# Looking for a transistor to power a load with PWM from a Bluetooth module

My project is a heating load power with a PWM from a Bluetooth module (BLE113). I managed to do the PWM I want with this module and now I would like to connect the load. The module give a PWM with 3.3V and 20mA or 4mA (I can choose pin with 20mA or 4mA) and the load is connected to a 3.7V battery, Rl=9O Ohm so the current needed is ~400mA. The output current of the module isn't enough so my research made me think of a transistor to switch the power of the load. I think it's the best solution. A good help from electronicsclubs.info give me some specs for the transistor :

• if I use the 4mA output : I need a h_FE > 5*Ic/Ib_max = 5*0.4/0.004 = 500
I also need a load for the module output of Rb = 3.3*h_FE/(5*Ic) = 825 Ohm

• if I use the 20mA output : I need a h_FE > 100
I also need a Rb = 165 Ohm

I'd like to know if I'm right for this, and which transistor could give me this specs. I looked in the web and I'm a little lost.

I also have some concerns about the transistor tension, because if it's too high, I won't have the 3.7V for the load.

You are right that you need a transistor to switch lots of current from a GPIO pin.

It is often better to put the transistor at the negative side of your load. It means that the GPIO pin does not have to see the 3.7 volts, or have to go that high to turn your load on and off.

It looks like your research was around BJTs, (bipolar junciton transistors, i.e. PNP, NPN). If you are switching large currents with these you sometiems use 2 of them to get enough hFE (this is called a Darlington pair), although an hFE of a few hundred is very common.

Have you looked at FETs (field effect transitors)? These have the advantage of not having a diode drop across them, so they don't get so hot and you don't lose so much power in the switch.

There are loads to choose from. E.g. FDN335N from fairchild.

• Ok thanks, my research was BJTs. So I had a look at FETs. If I understand, I turn on the MOSFET with the tension when the PWM is ON, and after that the MOSFET leave the current I need (so the .4A) going through the load. I don't care about the current in the gate like the BJTs ? Do I use like this the transistor in linear mode ? Apr 2, 2014 at 13:30
• Yeah you can pretty much think of it as a switch. Put some charge into the gate and it comes on, letting current through your load and to ground. Then to turn it back off connect the gate to ground.
– Will
Apr 2, 2014 at 13:40
• Ok so if I go with a componant like the FDN335N, do I need to put a resistance between the output and the gate ? Because with Mosfet, I don't see how the current in the gate is involved. Apr 2, 2014 at 16:00
• You put some charge into the gate to create an electric field (which is what turns it on) , and after that no current flows into the gate (apart from a bit of leakage etc etc). So in theory you don't need any resistor. In practice you can stick one in if you like (e.g. 1K), to make it a bit harder to accidentally short something out.
– Will
Apr 2, 2014 at 16:06

I think that N-channel MOSFET is a good choice here but you have to take care for Threshold voltage (Vgs). This is minimum gate-source voltage necessary to open transistor. In your case you may search for MOSFETs with Vgs=2.5V or lower.

• Do you have a standard/favorite one you can recommend? I've generally used ones the relevant client uses in other projects etc before, and none in particular spring to mind.
– Will
Apr 2, 2014 at 11:27
• No special preferences but i used IRLML2030 this or this RTR020N05 one for similar PWM design. As I see from the spec's FDN335N suggested by you should also work good Apr 2, 2014 at 11:52
• The FDN335N is just the first one that came up in a search on Farnell from a brand I've heard of.
– Will
Apr 2, 2014 at 12:28