# Gas station without pumps

## 2016 July 16

### pFET Ron_vs_Vgs

Filed under: Circuits course,Data acquisition — gasstationwithoutpumps @ 22:25
Tags: , , ,

I next need to look at whether a similarly simple pFET circuit will work for them to measure Ron  for pFETs.

That turned out to be even easier than I expected.  All I needed to do was change the source wire from Gnd to 3.3V, the voltage divider for the gate to report the voltage halfway between Vg and Vs, and the flip the function generator leads to drive Vgs negative instead of positive (also flipping the polarity of the electrolytic capacitor at the same time).

I had to adjust the gnuplot scripts somewhat to fix the computations, as Vgs is now -2*(\$4-\$2) rather than 2*\$2 and Ron is now (Rload*(\$4-\$3)/\$3) rather than (Rload*\$3/(\$4-\$3)), with the three channels still connected to Vg, Vd, and 3.3V.

With a drain-source short the wiring resistance came out to 24mΩ this time.

I have 3 pFETs:

• NTD2955-1G (39.87¢ in 100s) used this year in class
• IRFU9024NPBF (53.13¢ in 100s)
• IPP45P03P4L11AKSA1 (66.93¢ in 110s)

The two cheaper pFETs are very similar, but the bigger one (IPP45P03P4L11AKSA1) has a much smaller threshold voltage and a much low on resistance.

Would it be worth specifying IPP45P03P4L11AKSA1 for the class next year?  The class-D amplifier could run at a lower voltage if the pFET threshold is not so large, which would make testing easier—lots of students were confused by the behavior they got when they used a modest supply voltage that did not allow the pFET to turn on.

The gate-to-source charge plus gate-to-drain charge for the IPP45P03P4L11AKSA1 pFET is 16–24nC (that’s about how much charge is needed on the gate to finish turning on the transistor), while for NTD2955-1G it is only about 11nC, so IPP45P03P4L11AKSA1 will take longer to turn on (and to turn off, though that depends on the total gate charge).  At 10V, the total gate charge is 42–55nC vs 15–30nC for the NTD2955, so the turn off time is probably also about doubled.

The doubling in time will not matter much in the H-bridge circuit we used this year, since we were using 74HC04 inverters to drive the gates, with one inverter dedicated to each gate—the gates were being driven very fast, and a doubling in time would not affect performance.  It would have been a huge problem when we were trying to drive the gates directly from the output of the comparators.

Although I like the IPP45P03P4L11AKSA1 better for computer-controlled circuits, for the class-D amplifier it probably doesn’t offer enough of an advantage to justify the extra 54¢ per student. It might be worth having a difference in packaging between the nFET and the pFET, though, as several students swapped pFETs and nFETS in their breadboards this year. Because the IPP45P03P4L11AKSA1 comes in a larger TO220 package, the visual difference may be enough to justify the increased price.