Current source

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
erkanyigiter
Posts: 29
Joined: Wed Sep 08, 2004 1:01 am
Location: ISTANBUL, TR
Contact:

Re: Current source

Post by erkanyigiter » Wed Oct 06, 2004 9:43 pm

yeah, you are right that is a 12V battery charger.i want to charge it using an adjustable current source. i have a 24V power supply seperate from this battery.. i have not a transformer for isolated power supply and battery will work out of car.the project will be a microprocessor controlled battery charger. with a feedback of battery voltage i will adjust the current value. and if i can connect the negative side of battery to the ground my design will be minimized.
-Erkan

erkanyigiter
Posts: 29
Joined: Wed Sep 08, 2004 1:01 am
Location: ISTANBUL, TR
Contact:

Re: Current source

Post by erkanyigiter » Wed Oct 06, 2004 9:56 pm

RonH:
i want to control that current with a micrprocessor controlled DAC. i will use a voltage divider to divide voltage and a comparator to compare it with my referance voltages. so i can understand the voltage of battery and i will adjust the current. if it is higher then 13,7V, the charger will stop.
-Erkan

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

Re: Current source

Post by fsdenis » Thu Oct 07, 2004 8:46 am

Image<p>yigiter:<p>There is a way to build a current regulator with your load grounded and control circuit related to ground.<p>It uses a difference amplifier like that in Forest Mims' Engineer's Mini-notebook Op Amp IC circuits. Except:<p>1)Use LM324 opamp instead of 741 because the output will go to ground when needed.<p>2)Use two stages of this difference amp circuit, one following the other, to get rid of fairly large output offset voltage present if just one used.<p>This two-stage difference amp will measure the voltage across Re, the current sensing resistor, and relate it to ground.<p>This measured voltage goes into the (-) opamp input as if we were using the original current regulator circuit with Re from the opamp driven transistor to ground.<p>Now the opamp driven transistor operates the PNP
Q1 transistor. <p>The R1 might be 100 ohms or more. It serves to turn the transistor off if Vin is 0V and give the Q2 collector the voltage it needs to operate.<p>Will you look this circuit over and tell me if it makes sense to you? Let me know about whatever problems you have understanding it or making it work. <p>At least, I seem to be getting a little more accurate idea of your project.

Ron H
Posts: 360
Joined: Tue Dec 18, 2001 1:01 am
Location: Boise, ID
Contact:

Re: Current source

Post by Ron H » Thu Oct 07, 2004 10:19 am

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by yigiter:
RonH:
i want to control that current with a micrprocessor controlled DAC. i will use a voltage divider to divide voltage and a comparator to compare it with my referance voltages. so i can understand the voltage of battery and i will adjust the current. if it is higher then 13,7V, the charger will stop.
<hr></blockquote>
Image<p>There are lots of ways to skin this cat, as Windmiller has demonstrated. Above is one. I used a "compound Darlington" because I think 2N3055s are probably easy to get, and they can handle lots of power. You could use a PNP Darlington instead. I like the Darlington because it reduces the load on the op amp, and in this case allows us to use a low-power regulator (the 5.1V zener). The op amps both need to be basically rail-to-rail IO types, which tends to limit the choices for op amps that could handle the full 24V supply.
As I noted on the schematic, you can scale R4 to allow you to use a wide range of peak input voltage.<p>[ October 07, 2004: Message edited by: RonH ]<p>[ October 07, 2004: Message edited by: RonH ]<p>[ October 07, 2004: Message edited by: RonH ]</p>

erkanyigiter
Posts: 29
Joined: Wed Sep 08, 2004 1:01 am
Location: ISTANBUL, TR
Contact:

Re: Current source

Post by erkanyigiter » Fri Oct 08, 2004 2:07 am

windmiller:
i want to ask two question about your drawing.
1)what is the equation about current?
i=Vin/Rsense?
2)How can i find difference amplifier that you mentioned.
-Erkan

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

Re: Current source

Post by fsdenis » Fri Oct 08, 2004 4:16 am

yigiter:<p>Yes. The equation for current through Rsense is
I = Vin / Re. <p>Re is Rsense.<p>The difference amp I have in mind does not amplify the voltage across Re nor does it diminish it. So the current control circuit operates the same way as does the simple version.<p>Somewhere in my notes, I have built this difference amp circuit or one very much like it and have worked it out in detail. It has been a couple of years, though, and my filing system is good for putting notes into, but not so good for getting notes out of.<p>I will try to find this circuit and draw it for you. It could take a while, but I will get back to you when I find it.<p>[ October 08, 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 » Fri Oct 08, 2004 4:27 am

RonH:<p>NICE drawing!

rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: Current source

Post by rshayes » Fri Oct 08, 2004 4:43 am

An alternate approach is to generate a current signal at ground level, as some of the circuits show, and then use a current mirror floating on the 24 volt line. A current mirror can be built with gain, possibly in the 10 to 100 range. It would use 3 or 4 PNP transistors, depending on the gain and accuracy you need. For example, a 1 to 10 milliamp signal might control a 100 milliamp to 1 amp charging current.

Ron H
Posts: 360
Joined: Tue Dec 18, 2001 1:01 am
Location: Boise, ID
Contact:

Re: Current source

Post by Ron H » Fri Oct 08, 2004 7:26 am

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by windmiller:
RonH:<p>NICE drawing!<hr></blockquote>
Thanks! Linear Technology's <p>SwitcherCAD III.

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

Re: Current source

Post by fsdenis » Fri Oct 08, 2004 10:15 am

Image<p>yigiter:<p>I got lucky and found my notes on difference amps.<p>Here is a detailed drawing adapting a difference amp to your circuit.<p>The LM324 comes with four opamps on the chip. I'm using two of them as you see.<p>I think this will work if you build it (and I haven't screwed up somewhere).<p>There are two more opamps on the chip in case you want to use one for changing calibration:<p>1)Take out offset voltage if desired.
2)Add gain to make other parts of this design easier.<p>You would put the calibration (non-inverting) amplifier between the difference amp and the current controller amp.<p>You may use any resistance from 1k ohms to 100k ohms as R2, R3, R4, R5 so long as they all have the same resistance. You might have better luck with accuracy of the difference voltage measurement using 1% tolerance resistances.<p>But, you might find accuracy adequate with the first amp you build, depending on how tightly you want to specify your battery charger.<p>Please let me know if you have problems understanding this circuit or making it work.<p>Fred-

Ron H
Posts: 360
Joined: Tue Dec 18, 2001 1:01 am
Location: Boise, ID
Contact:

Re: Current source

Post by Ron H » Fri Oct 08, 2004 1:46 pm

I don't like to be critical, but I also hate to see a guy burn up his parts. There are several problems there:<p>1. The loop gain is way too high (with no compensation), which will cause the circuit to oscillate.
2. There is no current limiting through Q1 and Q2.<p>I can post a circuit that uses a high-side sense resistor and a diff amp (different kind), but I sorta feel like I'm horning in here. Yigiter seems to prefer Windmiller's circuits. No skin off my nose. :D<p>[ October 08, 2004: Message edited by: RonH ]</p>

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

Re: Current source

Post by fsdenis » Fri Oct 08, 2004 11:51 pm

RonH:<p>Congratulations! You are my first literary critic as an electronics writer wannabee.<p>I've built a number of circuits like this and haven't smoked one yet. It's a smoothie.<p>In any case, this is a tiny circuit with 9 parts, not counting the battery, and costs me less than two dollars to build as shown for low currents.<p>It can be built and tested in minutes on a solderless breadboard.<p>Not much to risk here in either time or money.<p>In the world of science and engineering, "Experience is king". <p>Even perfect logic can be wrong. That's why we do experiments. <p>I've done the experiments that indicate this circuit will work. So, unless there is a typo in
my drawing that I don't see yet, I expect this one to work.<p>I can't accept logic alone as a legitimate criticism, whether or not it is perfect.<p>I will worry if you build it and smoke it though.<p>[ October 08, 2004: Message edited by: windmiller ]</p>

Ron H
Posts: 360
Joined: Tue Dec 18, 2001 1:01 am
Location: Boise, ID
Contact:

Re: Current source

Post by Ron H » Mon Oct 11, 2004 6:56 am

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by windmiller:
RonH:<p>Congratulations! You are my first literary critic as an electronics writer wannabee.<p>I've built a number of circuits like this and haven't smoked one yet. It's a smoothie.<p>In any case, this is a tiny circuit with 9 parts, not counting the battery, and costs me less than two dollars to build as shown for low currents.<p>It can be built and tested in minutes on a solderless breadboard.<p>Not much to risk here in either time or money.<p>In the world of science and engineering, "Experience is king". <p>Even perfect logic can be wrong. That's why we do experiments. <p>I've done the experiments that indicate this circuit will work. So, unless there is a typo in
my drawing that I don't see yet, I expect this one to work.<p>I can't accept logic alone as a legitimate criticism, whether or not it is perfect.<p>I will worry if you build it and smoke it though.<p>[ October 08, 2004: Message edited by: windmiller ]
<hr></blockquote>
This is a civil forum, and I don't want to change that, but I have to respond. This is not meant as a flame.
That wasn't a literary critique. It was a circuit design critique. Your writing style is fine.
Since you seem to be certain this will work, perhaps you can provide component values. Yigiter will need them anyway.
I can't deny that hardware trumps simulation (which I did), but if your circuit actually works, this might be the first time that I've heard of when the hardware works but the sim doesn't. If they differ, it is almost invariably the other way around.
My transient sim oscillated at 65kHz, 20 amps peak-to-peak. An AC Bode plot shows 40dB (must be below 0dB for stability) of loop gain when the excess phase shift is 180 degrees (this is at about 100kHz).
This is so far from being stable that I have to stand by my critique. Show us a real schematic with real parts values so someone can actually build and test it. I might point out that building it and testing it in minutes will also require that you have on hand a 24V, 2 amp supply, a car battery (or some other fairly big SLA battery), a BIG PNP,and a major heat sink.

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

Re: Current source

Post by fsdenis » Mon Oct 11, 2004 6:30 pm

RonH:<p>Thanks for your fascinating reply suggesting a comparison of an actual circuit build performance against a circuit simulation performance.<p>I am setup to build and test a 0-2A battery charger by the conceptual scheme shown in the last drawing.<p>If I did build it and give you accurate and complete parts specifications and drawings, would you run your simulator and see what it predicts?<p>If it turns out that the experimental device works and the simulator predicts otherwise, would you be able to discover what model of bipolar transistors the sim contains?<p>I suspect that the simulator might not have a model of bipolar transistors operating in unsaturated mode. This unsaturated mode operation being all but unknown.<p>Would this experiment series interest you for its own sake? <p>Fred

Ron H
Posts: 360
Joined: Tue Dec 18, 2001 1:01 am
Location: Boise, ID
Contact:

Re: Current source

Post by Ron H » Tue Oct 12, 2004 6:24 am

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>Originally posted by windmiller:
RonH:<p>Thanks for your fascinating reply suggesting a comparison of an actual circuit build performance against a circuit simulation performance.<p>I am setup to build and test a 0-2A battery charger by the conceptual scheme shown in the last drawing.<p>If I did build it and give you accurate and complete parts specifications and drawings, would you run your simulator and see what it predicts?<p>If it turns out that the experimental device works and the simulator predicts otherwise, would you be able to discover what model of bipolar transistors the sim contains?<p>I suspect that the simulator might not have a model of bipolar transistors operating in unsaturated mode. This unsaturated mode operation being all but unknown.<p>Would this experiment series interest you for its own sake? <p>Fred<hr></blockquote><p>Fred, I think that's a fine idea. Bring it on!
Regarding bipolar spice models, I assure you that they cover both saturated and unsaturated modes. How well they are modeled is another question.
In the sims I ran, I used the component values shown below.
Edit: Oops! U1 is another LM324. I forgot to label it.
Image<p>[ October 12, 2004: Message edited by: RonH ]</p>

Post Reply

Who is online

Users browsing this forum: No registered users and 10 guests