Lark Board - Altera Cyclone V SoC Evaluation Kit - Review

Table of contents

RoadTest: Lark Board - Altera Cyclone V SoC Evaluation Kit

Author: iexpress

Creation date:

Evaluation Type: Independent Products

Did you receive all parts the manufacturer stated would be included in the package?: True

What other parts do you consider comparable to this product?: AT32UC3A3256 -Atmel SAMA5D4 Xplained - Mojo Board (FPGA control with Atmel Atmega32U4 MCU)

What were the biggest problems encountered?: This board assumes you are an expert with Linux, and from my perspective and experience with this board model so far, it is logical to state that one MUST have an equivalence of a Bachelors degree in computer science to make use of this board or bring understanding to the board's intended or unintended operations which may be improvised in a nifty manner with some cleaver shell scripting. Couldn't figure out the password for user LOGIN, typed in root as a guess, and then it loaded root without a password. The board environment is not programmed, you'll be spending a lot of time trying to figure out how to mount different types of devices to mount points, but once you mount them, it feels like an accomplishment. No internal volumes are mounted except for /root along with a tmpfs for volatile vars and ram. Figuring out the commands in the bash/vt100 after root login was horrific, but am implementing system features rather quickly through bash scripting and GPIO pin toggling for the HPS and FPGA Expansion slots to communicate with external hardware. Looking through thedocumentation and reading the manuals, it was not clear how to use the sys/class examples other than /sys/class/leds/hps_led brightness examples, which didn't work immediately until the system stopped offering auto/manual component testing upon bootup. The bash prompt has a chr(13) bug in it when mounting sda partitions, it's as if it gives itself the return key about 40 times per volume you mount. Testing is commencing, will update my progress over the next couple days, since the board is now operational. keywords shutdown and mkfs don't work other than mkfs.minix Had to format all devices on a machine running Windows XP. Yocto treats FAT and FAT32 the same, and mounts NTFS volumes as read only for some reason. Perhaps there is a 32768 MB barrier but it mounts FAT32 just fine if the partition is under 32768 MB in size. I have not been able to format any partitions in the text mode except for minix, and then it still couldn't comprehend what minix was when I tried to mount a minix formatted partition. Looked up tons of other external documentation such as Fedora and Ubuntu user manuals to figure out what the bash/vt100 commands were, before finding a list of commands somewhere in the /usr folder. One of the HPS buttons popped off the board when trying to desert a micro sd card from the TF slot. Others may place their finger on a button while inserting or deserting a card from teh TF slot, for leverage. If I popped a button off that easily, the components don't have enough solder on them. The bash shell sees my USB to ATA/ATAPI converter and DVD drive but fails to let me mount it. This system doesn't know what iso9660 is. It just states the device has input/output errors. Testing the unit with a multitude of hardware and hardware configurations to see what this board can do natively. Will be updating more on progress shortly.. Thanks.

Detailed Review:

Please check back often, I'll be leaving an amazingly detailed product review very shortly.




Shouldn't even need Quartus since we can talk to the board directly, or should be able to soon after I develop software for it.


Altera software gets an F- for difficulty of use and compatibilty. Never could get it to work.


I can say if their programmers were good enough, they could make their software compatible with x86 on up.

This is because I can implement 256 bit code on an Atmega328 8-bit MCU.


Their programmers aren't putting enough effort into their software.


Installing all the following on one partition.. Yep, it's possible.


The unsure nature of people having to guess which OS works with Quartus will lead to further Microsoft 64-bit OS sales.





Proper testing of the Lark Board is now in session.


  • The Lark is an impressive (and complex) bit of kit produced by a Premier Farnell subsidiary Embest.

    Why were you not able to have the unit checked, repaired or replaced under its 12 month warranty?

  • That's a great question. I was writing the manufacturer, Element14 team, and Altera at the then time, but Altera blocked me from requesting customer service through their service requests on their website regarding Quartus and the board's USB Blaster II.


    With a change to Altera after being acquired by Intel, they released a Quartus Prime Lite Edition which I still couldn't get to detect the board even though Windows saw it perfectly.


    Supposedly, the Quartus Prime installation included the CycloneV qdz support in the installer, but when I ran the program, it said no support for any device families were installed. Then I tried the Device Installer, and when I tried installing the qdz for CycloneV, it said it was already installed.


    Overall, the Altera side of things is terrible. Hopefully with Intel taking over, the company's software and customer service will improve.


    Will update more on the progress over the next few days.



  • Excellent.

    I'd concentrate my efforts with communicating with Embest and Element14 to discuss your issues. Altera's focus is upon its semiconductors only.

    Keep your questions simple to avoid them from being lost in translation.


    The Lark should be able to perform the functions of the Red Pitaya and more.

    The main difference is the vendor of the SoC - Altera vs Xilinx.


    We'd love to see the Lark operating as an SDR - (Software Defined Radio) amongst other things.

  • Certainly can do SDR, I found a way image An impressive way actually to hold premodulated waves in an array and then using an index pointer into the array, uints are sent to a DAC based on voltage level input. 1:1 ratio.


    It took custom programming to create the promod wav but I handed my array 4 samples per wave cycle per voltage level value.


    Starting with 8 bit on that, 256 values x 4 samples per voltage level inside the array so there's direct conversion on demand.


    Could have full frequency amplitude modulation, but also made a R2R hardware DAC to take parallel input. Can definitely hook that up to the HPS expansion port and do readouts on that.


    Perhaps we can even create WiMax or 4G signals with it. Just by studying the text mode bash prompt, I believe this board can do just about anything, in bash alone.


    The question is, are all of the hardware features of the Lark Board available as is right out of the box, inside the text mode bash prompt before starting an x-session?


    I haven't started X actually yet, going to figure out bash in its entirety for the board, and do believe all functions of the board can be accessed via the console outside of XFCE.


    cat urandom > fb0 - works perfectly filling the linear framebuffer, and theyalso gave different color/depth modes.


    Going to try to implement a bytecode interpretter, like Java, routing numbers to functions and hand the board USB CDC through the USB Blaster II jack to try to get bi-directional communication going.


    As far as Quartus, could never get it to work so developing a brand new driver for the Lark Board.


    Could just make a sript-based OS for the Lark Board and release it to all under GNU licensing along with all the other software I'm creating for the board during testing.


    Could even make a Verilog/VHDL compiler for it right inside the bash prompt through scripting alone.


    As for the FPGA portion, I don't see a way to access it unless there's a way to FORCE rewriting the FPGA in real time with the HPS2FPGA bridge, if they implemented it.


    I am more interested in the CPU portion of the part though.,

    The FPGA portion just creates a pathway for voltage to travel along.

    Through my understanding, one byte tells each FPGA cell what gate to enable.


    It is doubtful the FPGA cells have all 7 digital logic gates individually, but this would make it a superb part if they gave each individual cell all 7 gates.


    I have been soldering and making custom circuitry to go along with all this, and put a small logic board together, single cell FPGA board to select which gate you want to use and then have an asynchronous IO throughput of about 2 million bits per second through the selected gate. Will be posting pictures as I go along.


    Super Excited gain.

  • So the board is now up and working?

Comment Children
No Data