Computing stuff tied to the physical world

Let’s switch to a better MOSFET

At this stage, we have a working circuit, which correctly powers up the LPC810 µC. Now, under software control, we can turn on power to the RFM69 wireless radio.

But there’s a slight problem. The P-MOSFET chosen was picked at random from components which happened to be around at design time. From the datasheet:

Screen Shot 2015 05 05 at 16 58 58

When pulling the gate low by 10V, we can turn the MOSFET fully on. According to these specs it then has an equivalent resistance between drain and source of up to 150 Ω. That’s really a lot – it could lead to a voltage drop of perhaps 0.3V for the µC, but with an even higher transmit current of say 20 mA, this MOSFET is bound to be completely useless.

We have to find a P-MOSFET which turns on well with no more than 2..3V, and which then has a considerably lower resistance. Let’s fix this before trying to send radio packets.

Keep in mind that the Vgs(TH) “knee” at which a MOSFET turns on is not sharply defined.

There are no doubt better alternatives, but again rummaging through the P-MOSFETs in the large “lab component stash” at JeeLabs brought out this unit, which is still through-hole (the same TO-92 package and pinout, in fact), and can be expected to work better:

Screen Shot 2015 05 05 at 17 13 03

That’s a TP2104 from Supertex. And it’ll accept a 3V gate voltage:

Screen Shot 2015 05 05 at 17 15 27

It’s not perfect: 10 Ω at 20 mA is still a 0.2V drop, but at least this one should be able to feed an acceptable 2..3V supply voltage to the RFM69 while it’s in transmit mode.

Since this MOSFET is package- and pin-compatible, we can switch both of ’em to this type:

DSC 5047

… with lots of probes attached to monitor various supply voltages as well as the SPI bus.

Now – at last – we’re ready to try sending out wireless packets!

[Back to article index]