USB IR Toy preorder 1 defect

Unfortunately there’s a defect in the first batch of USB IR Toys. The infrared LED is backwards and does not emit properly. Ytsirk caught this defect and documented the issue in the forum.We’re really sorry about this bug, and we’ve done our best to track it and identify the cause.

We used a Cadsoft Eagle IR LED footprint in the USB IR Toy, specifically the SFH-482 from the default LED part library. The IR Toy design doesn’t require the SFH-482, almost any IR LED will work, but that is the package we found first. The silkscreen for the Cadsoft part has a tab on the anode (+ side) of the LED (bottom right image). The SFH-482, according to the datasheet, actually has the tab on the cathode side (top image). We used an IR LED on our prototype that looks like a normal LED without a tab, so we didn’t see the problem with the silk. We should have caught this and corrected it, or chosen a different LED package.

Seeed used SFH-480 IR LEDs (SFH-482 compatible) on the production IR Toys so they matched the Eagle part footprint. This is a miscommunication on our part because any IR LED would have worked, but this part is also fine. Seeed assembled the LED according to the silk, which was backwards, thus all the emitters are backwards.

There is a self-test in the USB IR Toy, and despite the backwards LEDs, the self-test still passes. We’re guessing that there’s enough current leaking though the IR LED that it blinks dimly despite the incorrect placement.

You can repair your USB IR Toy by unsoldering the IR LED, and re-soldering it so the metal tab is opposite the silkscreen (facing the indicator LED ‘I’ and R2).

We should have caught this, and we’ll do our best to clean up the mess. Please contact us if you need to arrange a repair or a replacement LED.

This defect will be corrected in all future USB IR Toys produced.

Join the Conversation


  1. Hey, nobody’s perfect. At least you found it and didn’t pull a Detroit or a Toyota.

  2. My de-soldering skills are still somewhat entry level. How concerned should I be about damaging the pcb in the process?

    1. If you have an adjustable iron there isn’t much worry of burning the board. If you have a $10 firestarter you should probably take a minute break for things to cool down every few minutes. The LED pads are nice and large, so they can that quite a bit of heat before they start to peel off.

      I usually use the alternating heat and yank method:
      First, clean up as much solder as you can with copper braid (desoldering wick).
      Next, alternate the iron on both leads until the solder melts, then continue to heat and pull on the LED until it slides out.

      Replacing the LED is easy if the holes are free of solder. If the vias fill with solder you may need to resolder it using the reverse of the heat and yank, the heat and force:
      Get the solder in the vias to melt, then gently force the leads through.

      But if you’re at all worried, feel free to send it to me. I’ll make the repair for you. I’m really sorry about this bug.

  3. I use a cheap non-autoranging DMM and a mobile phone camera, to quickly test IR LEDs.

    Switch the DMM to it’s <200ohm resistance measure (mine outputs ~2.7V). apply the probes to appropriate terminals and watch the LED through the phone camera. You should see it light up.

    The phone camera is usually cheap enough that it won't have an IR filter, but you can check by recording the business end of a TV remote control, while you press random buttons.

  4. I know how you feel. I was testing various batches on RGB LEDs, and made breakout boards for each type. One type refused to work, even though I’d carefully studied the spec sheet, and everything looked right.

    Surprise! The manufacturer in China had flipped the polarity on the devices, and never bothered to tell their US distributor. I caught the error when Google turned up the manufacture’s spec sheet on their Chinese web site.

  5. Glad I saw just as I was sitting down to figure out why the IR emitter wasn’t working! It was a little ironic that I had to remove the QC sticker to unsolder the IR LED… I noticed my board is missing C2, is this normal? Otherwise I’ve been having a great time playing with this thing. Thanks.

    1. Missing C2 is correct. On the prototype I used 2×0.1uF capacitors because I don’t have any 220pf to use for the USB voltage regulator in the PIC. Seeed has 220pF, and it’s cheaper to place 1 than 2 caps, so they just used one and left the other unpopulated.

  6. Welcome to Eagle! This is a common problem. I have found so many bad packages with pin numbers in the wrong order. On the one hand, it’s cool that they have a community of volunteers creating packages, but the problem is that the free libraries have their mistakes. To their credit, sometimes there are similar packages with no standard pinout, so it may not even be possible to have one package which can simply be used without double-checking.

    In my design work, I always have to double-check the size, positioning, and signal names for every package that I use in Eagle. I’ve actually created a lot of package in my own library with chips and parts that I’ve confirmed.

    All things considered, I’m not disappointed that there was a mistake like this. Of all the things that could have gone wrong, this one is the easiest to fix. It’s not even a surface mount part! For the $20 that it cost, even with a little labor I still feel like the USB IR Toy is a great bargain.

Leave a comment

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.