# Using a MOSFET as a switch

I would like to use the smallest component to control the power of a component (nano motor). I have seen on the web two main possibilities:

• DC converter : LTC3127 for example, I change the value of the resistor and I can get the 3.7 power supply I am looking for.

• The MOSFET : This technology looks really small and I am interested to make a simple montage but I don't know which component should I use.

Information:
* Power In = 3.7 to 4.2
* Power out = 3.7 to 6V
* Switch on = fast and last only 10 seconds
* Power = Extremely low power component please
* Motor Intensity = 160 mA pic to pic

If you could advise me, it would be very nice.

• I don't need 6V for power out but this value is the max value. Jul 16, 2012 at 15:19
• pic to pic... 160mA pk-pk? is that 0 to pk? or DC average mA with peak to peak mA spike on commutation. Please specify Peak current and Average. That motor must get "mega hot if ESR losses in nano motor are high. Jul 16, 2012 at 16:49
• It's 160 peak to peak and about 80 average. Jul 17, 2012 at 2:21

The MOSFET is like you say just a switch, and as such will switch on the 4.2 V, but on its own will never make 6 V of it. You'd get 4.2 V with the FET on, 0 V with the FET off.

The DC converter (called boost converter if the output voltage is higher than the input) has an oscillator which also switches a FET on and off, at a high frequency, and the current through an inductor will generate the higher output voltage. Very briefly said.

The question is whether you need the DC converter. You'll need it to get 6 V, but apparently the motor can work over the full range of the input voltage. And then you can do with a FET as switch:

FETs are low power: they have virtually zero input current, and can have an on resistance as low as a few 10s of mΩ, at 10 mΩ even at 1 A gives only a 10 mW dissipation.

The FDC855N can be controlled by a logic level, and has an on-resistance of maximum 36 mΩ at 4.5 V control voltage.

If you would rather have a PTH FET then this one is a low cost solution. At 160 mA it will dissipate 0.3 mW! There are literally thousands of FETs that will suit your needs.

• Thank you for your answer. I don't need 6V (maximum torque) to control my motor (I am only planning to use a range between 3.7 and 4.2). I think the FETs are the best solution according to your answer. Could you please advise me wich kind of FET would be more appropriate. Thank you very much. Jul 16, 2012 at 15:15
• @Mattew - Won't the FDC855N I mention in my answer do? 160 mA is not much, there are lots of FETs which can handle that. Like I said you want a logic level FET, and when I look on Digikey for those and which can also handle 160 mA it returns a list of more than 10 000. When I restrict to maximum 40 m$\Omega$ on-resistance there's still 8000 of them. (that gives you 1 mW dissipation) Jul 16, 2012 at 15:27
• The FDC855N isn't too much ? (MOSFET 30V, 6.1A) Because of this large amount of choice I am a little bit lost. Jul 16, 2012 at 15:37
• @Mattew - no the 6 A isn't too much. Nobody says you have to use its full power. But FETs like this, who are made for high currents, are also the ones with low on-resistance. At that high 6 A and a 10 V gate voltage the FDC855N still dissipates only less than 1 W, and therefore the package can remain small, and therefore cheap. If you would like a through-hole package instead of SMD, then this one is the cheapest I found on Digikey. Yes, 55 A is overkill, but don't worry about it. Jul 16, 2012 at 15:59
• The money is not an issue as long as the system is really small. Thank you for your answer. So if I understand, even if they say Vdss=30V it doesn't mean I need to power it with 30V, I can just use 3.7V. Thank you very much. Jul 16, 2012 at 16:27

There are millions of solid state switch choices.

THe best choice may depend on many application factors; Ones that matter here are;

1. size must be small
2. Offset voltage needs to be low (easy for FET switches and reed relays)
3. Switch on resistance only needs to be relatively low compared to motor ESR.
4. Switch capacitance needs to be considered if there is a charge transfer condition with high impedance load.
5. Stray electric fields of wiring may alter behavior of a nanoswitch, so shielding is essential or ground plane without excessive capacitance on the switch with charge injection needs to be understood.
6. Low ESR helps with the voltage source before the switch and path to load
7. Nanomotor's may in fact run faster than SEM capture rates of 30 or 60 Hz, so RPM is not easy to measure.
8. Better way is to use nano-current sensing resistor with high gain instrumentation amp to detect and filter spikes of conducted current as opposed to radiated fields and thus sense commutation nano currents and RPM.

The key factor in your selection ought to be switch ON resistance with 160mA pk-pk current which may or may not be bipolar (unstated) If you want to consider switch losses I^2R=V^2/R less than 5 or 10% power then you need to spec the ESR of the switch. as a rule of thumb, "big" MOSFETs get < 1mΩ ESR, small Mosfets ~ 1 Ω, FETs > 10Ω and IGBT << 1mΩ.

Consider you have V+ = 4V and motor spikes ∂I= 160mA pk-pk, look for MOSFETS with logic level drive @ 3mΩ or so in small packages. This will still drop 480mV and you may consider < 1mΩ if this is needed with a bigger part.

One possible solution of many; IRFH6200PBF 1.2 mOhm @ 50A, 4.5V specsheet Stock on hand currently >1K in stock @ \$2.43 and ships same day.

• Your solution is really nice. (Low Rds) My only problem is the size of the component, I don't really know how to solder it (I would prefer a surface mounting). Jul 16, 2012 at 16:47
• @Mattew - You mean not surface mount, because that's what this is. I just added the PTH (Pin Through Hole) part I mentioned to my answer, with a picture of the package. Jul 16, 2012 at 16:49
• The comment is really complete and complex, thank you Jul 16, 2012 at 16:49
• So if it's a CMS it's alright, I have never seen this shape before! Jul 16, 2012 at 16:50
• @Mattew - price may not be an issue, but note that Tony's "nice" FET costs 2.4 dollar, while mine is just 40 cents, and is solderable! The higher price is for the extremely low on-resistance, which you absolutely don't need! Jul 16, 2012 at 16:52