seeking: 1.5A+ low noise, high hFE transistor

[riotubes]

for 3.3V discrete regulator circuit. I've built a 5V circuit that uses BC550C transistors but those only are good for 100mA. So looking to keep the noise low and hFE (400+) as good or better but able to run 1.5A+ current. Any ideas?

 

[peranders]

D44H11 is a good type but the Hfe is rather normal. Why do you need so high Hfe?

 

[riotubes]

Thanks for the input Peranders. The circuit is here:

Low voltage discrete regulator

In the 5th paragraph, John recommends going with high hFE and low noise and TO-220 package if one needs higher current. I'm just going off this recommendation. If I read the datasheet for the BC550C correctly, the hFE is in the 420-800 range. As a beginner working with transistors I could benefit from understanding why this might be beneficial.

 

[peranders]

Since you use a darlington connection any good power transistors will do but right now I have idea how fast the teransistor should be in order to get a stable performance but the D44H11 is nice in lot's of respects.

The Hfe is not particulary important. 100 x 100 = Hfe of 10000 this was very ordinary transistors. This will mean 10 uA in base current. Not much if the LED has 1-5 mA current.

BTW: The symbol for the constant current generator is wrong....

 

[AndrewT]

Q2 should remain as a high frequency, high hFE transistor.
i.e. keep bc560c.
Q3 spec could be relaxed substantially.
Look for a 100MHz, hFE>100 medium power transistor.
The combined gain of Q2 & Q3 will be about 40,000 times.
The base current of Q2 @ 500mA output, will be about 10 to 15uA.
If you ask for 1500mA of output then the base current is likely to increase at least ten fold to ~150uA. Still very low, but well within the limits of the control and CCS.

 

[riotubes]

Thanks Guys. I found a Fairchild NPN transistor that is 3A with hFE of ~120. I didn't see any noise specs and it needs to be a low noise type. I think I should up the current to at least 2A cuz this is a replacement psu/ +3.3V reg for a squeezebox which is spec'd at 1.5A.

Any thoughts on whether this one is low noise?

 

[megajocke]

Low noise performance in the pass transistor is pretty irrelevant, noise of Q1, the LED and CCS diode is the only thing that will matter for output noise.

That circuit will have pretty bad temperature coefficient though. Output voltage will decrease about 10mV for every degree C ambient temperature increase.

A LM317 or a similar low dropout regulator if needed will have much better performance.

 

[aphile]

We have had excellent results from farchild KSC1173 (actually better than published specs)

 

[linuxguru]

Sanyo 2sc3807 can handle 2A with a Hfe of 600+ at 1A, and fT of ~260 MHz.

 

[Mooly]

I would prefer to use a regulator using a PNP pass transistor, particularly for high currents. The darlington NPN arrangement wastes a lot of headroom. You always have the 2 vbe volt drop. With the PNP type you are pulling the base to ground, much better for high current use. Just a thought
Regards Karl

 

[martin clark]

Using a sziklai pair is another way to solve both issues - low dropout and high gain.

Edit: Sorry, i think that's what yu are suggesting Mooly.

 

[Mooly]

Multiple posting, usually me that's last one in.
Great minds think alike