Current source

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
fsdenis
Posts: 90
Joined: Tue Mar 25, 2003 1:01 am
Location: San Manuel, Arizona
Contact:

Re: Current source

Post by fsdenis » Tue Oct 12, 2004 9:12 am

RonH:<p>After sleeping on the comparison of build with simulation notion I discover that it wouldn't be too tragic if the predicted 65khz oscillation does occur with the build:<p>This would be a self-starting oscillator capable of reliable operation while a LOT of power is being extracted from it continuously. <p>I might have uses for such an animal.<p>I want to design my experiment series to look for ways to both encourage oscillation and damp it.
This to suit both yitiger's purpose and the possibility of developing the above oscillator.<p>The drawing and components specification did come up on my computer. I'll study it for oscillator design hints.<p>Thanks,<p>Fred<p>PS: Might Q1, labeled MJ2955, be an MJE2955?
If so, is there such an animal as an MJ2955?<p>[ October 12, 2004: Message edited by: windmiller ]<p>[ October 12, 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 » Tue Oct 12, 2004 12:33 pm

If you wind up with a lemon, make lemonade. :) <p>A cursory look at MJ2955 vs MJE2955 datasheets looks like the differences are in package and power dissipation.

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

Re: Current source

Post by fsdenis » Wed Oct 13, 2004 10:12 am

RonH:<p>Congratulations! You are the proud father of a 10khz oscillator!<p>1)It has about a 60% duty cycle (60% "ON" time).
2)It oscillates between 0A and a max of 2.6A.
3)It is self-starting and reliable.
4)The transition times on and off are about 10 microseconds.
5)It is sweetly well behaved as built. Nothing heats up.<p>6)This sweet girl has NO explicit capacitors and NO explicit inductors needed to oscillate! <p>I developed it with stuff on hand and did a bit of detailed design to tame it from "first light wildness".<p>The circuit is the same as you gave me except:<p>1)I have a mechanical on/off switch in the positive lead of the 24V battery and a momentary on/off mechanical switch in the positive lead of the battery under charge. Both are closed (conducting) during operation of the oscillator.<p>2)I have used an N = 20 parallel array of TIP42C PNP transistors built by the base-resistor method with 10 ohm resistors in each base lead. This replaces Q1 with A1.<p>3)I have inserted a 4 ohm resistor in the collector of A1 (used to be Q1) before getting to the difference amp.<p>4)Vin is developed with a 50K ohm resistor tied to Vcc on one end and to a 5K ohm pot on the other. The other end of the pot is tied to ground.
Vin comes out the middle lead of the pot.<p>5)I have inserted a 1K ohm resistor in Q2 collector (the NPN transistor driving the base of A1).<p>6)I have inserted a 100 ohm resistor in Q2 emitter lead to ground.<p>No changes to the conceptual design.<p>My impression is that it is the difference amplifier that accounts for this circuit's ability to oscillate. It wouldn't oscillate when I was controlling A1 Ib with a simple current controller with no feedback from A1 Ic.<p>I'd love to find out:
1)How to square up the duty cycle.
2)How to speed up the transition times a bit.
3)How to adjust frequency to 10Khz exactly. (Frequency is actually close to 10.5 Khz.)
4)How to get rid of the battery under charge and still keep the oscillator.
5)How to change Vcc from 24V to 12V for convenience. But, without losing the option of a 24V Vce.<p>
This is an early report. I'll try to get some decent drawings together soon. Meanwhile, I think you might have enough stuff to input into your simulator to see if the as-built structure yields behavior close to prediction.<p>RonH, I need to lose quarrels to you more often. This oscillator could turn out to be useful.<p>Thanks,<p>Fred<p>[ October 13, 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 » Wed Oct 13, 2004 12:00 pm

Fred, can you post a complete schematic? I tried to duplicate your schematic, and I still get an oscillator, but not with the characteristics you mentioned. I'm not too surprised, because the oscillation frequency will wind up being a function of The Ft of the transistors, the beta of the transistors, and the bandwidth of the op amps, all of which will vary from mfr to mfr and from lot to lot. Oh, and I forgot to mention the most important parameter - the phase of the moon.<p>You have added resistors to limit peak currents - the peak currents that I predicted would let the smoke out of Q1, and probably also Q2.<p>If you had studied control theory and the stability requirements of feedback loops, you would understand that there are multiple poles (RC lowpass structures) in the loop and lots of gain, leading to too much phase shift around the loop in the presence of greater than unity loop gain. This is how oscillators are made.<p>I would think that, for a power oscillator to be useful, the frequency and duty cycle would need to be stable (and repeatable, if you want to make more than a few). This might be possible with some additional R's and C's, but there must be better ways to skin that cat.<p>Edit:
Sorry - I reread you post and saw that you said you will post a schematic.<p>Ron<p>[ October 13, 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 » Thu Oct 14, 2004 1:03 pm

RonH:<p>Thanks for your reply. The oscillator does still seem worth playing with. For instance I found a way of adjusting duty cycle over a wide range including 50%. I promise I will come up with a schematic for your simulator input.<p>But, I moved back to the original problem of yitiger's 0-2A battery charger on the bench and seem to have gotten lucky.<p>Here is the schematic. I hope there are no typos, all parts are clearly labled and there is enough here to input to your simulator.<p>The battery charger works as shown with clean and stable 0-2A control and no funny business. This is a working conceptual design. <p>Things like drift elimination and maybe boosting current capability well above 2A and so on might be done in a future detailed design cut.<p>Image<p>I would love to see what your simulator says about this circuit.<p>Thanks,<p>Fred<p>PS: Thinking about this thing became much easier when I realized that we really have a 3-rail (bipolar) power supply here. <p>B2, 24V, supplies V+ at positive terminal and v- at negative terminal. <p>B1, the 12V battery under charge has its negative terminal tied to B2 negative terminal.<p>The mid-0V (sometimes called ground) is the positive terminal of the 12V battery.<p>And we are just moving current through our controller from v+ through the battery under charge to v-.<p>The simple controller handles it well.<p>[ October 14, 2004: Message edited by: windmiller ]<p>[ October 14, 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 » Thu Oct 14, 2004 1:29 pm

Fred, I could sim that, and I'm sure it will look fine, but here's a quote from Yigiter:
<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>i think i can not do this. because my grounded load is a "car battery" so i can't connect the positive side of my load to 0V.(12V->0V) and also i have not got a isolated power supply. i have only a 24V dc source, thats all.
(sorry if i misunderstand your schematic)<p>-Erkan<hr></blockquote>
That's the reason I posted the schematic with the level shifter. If Yigiter (Erkan) can use your circuit, It couldn't get much simpler.

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

Re: Current source

Post by fsdenis » Fri Oct 15, 2004 2:01 am

RonH:<p>The circuit shown above is consistent with yitiger's battery under charge location and 24 volt power supply as charger requirements.<p>It should be usable with the battery either in or out of his car. I think.<p>This circuit is isolated from the rest of the world so we are free to recognize it as a 3-rail supply and pick any of the three rails as the reference (0V) point for measuring the voltage difference from that rail to the other two rails.<p>It is just a matter of convenience to pick the positive terminal of the battery under charge as my reference point for measuring all other voltages and name this point 0V. I could equally name it 100V if that were convenient. <p>In short, I found it intuitively stimulating for "cat skinning" a solution to yitiger's battery charger within his specified parameters to recognize this physical situation as a bipolar power supply and arbitrarily (legitimately) call the middle pole 0V. There is no absolute meaning to this name.<p>I think yitiger will be ok with this circuit as is with Vin controlled by a 0 to 10V dac referenced to the positive side of the battery under charge.<p>He should disconnect the Vin generating voltage divider from V+ and replace V+ with his dac.<p>He could have an 8K ohm in series with a 2K ohm resistor as his Vin voltage divider.<p>Connect the bottom of the 2K resistor to the positive side of the battery under charge. <p>Connect Vin to the mid-point.<p>Connect his dac reference to the positive side of battery under charge and output to the top of the 8K resistor. <p>Leave Re = 1 ohm for intuitive convenience.<p>Then 0V from his dac will yield 0A to the battery.<p>And 10V from his dac will yield 2A to the battery.<p>Yitiger can probably substitute an NPN darlington with a little heat sinking for the parallel array I used in this solution. I don't have any, so maybe someone will recommend one to him if he asks.<p>Unless I've missed something, I think that you and I have come up with a decent solution for yitiger to try.<p>I think I'll go back to playing with your oscillator for a while. Something about it has a good smell.<p>Fred<p>[ October 15, 2004: Message edited by: windmiller ]<p>[ October 15, 2004: Message edited by: windmiller ]<p>[ October 15, 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 » Sat Oct 16, 2004 6:24 am

Fred, I agree with all that. I just don't understand completely what Yigiter is trying to do (and I'm not sure he does). In this thread, I was guessing that he was trying to charge a car battery, and he wanted to use the car GND as GND for his controller (not to imply that he couldn't use +12V instead.) In another thread, he implies that it's a separate battery. Who knows?

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

Re: Current source

Post by fsdenis » Sat Oct 16, 2004 10:11 am

RonH:<p>Ron, my sympathies. I think we may have lost yitiger as a customer.<p>Nevertheless, I'm glad to have met you and had a chance to work with you on an interesting project.<p>I promised a schematic of your latest Dif Amp Oscillator version to play with on your simulator program. <p>This version yields very nearly a clean sine wave at 33Khz with 0-2A amplitude.<p>I'm going to put this odd creature in my curiosity box and wait for some notion to occur as to what I might do with it.<p>Image<p>I'm about done with this thread if you are and nobody else has anything to ask or add.<p>Until we meet again,<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 » Sat Oct 16, 2004 8:09 pm

Yeah, I think we pretty much beat this to death. I am posting the circuit I said I had that uses a diff amp of a different sort, along with waveforms. Note that the uotput is inverted relative to the input.
Image
Image

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

Re: Current source

Post by fsdenis » Sun Oct 17, 2004 4:48 am

RonH:<p>Thanks for your circuit, Ron. It will take a bit of study (possibly including your recommended control theory, new to me) to understand what the differences are between our approachs and why mine oscillates and yours does not.<p>The dif amp oscillator does have it's points, but just now it looks to be a complex beast to control properly, as you have noted.<p>Anyway, we have an interesting battery charger circuit, we didn't have before, and maybe we might build something useful on this development in future.<p>Fred

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

Re: Current source

Post by fsdenis » Wed Oct 27, 2004 6:22 pm

Anyone interested that the lead acid battery is a reactance in itself?<p>Since my last post, I've continued to try to find a difference amp design that won't oscillate when I try to control the charging current from V- as a reference. Several tried. All oscillate.<p>Additionally, a comparator for controlling battery voltage to 13.2 volts by turning off the current controller when it exceeds 13.2 volts and back on when it goes less than 13.2 volts results in an oscillator. (125khz using one of the LM324's as a comparator.<p>I have finally had it occur to me that nothing is wrong with my difference amps and nothing is wrong with my comparator circuits.<p>What is really going on is that the battery voltage snaps down very quickly when charging current stops and back up very quickly when charging current resumes.<p>Thus the battery is itself a kind of reactance.<p>It isn't similar at all to an inductance.<p>It is only remotely similar to a capacitor in parallel with a wildly voltage variable resistance.<p>Does anybody know what I'm talking about here? I'm
not sure I do.<p>Does a simulator program predict this reactive behavior of lead acid batteries?<p>Notions? Ideas? What?

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

Re: Current source

Post by dyarker » Wed Oct 27, 2004 8:49 pm

That's why in another thread with Yigiter I suggested charging to 14.6V, then switching to trickle till battery goes down to 12.6V (?memory?) (ie. charging is latched off).
Dale Y

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

Re: Current source

Post by fsdenis » Fri Oct 29, 2004 6:18 am

Thanks Dale,<p>It helps to know that other people have run into this battery reactance phenomenon.<p>I suspect I might use a window comparator to charge to 14.6V, then wait until battery voltage sags to 12.6V to resume charging, as you suggest.
This would seem to apply battery reactance to build a timer capable of long time periods. Interesting.<p>I've made an effort to discover definitions associated with lead acid battery charging, on the internet. So far, I haven't discovered any solid agreement and a number of contradictions among writers.<p>Might I ask your impressions?<p>1) Does "float charge" mean holding a battery at a given voltage, say 13.2V, maintaining whatever charge current accomplishes this?<p>2) Does "trickle charge" mean charging at a low enough current that battery voltage is free to drop to, say, 12.6V and hold there?<p>3) Is there an industry association or government body that develops and maintains definitions of terms associated with lead acid battery charging?
Or is disagreement among writers something I should expect?<p>Thanks for your reply.<p>Fred<p>[ October 29, 2004: Message edited by: windmiller ]<p>[ October 29, 2004: Message edited by: windmiller ]</p>

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

Re: Current source

Post by Will » Fri Oct 29, 2004 9:08 am

Why would you conclude that the battery is a reactance ? The battery is still discharging when it drops to 13.2 volts This means that there is more than 13.2 volts (The multi-cell e.m.f )driving the current and there is some drop across the cells which reduces the volts to 13.2. So, having dropped to 13.2 volts the charger switch on, the battery current is reversed and so the volts drop across the cells is now additive to the battery e.m.f. which was already more than 13.2 volts so that the charger immediately switches off again. The fact that the oscillation has a specific period will almost certainly be due to some reactance within the charger circuit - The transformer for instance ? Have fun
BB

Post Reply

Who is online

Users browsing this forum: No registered users and 8 guests