Page 1 of 3

Solar panel charger with auto shutoff. Solar LED drivers.

Posted: Fri Jun 14, 2013 9:37 am
by burger2227
Circuit monitors a battery being charged using a LM317T or LM338T TO 220 type voltage regulator:
A cheaper BC547 transistor can be substituted for the 2N2222 which cuts off charging voltage
When the battery voltage reaches a fully charged voltage set by P2, the Op Amp turns off the voltage
regulator. This circuit is similar to the tool charger circuit except that the op amp is driven by the battery as
a solar panel cannot be counted on to supply a steady voltage. Normally voltages will go up and down with
the brightness. The regulator should deliver the same current, but limit the voltage supplied to the battery
to 14 volts.

The voltage regulator may need a heat sink if it gets too hot to touch and could melt a breadboard!
Locate the voltage regulator on a PC board near an outside edge so that a heat sink can be easily added.

Tool charger circuit with circuit layout

The Shottkey diodes on both sides of the regulator protect it from reverse currents and the one on the output
keeps the battery from being drained when the transistor shuts the regulator off. Shottkey diodes drop less
voltage across them.

I tested this circuit in my car for a week and the battery voltage never got high enough to cause any damage:
Two 470 ohm in parallel for 235 ohm and two 4.7K in parallel for 2.3K on the regulator adjust to get 13.6 volts.

Adjust P1 to Set the regulator voltage on the output diode cathode to 14 volts without connecting it to
the battery. Use a power supply or the solar cell input voltage greater than 16 volts in full sun with a load on
the regulator output where the battery will be. An LED and resistor can be used as the load. The battery will
pull the output voltage down when it is connected to the cathode of the output diode.

Voltage Regulator Calculator

Once the regulator is set, connect the battery through a male cigarette type adapter plug or battery clamps.
When the battery voltage approaches 13 volts, set P2 so that the LED comes on then turn the pot the
other way a bit. The LED should stay on until the battery voltage drops about 1/2 volt due to the 2M and 47K
resistance feedback from the Op Amp. This also keeps the op amp from oscillating on and off!

This circuit can also be used to charge other voltage batteries:
P1 can be 2K for voltages from 6 to 10 volts, 5K for up to 28 volts. Output should be set for battery + 1/6 volts.
P2 can be any value up to 250K to reduce battery drain.
Z1 should be a Zener diode rated at about half of the battery voltage.
Op Amp most any kind. Increase the 2M resistor to reduce the voltage spread between shutoff and back on.
A cheaper BC547 transistor can be substituted for the 2N2222

Here is the above circuit I made to control the charging of a 12 volt hand vacuum:
Yellow to charging input AFTER the diode in the tool if the charger is AC. Add a diode when necessary.
Black to common which should be black in most tools. Only one common connection should be required.
Blue to the indicator LED + anode(long leg). A current resistor should also be in series to common.
Red to the + battery. Regulator voltage output can also be read before connecting to battery
Green normally to the battery with the red wire. Can also jumper to the input voltage when testing the
cutoff circuit without a battery.

To set the LM338 Output voltage: With volt meter on red and black and input to yellow, adjust the
5K trimmer to 1 + 1/6 battery voltage. Once set, the output voltage on the red wire should not vary.
The rectified charging voltage coming in to the regulator will normally be much higher than battery voltage.
The charger current should remain the same as the current the charger delivered previously.

To set the battery Cutoff voltage, connect the yellow wire to the input voltage and connect the
red and green wires to the positive battery. Adjust as necessary to the fully charged battery voltage.
When the red LED turns on, adjust the trimmer back down a half turn on a 25 turn trimmer. This may take
several adjustments to work properly!

When the rechargeable NiCad batteries die, I plan to replace them with 3 Lithium batteries for 11.1 volts.
I will need to keep the voltage to 12 volts max for the 3 lithium pack charging circuit.

Re: Solar panel charge monitor with auto shutoff

Posted: Sun Nov 10, 2013 7:21 am
by WarrenGrant
Awesome man... Thanks for sharing detailed information about the circuit.. I would love to work on similar project and share my own experience.

Re: Solar panel charge monitor with auto shutoff

Posted: Sun Nov 10, 2013 6:05 pm
by iamdenteddisk
a very nice project sir, thanks for sharing. If I here feedback data on this project I might duplicate it. how strong of a 14v out put is it?

I mean could you run say a power inverter with it or many of them in configuration?

Re: Solar panel charge monitor with auto shutoff

Posted: Fri Nov 15, 2013 4:09 pm
by burger2227
Frankly I don't think that the solar panel I have is good at doing anything. I tested the charger circuit, but the battery voltage never rose enough to stop the charging circuit. I don't think the panel can put out enough amperage. The best I got was about 80ma with full sun.

Re: Solar panel charge monitor with auto shutoff

Posted: Fri Nov 15, 2013 4:54 pm
by iamdenteddisk
not enough for my needs yet. need 6.5A minimum. but perhaps it could be used to charge a bank of batteries for intermittent use.. it may take a week or more to fully charge but then it would provide a invertible current.

cap banks may be the best answer there again to.

Re: Solar panel charge monitor with auto shutoff

Posted: Fri Nov 15, 2013 5:06 pm
by burger2227
Well charging is the exact opposite of discharging so for every hour at max amps you should get that back the same way in MAH with about 10% lost at minimum. Solar panels are getting better every day, so better ones will become available.

Re: Solar panel charge monitor with auto shutoff

Posted: Sat Nov 16, 2013 2:25 pm
by iamdenteddisk
yup, heat.
ever notice when your cell phone gets to the end and needs a re-charge it gets hot?

that heat is the loss most battery's give the heat off linearly as in a graph with a single straight diagonal line from min to max. but the Li-ion gives off its heat at the end or a curve or arc if pictured on a graph..

I don't know if I told you guy's about my dihalitic battery project but what I discovered was this loss factor is designed into the battery. what causes it is the dialysis process or the friction of electrons passing through the cell/plates.

the plate separator is imperfect having porosity or pores that work in a one direction way like a check valve. they are imperfect like I said and some electrons do catch a open pore and run in a reverse direction just because they can which act like opposition in the battery and that is the friction.

so there is no stopping this loss if using the same materials I have been using baked wafers of beach sand"separator" and aluminum salts"alum electrolyte" with aluminum/stainless plates and haven't seen the waste heat in the battery itself.

there is no open-share of info from manufacture's to know if this fail of design is for any reason but to create a internal resistance for the battery to offset it's natural capacitive effect to prevent oscillation in output?? anyone have that answer? wouldn't an oscillating DC still be useful?

any-who that's two cent's..

Re: Solar panel charge monitor with auto shutoff

Posted: Sat Nov 16, 2013 3:56 pm
by burger2227
Well, batteries are getting better too, just like computers are getting smaller and memory is getting bigger in the same space. That friction moves both ways. Hopefully you use it less than you charge it. I've never messed with batteries chemically. Too messy. I lost a suit to a spare battery in my trunk.

What we really need, especially with solar and wind energy is a battery that can act like a perfect capacitor. Capacitors can be charged instantly and hold that charge for a fair amount of time. Ideally forever!

Then we have to realize that electricity is all around us naturally. We just have to find a way to harness it. Imagine what we could run with all of the energy stored in the clouds floating above us all day!

Re: Solar panel charge monitor with auto shutoff

Posted: Sat Dec 14, 2013 12:40 am
by iamdenteddisk
right, I always wanted to dig a big hole lay thick plastic insulator then a ground plate then layer gravel as a separator then dirt/water electrolyte. then positive plate, then cement pad and antenna. using a large enough hole and a big spark gap for overload protection it could in theory store lightening or even collect static from the sky/clouds with antenna. but still what use is a super high voltage charge..
a lot to still figure out but in theory plausible..

best thing i seen yet is the "bloom box" youtube it you'll love it.

Re: Solar panel charge monitor with auto shutoff

Posted: Wed Dec 18, 2013 11:04 am
by burger2227
You'd need something far away from a ground...

Speaking of solar, I purchased a solar step light for my daughter's porch so people will not trip on it at night:
Two solar panels are wired in parallel to charge up a single 600 mah AAA rechargeable battery.

Naturally I had to see what was inside of it. Good thing I did, the white + battery terminal had slid halfway out:
All that is in the LED driver circuit board is the YX805 driver and one 56 uh coil wired to a replaceable NiMH battery.

The circuit board only has 3 components on it including the rubber push button switch:
The solder connections are +S(red) and -S(black) for the solar panels + and - for the LED's and +B(white) and -B(black)
for the battery.

I was able to find a schematic for the circuit, but the 4 legged YX805 chip is somewhat elusive:
The YX805 chip eliminates 2 transistors, a solar panel diode, uses one coil instead of two and requires no capacitors or resistors.

The value of the coil determines the current supplied to the LED('s):
The 2 LED's are wired in parallel so each gets about 10 ma of current when turned on by a lack of solar cell voltage.

Another schematic shows how a photocell could be added to turn it off when other lights are on.
Also the solar panel could be replaced with a resistor to just turn it on at night using a regular battery.

Now all I have to do is find a chip supplier from China... 12 to 18 yuan on the mainland.

Re: Solar panel charge monitor with auto shutoff

Posted: Thu Dec 26, 2013 12:45 pm
by burger2227
I found a place where I could order 100 instead of 10,000 for about $30 or 30 cents apiece. The IC may also be numbered 5252F.

I came up with this circuit for a power failure LED driver with photocell:
The rectifier diode, LED and capacitor may only be necessary for 120 volts AC input through a .068 uf 200 volt capacitor. I'll have to check to see if it needs the 33K resistor to pull the rectified AC voltage low too. That may be done internally.

It could be added to any device that has a DC supply too! Calculate the resistor to use using:

R = (Vin - 1.25) / .02 ma for the battery charging current. 20 ma is good for long trickle charging of AA and AAA.

This greatly reduces the number of parts necessary so I plan to make another 3 outlet power failure box using a V15H
battery. A 56 uH coil should run 2 bright white or blue LED's at about 10ma each. I will have to check the turn off
voltage for the Solar panel leg that can also be used as a switch by keeping it high during the day. Solar panels can
be used indoors for that kind of stuff even though they don't generate much current for charging.

YX805 or 5252F IC chip advantages
1) Chip stops drain on solar cell when voltage drops below .3 volts after charging has started.
2) Working battery voltages are from .85 to 6 volts. Up to 3, 1.25 volt rechargeable batteries.
3) Chip prevents blinking near shutdown voltage.
4) Current range is 1 to 300 milliamps with LED drive current depending on the coil value.
5) Can reduce LED light circuit board components down to 3, the driver IC, the coil and the LED.
6) Comes in TO-94 (4 pin), DIP 8 (8 pin chip) or SOT 25 (surface mount) packages.

The YX805 chip supposedly limits battery drain down to .85 volts so batteries last longer too!
I will have to test that out. I have some rechargeable AAA batteries that are greatly overrated at 1500 mah!
12 hours at 26 ma has drained them already. That's only 312 mah... very poor!
My current power failure circuit drains them down until less than .6 volts with a dim LED.

Re: Solar panel charge monitor with auto shutoff

Posted: Fri Dec 27, 2013 6:45 pm
by burger2227
Found an internal diagram of the chip. The Chinese company wants to send me the 5252F now.
The Shottkey diode from the solar panel to the battery was what I expected to keep the battery charged,
but after some thought, I think there should be an input to the control block just for the solar voltage.
The solar panel or photocell voltage is all that should be read by the chip so that the battery
voltage does not keep it off.

The battery should supply voltage to the controller at all times to keep it running too. If it gets too low
everything will quit until enough voltage reaches the controller from the solar cell diode the next day. It will
also keep the LED off until dark.

The FET on the right pulls the LED and coil voltage low as it oscillates on and off at speeds you cannot see.

The QX5252E schematic below shows R2 as a CDS photocell resistor on pin 5 of the DIP chip:
The LS pin can also turn the LED driver on or off w/wo a solar cell for room lighting. I already found a
way to add a photocell to the 4 pin chip without needing that extra connection.
The LS pin is really not necessary.

The LED can also be configured in parallel or series to the coil:
When the LED is parallel to the coil, the cathode goes toward the battery positive while in series the cathode
goes to battery negative. The LED anode always goes to the LX pin 4 as the coil charges and discharges through it.

Another version of the 4 pin chip is CL0116 which is used below to flash an LED:
Feedback voltage from the coil circuit to SBAT turns the chip on and off at intervals based on the values of the 10K resistor and 10uf cap.
A CDS photocell also keeps the circuit off during the day when its resistance value is low. The CL0116 normally only uses a 330 uH coil.

One thing I have noticed is that the SBAT pin must not have to be held low to allow the chip to turn on. That also may be done internally.

Re: Solar panel charge monitor with auto shutoff

Posted: Wed Jan 01, 2014 4:27 pm
by burger2227
Found a 3 pin LED driver numbered CL0117 which has no pin for solar panel control:
It looks just like an ordinary transistor and in this circuit acts similar to one in a Joule thief.

The Japanese circuit is similar to the other schematics without the solar panel pin:
There is also a CL0118 3 pin chip that uses 2 batteries(3 volts) and 2 series LED's in its schematic.
All of these IC's can run on one or two or even more batteries if necessary up to 5 volts.

I don't find these particularly useful as I need the fourth pin to control the LED whether it be with a solar panel,
a photocell, a motion sensor transistor or a logic chip pulling LX low to turn the LED driver on.

I plan to add a motion sensor to an LED driver to light up my kitchen clock when it senses somebody in the room...

Re: Solar panel charge monitor with auto shutoff

Posted: Sun Jan 19, 2014 2:00 pm
by burger2227
I finally received my shipment from China and all looks good!
Airmail no less...normally it takes over 30 days to get most stuff with free shipping. Shipping was $10.00 plus a Paypal fee.

Here are 3 of the one hundred I ordered next to a dime:
I will sell some of them if people are interested for .75 each plus $2.00 shipping per order. USA ONLY!

The chip makes one battery a powerhouse driving two LED's to over 20 milliamps if need be. 17.5 ma with a 56 uH coil :
Pin 1, on the left, is not even required to be used if your device has a power switch. Above I connected a power
supply voltage (lower red alligator clip) to see how much voltage was required to turn it off.

Pin 1 can turn off the LED circuit with just .3 volts. Pull the pin low to power it up or use a photocell from plus battery. The photocell may cause the LED('s) to flicker at dusk and dawn. A 33K resistor to common seemed to help avert that.

The first things I want to make are a simple LED circuit without a board and a module that can be inserted into a 2 battery LED device to replace one battery. If a device has 4 batteries, the chip circuit can replace two to drive the LED's only!

Switching must be between the driver chip and the battery, otherwise the chip will drain the battery:
This means that the circuit must have access to the other side of the switch, not the battery itself.

That also explains why many torch lights put the switch on the common end of the batteries. In that case, the metal case itself is OK to use as the chip ground because the switch is between the common battery end spring and the case.

A side common switch would require a connection to the lamp side of the switch or the LED cathode.

Re: Solar panel charge monitor with auto shutoff

Posted: Mon Jan 20, 2014 9:40 am
by burger2227
Here is my first AA battery eliminator. I was able to use this in my magnifier temporarily and it worked quite well:
From left to right, the positive power comes from the second battery to Pin 2 of the chip through the red wire.
Pin 3 goes to the ground spring on the side of the board which hits the metal case that is switched to common.
Pin 4 goes to the 1/2 inch brass fastener as the positive terminal of the battery combination. I marked it red.

Two LED's can be wired in parallel or series using the 56 uH coil to use about 18 to 20 milliamps of current.
Smaller coil values will increase the LED current. Series LED's use half the volts and parallel ones use half the current.
Note: Due to the coil from battery in to out to LED, the battery replacement may just act as a battery jumper!
Even if the circuit is grounded when inserted in a boosted device, it will not hurt existing booster circuitry or increase voltage.
Boosted devices may just work with one battery and will not be harmed by the battery replacement device!
You may wish to test devices first with a jumper to one battery before inserting this circuit.

Note that I also have a lot of room to add other components to this board. The brass fasteners could be cut
shorter for even more room. If the case is not metal or the switch is in the wrong place, the spring can be
replaced with wires.

Brass fasteners can be soldered and can be found in office supply stores. They are used to clip binder pages
together. The spring was clipped from a battery holder. If you only clip half of the spring like I did, the
holder may still work...In my case, I had some small 12 volt 23AE battery holders that I do not need. I also
made some into AAA holders.

The Rockwell SoniCrafter cuts circuit boards easily with the wood saw blade, but you need a vise to hold it
as the tool will vibrate your hand to sleep without cutting anything. It requires a solid cut object or it will
not cut anything. SAFE!

A 555 timer could be added to drive Pin 1 high or low to turn the LED's off or on respectively.
A low power 7555 timer could turn the entire chip circuit on with a tilt or proximity switch too.

Here is a circuit to do that:

The magnifier I tested the battery replacement on below controls 2 LED's, but from the positive end of the batteries. Switch is next to LED's
The entire battery tube is connected to the common end of the batteries so using the chip here is not good for the batteries.

I would have to gain access to the switch circuit inside of the top half. That would be very difficult!
The only other option would be to add a switch just to turn on the chip. Perhaps a tilt or proximity switch?

I may add the 7555 switching circuit to the magnifier so that it turns itself off after 5 minutes.
I've left it on before! I'm waiting on tilt switches that look like electrolytic capacitors with balls in them.
Mercury switches may be an option too.

Pin 1 allows .2 volts to turn off everything with only a .00002 amp draw.
However if Pin 1 is open or low it still draws 18 ma even when the LED is not in the circuit!
The only way to kill the driver chip's current draw is to disconnect the positive or the common battery connection.

Here I use a jumper to supply ground in a 2 AA LED light I made years ago:
I will redo this light also, but it will take a bit of re-configuring as the switch supplies positive voltage to the LED.
I'll have to keep the top battery to supply 1.5 volts through the switch to the 5252F driver chip to turn it on and off.
Hey look, I found another discarded spring!

Here it is redone without a circuit board! It is that easy!
I removed the spring and old LED, soldered a coil to pins 2 and 4, LED anode to 4 and cathode to 3, plus switch to 2, common to 3, done!
Add a 7555 timer and tilt switch to Pin 1 and it will turn itself off in 5 minutes too!

If anybody would be interested in purchasing some of these these chips, I have plenty more than I need!
Contact me and I will ship some to you for 75 cents each and $2 shipping per order US ONLY.

Re: Solar panel charge with auto shutoff. Solar LED driver c

Posted: Tue Jan 21, 2014 7:43 pm
by burger2227
Pin 1 does all of the charging, so let's try to use it. I ran the chip for 12 hours with a AAA battery at 18 ma current
and the 5252F finally shut off the LED's at .9 volts. So I know that the battery discharge protection works OK!


Here is the 5252 circuit I am testing now using AC house voltage with a 250 volt rated capacitor:
All of the charging voltage goes through the Pin 1 of the 5252F driver chip before it is sent to the battery through Pin 2.

With a .068 uf, 250 volt capacitor on 120 volts there isn't a lot of DC voltage so we have to rectify and
smooth the AC pulses or the chip will ignore the AC and keep the LED's on. The Shottkey diode and the green
LED send DC pulses twice as often as just one rectifier and the 10 uf electrolytic capacitor smooths it all out.
The 10 uf capacitor must be used!

That's all folks! Well the 250 volt capacitor, and 330 ohm resistor are isolated and insulated elsewhere, but :
I count 7 components including 3 LED's! The 5252 chip eliminates a lot of components and takes better
care of the battery. A good deal I think. The capacitor, re-settable fuse, MOV, and resistors most likely won't
be on a PC board anyhow!

DC voltage to Pin 1 starts out at about 1 volt DC when the battery is low and increases as the battery charges.
The battery voltage shut off the chip at .9 volts and it bounced back up to almost one volt before I started the AC trickle
charge to Pin 1.

When the battery voltage reached 1.25 volts, I took this picture of the Pin 1(left) and Pin 2(right) voltages:
Sorry about the mess. I like building more than cleaning! The high voltage capacitor and resistor are
in the wires on the left.

So now we know what it takes to charge one battery and how to do it with 120 volts. All DC charging
requires is a current limiting resistor to reduce the Pin 1 voltage to no more than 1.5 volts and perhaps
a charging indicator LED!

1) If the battery completely dies or is disconnected, the AC trickle charge voltage will run the LED's.
2) The photocell can be replaced with a photo transistor or diode with the cathode to Pin 1.
3) Trickle charging a AA or AAA battery takes a long time! For a power failure circuit that will work fine.

Re: Solar panel charger with auto shutoff. Solar LED drivers

Posted: Fri Jan 24, 2014 12:15 am
by burger2227
While working on another power outlet with night light, I came across this light sensitive diode:
While the board indicates CDS for a photocell, it appears to be a flat topped diode.

I tried it in my latest power failure prototype and it works like a charm and is very sensitive to changes in light:
Also note that I have changed to a smaller 10 uF Tantalum capacitor. They are 4 times as expensive and only
rated 16 volts. Since a capacitor has been added to Pin 1 , no more LED blinking has occurred, but there is a
slight delay in circuit reaction. A 100 uF capacitor would make the LED turn on delay longer.

So far the AAA battery has been charging 2 days and the voltage increase has slowed at 1.325 volts.
I shall see how long it lasts tomorrow night.

Here is the final power failure circuit board I installed in a 6 outlet surge suppressor:

Re: Solar panel charger with auto shutoff. Solar LED drivers

Posted: Tue Jan 28, 2014 8:20 am
by burger2227
I had to rethink my capacitive reactance formula because it did not reflect reality! I found out that cap
current lags the voltage going through the capacitor by 90 degrees in the AC wave form and my real currents
were half of what the calculator was calculating. So I divide the current by 2 to get a closer approximation:

Code: Select all

'Inductive Reactance = 2*pi*frequency*inductance
'Capacitive Reactance = 1/(2*pi*frequency*capacitance)

  INPUT "Enter AC supply voltage: ", voltage& '= 120
LOOP UNTIL voltage& > 0

ACfreq = 60 ' 50 'in Europe and Asia?
  INPUT "Capacitor farads(1uf = .000001) or microfarads >= .001:"; capacity ' EX:.068 = .000000068
LOOP UNTIL capacity > 0

IF capacity >= .001 THEN capacity = capacity / 1000000

PRINT USING "Capacity = .############ farads"; capacity

Creact& = 1 / (8 * ATN(1) * ACfreq * capacity)

PRINT USING "Reactance = ########,.## ohms"; Creact&
INPUT "Enter other series resistance or hit enter to continue: ", add
RMScurrent = voltage& / (Creact& + add)
current = .9 * RMScurrent * .5 'cap current is 90 degrees off of voltage sync
PRINT USING "Average supply Current = ##.##### amps"; current

The .068 capacitor with the 330 ohm resistance returns 1.2 ma in charging current to the battery whether
it goes directly to the battery or through the 5252F solar driver chip. That means that it takes 20 hours to
recharge for every hour of actual use of a battery since 2 LED's use 20 ma per hour in the circuit.

When a photocell keeps the circuit off on Pin 1 during a power failure, the circuit only uses 35 ua or .000035 amps.

AA and AAA batteries can be trickle charged with up to 10 ma using a .68 uf to 1uf capacitor but so
far .068 has worked OK.

The AC trickle charging voltage must be rectified to DC with at least 2 diodes (or diode and LED) to line and
common and be filtered with a 10 uf capacitor to Pin 1 of the 5252F chip in order for it to turn on and off properly.

Re: Solar panel charger with auto shutoff. Solar LED drivers

Posted: Sat Feb 01, 2014 12:17 pm
by burger2227
Took apart a cheap LED night light because it worked intermittently:
The LED may last thousands of hours, but the solder joints to the prongs don't!

The circuit runs an LED at 5 volts when it is dark. A 5 volt C5V Zener diode regulates voltage:
The S9014 NPN transistor shorts out the LED during the day from what I can see.
The large 250 volt capacitor is marked 564J or 560,000 pf or .56 uf.

I was able to run the LED by using 2 batteries with the 56 uH coil on the previous breadboarded 5252F LED driver circuit:
With the 56 uH coil, the LED pulled 40 ma. with a 33 uH coil, the current went up to 56 ma and the LED was
brighter. The transistor pulls over 100 ma when a bright light shines on the nite lite photocell so the batteries
would not last long.

I had to shine a light on BOTH photocells to turn the LED completely off because the 5252 driver sends AC
pulses that turns the cutoff transistor off and on so the LED stays partially lit. Light to the night light
photocell turns it off.

It appears that the nite lite uses more power when it is off than when it is on thanks to the shorting transistor during the day!

Re: Solar panel charger with auto shutoff. Solar LED drivers

Posted: Sun Feb 02, 2014 6:32 pm
by burger2227
I found an older nite lite that used a real light bulb so I decided to redo that:
It uses a PR606J SCR to turn an Ac light bulb on and off with a photocell. The SCR only rectifies half of the AC voltage.

Here's a similar circuit for an LED using the same SCR:
I'd like to figure a way to turn off the AC LED during the day in some of these circuits.
Maybe have the SCR jumper across the LED so that battery charging could be continued?

I use two bright flat top LED's, one that always runs on AC but only uses about 10 ma:
The AC also uses 1.2 ma to charge the button battery.

and the other that runs without AC using a V40H battery which could last almost 2 hours:
The .068 capacitor gives the V40H battery exactly 1.2 ma of trickle charging as required by Varta too!

I bought 6 of them as part of a power pack and broke the pack into parts for these kinds of projects.
I kept the remaining 40H batteries as pairs for higher voltage projects using hotter LED's. You can solder
wires to the tabs even if they don't look too pretty now. The 2 brand new tabbed 80H 80 mah batteries
on the right cost $3 apiece from China while the 6 40H batteries cost $5 in a battery pack for a Dell Latitude....

I mount and coat the batteries with clear silicone calking to protect the wiring and prevent accidental shorts.

It was a tight fit, but I manages to fit it all inside of the small compartment. I made the circuit with perfboard
point to point and the two biggest problems were the AC prongs. which were really hard to solder to.
I ended up wrapping solid wire around the prongs on the top and bottom of the board and soldering the
circuit to that. The prongs moved. but kept a connection pretty well. Once the prongs are reinserted into
the shell, they don't move at all.

Here's a look at the LED's with the AC power:

Comparing the AC powered LED with the more powerful LED nite lite, the one I made on the left compares pretty well using half the current:
And it doesn't use 100ma all day doing nothing but supposedly saving energy! The 5252F chip keeps current near zero!