Transistors in parallel to increase amps?

9 posts / 0 new
Last post
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi,

I'm creating a 8x8x8 LED cube.
Each 8x8 layer of the cube, will have a transistor to switch the cathodes for the 64 leds of each layer.

64 leds at 19mA adds up to 1216mA.
The highest current transistor I have enough of is a 2n2222a.
It can switch 800mA.

By how much can i overshoot the max value when the transistor is running at a 1/8 duty cycle?

Can i connect two transistors in parallel to get a total of 1600mA?

It's a sunday, the store is closed, so suggestions to go and buy something are not helpfull at this point :p

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

FETs are better for using in parallel - but like you say it's Sunday :)

Transistors don't match that well when in parallel - one of them always gets a greater amount of current - a small resistor in each collector will help - the value depends on the voltage headroom.

-=mike=-

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You could split each 8x8 layer in two 4x8 layer.
Use one transistor for each 4x8 layer.

Regards
Sebastian

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes, you can put transistors in parallel. Connect the collectors together, the bases together, and put a resistance in the emitter line of each transistor (not just one resistor for the pair) that will drop 0.1 to 0.2V at the nominal current through the transistor. This will create negative feedback to help balance the current through the transistors.

Mike

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Jokerman wrote:
FETs are better for using in parallel - but like you say it's Sunday :)

Transistors don't match that well when in parallel - one of them always gets a greater amount of current - a small resistor in each collector will help - the value depends on the voltage headroom.

-=mike=-

Hi,

I think you wrote "collector" when you meant to say "emitter".

Mike

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yep, your quite right, ballast resistors in the emitters - faulty memory :)

Confirmed it by looking in Horowitz and Hill :)

Regards,

-=mike=-

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you can afford CMOS transistors (they cost more) usage instead of bipolar, you can connect them in parallel without ballast resistors (I am talking about additional ballast resistors for transistors current alignment, not about LED ballast resistors - those you need anyway).

Because of their positive Rds/temperature dependency (drain-source resistance is going high with increasing temperature), they are self regulating current flow.

Another good thing about CMOS transistors is, that voltage drop across drain-source can be lower (with appropriate type selection) then in bipolar transistor, so less power you will waste with CMOS.

Just make sure you will use "logic level" (with lover gain opening voltage) type if you are connecting them directly to the MCU.

Best regards,
Gintaras

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

chrmoe wrote:

By how much can i overshoot the max value when the transistor is running at a 1/8 duty cycle?

Look in the datasheet, it will usually have a max pulsed current, or something along those lines - and tell you under what conditions you can hit that peak.

Like others say, if you can use FETs instead, you're set. In fact, you probably wouldn't even have to put them in parallel. Find some old motherboards, and take the DPAK FETs off of them, they're all logic-level, and could handle far more current than you're talking about using.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Btw,

Using 2 2n2222a in parallel worked just fine :) solved my transistor troubles.