] bgamari opened pull request #100: hardware: Make Fab layers more readable - https://git.io/fhbDv
<_whitenotifier-9> [Glasgow] bgamari opened pull request #101: software: Don't request RAM writes of > 4096 bytes - https://git.io/fhbDU
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/498215961?utm_source=github_status&utm_medium=notification
<whitequark> thanks!
<_whitenotifier-9> [Glasgow] whitequark closed issue #99: FX2 firmware too large? - https://git.io/fhbwW
<_whitenotifier-9> [whitequark/Glasgow] whitequark pushed 1 commit to master [+0/-0/±1] https://git.io/fhbDI
<_whitenotifier-9> [whitequark/Glasgow] bgamari 390955a - device.hardware: don't request RAM writes of >4096 bytes.
<_whitenotifier-9> [Glasgow] whitequark closed pull request #101: software: Don't request RAM writes of > 4096 bytes - https://git.io/fhbDU
<_whitenotifier-9> [Glasgow] whitequark commented on pull request #100: hardware: Make Fab layers more readable - https://git.io/fhbDt
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/498216418?utm_source=github_status&utm_medium=notification
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/498217199?utm_source=github_status&utm_medium=notification
setrofim has joined #glasgow
<bgamari> it's quite unfortunate that github doesn't support svg uploads
<_whitenotifier-9> [Glasgow] bgamari commented on pull request #100: hardware: Make Fab layers more readable - https://git.io/fhbDS
<bgamari> by the way, for any nix users out there: https://github.com/bgamari/Glasgow.nix
<_whitenotifier-9> [whitequark/Glasgow] whitequark pushed 1 commit to master [+0/-0/±1] https://git.io/fhbyv
<_whitenotifier-9> [whitequark/Glasgow] bgamari 4694b59 - hardware: make Fab layers more readable
<_whitenotifier-9> [Glasgow] whitequark closed pull request #100: hardware: Make Fab layers more readable - https://git.io/fhbDv
<_whitenotifier-9> [Glasgow] whitequark commented on pull request #100: hardware: Make Fab layers more readable - https://git.io/fhbyf
<whitequark> bgamari: cool!
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/498236331?utm_source=github_status&utm_medium=notification
<bgamari> sigh, FPGA configuration failed
<whitequark> interesting
* bgamari suspected that this FPGA would be harder than it seemed
<bgamari> this is my first time doing a BGA
<whitequark> how did you solder it exactly?
<bgamari> it looks like it sat down well
<whitequark> you could try heating it up and pushing down on it gently
<bgamari> steel stencil, newish paste and a steady hand
<whitequark> alternatively, remove and reball
<whitequark> IME you can reball a c hip maybe 3-4 times before the chip and board are FUBAR
<whitequark> sometimes more
* bgamari doesn't have a reballing stencil
<bgamari> nor balls
<bgamari> i can try giving it a tap though
<whitequark> you can place the balls manually
<whitequark> in principle
<whitequark> but both a 0.8 mm stencil and balls are really cheap
<whitequark> look for cellphone repair shops
<_whitenotifier-9> [Glasgow] DurandA commented on pull request #100: hardware: Make Fab layers more readable - https://git.io/fhby3
<bgamari> whitequark, would you recommend flooding it with flux before doing so?
<_whitenotifier-9> [Glasgow] whitequark commented on pull request #100: hardware: Make Fab layers more readable - https://git.io/fhby8
<whitequark> bgamari: from the sides? nah, pointless
<whitequark> you won't get flux under it anyway
<bgamari> sure
<whitequark> i guess it doesn't hurt
<bgamari> I have a super-aggressive liquid flux
<whitequark> oh
<bgamari> but it's conductive
<bgamari> so I never use it
<whitequark> yeah no never use that with bga
<bgamari> it's essentially impossible to clean
<whitequark> i don't use conductive fluxes in general
<whitequark> yep
<whitequark> they are shit
<bgamari> yeah
<bgamari> that purchase is now considered a mistake
<_whitenotifier-9> [Glasgow] DurandA commented on pull request #100: hardware: Make Fab layers more readable - https://git.io/fhbyB
<bgamari> whitequark, why is the pull-up on R9 so strong?
<whitequark> R9 is 100k
<bgamari> whoops
<bgamari> whitequark, rather on SYNC
<bgamari> R10
<whitequark> mmm, because on revAB it was open collector
<whitequark> open drain*
<bgamari> I see
<_whitenotifier-9> [Glasgow] whitequark opened issue #102: Reconsider SYNC pullup value? - https://git.io/fhbyA
<bgamari> hmm, tap didn't seem to help, sadly
<bgamari> perhaps it's time to just try another device
* bgamari has a few spares
<whitequark> you could give it to phone repair guys
<whitequark> and tell them you want it to be reballed and replaced
<bgamari> fair point
<whitequark> they are usually very good at it
<whitequark> one time, i tore off a pad on the PCB
<whitequark> with ViP
<bgamari> thankfully while the ICE40 isn't cheap, it's not devastatingly expensive either
<whitequark> drilled out the PCB substrate to get at the copper via
<whitequark> gave them the device
<whitequark> it worked afterwards
<bgamari> heh
<bgamari> whitequark, should the glasgow tools cache the bitstreams that it generates?
<whitequark> not really
<whitequark> it "caches" the bitstream in the actual FPGA
<whitequark> this achieves a good balance of convenience with cache invalidation bugs
<whitequark> you can e.g. update yosys/nextpnr and get a different bitstream, which will not get reuploaded because the cache key is the verilog sha
<whitequark> but the stale bitstream will persist no longer than you unplugging the device and plugging it back
<bgamari> sure
<bgamari> that is fair
<bgamari> whitequark, what is your typical success rate with BGA mounting?
<whitequark> bgamari: i do not assemble boards with BGA
<whitequark> i only ever rework BGA
<whitequark> often on 15+ year old devices
<whitequark> my statistics is not representative
<whitequark> in fact i don't think i have *ever* mounted a completely new BGA IC
<bgamari> fair enough
<bgamari> lucky you :)
<whitequark> BGA rework is... not easy
<bgamari> huh
<bgamari> apparently placement isn't either :)
<bgamari> chip 2 met a similar outcome to chip 1
<bgamari> definitely have IFCLK
<bgamari> and CLKREF
<bgamari> not sure about the remainder of the FX2-FPGA interface
<bgamari> sigh
<bgamari> FPGA configuration happens with the FPGA behaving as an SPI slave, yes?
<whitequark> yes
<whitequark> are you sure the FX2 is soldered correctly?
<bgamari> I just did another pass over the FX2 pads to be certain
<bgamari> it looks fine
* bgamari is trying to work out how to probe the relevant nets
<whitequark> you can probe the pins of FX2 at the side
<bgamari> squeezing both chips doesn't help, FWIW
<bgamari> hmm
<bgamari> CDONE is actually high
<bgamari> I think
<whitequark> interesting
<bgamari> indeed
<_whitenotifier-9> [whitequark/Glasgow] whitequark pushed 1 commit to master [+0/-0/±1] https://git.io/fhbQ1
<_whitenotifier-9> [whitequark/Glasgow] whitequark 7af5478 - README: update to account for ECP5 RE effort.
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/498306495?utm_source=github_status&utm_medium=notification
<_whitenotifier-9> [whitequark/Glasgow] whitequark pushed 3 commits to master [+0/-0/±5] https://git.io/fhbdm
<_whitenotifier-9> [whitequark/Glasgow] whitequark 83dbf5e - applet.yamaha_opl: try to enable missing features.
<_whitenotifier-9> [whitequark/Glasgow] whitequark 0a8722a - applet.yamaha_opl: web: clean up player queueing.
<_whitenotifier-9> [whitequark/Glasgow] whitequark e997514 - applet.yamaha_opl: web: add YM3526 support.
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/498372249?utm_source=github_status&utm_medium=notification
xorrbit has joined #glasgow
<_whitenotifier-9> [Glasgow] whitequark opened issue #103: Changing pin assignment only doesn't change bitstream ID - https://git.io/fhbdZ
<_whitenotifier-9> [Glasgow] whitequark commented on issue #73: Pipeline USB reads - https://git.io/fhbdn
davidc__ has joined #glasgow
<davidc__> anybody thinking of doing a glasgow revC production run (either for sale, or that wants to tack a few more onto an order?)
<whitequark> esden will sell revCs once i'm done with revC1
<davidc__> Cool. Are there major known Errata with revC?
<whitequark> not major
<whitequark> a few missing resistors for discharging output capacitors
<whitequark> and i've still not qualified the configurable pulls
<whitequark> revC0 is already usable and much better than revAB
<whitequark> but it's not yet what i'd call production ready
<ylamarre> I don't know if this was ever noted, but I find it rather ironic that Glasgow was built avoiding FTDI chips at all cost considering FTDI HQ is located in Glasgow...
<ylamarre> ... and this has been on my mind for way too long.
<sorear> that is in fact the reason it was named Glasgow
<sorear> it was a USB-JTAG interface that grew a bunch of features
<whitequark> yep
<ylamarre> Oh, it wasn;t clear from wq's old tweets, or maybe I skiped that part, my bad.
<whitequark> it was never particularly public
<whitequark> in fact at one point i forgot why i chose the name
<ylamarre> Ok, so I wasn't insane... I did remember you saying you forgot about the name.
<ylamarre> I mean, I don't have memory issue in that regard...
<ylamarre> But thanks for confirming the fact!
<_whitenotifier-9> [whitequark/Glasgow] whitequark pushed 2 commits to master [+0/-0/±2] https://git.io/fhbNE
<_whitenotifier-9> [whitequark/Glasgow] whitequark 7c034f0 - applet.jtag_pinout: fix a confusing message.
<_whitenotifier-9> [whitequark/Glasgow] whitequark 96ac88e - applet.jtag: scan: improve output for complex chains.
<_whitenotifier-9> [Glasgow] Success. The Travis CI build passed - https://travis-ci.org/whitequark/Glasgow/builds/498438611?utm_source=github_status&utm_medium=notification