Was stuck on creating libraries - Now just general KiCad banter

[photomankc]

I might just have to get that!

[photomankc]

Uggghhh. With the transistors on the back it shrinks easily by another 20%. That’s tempting… 500 mils wide at the pins and 1200 mils long from top to bottom pin. That may be enough to pay the price of hot air for the transistors on the back.

[photomankc]

pic_smart_switch_small.zip (2.0 MB)

I hope I don’t have any glaring stupid errors in this. It does have some new circuitry from my original version which had no transistor on the PWR_OFF signal so that had to be a 3.3V input and I swear to god I can’t keep the symbols for MOSFETs straight for source/drain no matter how much I try.

[SembazuruCDE]

@photomankc

MOSFETs straight for source/drain no matter how much I try.

Just remember, the source and drain are the pins that aren’t the gate.

[qu1ck]

Here, a smaller 2 layer board that I quickly put together. Nothing to be proud of, it can be polished further, but just to show that there was quite a bit of overhead. With smaller smd button you can easily shave off another 4-5 mm off one dimension.

pic_smart_switch.kicad_pcb (94.5 KB)

[photomankc]

Wow, that’s pretty tight. I see you dumped the ref numbers off the silk which seems to get it more compact and you’ve got a better eye for tight routing. I’ll probably stick to my second version but your’s gets dang close to the same size as some of the Robot Shop and Pololu units that don’t have an MCU so that’s pretty impressive to me.

Thanks for the inspiration though!

[Rerouter]

You appear to have 1 glaring issue, the body diode bypasses your switch,

So I had my own go at it, nudged it further down, If you used an SMD switch, you can rotate Q1 and likely trim it under 25mm long, for less than 24mm long, the power Diode would need a new home, but its possible

pic_smart_switch.kicad_pcb (110.6 KB)

[photomankc]

I don’t think so… it’s a P-Channel and the cathode of the diode is facing VIN at the source of the MOSFET. The load would be coming off the drain at VOUT and the diode would only conduct if positive voltage appeared at the drain with the source open or grounded.

I know this works with the drain pins facing VOUT because it’s operating on my breadboarded RaspberryPi system right now.

[dennischi]

Your schematic has the source of the P-FET connected to Vout and the drain connected to Vin. The body diode will be forward biased when the FET is off so Vout will be a diode drop below Vin. You need to reverse the source and drain conections in your schematic. Either that or your library part is drawn incorrectly. I couldn’t find a datasheet for that part to check the pinout.

[dennischi]

Based on the datasheet for a similar part Si4403CDY, I can see your schematic symbol is incorrect.

The source connects to pins 1-3, the drain connects to pins 5-8. You just need to correct your schematic library. The footprint is correct and the board will work as you have it connected.

I just realized you posted the required section of the datasheet for your part. That confirms the error in the schematic symbol you created.

[photomankc]

It has been corrected in my current library. I Must have zipped before I caught it. As I said earlier, whatever I draw the first time for Transistor or MOSFET is ALWAYS wrong!