SparkFun Forums 

Where electronics enthusiasts find answers.

Have questions about a SparkFun product or board? This is the place to be.
OK very simple requirement yet I am struggling to think of an elegant solution with the smallest footprint possible.

I have 4 LEDs in parallel, basic blue with approx 3.2 Vf. I need to switch them on from a 1.6V trigger source.

The trigger can have very little current drawn from it so ideally less than 500nA max.

A standard transistor with 0.3mV Vce was ok and could drive the LEDs with 0.3V loss without too much notice. But the DC gain was then too low for full drive too, so wasn't suitable.

I have then thought of a darlington pair, but the Vce lowest I can find is 0.8V which is far too much drop over the LEDs.

My first thought is to find a darlington pair with a 0.3V Vce if one even exists? Failing that, I am either looking for a solution that is no larger than a SOT-23 due to space limits, or failing that, ignoring space limits I would create a separate board that buffers the 1.6V source to then connect to the buffer instead, so what would be the best thing to do that - perhaps an op-amp unity gain?

If you don't want to draw current from the trigger source, use a FET rather than a bipolar transistor. It'll draw a small pulse of current when the trigger switches, because the gate has a little capacitance, but when it's switched on it draws almost no current at all.

Driven from 1.6V, the device above will have a drain-source resistance of about 0.1 Ohms. So if you're driving the LEDs with, say, 10mA, the voltage across it will be just 1mV.

I never knew that about FETs, thanks for that it sounds ideal and should do the trick. I guess I will do some more detailed reading up on MOSFETs, from a quick Wiki it sounds like BJTs are a bit out-dated and almost replaced by FETs, whereas I previously thought FETs were more aimed at power (1A+) requirements and BJTs were for lesser more small current requirements.


From what power source are you feeding the LEDs?

FETs are probably a good choice for your application, because of your extremely low (Did you really mean nanoamperes?) control current.

That said, where are you getting the 3.2 V and what is controlling its voltage? If you're regulating the voltage to accomodate the LEDs, why not just change the set point to include the voltage drop across the transistor? If you're regulating the current to the LEDs, what's the worry?


I'm integrating into an existing board and hijacking it's analog point and regulated (voltage) power supply - but from experience the analog source doesn't like it when you start taking more than a milliamp or so and the board starts using a lot of power itself and I want to leave room for more additions too it.

I have some FETs coming today and will test them tonight but from reading up yesterday they sound ideal... almost puts BJTs to waste - what is then the reason people would chose a BJT over a FET other than not knowing their benefits as I didn't?
angelsix wrote:I'm integrating into an existing board and hijacking it's analog point and regulated (voltage) power supply
It's not clear that you've answered the question I was trying to ask. I understand that the control signal doesn't have much current to spare (although nanoamperes still seems absurdly low). My question is: Is the 3.2 V you need to power the LEDs (That is, the power that is switched by the transistors and converted to light) coming from a power supply that, by amazing coincidence, happens to be regulated at 3.2 V? If not, at what voltage is it regulated? Also, how many mA can it provide at that voltage?
angelsix wrote:what is then the reason people would chose a BJT over a FET other than not knowing their benefits as I didn't?
I'm sure a bit of web searching would get you a better answer, but in brief:
  1. BJTs are controlled by current and FETs are controlled by voltage.
  2. Analog amplification is usually done with BJTs
  3. Switching and logic are usually done with FETs (although these functions can and are done with BJTs operated in saturation)

In the dark ages (~30 years ago), when I was a manufacturing engineer in a IC plant, we made BJT-based memory chips for very high speed applications. They were larger, more power-hungry, and more expensive than FET memory, but offered speed that FET couldn't provide.

Have Fun,