Unidirectional motor control

I am interested in controlling the speed of multiple small DC motors that will draw at most 1 amp each. I only need to run them in one direction, so an H-bridge IC like the SN54410 seems like the wrong choice. However, for my application, weight is a premium, and the SN54410 is the lightest method (per motor) for controlling motor speed I have found.

When I have looked into using individual transistors, they all seem to require a heat sink (either an external one or a metal back), which makes them heavier than the SN54410. In fact, I don't understand how the SN54410 is able to drive two motors at 1 amp each without a heat sink (but it does). Can someone suggest a good component for the job? I am hoping for either an IC that is designed for unidirectional motor control or a small (TO-92) transistor that could be used to control an individual motor at 1 amp.

EDIT: BTW, I am going to be using PWM to control the speed.

• Infineon sells HITFET smart switches, very low rdson, overvoltage, overcurrent, overtemperature protection. – Marko Buršič Feb 6 '16 at 9:44
• Are you assuming linear control, which will definitely need a heatsink at 1A, or PWM, where you won't if you're careful, whether you use 544410 or discrete transistors. Surely PWM, if you want your precious battery weight to give you a decent endurance. – Neil_UK Feb 6 '16 at 10:06
• I am going to be using PWM. – Ari Herman Feb 6 '16 at 17:12

You need to calculate the worst-case on voltage drop for your transistors and do a power calculation. For a BJT it will be the collector-emitter saturation voltage. For a MOSFET it will be determined by the on-resistance.

With this information the power is calculated by $P = V \cdot I$ or $P = \frac {V^2}{R}$. Compare this with the device ratings and calculate the temperature rise.

Given that weight is a premium constraint it sounds like a flying device. Given that you're using multiple motors I'd guess it's a quad-coptor. In that case you may be able to mount the switching transistors under the propellers and get some free forced cooling. A small TO-220 finned thermalloy heatsinks, for example, are very light and with forced cooling would dissipate a lot more heat than in a static environment.

Figure 1. Thermalloy TO-220 heatsink.

• It's funny you say that: I just thought of putting the switches below the propellers, too. I don't know if that will be enough to do it, but it's certainly easy enough to test! – Ari Herman Feb 6 '16 at 17:12
• Also, the TO-220 would add an unacceptable amount of weight. As you guessed, this is a quadcopter, and the motors/props can lift a total of about 50g. I have weighed out all of components, and I am at the point where every gram counts. – Ari Herman Feb 6 '16 at 17:27

Actually, it is spec'ed using a heat sink of sorts. See the data sheet, Section 12.2, Layout Example. The ground connection is specified to be a certain size, and all those vias are there for thermal conductivity rather than electrical function. So the ground connections do double duty as heatsinks.

• Good point. Do you know if the ground of a transistor or IC can be used for heat dissipation in general? – Ari Herman Feb 6 '16 at 17:29
• @AriHerman - It depends on the chip. In general, to use leads as heat sinks requires specific intent on the part of the chip designers to provide adequate heat flow. Some semiconductor processes do not place the substrate at ground. – WhatRoughBeast Feb 6 '16 at 17:31