Current source

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Current source

Post by rosborne » Tue Aug 17, 2004 5:15 am

Does anyone know of a way to reduce or eliminate the beta dependence of an emitter follower current amp? I'm looking for about 250mA. 24VDC source voltage. ~95 ohm load.
Purpose: I'm testing a coil with a variable current source controllable from a computer controlled 0-10VDC output.
I'm doodling and looking for a simpler design than the 3 OP-AMP current source I found in a textbook (Sergio I think).
-Rick

toejam
Posts: 253
Joined: Wed Dec 05, 2001 1:01 am
Location: n.c.
Contact:

Re: Current source

Post by toejam » Tue Aug 17, 2004 11:01 am

you can use an lm317 as an adjustable current source within those specs. You should be able to get how via texas instruments app data.

rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Re: Current source

Post by rosborne » Tue Aug 17, 2004 12:37 pm

toejam,
I don't think that will work. I need to smoothly change the output current from 0-250mA with a control input of 0-10 volts (the source of the 0-10 volts input can handle maybe 30mA max). I think the lm117 device would give me whatever current the load demands while maintaining a constant voltage. <p>I have a constant voltage source of 24 volts and as an output I want 0-250mA of current. Currently (no pun intended) I have a current amplifier (emitter follower) supplying the variable current, but the output is heavily dependent on the beta(hFE) of the transistor and I don't like it. It might add a variable resistor to my circuit and I don't want that iff someone can think of a way out of it for me. :) <p>Does anyone know of a way to amplify current that reduces or eliminates beta dependence?

rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Re: Current source

Post by rosborne » Tue Aug 17, 2004 1:21 pm

I figured it out a different way, but am still interested in other ideas.
-Rick

Engineer1138
Posts: 458
Joined: Thu Feb 05, 2004 1:01 am
Location: Minneapolis, MN
Contact:

Re: Current source

Post by Engineer1138 » Tue Aug 17, 2004 1:46 pm

Just off the top of my head here, but can you use an op amp followed by an NPN transistor, with feedback from a small resistor that goes from emitter to ground? Your load would then be connected between V+ and the collector with the opamp output driving the base. A pot at the amplifier's non inverting input is the setpoint.
The negative feedback should remove any dependence on the transistor's beta. I believe I've done this before, but don't remember exactly why :-)<p>I hope I managed to explain all this without resorting to really bad ASCII diagrams .
If you have Visio I could draw a quick diagram and post it on my website.

dyarker
Posts: 1742
Joined: Fri Aug 22, 2003 1:01 am
Location: Izmir, Turkiye; from Rochester, NY
Contact:

Re: Current source

Post by dyarker » Tue Aug 17, 2004 9:36 pm

Its not dependant on beta because it is dependant on alpha. Sort of common base turned sideways. For a fixed constant current I'd put a resistor emitter to common and a 3.3V to 5.1V zener base to common. Then the emitter current is set with the resistor,
Iemitter = (Vz - ~0.6V) / R
That current is constant regardless of reasonable impedance changes in collector circuit.
Icollector = Iemitter * alpha (alpha doesn't vary as much as beta transistor to transistor, or by temperature.)
(If I remember correctly beta 100 = alpha 0.99 and beta 300 = alpha 0.99666666666666666666666666666667)<p>A simple one transistor current regulator -
collector to DUT
emitter to 37.4 Ohm 1% to common
0V to 10V from PC on base<p>The trouble is it needs a supply voltage of 35V for a DUT up to 95 Ohms. If the 0V to 10V was sent thru an op amp with a gain of 0.5, the emitter resistor could be 18.7 Ohms, and supply voltage requirement reduced to 30V.<p>In your first post you said load ~95 Ohms
0.25A * 95 Ohms = 23.75V
A quarter volt isn't a lot of "room to play in". I think you're going to have to raise the supply voltage whatever you do.<p>Cheers,<p>[ August 18, 2004: Message edited by: Dale Y ]</p>
Dale Y

rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Re: Current source

Post by rosborne » Wed Aug 18, 2004 4:22 am

Engineer, I think that might work, but I'd still like to avoid use of a pot.<p>Dale, Don't forget that alpha is beta dependent.
α = ß/(ß+1)
and the input of the current amplifier is the base the output of which is on the collector or emitter (Ic and Ie are nearly equal because as you pointed out α ~= 1), however Ic = ß*Ib. And to bring it full circle Ie = (ß+1)*Ib which if you
Ic/Ie = ß/(ß+1) = α
which confirms what you said Ic = Ie * α. <p>My solution is a current mirror with four output transisors in parallel. The variable 0-10VDC goes to the base of an NPN transistor that has collector to 24VDC the emitter to the top of the current mirror's set resistor(120ohms). The (95ohm) load is connected between the collector of the output transistors and the 24VDC. The simulation I ran shows a smooth current increase from 0-250mA. Now I will build it, and see what's wrong with it. <p>-Rick

fsdenis
Posts: 90
Joined: Tue Mar 25, 2003 1:01 am
Location: San Manuel, Arizona
Contact:

Re: Current source

Post by fsdenis » Wed Aug 25, 2004 9:28 am

Try this:<p>http://paralleltransistors.com/drawings/ed0-2_fin.jpg<p>[ September 21, 2004: Message edited by: windmiller ]</p>

fsdenis
Posts: 90
Joined: Tue Mar 25, 2003 1:01 am
Location: San Manuel, Arizona
Contact:

Re: Current source

Post by fsdenis » Wed Aug 25, 2004 9:59 am

The LM324 can run on +24v to pin 4 and 0v to pin
11.<p>The "battery +" voltage can be +24v with load
resistance 95 ohms.<p>To avoid a pot on the input VIN, use a 39K and a
1K resistor in series with your computer generated
10V max at top of the 39K, the VIN input
connected at the joint between them and 0v to the
bottom of the 1K.<p>Note that the computer generated 10v will need its
0v connected to the 24v supplie's 0v to relate the
input to the output. Another way of saying this is
connect the two supply grounds together.<p>You won't need to parallel additional stages as
shown in the drawing on the website for this small
current.

fsdenis
Posts: 90
Joined: Tue Mar 25, 2003 1:01 am
Location: San Manuel, Arizona
Contact:

Re: Current source

Post by fsdenis » Wed Aug 25, 2004 10:27 am

Image<p>Sorry about the disjointed mess I'm making here.
I haven't gotten the hang of posting replies yet.<p>[ September 21, 2004: Message edited by: windmiller ]</p>

Engineer1138
Posts: 458
Joined: Thu Feb 05, 2004 1:01 am
Location: Minneapolis, MN
Contact:

Re: Current source

Post by Engineer1138 » Wed Aug 25, 2004 12:17 pm

windmiller: that's exactly the solution I proposed, except mine had only one "channel" of current control and I attached a trimpot to Vin for manual control.
I'd like to see what rosborne ended up using.

fsdenis
Posts: 90
Joined: Tue Mar 25, 2003 1:01 am
Location: San Manuel, Arizona
Contact:

Re: Current source

Post by fsdenis » Thu Aug 26, 2004 6:09 am

Engineer: We think alike, it seems. I took a look
at your website. Might I be right to think that
you are into control of circuits and devices with
the tiny chip computers?
Once I've worked out a few more applications in
power electronics to the 2 horsepower range, I
hope to get into controlling them with these little computers.
Maybe we will be talking again soon.
I too am interested in seeing what Rosborne finally makes work. This class of circuits fascinates me.

Will
Posts: 310
Joined: Tue Feb 11, 2003 1:01 am
Location: Katy Texas
Contact:

Re: Current source

Post by Will » Thu Aug 26, 2004 8:47 am

Windmiller,
Would be grateful if you could tell me how to get those diagrams on to the forum
BB

rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Re: Current source

Post by rosborne » Thu Aug 26, 2004 10:18 am

Me, too.
-Rick

Engineer1138
Posts: 458
Joined: Thu Feb 05, 2004 1:01 am
Location: Minneapolis, MN
Contact:

Re: Current source

Post by Engineer1138 » Thu Aug 26, 2004 12:13 pm

windmiller: the microcontroller stuff is the hobbyist aspect of it. <p>My day job involves developing control software for a medical instrument that uses a 1.5GHz Pentium as its CPU. I can't quite afford the $750,000 for my own instrument, so I make do with $5.00 Atmel microcontroller chips for my home projects ;)

Post Reply

Who is online

Users browsing this forum: Google [Bot] and 10 guests