MOSFET Discussion

=Question:=

On the replacement mosfets you recommended -

http://openfarmtech.org/index.php/CEB_Automation_Prototype_II#Solenoid_Valves

Given the 3A current requirement of the 12V solenoid valve, do we need the larger MOSFET you suggested if the Arduino drivers can handle 5A? Do you think there will be startup spikes? Do you think that the Arduino driver setup is built to handle spikes?

=Response:=

(Fairly pure) inductors (such as solenoids and relays) don't show a current spike at turn on: The inductor resists changes of current. (see http://www.allaboutcircuits.com/vol_1/chpt_16/3.html) The TIP120 has two disadvantages:
 * It has a peak current handling of only 8A, though this isn't a problem for the CEB application
 * If I'm reading the data sheet correctly, it has a voltage drop of 2V at 3A and 4V at 5A. That means that your 12V supply only delivers 10V into a 3A load in series with the TIP120, and the TIP120 needs a heat sink to dissipate the 6W (2V x 3A) that it's wasting.

With the MOSFETs you can do without a heat sink.

=Question:=

Please point me to where you are getting this V_ds. I'm not sure which figure of merit to look at on the TIP120 spec sheet.

=Answer:=

V_ce(sat). For a bipolar transistor the three terminals are labelled differently from those on a FET: Gate -> Base, Drain -> Collector, Source -> Emitter. Bipolar transistors are current mode devices: they provide current gain. FETs are voltage mode: they provide voltage gain. The Darlington Driver is a pair of bipolar transistors chained together to multiply their current gain. Bipolar transistors are not approximately resistive, so they don't have an equivalent of R_ds(on).

If I understand it correctly, the V_ce(sat) is the collector - emitter voltage when the charge carriers available in the device are all in use, which is to say that the device is carrying all the current that it can under the prevailing conditions. This voltage will depend on various things including the saturation current and the I_b (the current flowing into the base), but under the conditions noted on the datasheet, V_ce(sat) is 2V for I_ce = 3A. Up to a point, we could reduce V_ce by operating at less than the I_ce(sat): If we increase I_b above the level required to allow I_ce = 3A (ie: above 12ma), V_ce should fall. I_b(max) = 100mA, but the Arduino cannot provide that much without damage, and I don't think V_ce will ever be as low as the V_ds for the MOSFET. With the MOSFETs you can do without a heat sink.

=Question:=

For the MOSFET, we have R_ds = 0.045, so power loss is 1.1W at max.

=Answer:=

And at 3A dissipation is only 0.135W. =Question=

Why don't they use MOSFETs in the RepRap driver to begin with, if heat sinks are not required? That would save on a lot of fabrication time.

=Answer:=

I'm not sure. I suspect that whoever designed the board was familliar with Darlington Drivers but not MOSFETs, which are a more modern technology. I'm not aware of any advantage that a Darlington Driver might have over a power MOSFET except a slightly improved tolerance for static electricity. Perhaps we'll find out. :-)