| *** tpb has joined #timvideos | 00:00 | |
| mithro | seaLne: if you are around this evening we can see if we get the Opsis board working for you. | 00:17 |
|---|---|---|
| mithro | tumbleweed: Did you sort out CarlFK? | 00:18 |
| mithro | tumbleweed: I feel asleep when he was asking me questions | 00:18 |
| mithro | It was 4am in the morning though | 00:18 |
| mithro | seaLne: I'm around from now | 02:04 |
| *** hamster_ has joined #timvideos | 03:12 | |
| mithro | Hi hamster_! | 03:13 |
| mithro | Just replying to your email right now | 03:13 |
| hamster_ | @Mithro - Dropping by while I give the boy afternoon tea... might be a bit intermittant. | 03:14 |
| mithro | hamster_: no worries! | 03:14 |
| mithro | As I've said in my email, don't feel like you have to drop in on IRC, it is only a suggestion if you think it will help! | 03:14 |
| mithro | I find that when things just aren't working it can be helpful to have people to chat with / bounce ideas with or even to try and replicate my work and confirm that I'm not going insane :). | 03:14 |
| hamster_ | Oh, and having a finger in the splint doesn't help my typing! | 03:14 |
| mithro | hamster_: Ouch! :-( | 03:15 |
| mithro | hamster_: BTW Thanks for all your work on this! It's super awesome | 03:18 |
| hamster_ | Just thinking that my test video source code has something that works in Vivado VHDL, but not in ISE so about to re-verify everything from end-to-end again. | 03:18 |
| mithro | hamster_: I would be suspect of our hardware too... | 03:19 |
| mithro | hamster_: If I had the money, I would so send you a high speed oscilloscope :) | 03:20 |
| hamster_ | Well, the link stands up for a good few ms, with 8b10b working OK, so that is 2M symbols or so... can't do that on bad hardware :) | 03:21 |
| mithro | hamster_: ha okay | 03:22 |
| hamster_ | If I can find a Osc with the footprint I'll snip the old one out and try it.... | 03:23 |
| mithro | hamster_: problem with these prototype boards is that they were mostly hand loaded which means higher number of errors | 03:24 |
| hamster_ | I wonder if I can bodge something up using the 135MHz on the Nexys Video, out the differential PMOD, and into the SMA... might need a short trip to Jaycar. | 03:27 |
| mithro | hamster_: I wonder if you could generate the clock out the DisplayPort on the Nexys Video and recover it on the DisplayPort RX | 03:31 |
| mithro | hamster_: I'm going to owe you sooooooo much beer / whisky / drink of your choice at the end of this :) | 03:33 |
| hamster_ | Man, did I pick the wrong year to give up alcohol", nah seriously I'm loving the learning experience, and finding quite a few "it works only because of luck" bugs (like the noise on the idle AUX channel) | 03:37 |
| hamster_ | GTP_DUALs are a pain in the bum - how does anybody ever get them to work correctly without endless respins of the board? | 03:38 |
| hamster_ | Anyhow, must go and do household stuff... will hopefully be more fruitful now that I can stop going bouncing around in the refclock cul-de-sac. Seems I can't put the board aside at the mo. | 03:41 |
| *** hamster_ has quit IRC | 03:44 | |
| *** Bertl_zZ is now known as Bertl | 04:00 | |
| mithro | hey Bertl - you just missed hamster :( | 04:22 |
| Bertl | yeah, cool that he shows up on IRC now too | 04:24 |
| mithro | Bertl: Very, very, rarely | 04:24 |
| mithro | Bertl: the GTP tiles are just driving him mad | 04:24 |
| Bertl | I can imagine | 04:24 |
| Bertl | btw, replied to your comment on the zybo dual role HDMI | 04:25 |
| Bertl | if you want to discuss it, I'm available | 04:25 |
| mithro | Bertl: sure! | 04:26 |
| mithro | Bertl: I'm no expert here | 04:26 |
| Bertl | well, some "potential problems" I didn't mention are that TMDS requires 3.3V IO voltage on the bank involved | 04:27 |
| Bertl | so that needs to be considered as well (this one actually bit us with the inital HDMI out module) | 04:27 |
| mithro | Bertl: yeah - it's also restricted to the HR banks | 04:28 |
| Bertl | in general, I would avoid dual role directly connected to the FPGA if possible, better have dedicated inputs and outputs | 04:28 |
| Bertl | we decided to use LVDS and add a cheap DP/HDMI dongle chip to get good output quality (even outside the DDR limits :) | 04:29 |
| mithro | Bertl: I was hoping to have reconfigurable because sometimes you want 3 inputs, 3 outputs, sometimes 2 inputs, 4 outputs, etc | 04:29 |
| mithro | Bertl: define "cheap" :) | 04:30 |
| Bertl | we are using the PTN3363, but the PTN3361 (or so) would do for full HD | 04:30 |
| Bertl | PTN3360 ~ 70 cent, PTN3361 ~ 80 cent | 04:31 |
| Bertl | (at single quantities :) | 04:32 |
| mithro | Bertl: interesting | 04:32 |
| Bertl | the PTN3363 is about a dollar or so | 04:32 |
| Bertl | it has a nice equalizer and signal boost | 04:32 |
| mithro | Yeah | 04:32 |
| Bertl | so you can drive longer cables and get good quality with low power | 04:33 |
| mithro | Only uni-directional though, right? | 04:33 |
| mithro | power is not a huge concern for me | 04:33 |
| Bertl | yep, we are only doing output so far | 04:33 |
| Bertl | last time I asked if _florent_ is working on the HDMI input | 04:33 |
| mithro | Bertl: I can't see a reason you'd do HDMI input | 04:33 |
| Bertl | we will do that for our bridge solution | 04:33 |
| Bertl | I'm inclined to go for an analog devices chip there though | 04:34 |
| mithro | Bertl: the analog chips are *sooooo* expensive | 04:34 |
| Bertl | yes, I know | 04:34 |
| mithro | Bertl: does the PTN3360 have ESD stuff inbuilt? | 04:34 |
| Bertl | yep | 04:35 |
| Bertl | also back drive protection | 04:35 |
| mithro | Bertl: if they had a 5V regulartor in them, they wouldn't be priced badly | 04:36 |
| Bertl | well, for the 5V we use a tiny charge pump solution | 04:36 |
| Bertl | I'm curious what HDMI input data rates you will be able to achieve | 04:39 |
| Bertl | do you expect that you can recover data without GTs at high data rates or will the inputs be around 600Mb/s per lane? | 04:43 |
| Bertl | (maybe a question for _florent_ though) | 04:44 |
| mithro | Bertl: what do you mean? For the Opsis we get 1080p30 | 04:46 |
| Bertl | how do you synchronize to the individual data lanes? | 04:47 |
| mithro | Bertl: I don't quite know exactly how it works in the misoc based firmware, in the old firmware we used the SERDES has shown in XAPP495 | 04:51 |
| Bertl | IIRC, that does a max 666Mb/s on the fastest speed grade | 04:54 |
| Bertl | (which isn't enough for 1080p30, or?) | 04:55 |
| Bertl | ah, no, that was spartan 3 (the limits) | 04:55 |
| mithro | Bertl: yeah the S6 can do 1280×1024 @ 60Hz | 05:01 |
| mithro | shenki - apparently you know Bertl :) | 05:06 |
| Bertl | yeah, but I don't know if he remembers me ... | 05:07 |
| mithro | Bertl with how slack he has been to contributing to the project lately, I wonder if he remembers me ;-) | 05:15 |
| mithro | He did just buy a house, so I guess he does have an excuse | 05:16 |
| tumbleweed | mithro: I couldn't reproduce the problem | 05:23 |
| mithro | tumbleweed: It was something with his setup | 05:23 |
| mithro | tumbleweed: sometimes I wonder if Carl understands half the things he is doing :P | 05:24 |
| tumbleweed | heh | 05:24 |
| tumbleweed | yeah, anyway, we never got to the bottom of it | 05:24 |
| mithro | Bertl: so, that Opsis V2 "dream" would be interesting for you guys right? | 05:36 |
| mithro | Bertl: you could stream your video out your 3 HDMI ports in any format you wanted then use the Opsis V2 to reconstruct it into a 4k stream | 05:37 |
| Bertl | for development every reasonably fast (bandwidth wise) solution to move, convert and combine video data is very interesting | 05:38 |
| mithro | Bertl: actually - you could do that with the Opsis V1 now | 05:38 |
| Bertl | do you have the TOFE google spreadsheet at hand, I would have to dig out the link | 05:38 |
| mithro | Bertl: hdmi2usb.tv/tofe | 05:39 |
| Bertl | excellent, tx | 05:39 |
| mithro | Bertl: sorry - that has a "fixme: embed the spreadsheet" :P | 05:39 |
| Bertl | yeah, just saw it :) | 05:39 |
| mithro | Bertl: I was sure I had fixed that... | 05:39 |
| Bertl | it says, up to 34 LVDS pairs, how many on the Opsis v1? | 05:40 |
| mithro | Bertl: Hrm this is definately out of date... Now I have to track down where I committed the extra information too.... | 05:41 |
| Bertl | okay, take your time, drop me a note when it's done | 05:42 |
| mithro | Bertl: I believe the Opsis has the full 34 but I would need to check | 05:42 |
| mithro | Bertl: problem with git, you forget where you did stuff :) | 05:42 |
| Bertl | so I can assume at least 12 pairs are available, yes? | 05:42 |
| mithro | Bertl: yeah | 05:42 |
| Bertl | okay, so I guess that would be the best input choice or a 4k combiner then | 05:43 |
| mithro | Bertl: I was actually trying to figure out if you could use the HDMI Outputs as inputs | 05:43 |
| mithro | Bertl: there is nothing "in the way" | 05:43 |
| Bertl | i.e. make a dual slot plugin module for the Axiom Beta which has e.g. USB3.x or DP/Thunderbolt connectors and a similar part on the opsis | 05:43 |
| mithro | Bertl: I wouldn't recommend using the S6, the Artix-7 is a much better choice | 05:44 |
| Bertl | obviously, but that will take some time I guess | 05:45 |
| mithro | Bertl: you have 3 HDMI output ports currently, right? | 05:46 |
| Bertl | we have 12 LVDS pairs and 20 GPIOs currently :) | 05:47 |
| mithro | Bertl: ahh okay :) | 05:47 |
| Bertl | there are plugin modules for single HDMI at the moment | 05:47 |
| Bertl | and a triple HDMI dual slot module is in the works | 05:47 |
| Bertl | but avoiding the TMDS path might be beneficial | 05:48 |
| Bertl | i.e. just use LVDS (re)driver or similar to get reasonable cable lengths | 05:48 |
| Bertl | (we do not need to adhere to any specific protocol, as long as we can keep the noise low, i.e. coding) | 05:49 |
| mithro | Bertl: yeah | 05:49 |
| mithro | Bertl: more stuff pushed! | 05:49 |
| mithro | Bertl: it'll take about 5 minutes for the page to update | 05:49 |
| mithro | Bertl: https://docs.google.com/spreadsheets/d/15rivtOcSG4kqKf0AUK_OwBfs67XQtW8byr8kc5FUTgE/edit#gid=1919115012 | 05:51 |
| tpb | Title: Tims Open FPGA Expansion (TOFE) board connector interface - Google Sheets (at docs.google.com) | 05:51 |
| mithro | Bertl: Unless I've stuffed up somewhere - the Opsis should have 17 DIFF IO pairs and 5 DIFF IO CLK pairs for a total of 22 pairs | 05:54 |
| Bertl | so, if we use a low speed clock, we can get 11 pairs from the x1 | 05:54 |
| Bertl | and 16 pairs from x4 | 05:55 |
| Bertl | or even 17 (depending on the clock) | 05:55 |
| mithro | Bertl: The IO CLK pairs are just IO pairs which are connected to the GCLK pins | 05:55 |
| Bertl | that should be more than fine for our purpose | 05:55 |
| mithro | Bertl: awesome! | 05:56 |
| Bertl | you have a bunch of "expansion boards" listed on the page you linked | 05:56 |
| Bertl | any of those in a usable state yet? | 05:57 |
| mithro | Bertl: I should have KiCad templates for the TOFE interface on Monday | 05:57 |
| Bertl | i.e. schematic/prototype done, etc? | 05:57 |
| mithro | Bertl: Nope, I have a spreadsheet which shows the pin mappings | 05:58 |
| Bertl | okay, what have you planned for the 3x Gbit Ethernet? | 05:58 |
| Bertl | and you are missing the Axiom Plugin Module Adapter :) | 05:59 |
| mithro | Bertl: Basically just 3xRGMII Phys + connectors | 05:59 |
| mithro | Bertl: the pin mappings have started at https://docs.google.com/spreadsheets/d/15rivtOcSG4kqKf0AUK_OwBfs67XQtW8byr8kc5FUTgE/edit#gid=624502438 :P | 06:00 |
| tpb | Title: Tims Open FPGA Expansion (TOFE) board connector interface - Google Sheets (at docs.google.com) | 06:00 |
| Bertl | any specific devices/connectors planned there (3xGbit)? | 06:00 |
| Bertl | I've been looking for suitable parts for some time with litte success | 06:01 |
| mithro | Bertl: we'll probably just reuse exactly the same design as the Opsis uses for the GigE already onboard | 06:01 |
| mithro | Bertl: except have them all share a single clock | 06:02 |
| Bertl | yeah, makes sense | 06:02 |
| mithro | Bertl: we used an Realtek part which only has specs under NDA because Numato already had a supplier | 06:03 |
| mithro | Bertl: the RGMII interface is pretty standard and there isn't much you need the datasheet for | 06:03 |
| Bertl | I think I asked this before, is there a pdf of the opsis schematic somewhere? | 06:03 |
| mithro | Bertl: there is the full schematic + PCB in kicad format | 06:04 |
| Bertl | I know that the design files are on github | 06:04 |
| mithro | http://roger.mithis.com/~tim/opsis-schematic.pdf | 06:04 |
| Bertl | but sometimes I do not have a full kicad install with me | 06:04 |
| Bertl | ah, great! thanks | 06:04 |
| mithro | Bertl: yeah - we need to publish a PDF version in the opsis docs somewhere | 06:04 |
| mithro | Bertl: but need to fix some formatting issues first | 06:04 |
| mithro | Bertl: I'm really excited to have you guys interested :) | 06:05 |
| Bertl | we are glad that we found the opsis and the folks behind it | 06:07 |
| Bertl | where do you get the RTL8211E, a quck check with farnell/mouser/digikey gave no results | 06:08 |
| mithro | Bertl: maybe with the Opsis V2 Dream, we do 2 "long reach" output only and make the other 4 "dual mode" | 06:08 |
| mithro | Bertl: I don't know where Numato source it from, sorry :( | 06:08 |
| mithro | http://www.findchips.com/search/RTL8211E | 06:08 |
| tpb | Title: RTL8211E Stock and Price by Distributor (at www.findchips.com) | 06:08 |
| Bertl | okay, because that is a problem for us, i.e. we are trying to use components which are widely available | 06:09 |
| mithro | Bertl: yeah | 06:09 |
| Bertl | but nothing which concerns you, just saying | 06:09 |
| Bertl | not convinced of the dual role hdmi ports yet :) | 06:09 |
| mithro | http://www.digikey.com/product-detail/en/KSZ9021RLI/576-3636-ND/2234026 | 06:10 |
| tpb | Title: KSZ9021RLI Micrel Inc. | 576-3636-ND | DigiKey (at www.digikey.com) | 06:10 |
| mithro | That is *super* expensive for a RGMII IC | 06:10 |
| Bertl | in my experience, digikey is always a little more expensive than alternatives | 06:11 |
| mithro | Bertl: With 6 HDMI ports, I can't figure out what the correct configuration would be -- you need at least 2 inputs and 2 output | 06:11 |
| Bertl | so why not do exactly that and have two dual role then? | 06:11 |
| mithro | Bertl: Yeah - I guess | 06:12 |
| Bertl | although I'd prefer three in, three out (obviously) | 06:12 |
| Bertl | with proper LVDS/TMDS conversion, equalizers and drivers | 06:12 |
| mithro | Bertl: I guess we might just have enough IO to do 4 in, 4 out | 06:13 |
| Bertl | don't forget the huge number of additional low speed lines | 06:14 |
| Bertl | DDC (2) HPD (2) CEC(1) | 06:15 |
| Bertl | and if you use equalizers/drivers, you probably also want 2-3 IOs for controlling that | 06:15 |
| Bertl | of course, can be done via some I/O expander as well | 06:17 |
| mithro | Bertl: yeah - that is all low speed IO | 06:18 |
| mithro | Bertl: its kind of weird the Kintex-7 and Artix-7 are the same price on Digikey once you go up to the 676 pin parts | 06:23 |
| Bertl | yeah, xilinx pricing is strange in many ways | 06:24 |
| Bertl | in larger quantities they become surprisingly cheap | 06:24 |
| mithro | Bertl: yeah - I think that is more a Digikey strangeness | 06:25 |
| mithro | Bertl: dammit, you distracted me with thinking about V2 stuff again | 06:31 |
| mithro | Bertl: I should be working on getting the mode-switching stuff working so that seaLne can use his Opsis board :P | 06:32 |
| Bertl | I sometimes can be distracting :) | 06:33 |
| mithro | Bertl: have you looked at migen / misoc at all? | 06:39 |
| Bertl | yes, for a brief moment, but I'm more a VHDL person and I like to fine tune stuff once it is working (similar to Mike) | 06:40 |
| Bertl | but I see the advantages of a system like migen | 06:41 |
| mithro | Bertl: the big advantage I see is you aren't writing VHDL or Verilog :) | 06:53 |
| Bertl | yeah, maybe I give it a second chance in the not so distant future :) | 06:57 |
| mithro | Bertl: I really just want more people contributing to the video modules in misoc :-) | 07:04 |
| *** rohitksingh has joined #timvideos | 07:08 | |
| mithro | hey rohitksingh! | 07:15 |
| mithro | rohitksingh: Had any luck on getting the misoc firmware running? | 07:16 |
| rohitksingh | mithro: hi! | 07:18 |
| *** travis-ci has joined #timvideos | 07:19 | |
| travis-ci | [mithro/HDMI2USB-misoc-firmware/master#15] (cafb6a1): The build passed. (https://travis-ci.org/mithro/HDMI2USB-misoc-firmware/builds/84631355) | 07:19 |
| *** travis-ci has left #timvideos | 07:19 | |
| rohitksingh | mithro: upto now, I can make a custom misoc target and write a custom firmware for it, and interact with HDL and LM32 using CSR registers | 07:19 |
| mithro | rohitksingh: great! | 07:20 |
| mithro | rohitksingh: did you see the "soft" I2C thing that _florent_ added? | 07:20 |
| mithro | rohitksingh: seaLne was interested in helping with the VGA stuff too | 07:21 |
| rohitksingh | I think that I can use _florent_'s hdmi in code for pushing captured vga data to colorspace converter and and then to memory through dma | 07:21 |
| mithro | rohitksingh: yeah | 07:21 |
| rohitksingh | mithro: oh that would be great! | 07:21 |
| mithro | rohitksingh: He doesn't have an Atlys nor one of your VGA boards though | 07:22 |
| rohitksingh | mithro: My progress has slowed down in last 2 weeks as I have a recruitment test for Indian Space Research Organization tomorrow | 07:22 |
| mithro | rohitksingh: oh - that sounds exciting! | 07:23 |
| rohitksingh | mithro: My main issue is Clock domain crossing...and clocking...I've have been reading _florent_'s hdmi-in code for that | 07:23 |
| rohitksingh | some of the parts of code is too advanced for me to understand at one-go :-| | 07:25 |
| mithro | rohitksingh: yeah - the HDMI code could use some better documentation | 07:26 |
| mithro | rohitksingh: did you want to start with just getting the lm32 to detect / setup the VGA board? | 07:26 |
| rohitksingh | mithro: that part is now easy with I2C and LM32, just write a function for initializing AD9984A's registers. So, i was focussing more HDL part | 07:28 |
| mithro | rohitksingh: do the easy part first :) | 07:29 |
| rohitksingh | mithro: okay! would do that first! :) | 07:29 |
| mithro | rohitksingh: then maybe we'll have some documentation or something ready by the time you get to that stage | 07:32 |
| rohitksingh | mithro: sounds awesome! will do the setup part tomorrow itself! | 07:33 |
| rohitksingh | *after the exam :p | 07:33 |
| mithro | rohitksingh: cool | 07:36 |
| *** travis-ci has joined #timvideos | 07:42 | |
| travis-ci | [mithro/HDMI2USB-misoc-firmware/opsis-fixes#16] (8c74c98): The build passed. (https://travis-ci.org/mithro/HDMI2USB-misoc-firmware/builds/84632723) | 07:42 |
| *** travis-ci has left #timvideos | 07:42 | |
| seaLne | rohitksingh: hi | 08:06 |
| mithro | morning seaLne! | 08:06 |
| seaLne | lo | 08:06 |
| rohitksingh | seaLne: hi! | 08:06 |
| seaLne | rohitksingh: so are you focusing mainly on the software side of the vga just now? what about converting the board to tofe have you started on that? i can use kicad but not great at programing | 08:08 |
| seaLne | mithro: if you have time i'm around for the next ~3 hours | 08:08 |
| mithro | seaLne / rohitksingh: I'm hoping we have KiCad templates for the TOFE boards in the next couple of days | 08:08 |
| mithro | seaLne: I was hoping to be further along than I am by the time you got up | 08:10 |
| mithro | seaLne you can blame Bertl - he distracted me with day dreaming :) | 08:10 |
| rohitksingh | seaLne: yeah, my focus is right now on porting it to misoc firmware. it would be awesome if you could help with converting the design to TOFE. let me know if you need any help with that! | 08:10 |
| seaLne | heh, no worries i've not fully woken up yet | 08:11 |
| *** travis-ci has joined #timvideos | 08:11 | |
| travis-ci | [mithro/HDMI2USB-misoc-firmware/master#17] (2018ff2): The build passed. (https://travis-ci.org/mithro/HDMI2USB-misoc-firmware/builds/84632993) | 08:11 |
| *** travis-ci has left #timvideos | 08:11 | |
| seaLne | rohitksingh: great yeah i'll work on that when the tofe template is done | 08:12 |
| mithro | seaLne: But - take a look at my opsis-fixes branch https://github.com/mithro/HDMI2USB-misoc-firmware/tree/opsis-fixes | 08:12 |
| tpb | Title: mithro/HDMI2USB-misoc-firmware at opsis-fixes · GitHub (at github.com) | 08:12 |
| seaLne | i'll be away for the next few days till around the middle of the week | 08:13 |
| rohitksingh | seaLne: great! schematic would need just few modifications apart from changing from VHDCI to TOFE connector. Main effort would be needed in laying out the PCB | 08:14 |
| seaLne | yeah | 08:14 |
| mithro | seaLne: you could start the schematic now I think? | 08:19 |
| seaLne | well i'm going away for a few days holiday in about 2.5 hours so not really much point starting | 08:23 |
| mithro | seaLne: When will you be back? | 08:24 |
| seaLne | most likely wednesday | 08:24 |
| seaLne | bit annoying timing after just getting the opsis | 08:27 |
| *** rohitksingh has quit IRC | 08:28 | |
| mithro | seaLne: yeah :-( sorry about the shipping delays | 08:31 |
| seaLne | mithro: not blaiming you in anyway. more just that this week was only i had clear to get away for the next while | 08:33 |
| mithro | seaLne: should tell me next time and I'll know to knock some more heads | 08:33 |
| *** rohitksingh has joined #timvideos | 08:42 | |
| *** hamster_ has joined #timvideos | 08:55 | |
| hamster_ | Tim! Great news - stable DisplayPort picture! | 08:56 |
| mithro | hamster_: \o/ \o/ | 08:56 |
| hamster_ | I was shooting myself in the foot... the link status check was toasting the link. Sent you a bit file to test. It's only 800x600, not that that means anything.... | 08:56 |
| hamster_ | A link is a link is a link, even if it is only 800x600 | 08:57 |
| mithro | hamster_: awesome! | 08:57 |
| mithro | hamster_: yeah | 08:57 |
| mithro | hamster_: did you see my FAQ? | 08:57 |
| mithro | hamster_: https://docs.google.com/document/d/1noNqTdLYlEru_6gWOlFqFvYe-vctUE23Tp42oRoKKKo/edit | 08:57 |
| tpb | Title: Numato Opsis - Video Information FAQ - Google Docs (at docs.google.com) | 08:57 |
| hamster_ | Looking now... but I should check this code in (before HD crashes just to piss me off), then leave it running for an hour to check that the link stands up. | 08:59 |
| mithro | hamster_: SGTM! | 08:59 |
| hamster_ | Once I am sure the link is stable (e.g. tomorrow) I'll put the link check back in. | 09:00 |
| mithro | hamster_: commit early, commit often :) | 09:00 |
| mithro | hamster_: how are you clocking this? | 09:00 |
| hamster_ | 100MHz => PLL / 8 * 54 / 5 => 135MHz. | 09:01 |
| mithro | hamster_: Do you hate the poor 27MHz crystal or something? :P | 09:02 |
| hamster_ | Well, yes :-D | 09:02 |
| hamster_ | I'm at least sure that the 100MHz is ticking :- | 09:02 |
| hamster_ | 8-) | 09:02 |
| mithro | hamster_: fair enough | 09:03 |
| mithro | In theory there are 100MHz, 27MHz and 135MHz clocks | 09:03 |
| mithro | hamster_: how much work do you think it is to do the opposite direction? IE Receiving DisplayPort? | 09:04 |
| mithro | hamster_: I'm working on the tool to switch the Opsis between the different programming modes | 09:06 |
| hamster_ | At the moment I'm only working with a subset of AUX channel registers, so RX will need a full register set, and if they don't work right it will be.... um.... problematic. Once the link is stood up it will be fine - it's just a stream of bytes then... | 09:06 |
| mithro | hamster_: could we start with just doing pass through? | 09:06 |
| Bertl | hamster_: hey Mike, how's going? | 09:07 |
| Bertl | (Herbert here) | 09:07 |
| hamster_ | Bertl - great. Been bashing my head against a brick wall with only one pin for debug.... | 09:08 |
| Bertl | you need a debug module :) | 09:08 |
| mithro | hamster_: btw - Sorry about that | 09:08 |
| mithro | hamster_: the TOFE->Slow Speed I/O module should be ready in a week or two | 09:08 |
| hamster_ | I need a full Vivado License - but I don't need it $3k. | 09:09 |
| hamster_ | Oh, and a full ISE license too... | 09:09 |
| mithro | hamster_: you want Chipscope? | 09:09 |
| Bertl | there is an unlimited license for debugging/field testing | 09:10 |
| hamster_ | ?????? | 09:10 |
| Bertl | haven't tested it yet, sec, let me dig out the link | 09:10 |
| hamster_ | Pass-though would be problematic - there is a man-in-the-middle issue with link negotiation. Swings and Preemphasis and so on... | 09:11 |
| mithro | hamster_: have you thought about getting HDMI working on the GTPs? | 09:11 |
| Bertl | Vivado Lab Edition - 2015.3 Full Product Installation | 09:12 |
| Bertl | http://www.xilinx.com/support/download.html | 09:12 |
| tpb | Title: Downloads (at www.xilinx.com) | 09:12 |
| hamster_ | Output It would be a piece of cake, given that you can jam the output of a PLL as the reference clock into the GTP. Not so sure if you could pass the clock from HDMI in to HDMI out without jitter adding up.. | 09:14 |
| mithro | hamster_: the DisplayPort connectors on the Opsis are dual mode, which means with something like http://smile.amazon.com/gp/product/B00Y1DDNN8/ref=s9_simh_gw_p147_d10_i1?pf_rd_m=ATVPDKIKX0DER&pf_rd_s=desktop-1&pf_rd_r=173WJK3ZWRV4K23AGP74&pf_rd_t=36701&pf_rd_p=2079475242&pf_rd_i=desktop you can output HDMI instead | 09:15 |
| mithro | Bertl: Vivado doesn't support the Spartan 6 sadly | 09:15 |
| Bertl | yes, but the GT tiles are reasonably similar, no? | 09:16 |
| hamster_ | About debugging, I was thinking of passing frames out the HDMI port to pick up on another board for analysis. Crude but would work... | 09:16 |
| Bertl | and hamster_ has a huge selection of development boards I guess | 09:16 |
| hamster_ | Bertl:I know more about GTP tiles than I ever want to know - stupid things like RESETDONE[01] not working as expected, and that the REFCLOCK powerdowns are active low (WTF??? the only active low powerdown?) | 09:17 |
| hamster_ | Bertl: And got one more yesterday Kintex-7 KC7325T, with dev-locked Vivado license 8-) | 09:18 |
| Bertl | ah, good, so you have a chipscope license on that one, at least for some time | 09:19 |
| mithro | hamster_: _florent_ created a thing called "litescope" which is pretty similar to chipscope but completely open | 09:19 |
| mithro | hamster_: not as nicely integrated as chipscope | 09:19 |
| Bertl | mithro: does it use jtag or a different transport? | 09:20 |
| mithro | hamster_: Now I need to figure out what to plug it into | 09:20 |
| mithro | Bertl: it uses ethernet or serial | 09:20 |
| mithro | Bertl: the reason we have working Ethernet is because _florent_ wanted to use it when doing other bring up and hence decided to get it working first :P | 09:20 |
| Bertl | I see, nice | 09:21 |
| mithro | Bertl: he has a wishbone to ethernet adapter thing too | 09:21 |
| mithro | Bertl: lets you write Python scripts on your computer which do things to the wishbone bus inside the board through the ethernet | 09:21 |
| hamster_ | I was thinking of just an HDMI video output with 36 channels, and maybe RS232 input for control (arm, scroll, zoom). Would be enough for FSM | 09:22 |
| hamster_ | GOt to go | 09:22 |
| mithro | hamster_: thanks for all your work | 09:22 |
| Bertl | cya | 09:22 |
| mithro | hamster_: I'll email when I have replicated it! | 09:22 |
| *** se6astian|away is now known as se6astian | 09:29 | |
| hamster_ | oh, back - hot drinks made, dishes done. Just trying out 4k30 now... | 09:30 |
| hamster_ | (dumb qn - why the '_' under some usernames? | 09:31 |
| hamster_ | e.g. "hamster_" | 09:31 |
| mithro | hamster_: the _ normally means there is another user who has the name without it | 09:33 |
| mithro | hamster_: Often it is caused by reconnecting to IRC before your previous connection has timed out | 09:34 |
| mithro | hamster_: I'm trying to find a DisplayPort compatible monitor which won't break my work desktop setup | 09:34 |
| Bertl | special requirements? | 09:35 |
| Bertl | the iiyama ProLite is quite nice | 09:35 |
| Bertl | http://www.iiyama.com/gb_en/products/prolite-b2888uhsu-b1/ | 09:37 |
| tpb | Title: iiyama (at www.iiyama.com) | 09:37 |
| mithro | Bertl: requirement is that it is within walking distance of my desk :P | 09:38 |
| mithro | https://goo.gl/photos/YT2yzt9YHFL5kMRT7 <-- these are sitting just outside my offices at the moment.... | 09:38 |
| Bertl | ah, I see, you are on a scavenger hunt | 09:39 |
| hamster_ | Bah! 4k30 not starting straight up. A bit more debug..... | 09:40 |
| mithro | hamster_: you have to run all 4 lanes for that right? | 09:42 |
| mithro | I'm surprised at the lack of DisplayPort on all my monitors | 09:42 |
| hamster_ | mithro: Nope - 265 MHz pixel clock, 422 YCC just fits on two channels... | 09:45 |
| mithro | 422 is cheating ;) | 09:46 |
| mithro | I borrowed a coworker's monitor | 09:47 |
| mithro | Can everyone shout at me to put it back when I'm finished | 09:47 |
| hamster_ | What? HDMI 2.0's 420 is cheating - 422 is just optimizing for the device at the far end of the link (aka. eyes) 8-D | 09:48 |
| mithro | RGB or bust! ;-) | 09:50 |
| hamster_ | http://pethamstercare.com/wp-content/uploads/2009/10/dancing-hamsters-4.gif - 3920x2160@30 422 works, Tim's turn to jam the extra byte per pixel in there - how hard can that be? | 09:51 |
| mithro | hamster_: quick, commit and push before the people who make commercial DisplayPort cores hires a hitman :P | 09:54 |
| *** se6astian is now known as se6astian|away | 09:55 | |
| seaLne | mithro: where is a good place to collect wish list/feature request type items? | 09:56 |
| mithro | seaLne: depends on what the wish / feature is | 09:57 |
| Bertl | a pony? | 09:57 |
| seaLne | swapping automatically from presenter input to another input after delay period of no signal for the projector output (show generic slide show of announcements/sponsors etc) | 09:58 |
| mithro | seaLne: misoc firmware would be the correct place for now | 09:58 |
| seaLne | issues ok for that sort of thing then? | 09:59 |
| mithro | hamster_: I have lines on my coworkers monitor! | 09:59 |
| mithro | seaLne: yes | 09:59 |
| mithro | tag it "type-enhancement" | 09:59 |
| mithro | hamster_: got a 3920x2160@30 version for me? | 09:59 |
| hamster_ | Sweet - so far - DP 1.1a: 800x600 over one lane, 4k30 over two lanes - next step 800x600 over 4 lanes? That would prove the PCB as good, and allow for testing? | 09:59 |
| mithro | hamster_: yeah | 10:00 |
| mithro | hamster_: would prove the PCB works in many TX cases :) | 10:00 |
| mithro | Bertl: do you have a high speed oscilloscope at all? | 10:00 |
| Bertl | nope | 10:01 |
| mithro | Bertl: oh well | 10:01 |
| hamster_ | I can sneak a few cycles on a 5GS/s Tek if really needed... | 10:01 |
| mithro | hamster_: it would be nice to get an eye diagram of the signals | 10:02 |
| hamster_ | Oh... that fast - no chance :( | 10:02 |
| mithro | Numato are going to have to buy one of these if they continue to do this stuff | 10:03 |
| mithro | hamster_: my monitor isn't good enough | 10:03 |
| mithro | hamster_: Input Signal out of range! | 10:03 |
| mithro | Current Settings: H=65kHz V=30Hz | 10:03 |
| Bertl | maybe 30Hz is too low | 10:04 |
| Bertl | montors are not always happy with refresh rates below 50Hz, TVs are more forgiving in this regard | 10:04 |
| hamster_ | yep = 65kHz/30Hz. Anyhow. It's way past my bedtime again - I'm off to bed. Tim, drop me an email if I can do anything to help the cause... my pace might slow a little as I drops from "all consuming" to "quite consuming". But as a plus, the next bit I will be working on will be a VGA-like (pixel data + sync + blanks) front end.... | 10:05 |
| hamster_ | Is there some video mode that is close to your heart for implementation (bearing in mind my 'cheap' 4k monitor only goes to 4k30). | 10:06 |
| mithro | hamster_: I can take it from here | 10:07 |
| mithro | hamster_: the resolution fiddling stuff I mean | 10:08 |
| mithro | hamster_: getting the RX also working would be super awesome :) | 10:08 |
| mithro | hamster_: I need to go find some dinner myself | 10:08 |
| mithro | hamster_: I'm going to connect your output up to the Chameleon here and see what it thinks of your signal :) | 10:09 |
| Bertl | Chameleon? | 10:11 |
| mithro | hamster_: you are mainly on windows right? | 10:13 |
| mithro | hamster_: I might have a thing for you to try tomorrow or early next week | 10:15 |
| mithro | hamster_: getting the dual-mode functionality working would be another thing to do | 10:18 |
| mithro | anyway, I'm going to go get food | 10:20 |
| mithro | bblr | 10:20 |
| hamster_ | Bye alll. | 10:21 |
| *** hamster_ has quit IRC | 10:21 | |
| *** se6astian|away is now known as se6astian | 10:44 | |
| mithro | back now | 10:48 |
| mithro | se6astian: you see we had a success? | 10:48 |
| Bertl | mithro: btw, you know that the 7-series chips have support for the eye scan feature in tranceivers? | 11:24 |
| mithro | Bertl: nope! | 11:24 |
| Bertl | https://www.youtube.com/watch?t=19&v=zSnGAXoQ724 | 11:25 |
| Bertl | Vivado offers similar with the Serial I/O analyzer | 11:26 |
| mithro | I'm not sure I trust them :P | 11:32 |
| mithro | Bertl: want a job? :P | 11:32 |
| Bertl | currently I have jobs for three, and I still haven't figured out how to clone myself properly :) | 11:34 |
| mithro | Bertl -> https://www.youtube.com/watch?v=eR4i-XJDGCM :) | 11:34 |
| mithro | Bertl: I know that feeling well :) | 11:40 |
| mithro | Bertl: I think I asked you this before - do you have a mechanical spec for your expansion boards? | 11:46 |
| Bertl | I have one now, sec | 11:46 |
| Bertl | http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_plugin_module_1xHDMI_v0.5.pdf | 11:47 |
| mithro | Bertl: thats public right? | 11:47 |
| Bertl | yup | 11:48 |
| Bertl | that's the single HDMI plugin | 11:48 |
| Bertl | the Beta has two side by side with no distance inbetween | 11:48 |
| Bertl | the distance is created by the 5mil margin on the sides | 11:48 |
| mithro | Bertl: okay cool | 11:48 |
| Bertl | there are single and dual slot plugins | 11:49 |
| mithro | Bertl: do you have a spec for that? | 11:49 |
| Bertl | naturally the dual slot plugins use both connectors, i.e. have two blocks of fingers | 11:49 |
| mithro | Bertl: yeah - just want to see it to make sure I'm understanding you correctly :) | 11:50 |
| Bertl | they are identical to two single modules combined with the 5mil removed in the middle | 11:50 |
| Bertl | or to put it the other way round, take to single slot modules with a spacing of 10mil | 11:50 |
| Bertl | the height is not fixed yet, although the 875mil look like a good compromise | 11:51 |
| Bertl | (but that's not relevant for an adapter I guess | 11:51 |
| Bertl | ) | 11:51 |
| mithro | Bertl: it might with how I'm thinking of doing this | 11:51 |
| Bertl | it should be okay to simply have the PCIE-36 slots at 90deg from the PCB | 11:52 |
| Bertl | the plugin modules are oriented to one side, as they are on the side of the Beta | 11:52 |
| mithro | Bertl: got a picture? | 11:52 |
| mithro | Bertl: of the beta? | 11:52 |
| Bertl | https://www.apertus.org/sites/default/files/PCB-Stack-Concept-V03-022b-transp.png | 11:53 |
| Bertl | http://media3.film-tv-video.de/pics/B_NAB_Axiom_Beta_1.jpg | 11:53 |
| Bertl | http://media3.film-tv-video.de/pics/B_NAB_Axiom_Beta_2.jpg | 11:53 |
| mithro | Bertl: that is interesting - not how I was picturing it at all | 11:54 |
| Bertl | https://apertus.org/sites/default/files/beta-cabled.jpg | 11:54 |
| Bertl | here you can see a dual slot PMOD module | 11:54 |
| mithro | Bertl: got a schematic for that? | 11:55 |
| Bertl | sure, sec | 11:55 |
| mithro | Bertl: why did you go with PCIe rather then mini-PCIe? | 11:57 |
| Bertl | because it is a cheap solution so that everybody can whip up a board and order it via OSHpark without caring about connectors and similar | 11:59 |
| Bertl | just add the fingers and you're good to go | 11:59 |
| Bertl | http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_plugin_module_3xPMOD_v1.1.pdf | 11:59 |
| mithro | Bertl: That would be the same with mini-PCIe? | 12:00 |
| mithro | https://github.com/mithro/kicad-mini-pci-express :-P | 12:00 |
| tpb | Title: mithro/kicad-mini-pci-express · GitHub (at github.com) | 12:00 |
| Bertl | ever tried to plug a 1.6mm PCB into mini-PCIe? | 12:01 |
| mithro | Bertl: nope :P | 12:01 |
| Bertl | well, it almost works :) | 12:01 |
| mithro | README.md on that does say | 12:02 |
| mithro | 1mm PCB width. -- Most cheap places seem to do 1.6mm by default. | 12:02 |
| mithro | Yay past mithro for documenting things :) | 12:02 |
| mithro | Bertl: opps, I meant gerbers / mechanical not schematic | 12:02 |
| Bertl | gerbers are no problem, we don't have mechanical specifications for that yet | 12:03 |
| mithro | Bertl: well it already exists in reality right? | 12:03 |
| Bertl | yes, it does | 12:03 |
| mithro | Bertl: so.... I guess at some point I'm going to have to get one of these.... | 12:04 |
| Bertl | http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_plugin_module_3xPMOD_v1.1.zip | 12:04 |
| Bertl | here are the gerbers | 12:04 |
| mithro | Bertl: thanks! | 12:05 |
| mithro | Bertl: I think I know what I want to do mechanically for this | 12:05 |
| mithro | Bertl: but I'll print them out and confirm that it works | 12:06 |
| mithro | paper fit :P | 12:06 |
| Bertl | https://oshpark.com/shared_projects/ShbiDg6Z | 12:07 |
| Bertl | https://oshpark.com/shared_projects/lJ90ygq1 | 12:07 |
| Bertl | you can order them there for a few bucks and free shipping | 12:07 |
| Bertl | and for testing I can suggest this one *G*: | 12:11 |
| Bertl | https://www.apertus.org/pmod-debug | 12:11 |
| tpb | Title: PMOD Debug Board | apertus° - open source cinema (at www.apertus.org) | 12:11 |
| mithro | Bertl: perty :P | 12:12 |
| Bertl | so, no more distractions for a few hours ... I'm off to bed now :) | 12:13 |
| mithro | Bertl: I thought that was doing something nifty with delay lines - but then I saw the IC on the back | 12:13 |
| *** Bertl is now known as Bertl_zZ | 12:14 | |
| Bertl_zZ | simple is good ... | 12:14 |
| Bertl_zZ | (for debugging) | 12:14 |
| mithro | Bertl_zZ: true | 12:14 |
| mithro | Bertl_zZ: the fact you don't have a standard height on your boards makes me sad | 13:04 |
| *** rohitksingh has quit IRC | 14:13 | |
| *** rohitksingh has joined #timvideos | 14:21 | |
| *** rohitksingh has quit IRC | 14:29 | |
| *** rohitksingh has joined #timvideos | 14:44 | |
| *** CarlFK has joined #timvideos | 14:58 | |
| *** ChanServ sets mode: +v CarlFK | 14:58 | |
| *** rohitksingh has quit IRC | 17:54 | |
| *** rohitksingh has joined #timvideos | 18:08 | |
| *** rohitksingh has quit IRC | 18:13 | |
| *** rohitksingh has joined #timvideos | 18:27 | |
| *** rohitksingh has quit IRC | 18:33 | |
| *** CarlFK has quit IRC | 19:42 | |
| *** CarlFK has joined #timvideos | 20:16 | |
| *** ChanServ sets mode: +v CarlFK | 20:16 | |
| *** se6astian has left #timvideos | 20:53 | |
| CarlFK | mithro: make gateware ... | 20:55 |
| CarlFK | Saving bit stream in "atlys_hdmi2usb-hdmi2usbsoc-atlys.bin". | 20:55 |
| CarlFK | Bitstream generation is complete. | 20:55 |
| CarlFK | ERROR:Portability:50 - Execution of the "/home/juser/HDMI2USB-misoc-firmware/third_party/misoc/build/wbtc" failed. | 20:55 |
| CarlFK | ls: cannot access /home/juser/HDMI2USB-misoc-firmware/third_party/misoc/build/wbtc: No such file or directory | 20:55 |
| *** wanig_ has joined #timvideos | 21:38 | |
| *** kaalia has quit IRC | 21:45 | |
| *** wanig has quit IRC | 21:45 | |
| *** kaalia has joined #timvideos | 21:50 | |
| *** Bertl_zZ is now known as Bertl | 22:28 | |
| mithro | You can ignore that error | 22:46 |
| *** CarlFK has quit IRC | 22:47 | |
| mithro | The saving bit stream stuff is what is important | 22:47 |
| Bertl | yeah, save the bit stream! :) | 22:50 |
| mithro | Bertl: I don't think I understand your timezone :P | 23:12 |
| mithro | Bertl: your Austrian right? | 23:12 |
| Bertl | ah, I guess I have to explain that ... | 23:12 |
| Bertl | my timezone is BUT | 23:12 |
| Bertl | Bertl's Unique Timezone | 23:12 |
| Bertl | i.e. it is morning when I wake up, evening when I go to bed ... somewhere in the middle is noon/midnight :) | 23:13 |
| Bertl | sometimes that is in sync with the US, sometimes it is in sync with CET | 23:15 |
| Bertl | and yes, I'm from Austria and I live in Austria | 23:15 |
| *** CarlFK has joined #timvideos | 23:18 | |
| *** ChanServ sets mode: +v CarlFK | 23:18 | |
| CarlFK | mithro: make connect-lm32 ... HDMI2USB> ... not sure what the problem was yesterday. | 23:31 |
| mithro | Bertl: ahh, the joys of a remote worker :P | 23:53 |
| Bertl | yep | 23:56 |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!
