• rev0 assembled and lessons learned

    Gergely Imreh • 05/13/2016 at 10:04 • 0 comments

    Rev0 is back, and boy did we learn a lot in the process:

    • PCB factory somehow misaligned the solder mask just enough to enable accidental shorting between basically all the pins and the ground plane, but only sometimes... Need to find a new factory, I guess!
    • Forgot to add a noise-reject capacitor between the T-/T+ pins of the thermocouple sensor, so it basically didn't work. Soldered a small one directly onto the pins of the MAX31885 IC, and works like a charm! It's a 100nF instead of the recommended 10nF, not sure if it makes any negative difference
    • Figured out the voltage probe things. No wonder the parts were so expensive, those are not to plug the voltage probe receiving end, but to front - these should be used for the points that one wants to measure! So, need new connectors, found some in the local computer market, and redone it with some wiring hanging off. Works pretty well.
    • I've mixed up some of the silkscreen, wrong factual information (e.g. wrong pin number indicated), which needs to be fixed
    • Gave the diodes way too small footprint, were really difficult to solder, the rest of the SMD parts were just normal SMD difficulty. Maybe the MAX3378 with its 14-pin TSSOP is a trouble too, but still.
    • A microscope is super handy for SMD soldering! Been using one in the hackerspace, will have to get one in the future, with some good camera attachment too.

    There's the current assembled test package, for proof-of-concept use, before redesign.

    Tested the temperature sensors with some hot water cooling datastream. Simple but effective.

  • rev0 now being manufactured

    Gergely Imreh • 03/02/2016 at 13:13 • 0 comments

    Sent "rev0" to manufacturing yesterday with Seeed Studio's Fusion PCB service, and placed the order for the passive components, mainly on Mouser, but needed one piece from RS Components. Within 24h the board is now in processing and the parts have been shipped, so soon this should not be a Gimp mock-up, but a reality:

    Though, to be fair, I should try to do the modeling of the missing components (the voltage probe sockets, the thermocouple piece, and the missing Maxim ICs (U2 and U3). Looked into that in KiCad, and that's the only part I didn't yet dare to do (the 3D models). But that will have to come soon enough.

    Now waiting for the part delivery, the PCB will be the slowest (especially if I don't want to pay more for the shipping than for the actual manufacturing of this 10pc prototyping run).