Solar Power part deux
Archiver
Posts: 46,084
Hello Everyone,
Here is an update of my bell project and... a new set of problems
In my project I have a stamp powered by a Solarverter Panel (9V, 150ma) in
tandem with a Panasonic 6V 1.3 AH sealed acid battery. The stamp powers a
RF transceiver and the there is a servo hooked in parallel to the set up. A
LT1121CS8-ND regulates the power and avoids any power leakage back to the
solar panels at night... Additionally I have a NPN and PNP set up powering
the RF transceiver as needed. Why? the transceiver is very sensitive and
will not run on power supplied by a Stamp pin set to HIGH... it really needs
as close to 5V as possible. I am providing that out of the stamps pwr pin
and use another pin to turn it on or off via the PNP and NPN set up...
Finally, there is the servo. I am having trouble with it. It makes my
stamp reset whenever it tries to power up and the voltage has dropped over a
certain amount. When that happens the whole system spirals out of control,
and is unable to set itself to sleep and uses up all of my battery... so
taking a cue from the list, I am monitoring the voltage levels of the
battery using the RCtime command. The idea is that if I can figure out when
the voltage is to low the stamp will forgo activating the servo. AND in my
attempts to isolate the servo from the system I am driving it through a ULN
2803A, one pin controls the servo while another pin 'powers it up' through
the ULN 2803A... Phew! So much for a simple solution -but it continues to
be fun.
Here is the set up:
sleep
wake up
check voltage
if OK power up Transceiver and wait for transmission else go to sleep for 15
minutes
when transmission arrives...
check voltage again
if OK run servo, strike bell else Joto sleep for 15 minutes
LT1121CS8
| + | | |
solar | + | |
verter Battery |
|-1---ULN2803A
servo
| _ | _ | | |
| | |stamp|-0
|
| | | | PNP |
| | | |-2---NPN---Trans.----|
| | | | | |
| | |_____|--pwr-| |
| | | |
Is he use of the ULN 2803A misguided?
is there a better way to isolate the servo?
Does it sound reasonable to monitor the Voltage and decide to run the servo
based on that?
I am using the set up recommended by Tracy Allen:
470 ohm 100 kohm
P0 ----/\/\/\---o
/\/\/\----Vx unknown voltage
|
===== 0.1 uF film
|
Vss
I use his code without a loop test program is:
tx var word ' RCtime value
vx var word ' voltage value to be determined
Cn1 con 65000 ' first constant
Cn2 con 14 ' second constant
checkvolt:
low 0
RCtime 0,0,tx
low 0
vx=Cn1 / tx - Cn2 / 2
'if tx> than a certain amount return
'else sleep 15 minutes and back to checkvolt
I am not even looking at vx and I am simply logging tx... but the results
seem inconclusive. Sometimes the Stamp resets at a certain tx but at other
times it seems fine at the same tx?
Any ideas anyone? Is there something I am overlooking?
Thanks in advance
Paul Ramirez Jonas
451 Broome Street #9W
New York NY 10013
Tel/fax 212.941.9941
Mobile 917.744.2530
email prjnyc@m...
Here is an update of my bell project and... a new set of problems
In my project I have a stamp powered by a Solarverter Panel (9V, 150ma) in
tandem with a Panasonic 6V 1.3 AH sealed acid battery. The stamp powers a
RF transceiver and the there is a servo hooked in parallel to the set up. A
LT1121CS8-ND regulates the power and avoids any power leakage back to the
solar panels at night... Additionally I have a NPN and PNP set up powering
the RF transceiver as needed. Why? the transceiver is very sensitive and
will not run on power supplied by a Stamp pin set to HIGH... it really needs
as close to 5V as possible. I am providing that out of the stamps pwr pin
and use another pin to turn it on or off via the PNP and NPN set up...
Finally, there is the servo. I am having trouble with it. It makes my
stamp reset whenever it tries to power up and the voltage has dropped over a
certain amount. When that happens the whole system spirals out of control,
and is unable to set itself to sleep and uses up all of my battery... so
taking a cue from the list, I am monitoring the voltage levels of the
battery using the RCtime command. The idea is that if I can figure out when
the voltage is to low the stamp will forgo activating the servo. AND in my
attempts to isolate the servo from the system I am driving it through a ULN
2803A, one pin controls the servo while another pin 'powers it up' through
the ULN 2803A... Phew! So much for a simple solution -but it continues to
be fun.
Here is the set up:
sleep
wake up
check voltage
if OK power up Transceiver and wait for transmission else go to sleep for 15
minutes
when transmission arrives...
check voltage again
if OK run servo, strike bell else Joto sleep for 15 minutes
LT1121CS8
| + | | |
solar | + | |
verter Battery |
|-1---ULN2803A
servo
| _ | _ | | |
| | |stamp|-0
|
| | | | PNP |
| | | |-2---NPN---Trans.----|
| | | | | |
| | |_____|--pwr-| |
| | | |
Is he use of the ULN 2803A misguided?
is there a better way to isolate the servo?
Does it sound reasonable to monitor the Voltage and decide to run the servo
based on that?
I am using the set up recommended by Tracy Allen:
470 ohm 100 kohm
P0 ----/\/\/\---o
/\/\/\----Vx unknown voltage
|
===== 0.1 uF film
|
Vss
I use his code without a loop test program is:
tx var word ' RCtime value
vx var word ' voltage value to be determined
Cn1 con 65000 ' first constant
Cn2 con 14 ' second constant
checkvolt:
low 0
RCtime 0,0,tx
low 0
vx=Cn1 / tx - Cn2 / 2
'if tx> than a certain amount return
'else sleep 15 minutes and back to checkvolt
I am not even looking at vx and I am simply logging tx... but the results
seem inconclusive. Sometimes the Stamp resets at a certain tx but at other
times it seems fine at the same tx?
Any ideas anyone? Is there something I am overlooking?
Thanks in advance
Paul Ramirez Jonas
451 Broome Street #9W
New York NY 10013
Tel/fax 212.941.9941
Mobile 917.744.2530
email prjnyc@m...
Comments
> It makes my stamp reset whenever it tries to power up...
>
LT1121CS8
> | + | | |
>solar | + | |
>verter Battery |
|-1---ULN2803A
servo
> | _ | _ | | |
> | | |stamp|-0
|
> | | | | PNP |
> | | | |-2---NPN---Trans.----|
> | | | | | |
> | | |_____|--pwr-| |
> | | | |
>
Hi Paul,
Probably the most important thing is to route the the servo power supply
leads directly back to the battery. That goes for both the + and -. A
"star" configuration:
;
1121-5?
;
| +1uf |
LT1121CS8
o----o
|
| + | | | | |
solar \ | + | | |
verter \ Battery | |
|-1
servo
| _ \ | _ | | | |
| \ | | |stamp|-0
|
| \ | | | | PNP |
| \ | ===== | |-2---NPN---Trans.----|
| \ | |4.7 | | | |
| \ | | |_____|--pwr-| |
| \ | | | |
o----o----o
' |
gound|mecca |
'
The current that powers the servo should never flow in the same wires as
the power to the stamp. You need to provide a separate regulator for the
servo, because using the one on the BS2 is simply too problematic. The
reference point for the LT1121 should also connect at the "star" point, no
servo current in the reference lead. Include an extra capacitor across the
power supply near the stamp, something like 4.7 microfarads. Also,
connect a resistor of about 3kohms from pin 4 of the BS2 (atn) to ground.
That is a sensitive sneak path for resets. I doubt the ULN2803A helps.
>I am not even looking at vx and I am simply logging tx... but the results
>seem inconclusive. Sometimes the Stamp resets at a certain tx but at
other
>times it seems fine at the same tx?
> ..
> 2- I must have misunderstood something about Tracy Allen's
> setup because mytx does not seem to reflect the actual readings
> I am making with my meter.
The tx value is the inverse of the voltage, so when the battery voltage
goes up, the tx value goes down. Is that what you mean? The formula,
> vx=Cn1 / tx - Cn2 / 2
converts the RCtime value into a voltage, but you have to select the Cn1
and Cn2 values for your setup.
The resets due to ground-loop noise can seemingly be influenced by the
phase of the moon!
I hope that helps,
Tracy Allen
Electronically Monitored Ecosystems
http://www.emesystems.com
Thanks for the response... I am confused as to why a star configuration is
different than regular parallel, but I am trying it right now.
i.e. isn't:
X
+
| | | | |
+ | | - | | |
o
Y
o
the same as X Y Z
| | | | |
| | | | |
Z
-
Can't I get away without a 1121-5 for the servo? My worry is that it will
create one more place for power to get guzzled?
As for the RCtime problem... This is a sample of what I have:
tx returns my meter reads
1337 5.42 volts
1342 5.45 volts
1347 5.64 volts
1360 5.66 volts
1364 5.60 volts
1376 5.58 volts
1406 5.44 volts
1455 5.31 volts
hmmm... what am I doing wrong? or am I expecting too much? I am using a
470 ohm and a 100 K ohm resistor and a 0.1 uF Polyester Film capacitor...
Thanks so much for your help.
Paul Ramirez Jonas
451 Broome Street #9W
New York NY 10013
Tel/fax 212.941.9941
Mobile 917.744.2530
email prjnyc@m...
>From: Tracy Allen <emesys@c...>
>To: "INTERNET:basicstamps@egroups.com" <basicstamps@egroups.com>
>Subject: [noparse][[/noparse]basicstamps] Solar Power part deux
>Date: Thu, Mar 30, 2000, 02:40 PM
>
>Hi Paul,
> Probably the most important thing is to route the the servo power supply
>leads directly back to the battery. That goes for both the + and -. A
>"star" configuration:
>
> ;
1121-5?
;
> | +1uf |
>
LT1121CS8
o----o
|
> | + | | | | |
>solar \ | + | | |
>verter \ Battery | |
|-1
servo
>
> | _ \ | _ | | | |
>
> | \ | | |stamp|-0
|
>
> | \ | | | | PNP |
>
> | \ | ===== | |-2---NPN---Trans.----|
>
> | \ | |4.7 | | | |
> | \ | | |_____|--pwr-| |
> | \ | | | |
>
o----o----o
' |
> gound|mecca |
> '
>
>
>The current that powers the servo should never flow in the same wires as
>the power to the stamp. You need to provide a separate regulator for the
>servo, because using the one on the BS2 is simply too problematic. The
>reference point for the LT1121 should also connect at the "star" point, no
>servo current in the reference lead. Include an extra capacitor across the
>power supply near the stamp, something like 4.7 microfarads. Also,
>connect a resistor of about 3kohms from pin 4 of the BS2 (atn) to ground.
>That is a sensitive sneak path for resets. I doubt the ULN2803A helps.
>
>>I am not even looking at vx and I am simply logging tx... but the results
>>seem inconclusive. Sometimes the Stamp resets at a certain tx but at
>other
>>times it seems fine at the same tx?
>> ..
>> 2- I must have misunderstood something about Tracy Allen's
>> setup because mytx does not seem to reflect the actual readings
>> I am making with my meter.
>
>The tx value is the inverse of the voltage, so when the battery voltage
>goes up, the tx value goes down. Is that what you mean? The formula,
>> vx=Cn1 / tx - Cn2 / 2
>converts the RCtime value into a voltage, but you have to select the Cn1
>and Cn2 values for your setup.
>
>The resets due to ground-loop noise can seemingly be influenced by the
>phase of the moon!
>
>I hope that helps,
>
> Tracy Allen
> Electronically Monitored Ecosystems
> http://www.emesystems.com
>
>
>
>
>eGroups.com home: http://www.egroups.com/group/basicstamps
>http://www.egroups.com - Simplifying group communications
>
>
>
>
>different than regular parallel, but I am trying it right now.
>i.e. isn't:
>
>
X
+
> | | | | |
>+ | | - | | |
>
o
Y
o
the same as X Y Z
> | | | | |
> | | | | |
>
Z
-
>
Hi Paul,
I'd draw it as follows:
+
;
| | | / |
| | | + --o
; |
| | | | | |
X Y Z X Y Z
| | | | | |
| | | - --o
' |
| | | \ |
-
'
On the right, the power supply to "Y" and to "Z" come directly from "X",
without much shared wire. The difference is that wires are not just wires.
They are small but finite resistances and inductances.
+ ----/\/\
/\/\
;
/\/\---;
| | | + --o---/\/\----; |
| | | | | |
X Y Z X Y Z
| | | | | |
| | | - --o---/\/\----' |
-
/\/\
/\/\
'
=----/\/\----'
Suppose "Z" is a servo and "Y" is a BASIC Stamp and "X" is the battery.
In the diagram on the left, the servo current has a strong effect on the
power supply to Stamp--the servo current has to flow through the wiring
resistance (and inductance) that is in series with both of them. On the
right in contrast, the wiring resistance that goes to the Stamp does not
have to carry the servo's relatively heavy current.
When your servo is connected to the stamp, which in turn is connected to
the main power supply, you have a chain like the figure on the left. The
servo in operation draws large spikes of current, which amounts to lots of
noise on the power supply line.
One way to deal with this is to put a huge capacitor (say 3300 uf) in
parallel with the servo, in hopes it will smooth out the fluctuations. Try
it--it might work.
An even bigger issue is the 5 volt regulator inside the Stamp II. It is an
LM2936, which is only rated to supply 50 milliamps of current. Your servo
is going to draw at least twice that on the average, and lots more in the
spikes. The Stamp's built-in regulator is simply not up to the job. The
fact that it works at all goes to show that your LM2936 can be pushed, up
to a point, to deliver more than its rated current.
> Can't I get away without a 1121-5 for the servo? My worry is that
> it will create one more place for power to get guzzled?
In short, I think, no. Another LT1121, the 8-pin version, with the
shutdown pin connected to the stamp, would draw less than 20 microamps when
in shutdown. Maybe that regulator power the radio xmtr too. Put a large
capacitor on the 1121 output to supply current spikes to the servo. And
route the power wiring in a "star" directly back to the battery.
I hope that clarifies the issue of circuit layout.
-- Tracy Allen
Electronically Monitored Ecosystems
http://www.emesystems.com