one bit of memory?

[Thomas A. Horsley]

In my ongoing quest to waste as much time as possible on my silly
phone gadget (http://home.att.net/~Tom.Horsley/phonetale/markII.html)
I added a circuit to "remember" if the phone ever rang so I can
turn on a flashing LED as a "missed call" reminder.

For memory I used the LM339 as OR-gate sample circuit from the
NS pdf spec sheet, and fed back the output as one of the inputs
with the output from the 555 timer being the other input.

This worked great till I decided to switch to the much lower power
LM339 near-compatible TLC3704 chip from Texas Instruments
(http://focus.ti.com/docs/prod/folders/print/tlc3704.html).

Apparently the push-pull outputs work differently enough that the
initial power-up state of the chip is "on" rather than "off", so
my one bit memory circuit doesn't work with the new chip very
well (but I don't want to go back to the 339 because overall
the new circuit uses a lot less power than the old one and I don't
need pull-up resistors all over the place).

Anyone have any suggestions for a nice low power way to provide
the one bit of memory I need? Or maybe a way to hack around the
initial power-up state so I ignore that particular bit of feedback,
but notice it later when I really want it?

(Of course if I really wanted to save power, I wouldn't add all this
extra junk anyway - the important bits seem to be working fine
and only draw about 0.5mA - about 10 times less than the first version).
--

==>> The *Best* political site <URL:http://www.vote-smart.org/> >>==+
email: Tom.Horsley@worldnet.att.net icbm: Delray Beach, FL |

<URL:http://home.att.net/~Tom.Horsley> Free Software and Politics <<==+

 

[Michael Black]

Thomas A. Horsley (tom.horsley@att.net) writes:

In my ongoing quest to waste as much time as possible on my silly
phone gadget (http://home.att.net/~Tom.Horsley/phonetale/markII.html)
I added a circuit to "remember" if the phone ever rang so I can
turn on a flashing LED as a "missed call" reminder.

For memory I used the LM339 as OR-gate sample circuit from the
NS pdf spec sheet, and fed back the output as one of the inputs
with the output from the 555 timer being the other input.

This worked great till I decided to switch to the much lower power
LM339 near-compatible TLC3704 chip from Texas Instruments
(http://focus.ti.com/docs/prod/folders/print/tlc3704.html).

Apparently the push-pull outputs work differently enough that the
initial power-up state of the chip is "on" rather than "off", so
my one bit memory circuit doesn't work with the new chip very
well (but I don't want to go back to the 339 because overall
the new circuit uses a lot less power than the old one and I don't
need pull-up resistors all over the place).

Anyone have any suggestions for a nice low power way to provide
the one bit of memory I need? Or maybe a way to hack around the
initial power-up state so I ignore that particular bit of feedback,
but notice it later when I really want it?

(Of course if I really wanted to save power, I wouldn't add all this
extra junk anyway - the important bits seem to be working fine
and only draw about 0.5mA - about 10 times less than the first version).

An SCR. Once triggered, it will continue conducting until power is
removed from the anode, at which point it waits for another trigger.

Or for more complication, use two transistors as a flip-flop.

Michael

 

[Robert Monsen]

Thomas A. Horsley wrote:

In my ongoing quest to waste as much time as possible on my silly
phone gadget (http://home.att.net/~Tom.Horsley/phonetale/markII.html)
I added a circuit to "remember" if the phone ever rang so I can
turn on a flashing LED as a "missed call" reminder.

For memory I used the LM339 as OR-gate sample circuit from the
NS pdf spec sheet, and fed back the output as one of the inputs
with the output from the 555 timer being the other input.

This worked great till I decided to switch to the much lower power
LM339 near-compatible TLC3704 chip from Texas Instruments
(http://focus.ti.com/docs/prod/folders/print/tlc3704.html).

Apparently the push-pull outputs work differently enough that the
initial power-up state of the chip is "on" rather than "off", so
my one bit memory circuit doesn't work with the new chip very
well (but I don't want to go back to the 339 because overall
the new circuit uses a lot less power than the old one and I don't
need pull-up resistors all over the place).

Anyone have any suggestions for a nice low power way to provide
the one bit of memory I need? Or maybe a way to hack around the
initial power-up state so I ignore that particular bit of feedback,
but notice it later when I really want it?

(Of course if I really wanted to save power, I wouldn't add all this
extra junk anyway - the important bits seem to be working fine
and only draw about 0.5mA - about 10 times less than the first version).

According to your circuit schematic, your comparator has negative
feedback. That means it isn't acting as a comparators, it's acting as an
(uncompensated) opamp. This may well result in oscillations.

Here is a circuit that may work. It's a starting point for some
experimentation anyway. It's a flipflop, with the comparator used to
increase the impedance of the output. If your comparator is open
collector, you may need a pullup resistor on the output.

VCC=9V
.------------------o-------o-----------.
| | | |
| | | |
.-. .-. .-. |
| |100k 100k | | | |47k |
| | | | | | |
'-' '-' '-' |
| | | |\|
| o-------)---------|-\ Output
| ___ ___ | | | >------.
From 555 o-|___|-. .-|___|-o---. o-----o---|+/ |
___ | 1MEG | | 1MEG | | | | |/| |
->|--|___|--o----)-------o | |/ | | | ___ |
100k | \| '--)-----| | | '---|___|-----'
| |--------' |> | | | 1MEG
--- <| | | .-. |
1uF --- | | | | |47k |
| | | | | | | Rang=Low
| | | | '-' |
| | | | | |
'----o------------------o---)---o-------o---'
GND | |
| |
| |
| T |
| --- |
'----o o----'
Reset

(created by AACircuit v1.28 beta 10/06/04 www.tech-chat.de)


--
Regards,
Robert Monsen

"Your Highness, I have no need of this hypothesis."
- Pierre Laplace (1749-1827), to Napoleon,
on why his works on celestial mechanics make no mention of God.

 

[Tom Horsley]

An SCR. Once triggered, it will continue conducting until power is
removed from the anode, at which point it waits for another trigger.

Or for more complication, use two transistors as a flip-flop.

Rge SCR sounds good - I found a web page describing them with
an example equivalent circuit, and it looks like two transistors
setup as a flip-flop (or really just a "flip" is the same
as an SCR. I'll have to play around with some things and see if I
can get it to work.

Thanks.

 

[Tom Horsley]

According to your circuit schematic, your comparator has negative feedback. That means it isn't acting as a comparators, it's
acting as an (uncompensated) opamp. This may well result in oscillations.

Yeah, I think I noticed I got the diagram backwards, but the actual
breadboard has 'em hooked up the right way.

Here is a circuit that may work...

Thanks. I'll play around with it and see what I can get to happen.

 

[Thomas A. Horsley]

Just to let the folks who provided suggestions know, everything is working
now. The new and improved web page has the gory details:

http://home.att.net/~Tom.Horsley/markII/markII.html

Thanks for all the help! (Now I just need to decide on the mechanical
design and get a new box of some kind built and get the circuit board
all laid out and put together - heck I can still waste weeks on this
thing .
--

==>> The *Best* political site <URL:http://www.vote-smart.org/> >>==+
email: Tom.Horsley@worldnet.att.net icbm: Delray Beach, FL |

<URL:http://home.att.net/~Tom.Horsley> Free Software and Politics <<==+