MOSFETs dominate the transistor world these days, and I understand the practical reasons for this. Nonetheless, we don’t want to marginalize BJTs. I wonder if bipolar junction transistors are sometimes overlooked not because they are actually the inferior part for a given application but simply because we’ve become so accustomed to using MOSFETs.
I suppose this is why a recent product announcement from Diodes Incorporated caught my eye. They’ve released two new BJT families. The DXTN07 series consists of four NPN devices, and the DXTP07 series consists of four PNP devices. These transistors are intended for load-switching and power-supply circuitry. I find this interesting because it seems to me that, for these types of applications, development efforts are focused on MOSFETs.
BJT vs. MOSFET?
I’ve discussed this issue before, as has Lonne Mays when discussing which power semiconductor to choose for power stage design. Today, however, I’m going to take a new approach: I will make no attempt to compare these two types of transistors or make any recommendations regarding which is more appropriate for a given application. I’m simply not familiar with the latest techniques used in the design and manufacture of discrete transistors, and I’m concerned that the traditional approach to comparing field-effect devices and bipolar devices is not as relevant as it used to be.
The only advice that I will offer is the following: Don’t use field-effect transistors just because FETs are more common, or because the previous version of the design used a FET. If you really want to optimize performance, I think that it’s often worthwhile to consider and even evaluate a design that is based around a newly released BJT.
The DXTN07 and DXTP07 Devices
I would place these transistors in the medium-to-high-power category. They tolerate collector-emitter voltages ranging from 25 V to 100 V, and they have a continuous-collector-current rating of either 2 A or 3 A. Keep in mind, though, that many switching applications involve pulsed load current instead of continuous load current. Thus, you need to take a closer look at the specs to determine if these parts are a good choice for your system.
The plot shown below gives you an example of how maximum collector current varies according to the duration of the pulse.
Plot taken from the DXTN07025BFG datasheet.
One spec that I noticed is the collector-emitter saturation voltage, particularly for part number DXTN07045DFG. This device has a typical VCE(SAT) of 140 mV with a collector current of 1 A. This seems pretty low, and low saturation voltage is a good thing because it means less power dissipation for a given load current.
As shown in the following plot, the saturation voltage is affected by collector current and the ratio of collector current to base current; it can be significantly lower than 140 mV.
Plot taken from the DXTN07045DFG datasheet.
We can put this number in perspective by comparing the power dissipation of this transistor to that of a MOSFET under similar operating conditions. With a collector current of 1 A and a collector-emitter voltage of 140 mV, we have 140 mW of power dissipation. To achieve the same power dissipation with a drain current of 1 A, a FET would need to have on-state resistance of 140 mΩ.
I rarely miss an opportunity to emphasize the importance of incorporating thermal analysis into voltage regulators, motor drivers, and other circuits that involve a nontrivial amount of power dissipation. Looking at maximum-current specs is not enough; current leads to power dissipation, and power dissipation leads to heat, and heat leads to temperatures that can impair or even compromise the functionality of a circuit.
The DXTN07 and DXTP07 devices are housed in an interesting package (shown below). Diodes Incorporated calls it the PowerDI3333. They describe it as more thermally efficient than an SOT223, and it requires 70% less board space. This is a major consideration, because package size and thermal performance are typically in conflict—a smaller package makes it more difficult for heat to move from the chip into the PCB and the surrounding environment.
Image courtesy of Diodes Inc.
Do you have any thoughts on the MOSFET vs. BJT issue? Do you think that a newly released BJT might replace a MOSFET in one of your upcoming designs? Let us know in the comments section below.