# Controlled Current Source

Discussion in 'The Projects Forum' started by AetherNZ, Jul 22, 2008.

1. ### AetherNZ Thread Starter Member

Mar 23, 2008
21
0
I am trying to design a triangle wave generator for a university project.

I am using the attached circuit to generator the wave and a similar setup to above for each current source. Currently by varying R2 of both current sources at the same time the frequency is changed.

The specs state that the generator must have 20-80% duty cycle control. Which would work for the original design if you vary R2 in different proportions but this isnt very elegant.

So I was thinking of the other attached design as a solution. The main current source would be varied the same way as before. Then the current could be proportionally split and the other two current sources would mirror each side. This way the frequency would same the same but the duty cycle would change with only 2 knobs.

Would this work? How could I implement the current-controlled current-sources?

Is there a more elegant way of accomplishing the task?

Thanks

Note: I accidentally didn't hook up /Q to the switch but you get the idea.

File size:
30.4 KB
Views:
75
File size:
44 KB
Views:
79
2. ### John Luciani Active Member

Apr 3, 2007
477
0
Your schematic is a current-sink not a current-source. A better way to do this would
be to use an op-amp. With a control voltage you could vary the current.
There is a schematic of a current sink (a.k.a load cell) at
http://www.luciani.org/geda/util/matrix/index.html

Just scroll down the page and look for the load-cell.

If you want a fixed current I would use a TL431. The TL431 current-sink schematic

(* jcl *)

3. ### AetherNZ Thread Starter Member

Mar 23, 2008
21
0

I forgot to mention that idea behind the project is to get back to fundamentals so the less ICs the better.

Ah I always get the current sources and sinks mixed up haha.

I looked into using an integrator but a lot of the manufacturers seem to use the current-source/sink method.

4. ### hgmjr Moderator

Jan 28, 2005
9,030
214
Are you saying that you are required to use discrete components only or can you use ICs?

hgmjr

5. ### AetherNZ Thread Starter Member

Mar 23, 2008
21
0
ICs are allowed but discrete transistor-type circuits preferred.

Eg where I have used comparators in the conceptual design I would actually just use a transistor differential amp in the detailed design.

Thanks

6. ### hgmjr Moderator

Jan 28, 2005
9,030
214
You may want to look into current-mirror circuits. You can construct them from dual matched transistors that can be found in a single package.

You could use one current mirror composed of dual NPN transistors in the same packake as a current sink and one current mirror composed of dual PNP transistors in the same package as a current source. I stress matched transistors because this is the best way to insure that the transistors track each other over temperature.

hgmjr

7. ### AetherNZ Thread Starter Member

Mar 23, 2008
21
0
Thanks for the reply. Yea I have been looking into current mirrors and have taken your advice and built the current-sink mirror using the NPNs which is working.

Im not sure how to wire the current source mirror since the control side of the mirror will be sinking and the driving side will be sourcing so their bases can't simply be connected(the top current-controlled source in theory diagram)??

8. ### AetherNZ Thread Starter Member

Mar 23, 2008
21
0
This is what I have got so far.

R2 & R3 are just variable test loads.

R4 sets total current.
R1 splits current in each branch.

Am I on the right track? Dont know how to do the other half??

• ###### Current Mirrors.jpg
File size:
169.6 KB
Views:
87
Last edited: Jul 22, 2008
9. ### hgmjr Moderator

Jan 28, 2005
9,030
214
This is along the lines of what I was thinking about.

You will need to tweak it a bit.

hgmjr

File size:
540.5 KB
Views:
64
10. ### Wendy Moderator

Mar 24, 2008
20,772
2,540
Love the drawing Hgmjr...

A straight op amp integrator is a constant current source to the cap.

11. ### SgtWookie Expert

Jul 17, 2007
22,183
1,728
Bill,
Have you actually built that circuit?

I'm trying to simulate it, and without swapping the inverting and noninverting inputs to the opamp, it won't oscillate. With the opamp at unity gain, the triangle wave it puts out really looks dismally bad - more like a sine wave with a lopped-off negative peak.

Originally, I supplied the 555 timer with 10v, and the op amp with +/-10v, but both peaks were truncated.
In order to prevent truncation of the positive peak, I reduced Vcc to the 555 to 5v. It's still truncating the negative peak.

12. ### AetherNZ Thread Starter Member

Mar 23, 2008
21
0

That's the same circuit I came up with but I couldnt think of a way to scale both of the current mirrors so that you could adjust the frequency.