Using MOSFETs with non-microcontroller voltages isn’t always as easy as the Internet says it is.
Back in issue 32, we looked at using MOSFETs specifically with microcontrollers. We covered finding and choosing devices that are designed to be switched on from 5V or even 3.3V, with a bit of the theory behind how they work and why to choose logic-level MOSFETs rather than the more common automotive or general purpose versions when using microcontrollers.
However, not all situations the maker finds themselves in use logic levels, and for some makers, it’s the exception rather than the rule. MOSFETs have some specific advantages as high-powered load controllers, but they aren’t as straightforward as many Internet-published projects portray. While many design oversights may produce a working project, even one that works for some time, incorrect use can still destroy a MOSFET or its control circuitry over time. We have a link at the end to an article from a design engineer whose job is to investigate field failures of consumer goods, and while those design oversights are way beyond anything this article could address, it’s a good case in point.
We’ll revisit the theory of MOSFETs first, and expand it with details that were less relevant or just didn’t fit in the space last time, then work through how to drive general purpose and automotive MOSFETs effectively, safely and efficiently.
WHAT ARE MOSFETS?
MOSFET is an acronym for Metal Oxide Semiconductor Field Effect Transistor. Instead of amplifying current like a bipolar transistor does, FETs and MOSFETs are voltage-controlled devices. In other words, applying the right voltage allows a current to flow. Half the article in Classroom 32, March 2020, was dedicated to explaining what MOSFETs are and how they work so there’s little use covering it twice. If you don’t have a hard or digital copy of the issue, you can find the article on its own. It is titled ‘Making with MOSFETs’. That said, a short summary is still in order, although many makers will know at least a little about them.
MOSFETs are available in N-Channel and P-Channel options, which have similar roles to NPN and PNP types in the bipolar transistor world. N-Channel MOSFETs are low-side switches, where the voltage is applied to the gate and the device is connected between the load and ground. P-Channel MOSFETs are high-side switches, where the gate is connected to ground and the device is connected between the power supply and the load. Both types come in depletion or enhancement modes. Enhancement mode means that applying the voltage to the gate causes current to flow, whether that’s gate-to-source in an N-Channel or drain-to-gate in a P-Channel. Depletion mode involves devices which are effectively ‘normally open’, and applying negative voltages causes current to stop. These are much less common and almost no maker will use them. Their applications are fairly specialised, usually very low noise RF applications.
CHARACTERISTICS OF MOSFETS
A MOSFET is not an out-and-out switch. It still has a response to the amount of voltage applied. Datasheets and usually catalogue listings will quote a figure called VGS(th), which is the Gate-Source Threshold Voltage. The parameter carries the same name for N- and P-Channel devices but polarities reverse. This is the minimum voltage that needs to be applied to the gate for any current to pass across the Drain/Source channel, but the datasheets also give a current figure with the threshold. The voltage needs to be higher if more current is to pass. Anything less than the VGS(th), and the device is said to be in the ‘cut-off’ region.
The VGS(th) is usually given as a minimum and maximum number. This does not mean that the lower number is where the device starts to conduct and the upper number is where the device is fully on. It is actually manufacturing tolerance. So, your device may only begin to conduct any current at the larger number. The IRF540N, for example, quite a common retail MOSFET, has a 33A rated current capacity, but a VGS(th) of 2-4V, for a tiny current of 250μA. That means one device may conduct 250μA at a VGS(th) of 2V, while another may require 4V for the same current.
After this point, the device is in the ‘linear’ or ‘ohmic’ region. With increasing gate voltage, current across the Drain/Source channel increases. It is called the linear region because, by and large, the response is linear. Datasheets include a graph to show this. We’re using the IRF540N again, and the graph goes past the rated device voltage. You can see that for full rated current, the gate voltage is above 5V. The recommendation is to use 10V and above to drive this particular MOSFET.
There is a point where the device is fully turned on. This is called the ‘Saturation’ region, sometimes labelled the ‘active’ region despite the fact the device is quite active in the linear region too. In the saturation region, the device can pass its full rated current.
RESISTANCE AND CAPACITANCE
If only it were that simple. The gate voltage while operating in the linear region affects the resistance between the drain and source. This is where much of the heat generated in a MOSFET comes from. It is better to have the MOSFET fully switched on, rather than choosing a MOSFET with a big current capacity and using a smaller voltage, one within the linear region, to switch it partially on. This was our main point in Issue 32’s discussion of the value of logic-level MOSFETs. Many maker circuits use 30A or above MOSFETs, with a VGS within the linear region applied, in order to switch a few amps. These devices run this way get hotter than need be, and are prone to failure later in the project life. The resistance is the RDS, which is the Resistance, Drain/Source. There are graphs in the data sheet to show this at different VGS figures too.
The gate of a MOSFET behaves a lot like a capacitor. In fact, it literally is one, although not deliberately so. The way a MOSFET gate is constructed and operated means that it stores a charge. Gate capacitance is again found in the datasheets but not as a capacitance. It is given as a charge, labelled as Qg. This is the gate charge, and is in nanoCoulombs. This figure is the total charge needed to saturate the gate, and is made up of a Gate-Drain charge, a Gate-Source charge, and capacitance between drain and source that are all factored in.
This figure is critical, as it determines the maximum switching speed. Datasheets show yet another graph for VGS versus Drain-Source current. However, it is better to take the maximum value and always drive the MOSFET to this. Using the IRF540N, we find a Qg of 71nC. Capacitance in Farads is the total charge in Coulombs divided by the Gate/Source voltage VGS. The point of this Classroom is using MOSFETS on non-logic voltages, so let’s assume that the control circuit is 12V.
7.1 x 10^-10 ÷ 12 = 5.9nF. (See mathematical explanation at bottom of this page)
This means that there is a delay after voltage is applied as the capacitor is filled with this charge to reach Drain/Source saturation.
The short answer is: Switching time, and losses. The faster the charge can be delivered, the faster the device turns on. Considering that many uses of MOSFETS are repetitive switching applications like SwitchMode Power Supplies (SMPS) or motor speed controllers, this factor is important. With a Pulse Width Modulated controller running as a light dimmer, switching speed is less important - the slowest speed from a low drive current is still faster than the rate usually required, but not always. The same may be said for some motor speed controllers. However, other dimmers and PWM drivers use a fast switching speed.
The other challenge is losses. Any time the MOSFET spends in the linear region has much more loss due to resistance than in the saturated region. This wastes power generates a lot of heat, and reduces device life over a certain value. There are further losses involved in running a MOSFET in its linear region but they don’t particularly add anything to the discussion (though they do add to the losses). The upshot of this is that the gate needs to be driven with as much current as practical, to minimise switching time. That’s an issue for many logic circuits, which are usually high-impedance with minimal current. The graph is exaggerated over the time scale for visual clarity.
Further to this, when the gate voltage is removed, that charge has to go somewhere. Remember, the Gate/Source junction is a capacitor, not a current path as the Base/Emitter junction is in a bipolar transistor. This leads to a current wanting to dissipate when voltage is removed from the gate. In some circuits this is not an issue, as the outputs of many logic circuits switch between source and sink modes. In high-resistance drive circuits, however, this will mean the gate is slow to turn off. In sensitive circuits, it may cause damage.
THE ANSWER TO DRIVING MOSFETS
The solution to these problems comes in the form of an integrated circuit called a MOSFET Driver. These devices are specifically designed to use regular control, or logic circuit signals, which usually have low currents, and use them to switch a high current to the MOSFET. Many driver ICs also work as level shifters, meaning the MOSFET can be in a circuit running on a higher voltage than the control circuit. They are purpose-made to switch a MOSFET on and off quickly, and provide isolation between the MOSFET and the driving circuit. The current supplied is limited by the supply current. They have no internal supercapacitors or the like, so a 6A rated device needs a 6A power supply to deliver on its promise.
It is worth noting here that MOSFETS have two main voltage ratings. VGS(max) is the maximum value that the Gate/Source voltage can be, and is usually much lower than the maximum voltage rating of the device, the Drain/Source voltage or VDS. Back with the familiar IRF540N, the VGS is rated to a maximum of 20V, while the VDS is 100V. This means you can already use them to switch a higher voltage than the control circuit, within the limits of VGS.
Using a MOSFET Driver is fairly straightforward. The IC itself has Vcc and Ground connections, but besides that, the gate of the MOSFET connects to the driver’s output, and the signal from the control circuit that used to go to the gate, goes to the driver’s input. The same applies to half-bridge and H-bridge designs, too.
There is a bewildering array of MOSFET drivers out there. Some are simple, some complex. Some are single devices, others are more than one in a single package. Some are meant for P-Channel MOSFETs while others are meant for N-Channel MOSFETs. Some are meant to drive both and some are made to drive two of each in a H-bridge configuration. Many are surface-mount, however. The catch is, few if any are available at a retail over-the-counter level.
WHY NOT DIY?
You could design your own, and these would likely take the form of smaller MOSFETs with smaller gate capacitances driving the current to the bigger transistor. However, if you’re buying another device, it’s a matter of ‘same difference’ in many cases to just buy a driver. Many contain MOSFETs as their active elements and all are designed for switching speed. They’re also designed to isolate the input from the output. Add to that compactness and component count, and it’s really a no-brainer.
To buy a MOSFET driver IC, you will end up at online trade-based suppliers. There are several around, and some have a better reputation than others. DIYODE deals with Element 14, formerly Farnell. This website is huge, and searching for ‘MOSFET Driver’ returned 1,167 results. Narrowing it down to a DIP case using the search filters dropped the options to 147. We’ve chosen a few to examine but we’ll list the parameters you need to look for first.
When choosing a MOSFET driver, the data you’ll need to match is:
GATE VOLTAGE: Can it drive your MOSFET’s gate high enough?
OUTPUT CURRENT: This will be high enough on most drivers unless they’re meant for small MOSFETs. Double check anyway.
INPUT VOLTAGE: What voltage is your driving circuit? Some drivers have Zener clamp or other input limits.
OPERATING VOLTAGE: Some drivers have a different operating/supply voltage compared to the input voltage. It’s the input voltage usually listed in the device summary in the catalogue listing.
ORIENTATION: Do you want a low-side driver for N-Channel, high-side for P-Channel, or an arrangement of both such as a push-pull driver (half-bridge), or a H-bridge?
ADDITIONAL FEATURES: Some ICs have additional features or functions that will appeal to specific situations.
INVERTING: Whether or not the driver IC inverts the input signal.
SWITCHING TIMES: Driver ICs have their own rise and fall times, and propagation delays. This is an issue for very high speed switching but is generally outside the needs of most makers.
If that’s a little too overwhelming, we have chosen some for you. Having said that, many of the MOSFETs on the domestic market are from International Rectifier via Infineon Technologies. Their website features a gate driver finder function on the product page for each MOSFET.
We chose to buy our MOSFET Drivers from Element 14. Element 14 are a trade and commercial supplier who also sell to the public. While there are several companies online who sell a huge catalogue of electronic components and supplies, we like Element 14 because they operate a genuinely Australian warehouse and office, with Australian jobs and economic contributions attached.
While not all stock is housed in Australia because demand may be too low, Element 14 is very open about this, with expected shipping times displayed along with held quantities in each product’s line on the website. An item in Australia will show ‘n-quanity available for next-business-day delivery’, while items held overseas will show ‘n-quantity available for x-y business days delivery.’ The range of MOSFET drivers is mostly held overseas, but this is one of the few times we’ve found few local options.
We’ve purchased quite a bit from Element 14 recently, none of it subsidised or promoted, nor have we been paid to write any of this. Some of these items will be seen in the coming months, while others are supplies, such as components we can’t get at over-the-counter retailers like Jaycar, Altronics, and Core Electronics (we consider Core in the same category as Jaycar and Altronics based on the market they cater for).
So far, every time we’ve ordered something that says ‘next business day delivery’, we have received it the next day, even when the order was placed in the late afternoon. The one excePtion was a courier error and was fixed fast. All of this is more than we can say about our dealings with some of the other online suppliers who operate an ‘Australian’ website but have no actual operations here.
Prices published are correct at time of writing, and include GST. Element 14's website has both ex- and inc-GST pricing displayed at the same time, so make sure you're looking at the right line. Also check thier website for current free-freight minimum spends, but at time of writing, Orders $50AU and above were free, with a $15AU freight charge for orders less than this.
All schematics provided are examples only, usually from the datasheets but reformatted or altered slightly. There are usually many ways to use each IC, and specific applications may require alternative circuits.
LOW SIDE DRIVER
Microchip MCP1406 and MCP1407
These two drivers are complimentary: One is inverting (MCP1406), the other not (MCP1407). The inputs are rated from less than 2V to a maximum of the IC’s Vcc, which itself has a maximum value of 18V. They will supply 6A to the MOSFET gate. They can charge a gate capacitance of 2500pF in 20nS. They come in an 8-pin package, with two pins for each of ground, Vcc, and Output, with one input pin and one NC. The schematic is for both, as the only difference is the fact that one has inverted operation.
This is one for makers building with motors. It has four channels and is designed for use as a H-bridge. The inputs are a special AND gate; one AND input is inverted, while the other is not. You can tie one input high or low as needed, and use the other for your control signal. In doing so, control signals can be high or low going, or can be made to invert the output. We’ve only provided a schematic for one connection option, but there are many with this IC. The device has 14 pins and comes in through-hole DIP and SMD packages.
As a half-bridge option, we chose an offering from the same manufacturer of many of the popular MOSFETs. The IRS21864 can drive gate voltages of between 10 and 20V, at up to 4A, with control input voltages of 3.3V up to Vcc. Switching times are a little slower than some, at 170ns, but that’s plenty fast enough for many applications, particularly given that most maker uses for half-bridges are going to be power switching or motor control. In addition, the high-side section can be configured to float if needed, but this isn’t used in standard operation. This is another 14-pin device.
HIGH SIDE DRIVER
There are far fewer options for high-side drivers, partly because there is a much smaller demand given that most uses for high-side switching are in bridges. One of the options is Texas Instrument’s TSP2812P. There are two identical drivers in one 8-pin package, and both will dump 2A into capacitive loads, but continuous currents are limited to 100mA. That’s ok, because the peak current will do the job of turning on the MOSFET quickly, and 100mA is enough to keep it there. The device even has an onboard regulator, giving a stable 11.5V typical output for input voltages on 14V and above, at 20mA current. This can run the IC with input voltages of up to 40V.
- MCP1406/7 Datasheet
- TC4469CPD Datasheet
- IRS21864PBF Datasheet
- TSP2812P Datasheet
Of course there are dozens of other options and if you know the specifications of what you want, you can go looking. However, if you want a general purpose low or high side, h-bridge, or half-bridge MOSFET driver, without having to search through well over one thousand products, then those above will be an easy choice. As always, check the data sheets and have a good look at the tables, diagrams, and graphs within. While we normally choose an IC and unpack it in detail in the Classroom, we just can’t do that for so many ICs this time. Hopefully if you’ve read a few issues, you’ll be familiar enough with the process to go independently, and many makers are already comfortable with datasheets.
Unfortunately, we don’t have a build per se this month, particularly because there wasn’t enough time between our parts arriving, and deadline. It takes quite a bit of time to fact-check and research articles, so by the time we were ready to build, we had to wait another 5 days for stock arrival. Even so, there was probably not a lot that we could build to a simple enough standard that we haven’t already done.
MOSFET Drivers, in whatever flavour you need them in, will simplify the operation of MOSFET circuits, increase longevity, improve efficiency and reduce heat loss problems. Your control circuitry will be able to utilise low current, low power designs while still providing sufficient current to the MOSFET to achieve saturation, and enable separation of controlling and controlled voltages. You can even use them as level shifters. While we couldn’t build this month, we’re confident you’ll be able to build something that utilises a MOSFET driver, and hopefully this has helped eliminate potential (or existing) problems in MOSFET-based circuits.
EDITOR'S NOTE: There are some inconsistencies between some of the schematics for the driver circuits. The diagrams are from the datasheets, and are from different manufacturers with different design methodologies. It is best practice to have a gate resistor on all gates for reasons deeper than this article, but we didn't modify the manufacturer's schematics, besides adding the MCU.
Further ReadingAn interesting article from a design engineer regarding field failure of even seemingly well-tested designs:
We've added the following hand drawn formula to assist with calculating the answer to our capacitance formula above.