Archive for August 2014

GSoC Project Report: VGA capture expansion board for HDMI2USB on Digilent Atlys

GSoC Project Report: VGA capture expansion board for HDMI2USB on Digilent Atlys

GSoC Result Summary


Original Goals

The goal of this project was to develop a VGA Capture expansion board for HDMI2USB on Digilent Atlys. The deliverables for the project were usable product with associated hardware, software and documentation.

Fine details of original goals were:
  • Working PCB for VGA Capture
  • Working VGA Capture code for the hardware
  • Integration of this code with HDMI2USB
  • EDID Capability for the VGA Capture expansion board
  • Documentation for all the code and hardware
Original Proposal: http://goo.gl/rRtSrS

Final Outcome:

I achieved quite significant success, along with associated failures, in this project. While I failed to reach all the original proposed goal I have succeeded in getting the following working:
  • Nice and working PCB for VGA Capturing using Atlys
  • Working, but not perfect, VGA Capture HDL code
  • Integrated the VGA capture code into HDMI2USB.

The Expansion PCB features a VGA digitizing IC from Analog Devices, AD9984A and has support for auto-detection and as well as EDID. Its design is production-ready unless issues crop in future during continued development on this project.
top.pngbottom.png
DSC00446.JPG
The VGA capture code, on other hand, needs improvement and ruggedness to field working environment. It is currently working very well with the latest PCB version, but continued development is required on it.


Community Outcome

With this project. I was able to get VGA capture working, albeit not perfectly. Most of the project time duration was spent on PCBv01, which was later realized to be buggy, after testing the PCBv02. I created a new PCB design which was fabricated by hackvana and I tested it. It has significant improvements over previous design namely test pads, intuitive layout, length-matched traces, VGA splitter output, autodetection capability, EDID capability and friendly silkscreen. Most importantly, it works much better than previous version!
The code is rudimentary, but it works. It is able to capture digitized VGA sent by AD9984A and output it using HDMI. It also features a picoblaze core for I2C configuration of AD9984A. Missing  features are autodetection, EDID hacking capability and most important, automated test infrastructure.
Upon completion, the major benefit this project provides is giving VGA capture capability to HDMI2USB hardware using expansion board. Minor benefits it provides are autodetection feature for boards, UART-I2C bridge, EDID hacking & logging and test infrastructure (not yet

How to replicate results


Required:
  1. Digilent Atlys
  2. Properly set up Xilinx and Digilent toolchain. Refer to HDMI2USB Wiki for this
  3. VGA Capture Expansion PCB

Repositories:


One can either directly order PCBs using the Gerbers present in the repository or modify the KiCad files and generate own gerbers. The silkscreen is labelled with each component and values for easier assembly.

Once the PCB is successfully assembled, its time to test them. Start with testing power supplies at the labelled test points (VD, PD & OVDD). They should be 1.8V, 1.8V & 3.3V respectively.

Now one can load the bitfile from HDMI_Test_v06 repo or HDMI2USB-vModVGA-TestBitFiles repo. The name of bitfile is hdmi_test_v06.bit.
For those willing to modify/edit/improve the code, use the HDMI_Test_v06  repository and load the project into Xilinx ISE, modify code as per your ideas and generate bitfile for flashing into Atlys.
There is no need for any I2C initialization for AD9984A as I’ve added I2C core into the projects. Even then if you want to see I2C Initialization scripts for BeagleBoneBlack, then they are located in i2c directory.
For any issues or help you can contact me on IRC: rohitksingh on #timvideos channel or Email: rohit.singh@gmx.com. I would be very happy to help

To Do / How to help?

There are quite a few issues still remaining with VGA Capture HDL code. I encourage everyone to fork and develop code for the VGA capture. I have 11 spare Rev02 PCBs remaining with me, and would happily courier one to anyone interested in contributing in development.

Here are issues remaining:
  • Major:
    • Improve/cleanup the VGA Capture code
    • Add automated test & verification feature
  • Minor:
    • Add EDID hack feature
    • Add Autodetection code.
    • Add UART-I2C bridge

Documentation:


Learnings

I have to admit I learned *significantly* during the whole project duration. Learnings I would cherish lifelong are Git(Hub), KiCad, improved SMD soldering skills, vast VHDL experience especially on clock issues etc and finally basics of open-source development and community culture! I wonder how I survived without Github until now. Having used KiCad so much during these months, I feel like an outsider to Eagle now. I made many new wonderful friends during the project duration! And best of all, I’m now officially a hardware hacker of HDMI2USB!




Posted in | Leave a comment

[GSoC Daily Log]

Monday, 18th August 2014:

  • Tested VGA Capture code using Rev02 PCB.
  • My initial code (non-FIFO implementation) is working excellent!
  • But ironically, the FIFO implementation is not working on new PCB!

Posted in , | Leave a comment

[GSoC Daily Log]

Sunday, 17th August 2014:
 
  • Soldered the remaining components like Bypass caps and VHDCI connector.
  • IC is responding to I2C Hurray! :)

Posted in , | Leave a comment

[GSoC Daily Log]

Tuesday, 12th August, 2014:

  • Had to go to college. So, no work planned for today

Posted in , | Leave a comment

[GSoC Daily Log]

Monday, 11th August, 2014:

  • Tried using replacing the IC on assembled PCB from my spare AD9984A IC. In the process a pad on PCB got broken! :( Will have to do complete assembly from scratch on new PCB. Also got a burn injury on my finger when accidently touched the soldering iron. In short, Bad day!

Posted in , | Leave a comment

[GSoC Daily Log]

Sunday, 10th August, 2014:

  • Assembling completed. One PCB got damaged, due to its pad getting broken. (Either the PCB is slightly delicate or I’ve forgotten my soldering techniques :/ )
  • Power supplies working fine, but IC is not respnding to I2C.Very discouraging!  I was really excited to try work with my new board. Anyways, time for debugging.

Posted in , | Leave a comment

[GSoC Daily Log]

Saturday, 9th August 2014:

  • Received the PCBs!
  • Went shopping for some out-of-stock passives for assembling the PCBs. 10nf, 100nf, 22uF and 47uF SMD caps and 4.7K SMD resistors
  • Started assembling a PCB




Posted in , | Leave a comment

[GSoC Daily Log]

Friday, 8th August 2014:

  • Went to college. Wasn’t able to do any work

Posted in , | Leave a comment

[GSoC Daily Log]

Thursday, 7th August 2014:
  • Went to college. Wasn’t able to do any work

Posted in , | Leave a comment

[GSoC Daily Log]

Wednesday, 6th August 2014:

  • Went to college. Wasn’t able to do any work.

Posted in , | Leave a comment

[GSoC Weekly Overview]

This week although started great, ended becoming most unsatisfactory and least productive week of my GSoC duration!

I2C & VGA Capture

As mentioned in previous weekly overview, PCBs, have been sent for fabrication. After that I set on fixing the I2C part so that my mentor can try my bitfiles without much headache. I had purchased a logic analyser just a week back, specially for I2C debugging and, God!, it was really helpful in identifying problems in code. I fixed the I2C code and added it to VGA Capture code, thus eliminating need for any other setup for AD9984A initialization.
I also tried I2C using picoblaze, which interestingly used same resources as standalone I2C core.
Rest of the week was unproductive with respect to my project as my college has opened, and almost 4 days were lost to semester registrations, assignments,  and training conducted by Training & Placement (TnP) department.

Although, I was able to complete the 2-FIFO implementation, and simulate it. On, actual hardware, every alternate line is skipped by it on the screen, which it shouldn’t. As of writing this post, I’m currently testing the code on Atlys and trying to ascertain it reason.

Update: Just now I found out the reason for every alternate skipped lines. It was due to incorrect FIFO switching.
There is also a new weird issue. If the screen has only a background and no windows, then the video captured is very stable, but if I drag a window to the screen, the capture becomes unstable, and also some pixels are shifted! I’ll add a video of this phenomenon later today.

Upcoming Week:

I’ll be focussing on completing the code this week, so that the final week could be used for finishing touches and documentation. Also, upon receiving the PCBs, I’ll assemble and test them. They should hopefully offer better performance and color clarity than current PCB.

Posted in , | Leave a comment

[GSoC Daily Log]

Sunday, 3rd August 2014:

  • Simulating the 2-FIFO implementation. Problem in switching FIFO. Will have to look at code again
  • Update: Okay. It was the testbench’s fault;  never lowered down the reset! Silly mistake #facepalm

Posted in , | Leave a comment

[GSoC Daily Log]

Saturday, 2nd August, 2014:

  • Completed assignment, went to college for its submission

Posted in , | Leave a comment

[GSoC Daily Log]

Friday, 1st August, 2014:

  • Did the pending college assignment

Posted in , | Leave a comment

Search

Swedish Greys - a WordPress theme from Nordic Themepark. Converted by LiteThemes.com.