sampled some TI chips and bought some surfboards

One of the minor hurdles that I have to deal with is how to connect the output from the amiga video, which is 5v to the FPGA at 3.3v.

I was originally going to do a custom circuit board but that has turned into a pain in the butt, and there’s simply too much work involved for me to want to mess with it.  Maybe I’ll do a custom PCB as phase II or something.

Anyways, I ordered the SN74CBTD3384CDBQR – 10-Bit FET Bus Switches With Level Shifting with -2 V Undershoot IC to convert between the two voltages.

An application note at TI explains how it works.

But these damn chips can’t be sampled in DIP. I ordered some QSOP-24 packages, and then some surfboards.

I ordered the surfboard like things which convert to QSOP to DIP.   The only ones I could find that had great datasheets on it was this:

They were somewhat expensive. Almost $10 each shipped, but I’ve got to get moving on the project, and this is currently holding me up.

I’ve never soldered SMT with a small pitch (0.635″, 25 mil), so this is going to be a challenge. I have to order some flux, a loupe, and I eventually want to get a magnifier with light…..



Amateur Electronics Design Engineer and Hacker


  • If all your signals are unidirectional, could you use something simpler like a 74LVC244? That’s a pretty popular solution for going from 5V to 3.3V, and you can get it in a through-hole DIP package. Looks like DigiKey has them for 48 cents. For the Amiga VSC I’m not sure you’ll have any signals coming from the FPGA back to the Amiga, but if you do you can generally just connect those straight up without any level conversion.

    Good luck!

  • So it’s all coming back to me now.

    I wanted to build a semi-generic voltage conversion board that can be used in multiple different applications. I imagined creating a few PCBs and then using them to interface various different projects to my FPGA boards.

    For this application, the signals are only unidirectional.

    The 245’s parts are only unidirectional, and have a comparable long propagation delay. (which may or may not matter for certain applications)

    The IDT quickswitches are very popular for this, but they have a requirement of 4.3v Vcc. An external diode and resistor can bring 5v down to that, but I didn’t want to mess with extra parts.

    Enter TI’s parts, recommended by someone on Altera’s forums, that have the built in diode for dropping the voltage, very low propagation times, support bidirectional communications (even if the voltage is only translated in the H->L direction)

    They either weren’t available in DIP, or couldn’t be sampled in DIP, or both.

    I also planned on creating a PCB, and if I went that route, I wanted to go with SMT parts anyways.

    THEN I said, f*ck the PCB because there are too many details I don’t want to spend working with, and so now that I’m back at the solderless breadboard stage — I wanted to get the TI parts onto the breadboard. I wanted to practice my SMT soldering anyways, so ordering these surfboards will let me accomplish the task.

    However, now that I’m going back to the breadboard design, I may revert back to the 245 design. I do need a breadboard with multiple sets of busses to accomodate both the IDC-40 pin and the (2) 245’s.

    I’m brain-dead sometimes.

    P.S. I actually have a few of these at home now.

  • Late reply: I see you went with the 74LVC245’s, but I’m curious about your propagation delay comment. From the datasheet, I see both the 244 and 245 have propagation delays around 6 ns. I would have called that fast (relative to your clock speed), but I see that FET bus switch you linked has a tpd of 0.15 ns!!

    • Yeah, I double checked the data sheets before posting the comment about the propagation delay. Big difference. Now whether it matters, probably not. Hence my albeit in the parentheses. 🙂