Page 2 of 4

Posted: Thu Mar 28, 2013 11:10 am
by burger2227
I'm gonna have to put in another LED on the bottom because the shadow of the hand axis always makes it
look like something thirty. I can wire them in parallel and then cut down the current with a resistor a bit. I
tried it with two LED's on 3 volts and found that 75 ohms might work. They were still pretty bright, but the
power supply reads 1.9 milliamps when they are on. Best I can do. Hope it lasts.

Posted: Thu Mar 28, 2013 3:05 pm
by iamdenteddisk
what is the mAh rating of your battery? usually written on the side.

about 300-500 is best you will find but that can be doubled by using two in parallel.

that rating is how many miliamps it can charge/drain safely in an hr without getting hot or (dangerous) as if the sealed battery boils and builds pressure it will rupture (pow)..

this can also help in calculating for how long they will stay lit. the battery will continue to light normally until it reaches 50% drained. with use of a joule thief that number probably drops to 20 or 30%.

with a active current of 1.9 milliamp and 300 mAh you should have a burn time somewhere around 110.52 hrs. from full charge to dead.
that was with subtracting the 30%
300-30% =210.
210/1.9=110.52 hrs.

now this is how long the battery can make 1.9ma not how long the diode will stay lit, as voltage falls as it drains and a diode needs so much before it will light or conduct. that data should be listed in the diode datasheet.

with this you should have enough data to plan power needs for the project..

Posted: Thu Mar 28, 2013 3:51 pm
by iamdenteddisk
ok im am working up a circuit simulation for it now in Ltspice. give me the min cuton voltage & required current of the LEDs and the range of your ldr. I will give you back the "most efficient" values for other R's for the circuit.

if you can, find and install the free version of LTspice and I will post the schematic to you so you can play and simulate or even add components to it..

Posted: Thu Mar 28, 2013 4:10 pm
by burger2227
I get 22 ma with two LED's on at 3 volts. They stay bright enough until about 2.5 volts.

I adjust current down to zero on my power supply and they are still
bright with about 75 ohm resistance.
After that it is a matter of judgement on how bright the clock needs to be.

Posted: Thu Mar 28, 2013 4:23 pm
by iamdenteddisk
ok what is the light dark resistance of the ldr?

Posted: Thu Mar 28, 2013 4:33 pm
by burger2227
1 Meg at least. I've tried theory, but reality is better. Theory cannot tell you when I can see the clock well
enough either. My power supply cannot tell me milliamps as well as my meter does either.

The LED's have a knee in the current characteristics so that after a certain current, the brightness just
doesn't get any better and their effective lives get much shorter... :wink:

Posted: Thu Mar 28, 2013 5:10 pm
by iamdenteddisk
ok its simulating now will post the schematic and values to you.

did you find and install ltspice? its the only way to view the file..

it is super as a way to experiment with EET you just draw your schematic input values for components and hit run. from there you can poke different simulated meters in anywhere you want and even see an o-scope wave form or what ever.. makes it much more fun and reduces cost when you only have to buy what you really need..

Posted: Thu Mar 28, 2013 7:17 pm
by burger2227
I downloaded it, but haven't figured out where the symbols are. I found some sample circuits but no luck running them.

Posted: Fri Mar 29, 2013 12:06 am
by iamdenteddisk
i always said your one smart cookie, getting a replacement like that is the way it should be but you hear happy endings few and far between.

see this

and scroll down till you see "cutoff voltage", according to the datasheet this transistor isn't supposed to operate at 3v. the two fully charged batterys will hold around 3.3v in reality but the 2n2222 will cut off when it hits 3v even meaning good battery's wont power the circuit long..

kinda befuddled now as to how it is working but. if it is then more power to ya! he he he..

any way as I figure the large resistor should have been close to 2.7k and the small one roughly 270 leaving the last 30 ohms to be switched by the ldr.
that is in an "ideal circuit". and providing 1mA current to 1 diode.

3vdc/3000ohms=1mA the rest is voltage divider formula where you figure 2.3-2.7 is required to bias transistor and .8-1.5 to bias diodes
which will put you close with very little to no change to spare if using 3v

but your short by about 30mv to use 2 diodes so I suggest finding higher efficiency transistor or higher voltage. then even if adding a third battery to 4.5v those same resistor values should keep you in a good efficient range.

I will need to drop box the schematic for you but will post it in a bit.

p.s. in lt spice look for the and gate symbol you should get a popup that says "component" when you mouse over it click it once inside there is a search bar, type "npn" and it will give you a choice of transistors. same for most digital devices batteries n such. the analog devices are on the toolbar where you see the components button. to the left side of the tool bar is a little running man click him to simulate the circuit. you will no doubt get errors till you learn to setup the simulation settings under simulate/"edit simulation cmd" this is simulation timing and step timing toy with it till you get a setting that works because its hard to explain but the help file is good if you take time to read and loads of help on youtube..just search "LTspice" for walk troughs to get you started..

Posted: Fri Mar 29, 2013 3:53 am
by iamdenteddisk
clock looks great by the way very commercial with the light installation..can't tell it wasn't bought that way.. :wink:

the 75 ohm R is in the collector emitter circuit this is the supply the base emitter circuit is the valve circuit that controls flow. so long as the adequate supply is ready for use the valve controls its flow.

you done great in my book Ted, no finer work have I seen from collage graduates. for the most part I don't believe they learn the real meat but just the bare essentials of the text to pass the test. most don't even venture as far as you have now and even professors at my school couldn't tell me the stuff we discussed here already. that is why I ended up teaching while attending.

"forest mimms handbooks from the shack"and circuit encyclopedia, worth every cent. when I got to I.T.T. they where astonished and in my view, had no idea but was just teaching to the test. the real meat is in the lab and on the soles of worn-out shoes and burned fingers, youknow..

Posted: Fri Mar 29, 2013 11:05 am
by burger2227
I created the ASC file and it loaded, but said something about an error. Like I said, reality beats theory every time!
I saw the 2N2222 circuit on U-Tube so it works. Don't care why...

Now all I gotta figure is how long it will work. I ordered some reflective gold tape. 10 feet for $10, must be made of real gold...

Posted: Sat Mar 30, 2013 9:06 am
by burger2227
I made my first coil today based on this circuit:
It requires 10 turns, a center tap, 10 more turns, an end tap and 10 more turns for the flyback using #26 or #30 gauge
enamel coated wire. The taps and ends will require that the enamel be removed by scraping or burning with
a lighter or match to make clean solder connections. Each loop through the ferrite core center counts as a turn.

The voltage coming from the coil flyback is almost 5.2 volts with 10 turns on all 3 windings.
Voltage is measured from the cathode end of the diode to the LED. A Shottkey diode adds another .05 volts.

To increase the voltage, I had to use less turns on the flyback windings. 7 turns made the 2 LED's bright:
Two batteries are driving a 5mm and 10mm bright LED wired in series.

Here the same circuit uses ONE rechargeable battery generating 2.8 volts to light a 10mm bright LED:


With 330 ohms driving the transistor base, it gets quite hot so I tried higher values and settled
on a 4.7K resistor to the base. It doesn't heat up at all now and the LED is just as bright!

The number of turns on the flyback can be changed to suit LED voltage needs to a degree.

You can also add a light dependent resistor from the transistor base to ground now to turn it on at night.

I also finally got the box open with the ugly nite lite. It has a full back plate with 5 Y wing screws. Here is what I found:

The center is perfect for one battery. I may put the .068 600V capacitor right next to it. I rearranged the hardware
to set up the existing wiring for the power monitor high voltage circuit. The old circuit with the ugly orange LED is at top.
It looked like an old neon bulb.

Posted: Sun Mar 31, 2013 10:58 am
by burger2227
WARNING! Do not attempt to plug the electrical box into a wall socket
without the back of the box securely screwed back onto the box!

The box already had an area to put the circuit in, but it was a lot smaller than what I will need. After cutting away some
plastic and hacksawing a PC board to fit, I have this area to put the white LED circuit with the coil. I may need to stand
the coil up on its side and put it in the area on the left. Electrolytic caps could go on the right.

The copper clad circuit board came from Radio Shack. I had to cut the board in half to accommodate circuit requirements.

Here is the finished circuit with battery holder. I decided to use a single battery holder from Radio Shack and cut it
to fit. All I needed was the end contacts held together by the base. The 110 volt circuit consisting of the capacitor,
MOV, resistors and diode to the battery will go to the left of the battery.
I set the circuit board up so that components requiring plus voltage or ground were placed nearest to the appropriate
wires in the box. When I first started placing components, I figured that the ferrite coil and electrolytic capacitors
would take up the most room so I placed them on the appropriate ends and worked my way into the LED placed
in the center. The space in the middle fills up faster than you think!

Here is the first test which went off without a hitch. I always dread the first test because I often find I have missed
something and will have to spend hours figuring what I did wrong or find a bad connection.

Here is what the final product will look like when the power goes out. It gives off plenty of light!

I am still waiting on safety components to complete the 110 volt part of the project. Richard Cappels, the author
of this project, was nice enough to send me some 25 volt varistors to protect the battery circuit components.
He resides in Thailand... He has also provided knowledge and assistance in making this project work as designed.

I have not added the green LED(D3) that will run off of the AC voltage yet either. There is room for it next to the
bright LED and it will require a wire from the AC capacitor circuit as will the diode to the base of the 2N2907
transistor which is already mounted on the board. I will update this as I go.

Note the orientation of D3, the green LED. The cathode (short leg) goes toward the incoming AC voltage,
but since the AC is not rectified, the LED will light when there is AC voltage using the reverse duty cycle.
It will turn off when there is no AC power and the battery driven LED circuit should come on when the
PNP transistor's base no longer has voltage on it.

Note: The AC trickle charging .068uf 300 volt(or higher) Mylar capacitor value may be designated as 683K on the case.

K Capacitor conversion chart: ... _chart.htm

The prototype LED circuit has been running with a photocell added to the base of the 2N4401 transistor for 3 nights so far.
The LED circuit will run on the battery constantly without the high voltage trickle supply to the PNP transistor base.

WARNING! Do not attempt to plug the electrical box into a wall socket
without the back of the box securely screwed back onto the box!

Posted: Tue Apr 02, 2013 2:18 pm
by burger2227
I added the green LED and a green wire to go to the AC trickle charging circuit next to the battery.

Note: The 2N4401 base resistor R7 should be about 4.7K to keep the transistor running cool!

The cathode of the LED goes to the anode of D2 which goes to the base of the PNP transistor. The green wire
goes to the anode of D1 which taps into the red wire down at the battery case. R4 and D1 are the only parts
not on the circuit board. I did this to isolate the AC voltage from the low voltage circuit as much as possible.

Use 1N4004 or higher rated diodes to rectify the trickle voltage for safety sake. 1N4007 is rated for 1000 volts, 1 Amp.

If I had to do it all over, here is how I would arrange the components on the pre-drilled PC board:
Note that the required jumpers are numbered 1 and 2 for the two 100 uf electrolytic capacitors only.
Neither 100 uf capacitor is critical to the circuit. To conserve space, a smaller capacitor can be used
with the 1K resistor that pulls the base low on the PNP switch transistor to activate the light.

The two transistors are wired left to right as Collector, Base and Emitter as oriented by the tab of the
PNP can and flat side of the NPN. The metal can top of transistors is normally the collector and
contact with other components or wires should be avoided! The tab usually denotes the emitter side.

I also would make the coil connector wires longer when winding the coil to eliminate extra jumpers. The #26 gauge
wire is coated so that it can be used like insulated wire. I tried heating the coil wire ends with a match as
recommended, but I ended up having to scrape it too to hold the solder connections better.

Posted: Thu Apr 04, 2013 8:53 am
by burger2227
A prototype of the same white LED driver without the extra high voltage stuff is currently running two LED's in parallel
very brightly. The rechargeable battery lasted 2 days with one LED and one day with two of them in tests.
The circuit is drawing about 20 milliamps with one or two LED's. It uses .7 milliamps when the LDR shuts it off
during the day. It is so sensitive that it lights up if I walk past it during the day. A proximity sensor perhaps? :wink:

Flyback circuit with photocell (LDR) will come on when in a shadow.

While the one battery circuit is interesting and useful as an emergency light, 4 days or a week is not long enough
to use it to light my clock. The clock has been lit since March 20th with the simple transistor LDR circuit. I am testing
two flat headed 120 degree LED's and they do a better job. I am waiting for some warm white ones to come that
are four times as bright and rated at 6000 mcd. That may do the trick!

Clock lites

Posted: Thu Apr 04, 2013 10:42 pm
by burger2227
The disk diameter of a metal-oxide varistor (Varistor=Variable Resistor) determines the surge current
capability of the device to divert transients. A typical 20mm varistor is rated to handle a surge up to 6,500
amperes based on the standard 8 x 20 us waveform defined by IEEE C62.41. Likewise, a 40mm varistor is
rated to handle surge currents at 40,000 amps.

MOV labeled values:
Z250-20 = 25 * (10 ^ 0) = 25 volts 20 joules

Z151-20 = 15 * (10 ^ 1) = 150 volts 20mm = 80 joules

Leviton reference

Posted: Thu Apr 11, 2013 6:50 pm
by burger2227
Here is the power adapter in use. The bright green LED helps at night.
Always screw the back of the box on, to avoid a major accident. The battery should always be rechargeable
and should last a long time as it is trickle charged. Don't use normal batteries with the trickle charger!


If the green LED is not desired, use a 1N4004 or better in its place. When lighting LED's with AC voltage,
either use two in opposing directions or use a diode in the opposite direction to accept the reverse voltage cycles.
In the power failure circuit, the opposing AC rectifiers are the diodes to the battery and PNP switch transistor.

Otherwise an opposing LED will burn out from too much reverse current!

Posted: Sat Apr 13, 2013 1:47 pm
by burger2227
Investigating the old night light circuit from the box, I found that the 120 volt circuit board runs the LED circuit
with 1/2 watt resistors. I used a mirror photo to trace the circuit:

Components from left to right are the light sensing diode, the SCR and the LED. The alligator clip holds the
common white wire and the black lead on the right is the 120 volt lead. It goes directly to the LED anode
and the voltage divider resistors for the SCR gate. The LED cathode goes to the 33K resistor which goes to
the SCR anode. The light sensing diode controls the SCR gate by pulling it low when it senses more light.
Note how the SCR body blocks the ambient light from the LED...


The 33K resistor is the only current limiting resistance between 120 volts and ground so using V = IR
we get voltage = 120 - 2 volts for a red or yellow LED:

current = V / R = 118 / 33000 = .00357 amps or 3.6 milliamps

The resistor wattage is P = IE or power = current * voltage or

watts = 118 * .0036 = .4219

or using power = (I ^ 2) R = (.0036 ^ 2) * 33,000 = .4276 watts. Safely under 1/2 watt!

Substituting a higher current LED will require increasing the wattage and lowering the resistance.

I used this calculator to estimate the LED values: LED resistor calculator

Subsequent testing of the orange LED at 2.5 volts blew it up! It was ugly anyhow... :roll:

There are three main categories of miniature single die LEDs:
Low-current: typically rated for 2 mA at around 2 V (approximately 4 mW consumption).

Standard: 20 mA LEDs (ranging from approximately 40 mW to 90 mW) at around:
1.9 to 2.1 V for red, orange and yellow,
3.0 to 3.4 V for green and blue,
2.9 to 4.2 V for violet, pink, purple and white.

Ultra-high-output: 20 mA at approximately 2 V or 4?5 V, designed for viewing in direct sunlight.

The chart indicates that current draw goes up very quickly at certain forward voltages(Vf) for different colors." target="_blank

Here is a similar circuit using an SCR and an LDR as a night light:

The required 12K, 2 watt resistor limits the 6 LED current draw to 10 milliamps. The SCR also rectifies the AC current.
The CDS or LDR pulls the SCR gate voltage low during the day to cut off current through the circuit.

Interestingly the LED calculator says that it would be 2 watts no matter how many LED's are connected in series.
1 or 16, but the brightness would be affected I would imagine. 6 in parallel would require 12 watts!

Another project box

Posted: Sat Apr 13, 2013 6:43 pm
by burger2227
Here is another prospective container for a power failure light. This one is a 6 outlet surge suppressor with
2 status LED's built in. The LED's are wired after a fuse from 120 volts to ground and common with a 47K
1/2 watt resistor on each giving them about 3 milliamps of current using 2 volts each.


The left side has enough space for one battery holder. In the upper right corner is a screw pin to hold the
coil if necessary as shown. There is plenty of room for the power failure circuit board on the right side.
A bright 10mm LED could be placed on each side in parallel perhaps.

The surge suppressor has a green LED that indicated the suppressor is grounded while a red light indicates
it is protected. Two thermal fuses and 5 MOV's protect the circuit. I took the yellow heat wrapping off to see
the MOV ratings, but none are marked. The fuses are rated 5 Amp and set for 115 degrees.
Leave the heat wrapping on them or it may not work correctly!

I got it at Kmart for about $6.50. It is rated for 1140 Joules and comes with a limited warranty of up to
$100,000 for equipment damages if the unit fails to protect the equipment and it was not damaged through
cables or phone lines. You may want to keep the receipt and warranty information, but tearing it apart
probably voids the warranty too...

There is also a generic version available at Lowes for a few dollars more than Kmart's. No brand name on the unit.

Posted: Thu Apr 25, 2013 7:51 pm
by burger2227
Here is the surge protector with the PC board on the right and rechargeable battery on the left.
The high voltage capacitor with 1M ohm jumper is soldered to the top right power wire(black) and set behind
the crossing black wire. The 30 volt MOV(black) and polymer re-settable fuse(brown) protect the low voltage
circuitry. The smaller green wire carries the trickle charge current to the battery and switch transistor.

Since I did not need the green LED to indicate power, I just used 2 diodes on the left side of the surge box.
A 1N4007 diode goes from common to the green wire. A shottkey goes from the green wire to the battery.
There was room for two bright LED's above the surge suppressor board on each end right below the upper AC
buses. Yes you can run two bright white LED's with the power failure circuit and they work pretty well.

Here's what it looks like when there is no AC power. With power only the red protected and green ground LED's are lit.
The white LED's are mounted in clear diffusion lenses. This box worked out real nice and it has a lot of room for the circuitry!

The polymer re-settable fuses require the voltage to be removed to reset. They are available online with different amps.
The 30 is volts and 1.35 is the current rating of 1.35 amps. I did not find lower current ratings yet.

CAUTION! Make sure the buses are set correctly before closing the box up! Note how the top white bus
has moved over to the right. Fortunately the buses never touched each other, but the outlets did not work
until I took it back apart and realigned them. It could have been far worse! BOOM!

One benefit of using outlet boxes is that you have to put the back on to test the high voltage out.
Don't try to insert the boxes with the backs off of them! The buses can pull out and shock or explode!