Want independant pulse width variation on + and - side of gnd; square wave of about 1kHz

[jtaylor]

Subject says most of it; power requirements are small, frequency and some
pulse width drift/variation is no problem...

 

[Tim Wescott]

Subject says most of it; power requirements are small, frequency and some
pulse width drift/variation is no problem...

555 + diode + two pots + some thought.

First line says most of it; power requirements will be small, frequency
and pulse width drift won't be too bad.

 

[jtaylor]

555 + diode + two pots + some thought.

First line says most of it; power requirements will be small, frequency
and pulse width drift won't be too bad.

Are you sure about this?

In more obvious terms, I want to be able to vary the + width from 0 to 100%
independently from the - width (which should also be variable from 0 to 100%
independently from the + width).

 

[John Popelish]

Are you sure about this?

In more obvious terms, I want to be able to vary the + width from 0 to 100%
independently from the - width (which should also be variable from 0 to 100%
independently from the + width).

How can the positive and negative parts both be 100% (duty cycle?) at
the same time or 0% at the same time (infinite frequency)?

Or are you saying that you want the on time to be variable from some
small time to some maximum time and also the off time to be variable
from some small time to some maximum time and variations in the
duration of one state does not affect the duration of the other state?

 

[jtaylor]

How can the positive and negative parts both be 100% (duty cycle?) at
the same time or 0% at the same time (infinite frequency)?

Or are you saying that you want the on time to be variable from some
small time to some maximum time and also the off time to be variable
from some small time to some maximum time and variations in the
duration of one state does not affect the duration of the other state?

Um, no. I want ground in the middle, and the other line to go from say +5v
to -5v, at about 1kHz, with the + pulse varying in width, and the - pulse
varying in width.

 

[Spehro Pefhany]

Um, no. I want ground in the middle, and the other line to go from say +5v
to -5v, at about 1kHz, with the + pulse varying in width, and the - pulse
varying in width.

You can consider connecting the "GROUND" pin on the 555 to -5V. It'll
never know the difference, you know.


Best regards,
Spehro Pefhany

 

[jtaylor]

You can consider connecting the "GROUND" pin on the 555 to -5V. It'll
never know the difference, you know.

Yes, but how would I get a (say) drop from ground to -5V lasting for
1/4000th second , returning to 0V for 1/4000th second (1/2 cycle of 1kHz at
50%) followed by (say) a rise from 0V to +5V lasting for 1/8000th second,
returning to 0V for 3/8000ths of a scond (1/2 cycle of 1Khz at 25%)?

 

[Spehro Pefhany]

Yes, but how would I get a (say) drop from ground to -5V lasting for
1/4000th second , returning to 0V for 1/4000th second (1/2 cycle of 1kHz at
50%) followed by (say) a rise from 0V to +5V lasting for 1/8000th second,
returning to 0V for 3/8000ths of a scond (1/2 cycle of 1Khz at 25%)?

Okay. If you use a single 555 in astable mode you can steer the
currents through both ends of your two pots to get this effect by
using a FF (eg. 1/2 4013) and 2 complimentary pairs of transistors.

Driving the output to the 3 levels depends on how much drive you need.
One possibility would be to use 1/2 of a 4052. It could also be done
with discretes.


Best regards,
Spehro Pefhany

 

[Andrew Holme]

Subject says most of it; power requirements are small, frequency and some
pulse width drift/variation is no problem...

Two monostables?

| |
\ \
/ /
+--> \ VR1 +--> \ VR2
| / | /
| | | |
+----+ +----+
| |
+----o----+ +----o----+
| _ | | |
+--o CLK Q o-----o CLK Q o----+-------> out
| | | | | |
| +---------+ +---------+ |
| |
+---------------------------------+

 

[Rich Grise]

On Thu, 14 Oct 2004 15:06:03 -0300, the renowned "jtaylor"


Okay. If you use a single 555 in astable mode you can steer the
currents through both ends of your two pots to get this effect by
using a FF (eg. 1/2 4013) and 2 complimentary pairs of transistors.

Driving the output to the 3 levels depends on how much drive you need.
One possibility would be to use 1/2 of a 4052. It could also be done
with discretes.

I'd have done it with discretes. Eight of them, to be exact. (maybe ten,
if you don't want Rb to go to zero. ;-) )

Cheers!
Rich

 

[Andrew Holme]

Subject says most of it; power requirements are small, frequency and some
pulse width drift/variation is no problem...

Ignore my previous post - two monos might not start properly. Try this:

|\
+------| >O--------------+---> out
| |/ |
| Schmitt |
| |
| +---+ |
| | | |
| _V_ | |
+-----|___|-+---|<-------+
| VR1 D1 |
| |
| +---+ |
| | | |
| _V_ | |
+-----|___|-+--->|-------+
| VR2 D2
|
---
--- C
|
|
===
GND

 

[CFoley1064]

Subject: Re: Want independant pulse width variation on + and - side of

gnd;square wave of about 1kHz
From: Spehro Pefhany [email protected]
Date: 10/14/2004 1:58 PM Central Daylight Time
Message-id: <[email protected]>


Okay. If you use a single 555 in astable mode you can steer the
currents through both ends of your two pots to get this effect by
using a FF (eg. 1/2 4013) and 2 complimentary pairs of transistors.

Driving the output to the 3 levels depends on how much drive you need.
One possibility would be to use 1/2 of a 4052. It could also be done
with discretes.


Best regards,
Spehro Pefhany

I believe this is something close to what Mr. Pefhany is talking about. Unless
you want to change something else, this should do the job (view in fixed font
or M$ Notepad):



+5V
|
.-. +5V
| | | +5V 1/4 4066
| | .-. | .-----.
'-' 1/2 LM393| | | | |
| +5V| | '---oI Oo-.
.-. |\| '-' | C | |
| |<--------|-\ | '--o--' |
___ | | | >-o-----------' |
.--|___|----. '-' .-----|+/ |
| +5V | | | |/| |
| | | | | -5V o--o
| |\ | | | | |
.---o---|H>O----' o-----------o +5V | |
| | |/ | | | | 1/4 4066 | .-.
| | | | | | .-. .-----. | | |
| --- | | .----o | | | | | | | |
| --- -5V | | | | +5V| | .--oI Oo-' '-'
| | 1/6 74C14 | | .-. | |\| '-' | | C | |
| | |=== | |<--|-----|-\ | | '--o--' ===
| | |GND | | | | >-o-----|-----' GND
| -5V | '-' '-----|+/ |
| | | |/| |
|------------------' .-. -5V -5V
| | 1/2 LM393
| |
'-'
|
-5V
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Not elegant at all, but it will do the job. Split supplies, matched +/-5V. H
is 1/6 of a 74C14, which should oscillate at the frequency you determine with
the R and C. DC level of oscillations should be nearly 0V. Amplitude of
oscillation with +--5V supplies with a C14 (don't use a 40106 or other CMOS
schmitt trigger IC -- they're made differently and have lower hysteresis
voltage and less precise switching points) should be about +2V to -2V. The two
comparators pick off the signal from the cap, and are compared with the pot
settings. You can use those pots to set the time for +5V out (up to 50%) and
-5V out (up to 50%). These outputs go to half a 4052 or 4066, which are analog
switches that turn on the +5V or -5V output. When neither is on, you have 0V
(resistor to GND).

One immediately apparent way to make this better would be to use a quad op amp
instead of the dual comparator and the C14, to ensure symmetry around GND for
the oscillator and design in a fixed amplitude even if the supplies are not
matched. The OP is talking about a couple of KHz, so the rise time of a good
op amp shouldn't be too significant here. Use two of the other 4 op amps as
comparators, and lose the pullup resistors.

Oh, yes, and would you like fries with that?

(Note the trouble you can get yourself into on newsgroups while you're waiting
for a customer to call.)

Good luck
Chris

 

[CFoley1064]

I believe this is something close to what Mr. Pefhany is talking about.

Unless
you want to change something else, this should do the job (view in fixed font
or M$ Notepad):



+5V
|
.-. +5V
| | | +5V 1/4 4066
| | .-. | .-----.
'-' 1/2 LM393| | | | |
| +5V| | '---oI Oo-.
.-. |\| '-' | C | |
| |<--------|-\ | '--o--' |
___ | | | >-o-----------' |
.--|___|----. '-' .-----|+/ |
| +5V | | | |/| |
| | | | | -5V o--o
| |\ | | | | |
.---o---|H>O----' o-----------o +5V | |
| | |/ | | | | 1/4 4066 | .-.
| | | | | | .-. .-----. | | |
| --- | | .----o | | | | | | | |
| --- -5V | | | | +5V| | .--oI Oo-' '-'
| | 1/6 74C14 | | .-. | |\| '-' | | C | |
| | |=== | |<--|-----|-\ | | '--o--' ===
| | |GND | | | | >-o-----|-----' GND
| -5V | '-' '-----|+/ |
| | | |/| |
|------------------' .-. -5V -5V
| | 1/2 LM393
| |
'-'
|
-5V
created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

Not elegant at all, but it will do the job. Split supplies, matched +/-5V.
H
is 1/6 of a 74C14, which should oscillate at the frequency you determine with
the R and C. DC level of oscillations should be nearly 0V. Amplitude of
oscillation with +--5V supplies with a C14 (don't use a 40106 or other CMOS
schmitt trigger IC -- they're made differently and have lower hysteresis
voltage and less precise switching points) should be about +2V to -2V. The
two
comparators pick off the signal from the cap, and are compared with the pot
settings. You can use those pots to set the time for +5V out (up to 50%) and
-5V out (up to 50%). These outputs go to half a 4052 or 4066, which are
analog
switches that turn on the +5V or -5V output. When neither is on, you have 0V
(resistor to GND).

One immediately apparent way to make this better would be to use a quad op
amp
instead of the dual comparator and the C14, to ensure symmetry around GND for
the oscillator and design in a fixed amplitude even if the supplies are not
matched. The OP is talking about a couple of KHz, so the rise time of a good
op amp shouldn't be too significant here. Use two of the other 4 op amps as
comparators, and lose the pullup resistors.

And, if you want the + part to be able to go to 100% or the - part to be able
to go to 100%, you'll have to play with the pots a little. This is the basic
idea, though. Oh, yes. As always, you can use a PIC to drive these outputs at
any duty cycle you choose. If you need some power, just connect a rail-to-rail
op amp as a voltage follower at the output.

Chris

 

[Spehro Pefhany]

I believe this is something close to what Mr. Pefhany is talking about. Unless
you want to change something else, this should do the job (view in fixed font
or M$ Notepad):

<snip>

That's pretty simple. I was thinking of something more along the lines
of:

+5V +5V
o o
| |
| |
.-------. .-------.
| 4066 | Pin 6,2 | 4066 |
State 0-----| 1/4 | LMC555 | 1/4 |---- State 2
'-------' o '-------'
| | |
Ton + .-. | .-. Ton-
| |<-----+------>| |
| | | | |
Toff + '-' --- '-' Toff-
| --- |
.-------. | .-------.
| 4066 | o | 4066 |
State 1-----| 1/4 | -5V | 1/4 |---- State 3
'-------' '-------'
| |
| |
o o
-5V -5V

Plus the 555 itself, the decoding logic for the states and the output.

The above should require no adjustment (though if it was required, a
trim could be made using pin 5 on the 555, either a single trim to
compensate for capacitor variation or a dual-level trim to compensate
for the differences in the pot elements). It should be very linear
with pot resistance. Also no other passive components except perhaps
an output resistor to ground depending on how that bit is done. Using
1/2 a 4052 and a 4-resistor network could take care of the decoding
and the second half could drive the output positively to each of the
three levels. So, parts count:

1 - 4066
1 - LMC555
1 - 4052
1 - 4013 (1/2 used)
1 - 4-resistor network
2 - pots
1 - capacitor
-----------
8 (+ bypass caps)




Best regards,
Spehro Pefhany

 

[John Fields]

Yes, but how would I get a (say) drop from ground to -5V lasting for
1/4000th second , returning to 0V for 1/4000th second (1/2 cycle of 1kHz at
50%) followed by (say) a rise from 0V to +5V lasting for 1/8000th second,
returning to 0V for 3/8000ths of a scond (1/2 cycle of 1Khz at 25%)?

---
Here's a 1000 cycle square wave:

|<------500µs------>|<------500µs------>|
___ ___________________
|___________________| |_____


Then, plugging in your curious timing units, we get:

|<-----1/2000s----->|<-----1/2000s----->|
___ ___________________
|___________________| |_____

|----|----|----|----|----|----|----|----|

-->| |<---1/8000s



And, finally, if what you're asking for, above, is this:

___ ___________________
SQIN |___________________| |_____


____
+5 ___ _________| |______________
OUT |_________| |_____
-5

TIME |----|----|----|----|----|----|----|----|

-->| |<---1/8000s



This will get it for you:

+5V
|
+----+-----+--------------+
| | | |
+5V | | +--+--+ |
| +->[RV1] | | VDD | |
+--+--+ | | | |4538A| |
| Vcc | +----+-------|RC | |
|7555 | | | | | |
| OUT|---+----+-------|-------|IN | |
| | | | | | |__ _| E
| | [R1] | [C2] +-O|IN Q|---[1K]---B 2N4403
| __| | | | | | C
| TR|---+ | | | | |
| | | | | +-----+ [100R]
| TH|---+ | -5V |
| GND | | | |
+--+--+ [C1] | +--------+-->OUT
| | | | |
-5V -5V | +5V | [1000R]
| | +-----+ [100R] |
| +->[RV2] | | | GND(0V)
| | | |4538B| C
| +----+-------|RC Q|---[1K]---B 2N4401
| | | | E
| [C3] +--|IN | |
| | | |__ | |
+--------------O|IN | |
| | | | |
| | | VEE | |
| | +--+--+ |
| | | |
+----+-----+--------------+
|
-5

Select R1 and C1 to give you a 1000Hz output from the 7555.

Select RV1C2 and RV2C3 to give you a maximum pulse output width from
either 4538A or 4538B of 500µs. If you go over 500µs for either one,
or the output frequency from the 7555 is >1000Hz or the output of
either half-period from the 7555 is <500µs, The output totem-pole
will short, you'll get pretty close to zero volts out of the output,
and the only thing limiting the current through the transistors will
be the two 100 ohm resistors. With 10V across the transistors, if
they're saturated the resistors will limit the current through them to
about 50mA, which is no big deal but, if you need/want to change the
resitor values in order to get more current into your load, you'll
want to be careful not to fry the two output transistors with an
inadvertent input drive overlap.

The way it's set up now, with only the 1000 ohm resistor for a load,
the outout voltage will go from a high of +4.5V with 2N4403 on and the
2N4401 off, to ~0V with them both off, (or on!) to a low of -4.5V
with the 2N4403 off and the 2N4401 on.