[PaulH]

I have a background in electronics that I have not used in almost 10 years. Now I'm starting to get interested again and trying to re-learn some of the things that I have forgotten.

I would like to build a computer-controlled regulated power supply. I have found some simple swithing regulator circuits that are basically a Schottky diode, an inductor, a capacitor, and a power transistor that turns it on and off. You can feed the base of the transistor a PWM signal. The frequency determines the size of the inductor needed. The duty cycle determines the resulting voltage.

The resistance of the load may vary. I can easily obtain the current by measuring the voltage across a sense resistor of known value. Run that through an ADC and Bob's your uncle. I can also sense the voltage over the load itself with an ADC.

Now here's the stupid question: How can I regulate the current as well?

The best way to think of this circuit would be to look at it like a NiCd/NiMH battery charger. You need 1.5 volts per cell and a current that is based on the capacity of the battery (C/10, C/4, etc.)

Thanks!

[Al_The_Man]

You don't mention the use you are putting the supply to, Why does it need to be computer controlled?
I built my bench supply many years ago from a Australian Electronic mag. circuit.
It has an adjustable voltage out and an adjustable current limit.
It is built around an LM723, I am sure there are many other designs out there using voltage regulator I.C.'s
http://www3.telus.net/chemelec/Proje...466/MC1466.htm
Another one some used was a MC1466.
You mentioned NiCd battery chargers, there are some sophisticated IC's for that purpose my Maxim etc.
Al.

[PaulH]

I actually have a few projects in mind.

One example is a multiple-output battery charger that can accept packs of a variety of voltages and currents. I have seen a few available, but none of them have the feature set that I want.

You're right: I could use a Maxim chip for this purpose, but where's the fun in that?

A bench supply is something that I could really use. However I like the idea of one that is completely digital. I like pushbuttons and displays more than dials and gauges.

One thing that struck me after I wrote my original post is that one has to consider Ohm's law. If the resistance of the load is known and a specific current is desired, the voltage is a dependant variable in the equation.

In your bench supply, what happens if the current limit is exceeded? Does the supply just shut down or does the load continue to function at the supply's preset maximum current?

[Al_The_Man]

I have lost the original schematic, but the current just folds back IOW the current will be restricted to what ever the limit is set to. Obviously this will affect the voltage out also if a very low R. load is connected.
If a PWM output was used then the voltage could remain at the set level and the mean current would vary, but with some loads, a PWM signal is not desirable.
But there is not much point in maintaining the voltage if the current cannot be sustained.
I was looking for semi-intelligent battery chargers a while ago and came accross this one.
http://www.angelfire.com/electronic/hayles/charge1.html
MicroChip have some designs also if you want to go the µProcessor route.
Al.

[PaulH]

I like the battery charger on that page. I have run across it before from Google searches, but he doesn't make the source code available. ATMEL has something similar on their site and includes the source code.

One thing that just dawned on me is that nickel-chemistry batteries are designed for constant current charging. That means the voltage will rise or fall to whatever level is necessary to keep the current the same. That makes design easier. All that is necessary is a programmable voltage supply and a way of sensing the current. The processor can then vary the voltage to keep the current the same.

The R/C hobby has very few good trickle chargers available. Most use the "wall wart" that comes with the transmitter. In many cases, this is the worst thing that could be used, but there is no real choice. Trickle charging is the least damaging to the battery and is the only method that ensures a full charge.

I would like to design a charger that is modular. Modules could then be stacked, one per battery, with a master control module that allows the user to set the charge current for each charging module independently.

That is the pie-in-the sky dream. Right now I would settle for any decent slow charger with at least seven (7) outputs.

[NC Cams]

Pulse width modulation or current and voltage can be done quite easily with a number of PWM IC's. The LM3524 or LM3526 IC can readily be configured to create either a psuedo constant voltage or constant current circuit.

If you monitor the current across a sense resistor and then filter it a bit, you can use the error amps on the said IC's to automatically sense and regulate current and hold current QUITE constant from 1 to almost 10 amps depending on your current source 'stiffness'.

How you sense and filter a pulsing current is not intuitively obvious nor is it something that is published in any application notes on battery chargers.

The application notes for the IC's will readily explain how to do constant voltage regulation.

We built a peak detecting, PWM, constant current nicad charger over 10 years ago ago using variations of these IC's which I still own the rights to. The trick to doing NICADS is to sense peak voltage and cut off charging once you reach and just cross over peak voltage.

It is also critical as to when you sense the voltage (you have a choice of during charge or during rest). The circuitry we used could do this to within 20milliolts or so and was somewhat adjustable.

To get it to work does require very careful design and component selection. If you aren't careful, nicads do ugly things when they get severely overcharged.