Which transistor to drive a 12V 1.5W solenoid?

If I've done me sums right a 12V 1.5W solenoid will need 1.5/12 = 0.125A = 125mA to drive it.

Years ago I used BC109s for everything, but the datasheet says a max collector current is 100mA, so that would not be safe.

So what is a good transistor to use in this case? I don't want to risk anything, so a transistor with twice or three times 125mA would make me feel safe.

Of course as soon as I make the post I find something which could work, the aULN2803. According to the data sheet each of the darlington pairs can drive 500mA. 8 in a package is good because I'll be hitting several windchimes with several solenoids.

I suppose I'll have to be careful with the overall power consumption...

Anyway, any other suggestions are welcome. Its years since I've done this sort of stuff!

a

According to the data sheet each of the darlington pairs can drive 500mA. 8 in a package is good because I'll be hitting several windchimes with several solenoids.

Those are "marketing numbers", I wouldn't count on a sustained 500mA drive on an output nor on driving these kinds of currents on multiple outputs at a time. As you say, you will have to be careful with overall power consumption of the ULN2803 else it will be destroyed.

The modern way to drive >100mA currents is with logic-level MOSFET's (e.g., NDP6060L) or higher-current transistors like the 2N2222. Many other part numbers exist.

-- The Ruggeduino: compatible with Arduino UNO, 24V operation, all I/O's fused and protected

NDP6060L or higher-current transistors like the 2N2222

I'll be driving 125mA, so The NDP6060L seems a bit over the top, 48A and the 2N2222 seems more reasonable, 800mA.

The datasheet of the NDP6060L says "Rugged internal source-drain diode can eliminate the need for an external Zener diode transient suppressor". That is NOT the back EMF suppressing diode is it?

I'll be driving 125mA, so The NDP6060L seems a bit over the top, 48A and the 2N2222 seems more reasonable, 800mA.

Marketing numbers again. Most relevant is the on-resistance (25 milliohms at room temperature). It is overkill for this application (MOSFET's with on-resistance up to 1-2 ohms will be fine for this) but I find it easier to just keep a tub of NDP6060L's around for whatever purpose and not have to think about it. Same thing for the 2N2222 -- 800mA is a hard limit and the device may overheat at that current. Still, fine for 125mA.

The datasheet of the NDP6060L says "Rugged internal source-drain diode can eliminate the need for an external Zener diode transient suppressor". That is NOT the back EMF suppressing diode is it?

Meh...it could be, but I wouldn't recommend it. The back-EMF diode is supposed to allow current to recirculate in the solenoid coil when the transistor switches off, thus preventing a high voltage at the transistor collector/drain. Without it, the MOSFET's internal diode will avalanche at its rated voltage and will conduct that turn-off current....but diodes are cheap and I'm happier without unnecessary high voltages anywhere in the circuit. Who knows if the solenoid coil is rated for 60V (the minimum voltage at which the NDP6060L avalanches...probably higher)? If it isn't, then the "rugged internal source-drain diode" does you no good because the solenoid could be damaged.

-- The Gadget Shield: accelerometer, RGB LED, IR transmit/receive, speaker, microphone, light sensor, potentiometer, pushbuttons

2n2222 is probably OK - though there are better switching NPN transistors out there these days, capable of an amp or more continuous in a small package no problem - for instance ZTX851 - these have low Vsat values at significant currents and thus are more efficient - the awesome ZTX851 has a Vsat of 0.1V at 2A (so power dissipation only 0.2W) and handles upto 5A continuous!

The modern way to drive >100mA currents is with logic-level MOSFET's

The problem is that the modern way is surface-mount, and thus there are very few good switching MOSFETs in TO92 packages! TO220 package is overkill for this application really and the old-fashioned NPN transistor/darlington is available in handy DIP arrays for when you want to drive lots of outputs at once (no-one seems to make convenient MOSFET arrays - there are technical reasons for this).

If you are stocking up on general purpose switching devices its worth looking for more capable devices than the '1980's industry standard' choices, the world has moved on since the 2n2222, BC108, IRF610/630

(though the BC109C is an excellent low-noise transistor for audio amplifiers, note)

MarkT: for instance ZTX851 - these have low Vsat values at significant currents and thus are more efficient - the awesome ZTX851 has a Vsat of 0.1V at 2A (so power dissipation only 0.2W) and handles upto 5A continuous!

Thanks, I'll have a look at the ZTX851, which does seem a good halfway point between 800mA and 48A!

If one does a paramater search at Digikey n-channel mosfet > FETs Single > in-stock > thru hole > logic level > yields about 12 pages of results. If sort by price is then selected, one can browse the list to find a part with nice low Rds and good current capability in a workable package and affordable price. TO-92 struggles, the best current rating is 600mA. I like this part http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=NTD4960N-35GOS-ND nice low Rds so it will run very cool with no heatsink (.125A x .008ohm = 1mW), price is pretty cheap (48 cents), easy to work with TO-251 package. Just add a 1K pull down resistor from gate to ground so that it stays off at power up when the ATMega pins are defaulted to inputs until the sketch starts to control them.

That's surface mount only, not through hole.

Having said that D-PAK and D2-PAK devices can be soldered onto stripboard quite easily, not too hard for prototyping.

The lead spacing on the NTD4960N is 0.09", just a shade away from 0.1" for perfboard thru hole, is not difficult to bend the leads on the CASE 369D IPAK (Straight Lead DPAK) to install it.