Constant current source advice

[mishima]

Some of my advanced chemistry students would like to try this pharmacokinetic model circuit outlined in an AAPT publication (Cock & Janssen, The Physics Teacher, Volume 51, Issue 8, pp. 482-484 (2013)). It is a simple RC circuit, but I've never built a constant current source and am unsure of what things to watch out for.

http://scitation.aip.org/docserver/fulltext/aapt/journal/tpt/51/8/1.4824943.figures.online.f2_thmb.gif

The article suggests using a 72 uF cap, 27k/297k/420k/890k resistors, and 172/344/520/700 uA current. Then the procedure is to leave S1 open, and charge the cap for 0.15 s (how can I get timing like that!?). The switches' states are then exchanged and of course the charge discharges with the time constant (over 10 seconds).

I've looked at some schematics and I'm confident it can be done with just a few components (like this), but am less sure about how I can vary it over the 4 values stated above. I also don't know how to only allow it to run for 0.15 s...any thoughts?

 

[nsaspook]

Why do you need to build it? Couldn't you just model it using a electrical circuit simulation program like SPICE.

 

[NascentOxygen]

So that's the complete procedure of interest in its entirety--just a one-off 0.15 s charge of that capacitor followed by a 10 s discharge? You then toast to its success and pack it away in the cupboard?

 

[mishima]

Why do you need to build it? Couldn't you just model it using a electrical circuit simulation program like SPICE.

Great idea, we should definitely do that to get an understanding of what's happening. But, they really are amazed by soldering and want to try it out.

So that's the complete procedure of interest in its entirety--just a one-off 0.15 s charge of that capacitor followed by a 10 s discharge? You then toast to its success and pack it away in the cupboard?

Sorry, should have given more detail. The idea is that the charge on the capacitor represents the dose of medicine in the body. Charge decays over a day (10 s) before another "dose" is given, this represents the medicine being metabolized. The idea is to adjust the values so that after a few days, you don't have an huge/dangerous amount of medicine (by adjusting the time constant, the exponential function varies).

http://scitation.aip.org/docserver/fulltext/aapt/journal/tpt/51/8/1.4824943.figures.online.f6_thmb.gif

I'm sorry I don't have online access to the article, but in my notes I have that a tau value of around 30 is considered better by pharmacists than 64 for example. In the image above, the lighter blue line is what a pharmacist would try to achieve. The equation in that field is

dx/dt = -αx

where x is the amount of medicine in the body, and α is a constant. The jagged lines in the image above approximate the solution. See also wiki

 

[nsaspook]

Great idea, we should definitely do that to get an understanding of what's happening. But, they really are amazed by soldering and want to try it out.

That's great but judging from what you just explained the physical creation of this circuit is not the best use of their time IMHO.

 

[mishima]

Oh? I think science projects are great. Especially ones like this where they end up creating something, its a huge confidence boost for young girls especially. The 2 interested in this are really top of their class, and are considering pharmacology as a career. I think it would be neat for them to have a little hand held device that simulated the body which they could give medicine to in just the right way.

 

[nsaspook]

Oh? I think science projects are great. Especially ones like this where they end up creating something, its a huge confidence boost for young girls especially. The 2 interested in this are really top of their class, and are considering pharmacology as a career. I think it would be neat for them to have a little hand held device that simulated the body which they could give medicine to in just the right way.

I know you mean well and science projects are great but a understanding of exponential growth/decay and time constants (by easily changing results in a program) in general will help them much more than building a circuit unless they have a desire to explore electricity directly instead of indirectly in pharmacology.

I really don't think you appreciate the complexity of building a little hand held device with results that can be displayed instead of a static display of electronic parts. I can see building a static display of resistors and capacitors to teach soldering and to provide a prop for a computer simulation.

 

[mishima]

Well, I mean its an RC circuit with 2 switches and a multimeter hanging on...not that bad really (and I'm kind of a dummy when it comes to electronics). I was thinking they could use some block terminals, like the ones that resemble breadboard holes, to swap out the resistance values. Should fit in an Altoids tin or similar.

I don't think they would have chosen this project if they weren't a little interested in learning electronics. They've also already been exposed to exponential functions in science. Its really more to give them a platform on which they can use some error analysis techniques in a context they are interested in (error analysis doesn't exactly blow high school girls away). They have soldered once before but it was not a perf board kind of thing (they made some electrodes for a piezoelectric crystal they grew).

The article has some pictures of the commercial current source and I had to shamefully admit that I didn't own one/couldn't afford one. I'm just trying to find a way to make this happen, maybe there isn't.

 

[nsaspook]

Well, I mean its an RC circuit with 2 switches and a multimeter hanging on...not that bad really (and I'm kind of a dummy when it comes to electronics). I was thinking they could use some block terminals, like the ones that resemble breadboard holes, to swap out the resistance values. Should fit in an Altoids tin or similar.

It's a simple RC circuit with a sub-second timing/switching and micro-amp constant current control requirement to provide a linear time to voltage slope for the capacitor charging.

That makes it complicated in a circuit for beginners but trivial to program in a electrical simulator for them to manipulate.

 

[NascentOxygen]

You have in mind a single-IC current source, you'll also need a timer IC so you can set the variable charge and discharge times and to oversee the switching. If you use a breadboard or two there needn't be a single soldered joint in the whole project.

You'll need a DC voltage supply of some description to power the current source.

 

[jim hardy]

Does the body stop metabolizing during injection? Why do they open S1 during charge interval ?It's almost an integrator or lowpass filter with a pulse input
1/($\tau$s+1)
tau being rate of metabolism which could be a knob on 1 meg pot.
But 72 uf? maybe 10uf polypropylene and 10 meg pot.


But what sort of readout ?

Computing does seem natural - but i appreciate the value of actually building a piece of hardware.

Arduino anyone ?

 

[NascentOxygen]

You might wish to reconsider the 0.15s and 10s durations if you'll be monitoring with a meter. The predominant feature will be the exponential, that 0.15s part will show as just some unsteadiness in the reading. They'll just have to picture in their mind what is happening with your proposed time scale.

 

[mishima]

The article suggested the 0.15 s charge time, I have no real need for it. If it can be adjusted upwards then all the better. I simply don't know. Perhaps the values chosen by the author were convenient for the apparatus he had access to.

Let me ask this, let's change things. Let's say the charge time is 5 s. Assuming things scale linearly (which I don't know) then the day could then be about 5 minutes. They charge it by using a handheld stopwatch and incorporate their reaction time (around 0.30s) as an uncertainty. Does that effect the design of the constant current source at all? Isn't it kind of arbitrary?

I mean they are just taking measurements of voltage and calculating charge from capacitance, and getting 4 graphs of q vs t like in my post 4 above, for 4 different resistance values and 4 different currents. Wouldn't the current values just affect the time scale?

 

[NascentOxygen]

So the two switches will be manually operated push-button switches?

 

[mishima]

Any reason that wouldn't work given they recorded uncertainty? Or how about a double throw?

 

[NascentOxygen]

Any reason that wouldn't work given they recorded uncertainty? Or how about a double throw?

That should be okay.

 

[mishima]

So really, the design of the current source isn't critical at all? As long as I slap something together that can output 4 different values and not electrocute my students I should be fine? (thanks for bearing with me) What are the drawbacks of using say a voltage regulator vs a general op-amp vs a specialized IC?

 

[NascentOxygen]

I see no great dramas in getting this to work. There's room for plenty of maths.

Motorola boasts that chip as being almost blow-out proof. I guess that means blow-up proof.

I asked earlier how you intend to power it. What DC supply do you have?

 

[mishima]

Batteries would be convenient, but I've read current sources prefer a regulated voltage. I might see if I can get away with just a 5V wall wart without a regulator.

 

[jim hardy]

A decent if not precision current source can be made easily from a transistor , zener diode and a couple of resistors.

9 volt transistor radio battery might work fine.

Simplest current source is a transistor, zener diode and resistor.
http://www.ecircuitcenter.com/Circuits_Audio_Amp/BJT Current_Source/BJT_Current_Source.htm

This one uses two garden variety diodes instead of a zener. Maybe 1N4148 - abundant and cheap.
Replacing one of them with a 1N5231 5.1 volt zener should improve temperature performance at expense of output voltage.

For the zener version RE would be R = V/I = 5.1/desired currentand you could use a pot or a switch to select current.
RB just needs to conduct about two milliamps to get past the knee of the zener.
Its value would be R = V/I = (Vsupply - 5.1 )/ .002 ohms .
max output voltage would be about Vsupply - Vzener - 1 volt, only about 3 volts for a 9 volt battery supply.

For the two garden variety diode version RE would be 0.6/desired current
and RB ~( Vsupply - 1.2)/.002 ohms
and max output voltage about Vsupply - 1.2 - Vcb volts, about 7 volts.


A 12 or 15 volt wall wart would work well with the zener source.

2N3906 is a general use pnp transistor that should be easy to find.. i bought a hundred of them, lifetime supply for just a couple bucks.

 

[nsaspook]

Computing does seem natural - but i appreciate the value of actually building a piece of hardware.

I just want to avoid a Snark hunt.

All these are good questions and answers for advanced electronics students who need to understand the nuts and volts of building circuits to perform some sort of actual physical function but how is a physical circuit realization of a model with time spent on component selection and construction relevant to studying dosage, metabolism and a pharmacokinetic model for advanced chemistry students seeking to understand that electrical model as a tool to explore a bodies possible response to drug dosage over time?

 

[jim hardy]

Is there a Basic-like program they can use to simulate and graph this on their ipods or ipads or whatever are those things all the kids have? That's how i used to study systems, Basic on a TI99. Later Qbasic but it's gone now.

A snazzy looking output will be fun for the kids.

And watching integration & exponential decay by finite difference model might give some insight to the math they're probably taking.

I do get carried away by DIY projects. They always sprout mission creep.

 

[nsaspook]

http://www.appszoom.com/android_applications/ltspice
There are a few apps that might work.

 

[mishima]

Jim, thanks I have some 2907a pnps, I will try to make it work. It seems the main difference is the higher saturation voltage for Vbe.

nsaspook, I don't know what your experience with public education is, but you are right. It would be faster to just get a computer program that simulated the human body, and plotted out the changes in concentration over time for them. They've all done a hundred computer simulations before, and that would be adding one more. Speed isn't everything though, things like student interest and ownership also come into play. This is a fairly unique project for a high school student to engage in and presents a lot of opportunities they wouldn't otherwise come across. Have you never used a mass on a spring or a child on a swing to explain electrical resonance? I think instances like these where the mathematical model cuts across fields is rather worthwhile to explore.

 

[mishima]

We sometimes use Processing, not Basic, to observe real-time data. Processing is similar to Java or C++ but cuts away a lot of the complexity in displaying images. One very cool project they did last year was to turn key presses on an electric piano keyboard into a shape and color corresponding to its frequency. But it also works especially nice with physical sensors because its class for handling microphone input is incredibly simple.

But yes, I can imagine a program that shows a human body getting progressively grumpier when an incorrect time constant is chosen in their circuit. That would be rather easy with Processing.

 

[jim hardy]

Interesting. I know I'm old

but i never thought analog would become trendy again.

This circuit combines integration and exponential decay.
http://claymore.engineer.gvsu.edu/~jackh/eod/courses/courses-80.gif

But one step at a time. Get your current source working.

Do you have 555 timers? Are you familiar with them? And some small relays?

I can see this creeping already.

In high school electronics class we teamed up, each of us boys built a bistable flip-flop. Transistors were a new thing then, you see.
We cascaded them into a twelve or so bit counter driven by a ~2hz astable flipflop oscillator, and connected twelve Simpson 260s to indicate each flipflop's state.
We watched them count up - took nearly an hour to reach all one's - and cheered when they all swung back to zero at same time !

You sound like an involved teacher.

Good luck.

 

[mishima]

Thanks, got the current source going! As long as I keep the load resistance below Rb, it will work.

I do indeed have some timers and relays, as well as some counter ICs. I can at least say I am more familiar with them than op-amps and amplifiers. There are a lot of things about op-amps especially that are not intuitive for me; I regret not taking more EE courses in college (my physics major only required one) or joining a club. I've only really had electronics as a side hobby for 2 years, and while I'm sure I can build anything from a schematic (that isn't surface mount), the theory remains a mystery at times. The first few chapters of Sedra and Smith are doable but a challenge for me, for example.

But let me try giving the patient his first dose...

 

[nsaspook]

I think instances like these where the mathematical model cuts across fields is rather worthwhile to explore.

I think we both agree on this. If you just wanted to design and build this to teach students with the completed project I could completely agree with the construction.

 

[mishima]

Fair enough, sir.

 

[jim hardy]

But let me try giving the patient his first dose...

I started here:

http://my.ece.ucsb.edu/York/Bobsclass/2C/Tutorials/App%20notes/an-20.pdf

available many places. like the source, http://www.ti.com/lit/an/snoa621c/snoa621c.pdf

Here's a listing of many similar. http://my.ece.ucsb.edu/York/Bobsclass/2C/Tutorials/application_notes.htm

An31 is always handy
http://my.ece.ucsb.edu/York/Bobsclass/2C/Tutorials/App%20notes/an-31.pdf
also from the horse's mouth at http://www.ti.com/ww/en/bobpease/assets/AN-31.pdf
but it's not annotated, use AN20 until you get the hang of it.

old jim

 

[mishima]

I obtained a schematic of the entire apparatus from the journal article's (retired) author, if anyone is interested.

 

[jim hardy]

He certainly used good quality parts. Are you familiar with those IC's ?

OP07 is a precision opamp.
Left hand one causes voltage selected by selector switch E1 to appear across R10. That's what sets the charging current through your capacitor C13-C14. Observe when T2 is off no current can flow through R10 so capacitor charging current is zero.

Turning on T1 discharges capacitor through whatever resistance is across E4 terminals.

Right hand OP07 is just a buffer to repeat capacitor voltage over at Vout for your meter there. That way a few milliamps of meter current won't discharge the capacitor. .

74121 is a pulse generator, colloquially called a "retriggerable one shot" or "monostable", that makes a pulse of precise duration every time the 555 gives it a "go" pulse. P2-C15 set that pulse width. Its two outputs are complementary, that is when one is high the other is low.

Have fun !