Hackaday Prize Entry: BeagleLogic

A few years ago, [Kumar] created the BeagleLogic, a 14-channel, 100 MSPS logic analyzer for the BeagleBone as an entry for the Hackaday Prize. This is a fantastic tool that takes advantage of the PRUs in the BeagleBone to give anyone with a BeagleBone a very capable logic analyzer for not much cash.

This year, [Kumar] is back at it again. He’s improving the BeagleLogic with a BeagleBone on a chip. This is the BeagleLogic Standalone, a 16-channel logic analyzer at 100 MSPS using a single chip.

Like the BeagleLogic from a few years ago, [Kumar] is relying on those fancy PRUs in the BeagleBone that make reading GPIOs and blinking LEDs so easy and fast. Unlike the BeagleLogic shield/cape/whatever, the BeagleLogic Standalone uses the Octavo Systems’ OSD3358 — the BeagleBone on a chip — for the hardware. This incorporates everything in a BeagleBone into a single package, making for a compact unit that still has all the capabilities of the bigger BeagleLogic.

On board this pocket-sized logic analyzer is the OSD3358 itself, the logic analyzer frontend, a gigabit Ethernet port, USB, an SPI Flash, SD card slot and eMMC, and an RTC. An expansion header breaks out a UART, I2C, SPI, two PWM outputs, 6 GPIOs, and a clock to a PRU for experimental synchronous captures.

With a web-based frontend for this Logic Analyzer, this looks like it’ll be a fantastic tool for any hardware hacker, and something that should be reasonably inexpensive.

17 thoughts on “Hackaday Prize Entry: BeagleLogic

    1. I would say no, since its not a RTOS running on system, I could be wrong and the whole GPIO side is running Real Time, so in the OS its a set and forget, but since most OSs that run on the BB are not RT, if something deadlocks the core, you might have a injector stuck open till it gets a command to close..

      but yes, set it up right, you can run an engine on the BB, a V16 can be controlled on a 6502 even..

      1. The PRUs (Programmable Real-time Units) on the AM3358 are certainly hard real time, they don’t even run any RTOS but are rather programmed to the metal. It’s why reason that despite the age of the SoC you still see many new products with AM3358 due to the flexibility of the platform.

  1. A more interesting project would be to do the same LA specs (16-bit or greater @ 100 MS/s) at a fraction of the cost using an input buffer board and a $49 FX3 explorer kit leveraging hardware most hackers already have – a laptop or desktop w/ a USB3 port. The OMAP3 seems limiting here and only adds cost to add a feature most people don’t need (standalone operation).

    /shrug. I guess I should probably put up or shut up rather than critiquing some otherwise impressive work. Just my $.02

    1. It’s actually an AM3358 SoC which is a distant relative of OMAP3, and no, it’s hardly the bottleneck. In fact, the dual PRUs (two 200MHz custom RISC processors optimized for real time IO) on the chip is exactly what enables this application, and the whole thing should cost about the same as an FX3 Explorer.

      1. You cannot sustain capture at 16-bit 100 MS/s. The OMAP3 (general family name) does not have any interfaces that allow continuous output at 200 MB/s. Even if 80% of dedicated RAM was used, that only affords 4 seconds of capture time max. FX3 would allow sustained capture with a fast and large enough disk. That’s why I say it’s limiting.

        And it is impossible for a small run of these boards produced for less than $49 + the cost of a buffered wing board. Qty 100 pricing (w/ only 512MB RAM) from Digikey is $41/ea for the SoC alone….

        1. That’s why I said it’s not quite an OMAP3; the AM3358 has a gigabit Ethernet interface which was not used on the Beaglebone but which this board is planning to use. The FX3 is also not that cheap in small quantities costing around $30, with no DRAM whereas the OSD3358 comes with 512MB DDR3.

          1. Again 1 Gbps is less than half of 16 bit * 100 MHz. You can’t sustain capture rates anywhere near what an FX3 can nor the claim of 16-bit @ 100 MT/s. And if you are doing sustained capture, you don’t need any local RAM other than for simple buffering which the FX3 has.

            This project is a great eval board for an BBB clone. But call it just that. Say that it can be used for a 2s one-shot 16-bit 100 MS/s standalone LA if you want. But there are better/cheaper ways to skin the low-cost continuous LA cat. I’m just surprised it hasn’t been done yet.

          2. If all you want is to dump gigabytes of raw capture onto your hard drive (which I wager won’t even keep up with the sustained 200MB/s) then you might as well go with an FPGA connected to a PCI-E bus. The wonderful thing about having a full fledged Cortex-A8 is that you can do all sorts of processing and compression on the captured data so you wouldn’t need to constantly stream 200MB/s of raw data. But whatever floats your boat I guess.

          3. kaofishy and how much for fpga capable of implementing PCIE?
            not to mention even 10 year old $30 core2 laptop has 2-4 more processing power than AM3358

          4. Except the laptops in question would have limited IO and cannot do real time processing the way a Cortex-A8 + PRUs can. There’s a reason why we don’t run everything on GP x86 PCs. My point is that while this is not the *only* way it’s certainly a very flexible one and you get a lot of value for your money.

  2. * Open Workbench Logic Sniffer $50

    200 Msps 16 ch native. Expansion options available. Open source.
    Yeah this is a Dangerous Prototypes (DP) thing, so it’s going to take some work to get it running. But the DP site has people that will help. The problem is the documentation is poor/dated on the DP site and the damn thing keeps going out of stock at the likes of SeedStudio. Also, SeedStudio NEVER ships products like these with up to date firmware. So that’s your first obstacle after getting it working. But it DOES work! I use mine regularly and it shines :-)

    https://www.seeedstudio.com/Open-Workbench-Logic-Sniffer-p-612.html

    http://dangerousprototypes.com/docs/Open_Bench_Logic_Sniffer

    The DP support forum for this thing launches from here. As questions, you will probably get some answers not already covered, but it’s no panacea:

    http://dangerousprototypes.com/forum/

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.