Attached below is an outline schematic for a single chip circuit to drive your two LEDs with your Teensy 3.2.
The PSMN1R1-30PL is fantastic for switching my peltier tiles without requiring a heat sink (24V 8A). With the minimum VGS(th) being stated as a max 2.2V I thought this would be perfectly acceptable with the uC logic being >1V more, is there a standard or guideline I should look to follow in future?
I've never had any dealings with surface mount - aside from one LED about 8 years ago. I'm interested, just is this really practical for small runs? Space is extremely tight so I'm pretty certain surface mount components would help a bunch!
The MAX14874 looks massively overkill on paper but it's tiny!! To be fair though, if I were to go surface mount, then the component count wouldn't be an issue. Half of my board is peppered in analogue logic and precision resistors.
--
This is a single unit with 2 LEDs, with built in resistors.
If I were to go surface mount then I would definitely go simple with the NPN in series with the base of a PNP.
Cheers
Yes the PSMN1R1-30PL is a fantastic MOSFET. I was quite surprised how good it is.
There are quite a few characteristics to look for when specifying a MOSFET: VCE, SOA, RDss, Vth and Thermal resistance J/C being the main ones. IDmax, and especially Ptot max, have little practical use and are mainly sales inspired parameters.
When it comes to driving MOSFETs it is a common mistake to think that Vth is the only parameter to consider. While it is important it is not the end of the story. Paradoxically Vth only really tells you the point at which you can turn the MOSFET off, which is obviously important. But what really counts is the value of Rd, because Rd defines the voltage drop between the drain and source. So it is important to look at the MOSFETs graphs of Vg/Rdss or equivalent.
In summary, I would suggest that, as a general rule, choose a MOSFET that has a low enough Vth to be turned off by the logic low output voltage of your driving device, so that there is the largest possible positive swing from the logic 0V to the logic 1V. That way you will get the lowest Rd and can drive the most current.
Tthere is another factor though. In general, the lower the Vgth the worse are the other MOSFET characteristics, particularly VDSmax and Rdss. In addition the various parasitic capacitors go up to alarming values.
Taking all of the above into consideration I would suggest that a MOSFET suitable for driving with 3.3V MOS logic should have a Vth of around 1V. That gives an effective gate drive of 2.3V for logic 1 but still ensures that you will be able to turn the MOSFET off with logic 0.
I hate to say this, but describing any design/device as 'overkill' is very dangerous. If you applied the implications of that word generally very few items would pass the criteria, especially MCUs. I suggest that the only criteria that should be applied, in practice, are: cost, size, weight, performance. So it is with the MAX14874. I can remember when 16k byte memory chips were defined as overkill.
As you also say, you can squeeze an awful amount of electronics into almost no space at all using surface mount components. I prefer the bigger SM chips with around 1mm pin spacing rather than the teeny SM chips increasingly used. Like it or lump it though, SM is the only way to go if you want to use the latest and greatest chips. But from a performance point of view SM does have problems: heat dissipation being one, and ceramic SM capacitors have a worse performance than the larger through hole types. As you no doubt know, for breadboarding you can fit your SM components to DIL headers: especially useful for messing with opamps.