SOICbite: A Program/Debug Connector For An SOIC Test Clip

The problem is well-known: programming and debug headers consume valuable board space and the connectors cost money. Especially troublesome are the ubiquitous 100-mil pin headers, not because they’re expensive, but because they’re huge, especially along the z-axis. If you’re building miniature devices, these things can take up a ridiculous amount of space. With some clever thinking, [Simon Merrett] has found a way to re-use something many of us already have — an SOIC-8 test clip — to connect to a special footprint on the PCB without requiring another connector. He calls the system SOICbite.

The SOIC clip attaches to a footprint consisting of eight pads, four on each side of the PCB, plus five non-plated-through holes, which serve to anchor the clip in place. The idea of mating a PCB footprint directly with a removable connector isn’t entirely new — Tag Connect has been doing this for a while, but the connectors are expensive and single-sourced. On the other hand, SOIC test clips of varying quality are available from a number of vendors, including dirt-cheap deals on your favorite websites. The one disadvantage we can see is that the SOICbite footprint must be at the edge of the PCB to properly mate with the clip. The savings in space and cost may well make up for this, however.

[Simon] has made his KiCAD footprint available in a GitHub repo, and has offered to host footprints for any other CAD package there as well. So, fire up your preferred tool and draw one up for him to get these things widely adopted, because we think this is a great idea.

For the commercial alternative, check out our coverage of Tag Connect back in 2014.

 

33 thoughts on “SOICbite: A Program/Debug Connector For An SOIC Test Clip

    1. I use and like the Tag-Connect as well. However, I see the beauty in this method. If you added a jig to the setup so that you can quickly slide PCBs in and out without having to think about alignment, then this would be even better. One thing I don’t like about the tag-connect is that after 50 – 100 boards my hands start to get tired from holding pressure.

    1. I just use the standard thru holes and spacing then on the 2.54mm header I bend the pins slightly so that they go in the holes but they catch and friction holds them in. The same header is used on many boards and I have many headers.

      It just works.
      OK it’s no good for a production run of 10,000 but for a hobbist who can re-bend pins, it’s quick, and free and means no changes.

      For the production stuff in the day job it’s pogo pins and custom board testers because 10,000 units.

      1. instead of bending the pins I make the holes slightly offset from center in a zig zag so all the odd numbered pins are friction fit against one side of the hole and all the even numbered pins are against the other side

  1. I love the fact that I don’t need to buy a specific thing to make this work, the need to put it on an edge might be an issue occasionally, one solution would be to Make a small pcb and solder it on edge to the main board

    1. This part is rated “Minimum of 10 mating cycles”, but how many cycles it will reliably last in practice ? Because a programming/debug probe could reach hundred of cycles quickly in case of a small production.

  2. An engineer I worked with had two solutions to this issue. One was, of course, tag-connect. The other was to take a dirt cheap 8-pin board-edge connector (Think PCIe slot, but smaller), hacksaw out one edge, and design in a slot on the board beside the fingers. Tada, keyed low-profile 8-pin programming header.

  3. We use the Tag Connect Header. It works great and has lasted for years. The only thing I do not like about it is during development, It can be difficult as a debug header. It wants to come out of the board.

  4. Now, if Microchip would get rid of that ridiculous RJ11/RJ45 connector on the ICDx programmers and replace it with a 0.100″ connector then we could have a decent way to use ribbon cable on the programmer end of the cable.

  5. Idea: Use 4 Pogo pins and 2 Header Pins in a h-p-p-p-h-p configuration. Crappy visualization:
    https://i.imgur.com/HAOvEpP.png

    The Footprint (front) is in a straight 1.27 grid. The Pin Header on the left is shifted a bit, so friction may hold the rest down a little. The Pogos can be moved closer together if space is a constraint. And the footprint for the pogos don’t have to be vias but can be pads as well.

    Bad idea?

  6. This. Is pretty brilliant! The SOIC clip ain’t that standard in the toolkit, but it is widely available.

    It also solves the problem of Pogo-Pin/Tag-Connect systems in that they have to be pressed onto the target board constantly. This is just clipping on a Clothespin and done. Giving the same convenience as using a bulky 2.56mm pitched header.

    Definitely Gonna port this design over to Eagle

      1. Sure thing! I cleaned my rough copy a bit to be proper Eagle library and have submitted a Pull Request. Also included Eagle render of footprint along with the device entry as it appears in the library

  7. Sure thing! I cleaned my rough copy a bit to be proper Eagle library and have submitted a Pull Request. Also included Eagle render of footprint along with the device entry as it appears in the library.

Leave a Reply

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