Saturation? Or am I all wet?

Okay… so I’m sitting down trying to sort out what seemed to be simple… and of course it’s seeming not to be. I’m pretty sure I’m missing some critical piece of transistor theory… so here goes.

I’m powering LED’s.

I’m using LM317 as a current source, feeding three power LED’s in series (white, 3.2-3.4v @ 300mA rated). The current sources are fed from a switching power supply set for 13v, giving 2v for the regulator and another 1 volt for the transistor. As I have them configured, I am using 4 ohms as a setting resistor, which makes the output 313mA. The regulator output is then fed to the LED’s in series, then to the collector of an NPN 2N5550 transistor. I’m using the transistor to switch to ground, for PWM. I have a base resistor of 235 ohms, which should provide 21mA to the base, which should be MORE than enough to push the transistor into saturation.

So, when the LED’s are connected with an ammeter in series with them, things get confusing.

The first reading (250mA) is understandable due to resistance tolerances and such- that’s current when measured directly to ground. I’d be happy feeding the LED’s 250mA, it’s not going to be noticable in terms of brightness in most cases. However, what I find is that if I connect the base resistor to 5v (or logic high) about 18mA flows (okay) into the base, but the measured current is only 150mA. I’m “losing” 100mA to the transistor, and for the life of me it’s not making any sense to me.

The transistor’s HFE is supposed to be a minimum of 20. What’s missing?


"Am I all wet?" Welll, it has been raining a lot here ;)

Seems over controlled - you have the LM317 as current source, and also attemtping to use NPN 2N5550 for current control. 300mA isn't even on the charts for this part. I'd say looks like the wrong part for this job.

Right, but I'm using the transistor as a switch.. and it's rated for 600mA Ic - though the datasheets don't show the gain at that level. I just assumed the 20 carried forward from the highest shown level. I'm only using the transistor in terms of current control as a switch- PWM - so if the part is rated to be able to handle 600mA, but they only give gain ratings up to 50mA.. what do you do?

If the device is rated for 600mA continuous collector current, I would think that 250mA would require a reasonable base current-- with 20mA going in and 150 coming out, that's only a gain of 7.5.. That just doesn't seem correct, like I'm missing something. 2n3904 is going to be too small, I have pn2222...

Ok, then drive it on hard and let it act like one. Instead of 235 ohm base, go down to 150. With Vbe of 0.5, 5V from arduino (ideally), 35mA=> (5 - 0.5)/.035 = 128. So try 150. But we also know Arduino Vout drops with current, so may (4.2 - 0.5)/.035 = 105, so maybe try 120. Keep a series meter in, see what you are getting. Can also hook up darlington style, drive first and let it drive higher current into the 2nd, using ULN2803 style connections.

my best bet may be to slap another 470 ohm in there, that'll make it 156 ohms. The circuit board is getting pretty ugly hehe.. As usual, I cobble from what I have in the bins. I got 10ohm 1/2 watters (200 for $3 on clearance) that I use for the set resistance, and the 470's came free as "12v LED resistors" with a pack of LED's. You can get really funky resistances doing a bit of series and parallel resistor setups...

There's eight channels, identical. This was the "eyes" of the halloween rig, and for that purpose, 150mA going to the LED's was plenty. Feeding them 250mA provides MUCH more light.. and now I'm going back and righting the wrongs..