Alternate Current S...
 
Notifications
Clear all

Alternate Current Source

8 Posts
6 Users
0 Reactions
3,245 Views
Posts: 6728
Topic starter
(@p-short)
Illustrious Member
Joined: 20 years ago
[#16378]

As an alternative approach to using a CL2 or LM317, it's possible in new designs to use a darlington transistor to limit the current. This works best with new designs, but the basic idea is to use an NPN darlington-pair transistor, such as the MPSA29 with a 140Ω emitter resistor (to ground) and a 1KΩ (or so) base resistor to the micro output. The base-emitter voltage will end up around 1.4V, leaving 3.5V (or so) across the 140 Ω resistor, thereby limiting the current to 25mA. Since a transistor is needed to buffer the output of the micro anyway, it uses fewer parts than the other approaches. It's less accurate, but I don't think that the extra accuracy is really needed in this application.

I think that a similar approach was discussed some time ago, but not much came of it then.


7 Replies
budude
Posts: 10400
(@budude)
Illustrious Member
Joined: 18 years ago

Thanks Phil - perhaps DIYC Flood v2 can incorporate this...


Reply
Mactayl
Posts: 2491
Guest
(@mactayl)
Member
Joined: 17 years ago

I was trying something like that awhile back with with an MPSA14 and I believe I had biased the transistor wrong so I gave up on that but I do see your point with the MPSA29...will have to give it a try.


Reply
Matt_Edwards
Posts: 2664
(@matt-edwards)
Famed Member
Joined: 18 years ago

I played with some PNP sources for a short while, until work got in the way. My goal was 20 mA SMD and small foot print. See the thread here. I think your comment is spot on about the accuracy.

NPN works very well and replaces the need for an additional switching Transistor.
What are your thoughts for Flood light where you need multiple parallel LED string?


Reply
Posts: 2168
(@djulien)
Noble Member
Joined: 19 years ago
P. Short wrote:It's less accurate, but I don't think that the extra accuracy is really needed in this application.

What is the approximate accuracy? That info would help to know how much padding to use when calculating the expected current. I noticed that even the CL2's said +/- 10%, so I was interested in a relative comparison to that.

don


Reply
Posts: 6728
Topic starter
(@p-short)
Illustrious Member
Joined: 20 years ago

That 10% number surprises me, I expected it to be a lot more accurate.

The output current is approximately (Vhi - VBE)/Re, where Vhi is the high-level output voltage from the Micro, VBE is the base-emitter voltage of the transistor (around 1.4V), and Re is the emitter resistor. And Vhi depends on the power supply voltage for the microprocessor... The Vhi should be controllable to perhaps +/-5%, the resistor to +/- 1%, and VBE to perhaps +/- 0.2V (or +/- 7%), so I would figure that overall it should be +/- 10% (absolute). The matching from one output to the next should be a lot better than that, although that probably won't matter much.

If a finer tolerance is needed, there are additional techniques that could be employed, but I don't think that would really be needed.


Reply
Posts: 2168
(@djulien)
Noble Member
Joined: 19 years ago
djulien wrote:I noticed that even the CL2's said +/- 10%, so I was interested in a relative comparison to that.

I should have included a reference to that ... I was looking at the CL2 link from the DIYCFlood page on the Wiki:

http://www.supertex.com/pdf/datasheets/CL2.pdf

don


Reply
ErnieHorning
Posts: 4434
(@erniehorning)
Famed Member
Joined: 19 years ago
P. Short wrote:That 10% number surprises me, I expected it to be a lot more accurate.

From the charts, it looks like the 10% is over the entire voltage and temperature range. In the range that it would typically be used, it looks pretty steady.


Reply
Share: