Searching \ for '[EE] Electronic seal for microamp application?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=electronic+seal
Search entire site for: 'Electronic seal for microamp application?'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Electronic seal for microamp application?'
2007\01\18@190303 by Peter Todd

picon face
I've been tasked with making a very simple device that will turn an LED
on for one minute every day. I'm using a 16f683 with a DS32khz for the
timekeeping and have designed a nice circuit that seems to use about 5uA
sleeping and keeps accurate time.

The application also needs to have a button, that when pressed starts
the device up and starts it's led on, led off routine. Before that I
want the device to be completely off, drawing no current from the
battery, after pressing the button, it should apply power to the DS32khz and
PIC. This should be a latch, so pressing the button again has no further
effect.

I've come up with a little bistable flip-flop that seems to do this, see
http://gw.petertodd.ca/~pete/bistable.png When power is first applied
the capacitor along with the resistor makes sure that MOSFET T2 is off,
this causes T1 to be turned on. When the switch is pressed, T1 turns
off, T2 turns on and the load is activated. Pretty simple.

Of course this uses 0.3uA when off due to discharge of the 10Mohm
resistor. I'm also worried that EMF could trigger the circuit,
especially with that 10Mohm resistor.

Anyone built anything similar? I know a lot of Maxim RTC parts say that
the internal batteries are protected by an "electronic seal" that only
activates upon the first application of power.

--
http://www.petertodd.ca

2007\01\18@204752 by Jinx

face picon face
part 1 898 bytes content-type:text/plain; (decoded 7bit)

> Anyone built anything similar ?

Peter, I'm experimenting with the attached. A work-in-progress
so values aren't finalised. Why not do what I'm doing and have
the PIC control the load and also a time-out to ignore the button ?

I had two goals for this circuit - to reduce current (previously it
used 10uA when switch was left closed) and to increase niose
immunity

A problem I had was that one particular unit is installed outside
and would be falsely triggered all the time a neighbour was using
a very noisy 2-stroke weedeater. He weedeated again a couple of
days ago, after changes had been made to the circuit, and the unit
ignored it. It even performed its real function whilst being bombarded
at close range (< 1m) by spark plug EMF onto the switch wires

You could reduce the power consumption down to 60-ish nA by
using a kitchen clock module



part 2 2238 bytes content-type:image/gif; (decode)


part 3 1662 bytes content-type:image/gif; (decode)


part 4 35 bytes content-type:text/plain; charset="us-ascii"
(decoded 7bit)

2007\01\18@231214 by Sergey Dryga

face picon face
Peter Todd <pete <at> petertodd.ca> writes:

>
> I've been tasked with making a very simple device that will turn an LED
> on for one minute every day. I'm using a 16f683 with a DS32khz for the
> timekeeping and have designed a nice circuit that seems to use about 5uA
> sleeping and keeps accurate time.
>
> The application also needs to have a button, that when pressed starts
> the device up and starts it's led on, led off routine. Before that I
> want the device to be completely off, drawing no current from the
> battery, after pressing the button, it should apply power to the DS32khz and
> PIC. This should be a latch, so pressing the button again has no further
> effect.

Peter,
how about RS flip-flop using 74HCT logic or similar low power one? Off-current
consumption should be very small.

Sergey Dryga


2007\01\19@083844 by Peter P.

picon face

You did not say it has to be cheap. A PFET and a NFET can build you a nice power
switch with zero drain both when off and when on. E.g. D7000 + ??? PFET (small
too).

Peter P.


2007\01\19@122217 by M. Adam Davis

face picon face
Out of curiosity (I like to have examples for when standard practices
don't meet the requirements) what about this application precludes
using a small plastic strip installed between the battery or cell and
the contact that is pulled out when one wants to activate the unit?
Given the button, mechanical action is necessary to start it, so why
must it be a push motion instead of a grab, pull, and discard motion?

I don't know if this idea would work or not, but note that some PIC
pins don't have protection diodes, but can still be used as I/O.
Attach the power input to an I/O without protection diodes, then use a
button to temporarily power the PIC which turns the I/O output high.
The I/O then shunts the power into the positive rail.

I haven't thought through the physics of running the I/O backwards
like that - it may be a non-starter right out of the gate.  In other
words, "This exercise left to the reader."

-Adam

On 1/18/07, Peter Todd <spam_OUTpeteTakeThisOuTspampetertodd.ca> wrote:
{Quote hidden}

> -

2007\01\19@122812 by M. Adam Davis

face picon face
Nevermind, the only one without a protection diode is MCLR, and that
can only be an input with an enabled weak pullup.  If your current
needs were very very low, maybe that weak pull up would still work.
Would be interesting to check out and try, anyway...

-Adam

On 1/19/07, M. Adam Davis <.....stienmanKILLspamspam@spam@gmail.com> wrote:
{Quote hidden}

2007\01\21@003922 by Peter Todd

picon face
On Fri, Jan 19, 2007 at 02:47:34PM +1300, Jinx wrote:
> > Anyone built anything similar ?
>
> Peter, I'm experimenting with the attached. A work-in-progress
> so values aren't finalised. Why not do what I'm doing and have
> the PIC control the load and also a time-out to ignore the button ?

It's funny how you can totally ignore the obvious, simple, solution...

I did that and the circuit works fine with a very low storage current.
Schematic is here if you are interested:
http://petertodd.ca/persist/2007-01-21/moment-schematic.png

For the switch the PIC initially turns on the soft pull-up for GP2/INT
and waits for an interrupt on the falling edge. Once woken up it turns
the leds on to half brightness, then waits for 10 seconds of continuous
pressing. If the user doesn't let go, the leds kick into full brightness
and the rest of the program runs. Upon wake up I have the pic set GP2 as
an output again, pulled low in case the switch is touched again. I also
turn the DS32khz on via GP4.

I put this circuit next to the spot and arc welders at my school while
they were running... No activations. Looks the delay should be enough to
filter out noise. I assume in your circuit you put the resistor and
capacitor in there to do the same, but in hardware?


It's funny, I realised this is the first PIC device I've ever made who's
code doesn't consist of a configuration phase followed by a single while
(1) loop...

{Quote hidden}

Out of curiosity, what does your circuit do anyhow?

Judging by the alkaline batteries, I assume you aren't aiming for a 10
year battery life. :)

--
http://www.petertodd.ca

2007\01\21@004033 by Peter Todd

picon face
On Fri, Jan 19, 2007 at 01:38:13PM +0000, Peter P. wrote:
>
> You did not say it has to be cheap. A PFET and a NFET can build you a nice power
> switch with zero drain both when off and when on. E.g. D7000 + ??? PFET (small
> too).

Got a link to an explanation? I couldn't figure out what configuration
to put them in for zero drain when I was playing around.

--
http://www.petertodd.ca

2007\01\21@004431 by Peter Todd

picon face
On Fri, Jan 19, 2007 at 12:22:11PM -0500, M. Adam Davis wrote:
> Out of curiosity (I like to have examples for when standard practices
> don't meet the requirements) what about this application precludes
> using a small plastic strip installed between the battery or cell and
> the contact that is pulled out when one wants to activate the unit?
> Given the button, mechanical action is necessary to start it, so why
> must it be a push motion instead of a grab, pull, and discard motion?

Easier mechanical design really. I also figured the device might be more
reliable in the long term if once activated, there is absolutely no
contacts or what not that could concievably corrode. As it is the
battery will be one with solder tabs on it, directly attached to the
PCB, and the whole thing (minus the switch!) will probably end up
encapsulated in polyester resin.

Aw heck, just reminds me, gotta find out if polyester resin is at all
conductive!

> I don't know if this idea would work or not, but note that some PIC
> pins don't have protection diodes, but can still be used as I/O.
> Attach the power input to an I/O without protection diodes, then use a
> button to temporarily power the PIC which turns the I/O output high.
> The I/O then shunts the power into the positive rail.
>
> I haven't thought through the physics of running the I/O backwards
> like that - it may be a non-starter right out of the gate.  In other
> words, "This exercise left to the reader."

Great minds think alike, actually I think what you are describing is on
the piclist site already, under power supplies.

--
http://www.petertodd.ca

2007\01\21@042309 by Jinx

face picon face

> Aw heck, just reminds me, gotta find out if polyester resin is at all
> conductive!

It isn't, but free styrene is pretty aggressive towards most plastics.
I'd suggest pre-coating the circuit with lacquer (eg nitrocellulose)
first if in doubt

2007\01\21@045556 by Jinx

face picon face
> It's funny how you can totally ignore the obvious, simple, solution...

Oh many's the time I've started off with a simple circuit, made it
more and more complicated and eventually come full circle back
to a simple solution

> http://petertodd.ca/persist/2007-01-21/moment-schematic.png

Thanks

> I put this circuit next to the spot and arc welders at my school
> while they were running... No activations. Looks the delay should
> be enough to filter out noise. I assume in your circuit you put the
> resistor and capacitor in there to do the same, but in hardware ?

Well originally I had just a large (390k) pull-down and 10k + switch
to Vcc. That worked but INT was not being pulled down hard enough
and was prone to false trigering. In my product the switch may be
closed for a fraction of a second or for several days. Although 10uA
isn't really a killer (permanently closed, 2800mAh capacity -> 31 years)
there was still the problem of noise immunity

Simply reducing the 390k would have pushed up consumption. 100k
would have given reasonable battery life, but still too high. Then I
thought of using a cap to store Vcc and dump into a low-value pull-
down as a pulse. That actually functions as a debounce filter (although
bounce is not an issue here and wasn't the intention of the mods) but
most importantly blocks DC if the switch is left closed

I daresay the 3M3 could be substantially increased, decreasing current
further, but there's little point really. At 1.6uA worst-case, that's still
many times the life of these batteries, almost 200 years

> Out of curiosity, what does your circuit do anyhow?

Can't say (yet), sorry. When it goes on the shelves I'll tell

> Judging by the alkaline batteries, I assume you aren't aiming for a 10
> year battery life. :)

See above. With the consumption it has now, smaller batteries could
be used (AAAA or coins) but they're relatively expensive compared
with AA

2007\01\21@053518 by Vasile Surducan

face picon face
On 1/19/07, Peter Todd <.....peteKILLspamspam.....petertodd.ca> wrote:
> I've been tasked with making a very simple device that will turn an LED
> on for one minute every day.

What for ?

Vasile

2007\01\22@150027 by Peter P.

picon face
Peter Todd <pete <at> petertodd.ca> writes:


> Got a link to an explanation? I couldn't figure out what configuration
> to put them in for zero drain when I was playing around.

Use fixed font to view:

            PFET
batt+ --+--- S G D -----+--- Vsw
       |      |        |
       +-C1---+        |
              |        |
              |        |
              D        |
          NF  G--------+
              S        |
              |        C2
              |        |
              +--------+
              |
             GND

C1 prevents spurious switch on C2 is a filter cap on the switched side and
serves the same function for NFet as C1 for PFET does. The start switch bridges
the PFET DS. You can tie the gate of NFet to a microprocessor output (use a
pullup 10Meg). A bleeder resistor across Vsw and GND may be ncessary to achieve
reset (turn off).

Peter P.

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...