Voltage Scaling

Discussion in 'Electronic Basics' started by [email protected], Mar 22, 2008.

1. Guest

Hi all,

Can anyone help with the following,

I need to scale a control voltage which ranges from 0VDC to 5VDC, to:

1. 0V..15V (e.g. when Vcoltrol=2.5V, Vout should be 7.5V).

and

2. -3V to 3V (e.g. when Vcontrol=0V, Vout is -3V, when Vcontrol=2.5V,
Vout is 0V.. essentially a linear scaling..)

I have a power supply of +9V and nothing else (got suggestions to
connect +/-18V to an opamp but obviously, i dont have these voltage
ranges from the wall-wart, only +9V..)

How would you go about doing this?

Thanks,

2. JamieGuest

how about a simple OP-AMP using + input with a gain ratio of 3 network
on the (-) input?
and for the voltage supply issue.. You'll need a DC-DC converter to
up the level.
I don't know where your level of expertise is and how advanced you
want to make this but, if you're looking for a weekend hack job. You
could construct an inverter using a mini xformer like what radio shaft
sells in the audio output line with a CT on it. You can drive the
secondary side with something like a 555 timer chip and derive a +/-
supply that is actually isolated if you wish on the other side.
Input that into a post type regulator of 15 Volts and you have it.

But Like I said, this would be a weekend worrier project or something
that needs to be done in a hurry on a perf board or something etc..

3. ChrisGuest

Hi, Gilad. You aren't going to be doing anything with output voltages
above or below your power supply rails unless you do something with
your supply. By using a 555 and a handful of caps and diodes, you can
generate a +17V and -8V supply like this (view in fixed font or M\$

|
| VCC VCC
| + +
| | |
| .-. |
| | |2K2 | N.C.
| | | .---o----o---. VCC
| '-' | 8 4 | +
| | | | |
| | | | D -
| o----o7 | ^
| | | | C | 2*Vcc - 1.4V
| | | | ||+ | D (~+17V)
| .-.15K| 3o----o----||---o--->|---o-------o
| | | | | | || |
| | | | LM555 | | +|
| '-' .-o2 | | C ---
| | | | | | ---
| | | | | | |
| o--o-o6 | | ===
| | | | | GND
| | | | |
| | | 1 5 | |
| --- '---o----o---' | C -Vcc+1.4V
| --- | N.C. | +|| D (~-7.6V)
| |.01uF | '----||---o---|<---o-------o
| | | || | |
| === === | |
| GND GND D V C---
| - ---
| C = 100 to 220uF, 25V | +|
| D = 1N4001 or 1N4148 | |
| === ===
| GND GND
|
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

Now that you've got realistic supplies, get an LM324 (a quad op amp),
and go to work:

|
| 0 to +5VDC in +17V
| |\|
| o-------------|+\ 0 to +15V out
| | >--o-------------o
| .---|-/ |
| | |/| .-.
| | -8V | |10K
| | | |
| | '-'
| | |
| '---------o
| |
| .-.
| | |20K
| | |
| '-'
| |
| ===
| GND
|
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

VCC
+ ___
| .--|___|--.
| | 12K |
.-+2.5V|\ | |
10K| |<---|+\ ___ | |\ |
| | | >--o--|___|-o-|-\ |
'-' .--|-/ | 10K | >----o-----o
| | |/ | .-|+/ -3Vto +3V out
| | | | |/
=== | | |
GND '--------' |
|
|
|
0 to +5V in ___ |
o--------------|___|---o
10K |
|
.-.
12K| |
| |
'-'
|
===
GND
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

Assuming 1% resistors, the output of your 0-15VDC op amp should be
accurate to within 2% +/-20mV, probably much better. The second
circuit as shown is going to be mostly dependent on the stability of
your 9V supply. If it's regulated, all the better. If not, you
should consider getting an LM78L05 as a post regulator, generating a
solid stable +5V, and running the tweaker pot off that, or use two 1%
resistors instead of the pot to get your 2.5V from the 78L05.

Here are a couple of references and data sheets to give you some
background:

http://www.national.com/ds.cgi/LM/LM555.pdf
http://www.national.com/ds.cgi/LM/LM124.pdf
http://www.ikalogic.com/shm_voltage_inv.php
http://www.national.com/an/AN/AN-31.pdf

Chris

4. Bob MonsenGuest

You can't really go from 0 to 15V with only a 9V power supply unless you
play tricks. However, why do you need to go to 15V if you don't have a 15V
power supply? What are you controlling?

The tricks are using a dc-dc converter, or using one of the capacitative
voltage doublers like an ICL7660, which can get you 16V from your 9V input.
Then, use a rail-to-rail opamp configured to amplify 3x in the positive
direction, using a circuit like this:

20k
___
.------|___|---.
| |
| |\ |
o----|-\ |
| | >------o------ out
in -----------)----|+/
| |/ 0-15
0-5 |
|
.-.
| |10k
| |
'-'
|
GND
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

For the -3V, you can use another ICL7660 to give you a -8V rail from your 9V
input. However, I believe that getting the voltages to come out right using
an opamp will require some kind of external reference, in addition to the
ICL7660. However, there may be some kind of clever opamp circuit that can do
this with just resistors, and no reference. I don't know.

Regards,
Bob Monsen

5. Paul E. SchoenGuest

You will need a negative supply to get the -3V output, and you will need at
least 16VDC to get 15VDC output. So you need a voltage doubler and an
inverter from your wall-wart. There are many off-the-shelf DC-DC converters
that can do this, and there are countless circuits available if you want to
roll your own. It all depends on your budget of money, time, and ability to
build circuits. Let us know these details, and you can get the information
you need.

Paul

6. ChrisGuest

Hi, Jamie. You're right -- the diode is backwards. Good catch -- it
should look like this:

|
| VCC VCC
| + +
| | |
| .-. |
| | |2K2 | N.C.
| | | .---o----o---. VCC
| '-' | 8 4 | +
| | | | |
| | | | D V
| o----o7 | -
| | | | C | 2*Vcc - 1.4V
| | | | ||+ | D (~+17V)
| .-.15K| 3o----o----||---o--->|---o-------o
| | | | | | || |
| | | | LM555 | | +|
| '-' .-o2 | | C ---
| | | | | | ---
| | | | | | |
| o--o-o6 | | ===
| | | | | GND
| | | | |
| | | 1 5 | |
| --- '---o----o---' | C -Vcc+1.4V
| --- | N.C. | +|| D (~-7.6V)
| |.01uF | '----||---o---|<---o-------o
| | | || | |
| === === | |
| GND GND D V C---
| - ---
| C = 100 to 220uF, 25V | +|
| D = 1N4001 or 1N4148 | |
| === ===
| GND GND
|
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

Trust but verify.
Chris

7. JamieGuest

Something tells me the above circuit is flawed.
Something about a diode seems to be out of place in the + output.
Also the voltages don't look correct referenced from the GND.
I just can't see how that is going to generate 17 Volts ?
the (-) output side looks ok how ever.
I guess if you were to use the (-) output as your common to the LM324
, you could get a single output effect of ~ 17 volts maybe..
Maybe I missed something, just an observation.

http://webpages.charter.net/jamie_5"

8. JamieGuest

Yes,. it appears the diode is backwards in the + output multplier from
the VCC+ source.
I knew something didn't look correct, I came to me after I pushed the
send button.
It's a bitch getting old!

9. Guest

Thanks for the quick reply.. will definitely try this.

For the curious:
The project is to build something along the lines of the following
clock (i'm using arduino diecimila):

http://web.jfet.org/weston-clock/

the board gives several variable 0-5V pwm outputs, but i can't find
analog panel voltmeters for anything less than 0-15VDC (also i have a
beautiful pre-WWII -3V~3V panel one which i really want to use, which
explains the need one of the scaling circuits). If anybody knows where
i can find 0-3V (or something along those lines) analog panels, that
would save a lot of hassle..

Also, usingDS1307 RTC the proper voltage is delivered to the 3 panels
(for sec's minutes and hours display).
In the spirit of open soft/hard-ware i'll post links to circuits and
code for the whole project when it's done..

10. Guest

message (three messages down..)

11. Paul E. SchoenGuest

Analog voltmeters usually have internal resistors to provide scaling. You
should be able to open up the meter and change the resistor so the 0-15V
meter becomes 0-3 VDC or 0-5 VDC which can be driven directly from the PIC
PWM.

If you can connect the (-3V)-0-(+3V) meter to a bridge type circuit, you
can drive it without a negative supply rail. Or even directly between two
PWM outputs of the PIC.

It may also be possible to rotate the rear spring adjuster of the meter far
enough to make the pointer rest at the leftmost mark on the scale. Long ago
I worked as a meter technician and we would modify analog meters for
suppressed zero, live zero, and center zero. But these were more modern
meters, vintage 1975 or so, and not 1894 or 1940s. I'm not *that* old!

You can also look for switchboard DC ammeters, which are generally 50 mV or
100 mV full scale. They usually take a little more current to drive, but
probably 10 mA or less, and you can just add a series resistor. Or look for