Variable Power Supply Circuit Design Question

[Zach Zaborny]

Hello all,
I am working on a low amperage variable power supply for use with some of
my logic circuits that only require a few milliamps. The schematic for it is
located at:
http://random.nts-technologies.org/images/Hosted/powersupply.gif. The
circuit itself is based on the car battery charger circuit located at:
http://www.interq.or.jp/japan/se-inoue/e_car2_1.htm (because it supplies a
voltage and current similar to the one I want) with a few minor changes to
up the voltage and current ranges (resistor values). Upon setting up the
circuit on my breadboard, notice that when a load is put on the output,
like a cassette deck motor or an LED, the voltage meter returns to 0. Is
there any problems in the circuit which can cause that to happen? Could it
perhaps be a faulty voltage/current regulator? I have also seen many voltage
meter applications that have a resistor across the meter's terminals, would
that be needed? If so, what value?

I am a beginner and this is my first big project, so any ideas and
improvements on my circuit are extremely welcome.

Thanks in advance,
Zach, NTS

 

[Zach Zaborny]

Update: the problem was that I assumed the LM317 pinout was the same as the
7805. It was quite different, and upon connecting it correctly, things are
improving.

Any other improvement ideas?

-Zach, NTS

 

[Zach Zaborny]

Sorry for all the annoying posts, I wish I could edit previous ones as I
test the circuit.

Even when the voltage regulators, etc are in in properly, the voltage and
current meters go way down when there is a load applied to the output. Is
this just something that happens to all power supplies? Is there something
wrong with my circuit that is causing that to happen?

-Zach, NTS

 

[Mook Johnson]

For us to be of much help, you either need to describe the output
requirements in terms of voltage range and current limit range or describe
the type of load it will be seeing (4000 series CMOS, 74HC, analog, etc.)

You can add current limiting directly to the LM317 without having the 7805
(see lm317 datsheet). Make sure you use a heatsink large enough to handle
the power under a short curcuit situation. The 7805 and the output diode
will add voltage drop that will be dependent on the load (output impedance)
and cause errors in your output voltage setting.

LM317 datasheet probebly has what you need.

 

[Tim Wescott]

Mook Johnson wrote:

For us to be of much help, you either need to describe the output
requirements in terms of voltage range and current limit range or describe
the type of load it will be seeing (4000 series CMOS, 74HC, analog, etc.)

You can add current limiting directly to the LM317 without having the 7805
(see lm317 datsheet). Make sure you use a heatsink large enough to handle
the power under a short curcuit situation. The 7805 and the output diode
will add voltage drop that will be dependent on the load (output impedance)
and cause errors in your output voltage setting.

LM317 datasheet probebly has what you need.

-- snip --

That's stated very mildly. Let me amplify: The circuit behavior with
the 7805 as a series-pass element will _not_ be satisfactory. It's
probably great for charging batteries, there may even be some parasitic
variations in the charge voltage with temperature due to the 7805, but
it is _not_ what you want in a power supply for logic.

Either use Mook's suggestion for current limiting with the 317, or use a
circuit based on the 723 (I think they're still available).

 

[Clarence]

Sorry for all the annoying posts, I wish I could edit previous ones as I
test the circuit.

Even when the voltage regulators, etc are in in properly, the voltage and
current meters go way down when there is a load applied to the output. Is
this just something that happens to all power supplies? Is there something
wrong with my circuit that is causing that to happen?

-Zach, NTS

You can! Copy and past into your new post and with the old message selected
(ON IE6) select "Message" and on the pull down menu select "Cancel Message" and
click.

Then sent the new or updated message.

The old one will disappear in a short time (up to a couple of hours) and the
new message will appear. With the same topic and in the same thread.

 

[Zach Zaborny]

I do not think you guys know what I want from this power supply. Let me
rephrase myself.

I am building a benchtop power supply, for general use testing and powering
some circuits. It will be used for powering small motors, logic circuits of
all kinds, and more. A good range for all that is 1.5V-24V at 1mA-500mA
since most of my projects have ever required any more than that. I want to
be able to adjust output current and output voltage separately.

The circuit as it is now doesn't seem to be working. When it is powered on,
the volts meter is at 24VDC and the amp meter is at 0 because there is no
load and it will not work without one (see the schematic for an
explanation). When a load is applied (in this case, a cassette deck drive
motor), the votlmeter goes down to about 12v and the milliamp meter goes to
about 200mA. I have two rheostats, (seen in the schematic), one is for
voltage and one is for current, When I change the voltage one, the current
changes a little also.

What am I doing wrong in the circuit?

As for the LM317 datasheet advice, I saw nothing I wanted in there, unless I
didn't understand something properly.

Thanks,
Zach, NTS

 

[Chris Holmes]

I do not think you guys know what I want from this power supply. Let me
rephrase myself.

I am building a benchtop power supply, for general use testing and powering
some circuits. It will be used for powering small motors, logic circuits of
all kinds, and more. A good range for all that is 1.5V-24V at 1mA-500mA
since most of my projects have ever required any more than that. I want to
be able to adjust output current and output voltage separately.

The circuit as it is now doesn't seem to be working. When it is powered on,
the volts meter is at 24VDC and the amp meter is at 0 because there is no
load and it will not work without one (see the schematic for an
explanation). When a load is applied (in this case, a cassette deck drive
motor), the votlmeter goes down to about 12v and the milliamp meter goes to
about 200mA. I have two rheostats, (seen in the schematic), one is for
voltage and one is for current, When I change the voltage one, the current
changes a little also.

What am I doing wrong in the circuit?

As for the LM317 datasheet advice, I saw nothing I wanted in there, unless I
didn't understand something properly.

Thanks,
Zach, NTS

Zach,

Just off the top of my head, I think the LM317 has a minimum load
requirement to maintain regulation. Can you trying putting a resistor of
maybe 1k or 10k from the output pin of the LM317 to ground?

Chris

 

[Tim Wescott]

I do not think you guys know what I want from this power supply. Let me
rephrase myself.

I am building a benchtop power supply, for general use testing and powering
some circuits. It will be used for powering small motors, logic circuits of
all kinds, and more. A good range for all that is 1.5V-24V at 1mA-500mA
since most of my projects have ever required any more than that. I want to
be able to adjust output current and output voltage separately.

The circuit as it is now doesn't seem to be working. When it is powered on,
the volts meter is at 24VDC and the amp meter is at 0 because there is no
load and it will not work without one (see the schematic for an
explanation). When a load is applied (in this case, a cassette deck drive
motor), the votlmeter goes down to about 12v and the milliamp meter goes to
about 200mA. I have two rheostats, (seen in the schematic), one is for
voltage and one is for current, When I change the voltage one, the current
changes a little also.

What am I doing wrong in the circuit?

As for the LM317 datasheet advice, I saw nothing I wanted in there, unless I
didn't understand something properly.

Thanks,
Zach, NTS

Sounds like a job for a 723, or a discrete reference and a bunch of op
amps. On the plus side your transformer, rectifier and C1 are correct.
It isn't surprising that your circuit is working the way it is: the
topology is very innovative, it doesn't use the 7805 at all the way it's
original designers imagined. It may be _fine_ for changing a battery,
but _that isn't what you are doing_.

Since you appear to operate at the level of finding a schematic and
using it, why don't you look for a schematic for a bench supply with the
features you want? It's got to be out there.

 

[legg]

Hello all,
I am working on a low amperage variable power supply for use with some of
my logic circuits that only require a few milliamps. The schematic for it is
located at:
http://random.nts-technologies.org/images/Hosted/powersupply.gif. The
circuit itself is based on the car battery charger circuit located at:
http://www.interq.or.jp/japan/se-inoue/e_car2_1.htm (because it supplies a
voltage and current similar to the one I want) with a few minor changes to
up the voltage and current ranges (resistor values). Upon setting up the
circuit on my breadboard, notice that when a load is put on the output,
like a cassette deck motor or an LED, the voltage meter returns to 0. Is
there any problems in the circuit which can cause that to happen? Could it
perhaps be a faulty voltage/current regulator? I have also seen many voltage
meter applications that have a resistor across the meter's terminals, would
that be needed? If so, what value?

To increase the maximum output current limit, alter (reduce) R3, not
R4.

This is a very poor circuit, as far as voltage regulation goes, due to
all of the rubbish present after the output of the voltage regulator.

The LM317 could be used as a limiter, if the sensing voltage is
recalculated for the lower ~1V3 reference presented at its control
terminal.

Get a better circuit from the 317 app notes.

RL

 

[Terry Pinnell]

As others have already said, your requirements are not
fully-described. What range of voltage and current do you want to
cover? Must the current limit be variable? What is your source DC
supply? Do you anticipate needing a dual (+/-) output (e.g. for
working with opamaps)? What sort of metering indication do you want?

That basic circuit on which you've unwisely based your projectis is a
battery charger circuit (although it seems unduly complex at first
sight). So it's unlikely to be a good foundation for a general purpose
bench supply.

A single LM317 could be used as a sound basis for a bench power
supply. It has built-in, fast-acting current limiting, and temperature
sensing to close it down when it gets too hot.

Here are a couple of examples:
http://www.uoguelph.ca/~antoon/circ/vps.htm
http://ourworld.compuserve.com/homepages/Bill_Bowden/page12.htm#317.gif

Or you could use a 723, noted for its low noise, which I used in my
own bench unit, now about 20 years old. Here's a scan from one of my
old books:
http://www.terrypin.dial.pipex.com/Images/723PS-Babani.gif

I note that you're talking about DC motors, despite your opening
statement about "logic circuits that only require a few milliamps"! So
you might want to consider also the LM338K, such as in this example:
http://www.aaroncake.net/circuits/supply.htm

You'll find many other circuits by googling. All things being equal,
choose the simplest, as it's your first major project. Make the case
large enough to get easy access for later modifications, including
replacing the transformer by a marginally larger one when you
accidentally burn out the first!

 

[Bill Bowden]

I do not think you guys know what I want from this power supply. Let me
rephrase myself.

I am building a benchtop power supply, for general use testing and powering
some circuits. It will be used for powering small motors, logic circuits of
all kinds, and more. A good range for all that is 1.5V-24V at 1mA-500mA
since most of my projects have ever required any more than that. I want to
be able to adjust output current and output voltage separately.

The circuit as it is now doesn't seem to be working. When it is powered on,
the volts meter is at 24VDC and the amp meter is at 0 because there is no
load and it will not work without one (see the schematic for an
explanation). When a load is applied (in this case, a cassette deck drive
motor), the votlmeter goes down to about 12v and the milliamp meter goes to
about 200mA. I have two rheostats, (seen in the schematic), one is for
voltage and one is for current, When I change the voltage one, the current
changes a little also.

What am I doing wrong in the circuit?

The current control part doesn't look too good.
It looks like the input to the 7805 will have to
be about 7 volts higher than the output and at 500mA
you will be wasting 3.5 watts of power in the
7805 and 50,10 ohm resistors.

Try testing the circuit with the cassette load
and voltmeter connected directly to the LM317
output. The LM317 will not need a load since the
resistor R1 is drawing about 12mA which should be enough.
So the voltmeter should read a constant 12 volts
or whatever with the cassette on or off.

You can use a second LM317 as a current control
with much better efficiency since the drop on
current sense resistor will only be 1.25 volts
as opposed to 5 for the 7805.

-Bill

 

[Bill Bowden]

The current control part doesn't look too good.
It looks like the input to the 7805 will have to
be about 7 volts higher than the output and at 500mA
you will be wasting 3.5 watts of power in the
7805 and 50,10 ohm resistors.

Try testing the circuit with the cassette load
and voltmeter connected directly to the LM317
output. The LM317 will not need a load since the
resistor R1 is drawing about 12mA which should be enough.
So the voltmeter should read a constant 12 volts
or whatever with the cassette on or off.

You can use a second LM317 as a current control
with much better efficiency since the drop on
current sense resistor will only be 1.25 volts
as opposed to 5 for the 7805.

-Bill

If you put the current limit control in front of the
voltage control, the regulation will be better since
the output voltage will not fall off as the drop on
the current sense resistor increases. Something like
below.

The current resistor will be about 2.5 ohms for 500mA
and proportionally higher for less current.
5 ohms = 250mA, etc. The input voltage needs to be
about 2 volts higher than the output for each
regulator, plus the 1.25 drop on the current resistor,
so to get 24 volts out, you need about 30 volts in.

Current
+- - - - - - - -+ Adjust +- - - - - - - -+
DC in --| LM317 |---\/\/\---+------| LM317 |-----+----Out
| | | | | |
+- - - - - - - -+ | +- - - - - - - -+ \
| | | /
| | | 100 \
+-------------------+ | |
+------------+
|
\
V Adjust /
2K \
|
GND
-Bill

 

[Zach Zaborny]

Thanks Bill, that is exactly what I need. I'm picking up another LM317 at
the electronic surplus store today to try 'er.

-Zach