Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Red Light for Illumination - the "Joule-Thief"

  • Please log in to reply
59 replies to this topic

#1 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 07 March 2019 - 06:14 PM

On the next clear night I hope to make a sketch of the the Tau Canis Majoris Cluster - NGC 2362. Observing Tau CMa the other night as a double star I was taken by the beauty of the cluster of which Tau is a member. While I've done very little sketching it was apparent that I needed a handy source of illumination for my sketch-pad. Looking on-line to buy a cheap clip-on light, I didn't find any that utilized a dim red light. It seems like the focus was on bright white light. I know many amateurs use a headlamp type of illumination. But I decided to make my own light to attach to the top of my clipboard.

 

My electronic parts cache included a bunch of resistors and red LEDs. So a bit of research and testing showed a simple red light could be made from a 9-volt battery feeding a red LED with a 470-ohm 1/4-watt resistor in series. A little further online research revealed a neat little circuit that can power a red LED using a  AA battery, even a "dead" one. This is the "Joule Thief". A little explanation is in order.

 

An alkaline battery (AA or AAA) is considered dead for most uses when its voltage has dropped from 1.5 to around 1.2 volts. In a flashlight the light would be getting quite dim. In another device (like a small radio) it would just cease to work. But such a "dead" battery still has a fair amount energy left in it. One unit of energy is the joule. Thus comes the name for this handy circuit - the Joule-Thief. It doesn't really steal or create energy. It just utilizes the energy left in a supposedly dead battery. It functions by boosting the voltage from something as low as 0.4 volts up to what is needed to run a LED light or other electrical device. It works quite well on a fresh battery, which will just last longer. 

 

There are a lot of online references for this little circuit. So if you are at all handy with a soldering iron you might consider making your own. Here is a source for parts or assembled circuits that I have found:

Here is what the assembled unit looks like:

Joule-Thief.jpg

 

And here is a video describing how to assemble the kit:

My DIY kit is on the way. I contacted Alex, the seller of the kit, and inquired if a red LED could be substituted. He replied within hours saying the circuit would work fine with the red LED.  I'll report on the results when it is all put together. These kits even include one extra transistor and two white LEDs if you want to make other Joule-Thief circuits. Here's a link showing a minimalist circuit without the printed circuit board like what is included in the kits:

I believe you don't really need the commercial toroid form for the coil. You can just bend a nail around in a small circle and wind wire around that to make the coil. The DIY kit includes enough wire to make more coils.

 

The Joule-Thief has been around for decades. It is really an amazing invention. On a "dead battery it is reputed to power the light for a week! I'll run a test on a fresh battery to see how long it runs on that. Such would be useful as an emergency light when the power grid is down.

 

Edit:

Some research on the internet shows another promising source for a kit:

This is a very bright LED light source with attached battery holder. So the light output is way too bright and the wrong color for amateur astronomer use. I contacted the vendor, inquiring about availability of a red LED version. They replied there was a red version available.  At the very least the white light version would be useful for emergency use. It is interesting for that purpose, since it is a complete unit, ready for use once the kit is assembled - just add a battery and turn it on. The instruction-assembly manual can be examined on this link. Since the red LED version isn't listed on their website, I am inquiring how to order one in that color. Maybe it is just a note in the comments section when placing the order. I'll let your know what I find.

 

Edit:

A quick response from customer support for the red LED version:

  • "We have a joule thief that runs (2) T-1 or T-13/4 LEDs.
    Its not a listed product, so we would have to invoice you separately for it.
    $14.99 each"
  • The instruction/assembly sheet has a section on how to assemble the red version. There are a few changes compared to the white LED version.

 

According to the website:

 

"The EXTREME Joule Thief kit utilize a dedicated onboard 1W LED driver which can operate from voltages as low as 0.8V. There is an onboard potentiometer which can vary the output current of the driver from 0mA to 350mA. The Joule Thief kit can also run customer supplied LEDs with Vfwd voltages up to 4.4V. There is also a soft turn-on / turn-off feature incorporated into the design."

 

So it seems like the brightness of a red LED unit can be set to appropriate levels for astronomer usage. This is another tempting option, particularly since shipping/handling charges aren't out of line. I will likely order one of these in white light just for emergency use. 

 

In my thinking this unit could be packaged by the end user to be quite useful out under the stars. I'm thinking some shrink-tubing could be applied over the whole device. Then cut-outs could be made for the LED, brightness adjust, battery insertion and some holes for ventilation. It wouldn't be very "elegant", but might work OK.

 

Note to Administrator:

Perhaps this should be moved to the ATM, Optics and DIY forum.


Edited by Rustler46, 08 March 2019 - 01:45 AM.

  • Jon Isaacs, siriusandthepup, dswtan and 6 others like this

#2 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 09 March 2019 - 11:34 PM

I couldn't wait for arrival of the two Joule-thief kits on order. So I rooted through my electronics parts to find enough to make a demonstration light.

 

Joule-Thief-1.jpg

 

The ferrite toroid coil core is much larger than it needs to be. But it's what I scavenged from a 110 VAC power cord. Also I didn't have a single cell AA or AAA battery holder. But I found that a AAA battery would fit snugly inside the coil. In addition to the battery, toroid and wire there are just 3 electronic parts - a red LED, a 1.5K resistor and a small 2N2222 transistor. It so happened my deceased father-in-law had a dozen of these transistors in his kit. The LED and resistor came from my own cache. 

 

Right now the Joule-thief pictured above is under test to see how long it will run on a "dead" battery. This one had previously been discharged from 1.5+ volts down to 1.1volts. Without the Joule-thief, that battery wouldn't light the LED, even if directly connected to it. The LED needs in excess of 1.8 volts, more than a single alkaline cell can provide even when new. The circuit, oscillating at ~50 KHz, boosts the voltage up to around 2 volts to run the LED. All the time it is drawing energy (Joules) from an otherwise useless spent cell. Some 24 hours later the LED is still shining, though a bit dimmer. Battery is at 0.55 volts.

 

I'll be giving some more reports on the two Joule-thief kits I ordered:

  • EXTREME Joule Thief - ($22 ppd) - A very intense 1W Cree X-Lamp LED, useful for emergency situations. Works down to 0.8 volt. 
  • Joule-thief - DIY Kit - ($11 ppd) - Works for up to a week continuous on a weak battery, down to as little at 0.4 volt. 

My test light may be too dim to be useful at night. Time will tell. In any case I'll order some brighter red LEDs. Any that are too bright can be dimmed to a suitable level with some layers of red Rubylinth. I keep y'all posted.


Edited by Rustler46, 10 March 2019 - 06:50 PM.

  • siriusandthepup, PaulEK, pfile and 4 others like this

#3 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 12 March 2019 - 03:40 AM

One of my Joule-Thief kits came in today. Ordered on Thursday and received on Monday - that's fast service. The other kit (Extreme Joule-Thief) was ordered, but a day or so later the money was refunded via PayPal - no explanation. So I contacted customer-service. The reply was they are out of stock. They will occasionally be back in stock and again available. Too bad since that is a very bright LED, which would be good for emergencies. 

 

So there's a warning here:

 

  • Extreme Joule-Thief will allow you to pay on line for the item, even though they are out of stock. I would contact them to see if it is in stock before ordering, since the web site doesn't tell you it is not in stock.

In any case the other Joule-thief has been assembled with the circuit board attached to a single cell AAA battery holder. The included white LED is very bright - uncomfortable to look at directly. At night it easily illuminates a 25-foot hallway, using a "dead" AAA cell..

 

The kit included an extra transistor and white LED. Also there's enough extra red and green enameled wire to wind two more coils. So I scavenged a small ferrite toroid from the circuit board of a dead compact fluorescent light. The extra wire was fine enough to get 15 turns of red and green wire for the coil I wound. These parts were used to assemble another Joule-thief as shown in the test-bed below.

 

First here is the new Joule-thief kit assembled on a AAA battery holder (shown on the right below) :

Joule-thief-7.jpg

The assembled Joule-thief kit is on the right, and a test setup is seen on the left. It has components for another Joule-Thief, using the little coil previously described. That coil is above the back alligator clip.

 

Joule-thief-5.jpg

Same view as above, but with the LED more easily seen. The little black transistor is right below the LED.

 

The image below shows my test bench (AKA kitchen table). You can see my multimeter, variable DC voltage supply powering the test-bed Joule-thief.

 

Comments?

 

 

 

Attached Thumbnails

  • Joule-thief-6.jpg

  • Tenacious likes this

#4 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 12 March 2019 - 08:41 PM

Progress Report:

 

Here's an update on my Joule-thief project. First here is the basic circuit:

 

Screen Shot 2019-03-12 at 6.00.39 PM.png

  • Goal - to make a simple red light for illuminating a star atlas or sketching pad while out under the stars in the dark.
    I have been using an Orion red flashlight, which has worked well. But the desire to experiment and make an efficient light moved me to look into making a simple LED light. My research on LED lights showed a simple current limiting resistor with a 9 volt battery would work. Then I stumbled on the Joule-thief circuit for powering a LED with a single alkaline battery. Normally a 1.5 volt cell won't run a LED, since they require at least 1.8 volts to run. But the Joule-thief not only boosts the voltage of an A- or AA-cell up to the LED requirements, but it has the advantage of functioning with voltage as low as 0.5 volt. Such a cell would normally be considered "dead" and useless. So the Joule-thief circuit offers the advantage of utilizing more of the energy in an alkaline cell.
  • Design considerations - here are some factors that are driving my solution:
    • Should be inexpensive. So this appears to be doable since my original Joule-Thief light was made with some spare parts and junk I had on hand - dead battery, cheap transistor and resistor, wire, ferrite toroid, etc. Commercial sources for parts, kits and assembled circuits are of reasonable cost.
    • Its brightness should be adequate for the purpose, yet adjustable to accommodate changing conditions. I learn while using my first two circuits that a Joule-thief will continue running down to 0.5 volts and below. But a simple, unregulated circuit produces dimmer light as the voltage drops. A brightness adjustment would allow for continued use as the battery more closely approaches death.

      I have a couple of handfuls of  colored LEDs from the 1980s. While these work quite well in the circuit, they are quite dim. I learned that LED technology continues to improve in regards to power output. The standard white LEDs included with my first commercial kit proved to be very bright. Still the brightness decreased as voltage dropped. When I get a supply of current generation red LEDs I can experiment with what is needed to meet design requirements.
    • ​The circuit should be easy to build. While I have built a number of electrical devices over the years, I don't consider myself to be skilled at this sort of thing. Yet I has able get three Joule-thief circuits functioning within the last week - the first demo shown in the original post of this thread, the commercial DIY kit received yesterday and a test-bed circuit made from various extra parts. So it can't be that hard to make, since this old guy made three in a week.

      To be successful in building one of these, there are a few necessary points to consider:
      • Follow the directions carefully.
      • Use good soldering techniques.
      • Protect sensitive components from excess heat. Note the forceps in the photo below. This is used as a heat sink attached on the lead between the solder point and transistor body.
      • I recommend checking all solder joints with an ohm-meter to ensure low resistance.
    • The red light and its power source must be assembled into a package can be used at the telescope, in the dark. For example my test-bed light with experimenter's patch board and variable DC power supply wouldn't be of much use outside at night.

So that's where I'm at at this point. Here are a few photos showing how the test-bed light was assembled.

 

Joule-thief-3.jpg

 

The above photo shows where I scavenged some small ferrite toroids for the coil. Inside a dead compact fluorescent light is a circuit board with a small coil, the green thing with copper wires.  Just cut the wires to remove the toroid. This little guy needs very small wire to have enough room to wind the coil. But there was enough extra 30 AWG wire in the commercial kit for a couple of small coils like this. The finished coil with 15 turns each color is shown below. It seemed to work as well as the larger diameter toroid used in the kit. It's harder to wind the coil since it is so small. But even with my fumble fingers, it is doable.

 

Joule-thief-2.jpg

 

One important point on construction of the coil. There are two windings existing in the toroid - one with green wire, the other with copper colored wire. The two wires twisted together (copper and green color) come from different sides of the toroid - one coming from the top and the other coming from the bottom of the toroid. That leaves to two remaining colored wires exiting from opposite sides as well. These wires must be connected this way or the circuit won't work. The WRONG way to connect the twisted pair would be to just take green & copper colored wires that exit together (top or bottom) and use these to twist together. 

 

As for how to assemble a working red light into a useable package for outdoor use, I've come up with an inexpensive single cell battery holder with on/off switch with enough room inside for some electronics. These are on order from China, arriving in a few weeks. I'll then find out if it can be used like I intend. As for making the light intensity adjustable for conditions and changing voltage, I believe a small variable resistor on the transistor base connect will be needed. 


Edited by Rustler46, 13 March 2019 - 03:22 AM.

  • Oregon-raybender, DHEB and Tenacious like this

#5 Oregon-raybender

Oregon-raybender

    Optical Research Engineer

  • -----
  • Vendors
  • Posts: 493
  • Joined: 13 May 2010
  • Loc: Oregon, South Western Coast

Posted 13 March 2019 - 01:41 AM

I like the HP cal, good old RPN. I have several I still use after 35 years.

 

Starry Nightswaytogo.gif


  • tim53, LarsMalmgren and zyzyz like this

#6 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 13 March 2019 - 03:21 AM

I like the HP cal, good old RPN. I have several I still use after 35 years.

 

Starry Nightswaytogo.gif

Hi Robert,

 

It's good to hear from you. Yeah, that HP 32S II RPN Scientific is my favorite calculator. The RPN (Reverse Polish Notation) can be fun when you ask someone to add 2+2, and they find there is no "equals" key. Yet that form of calculation plus the stack makes finding the hypotenuse of a right triangle a quick set of key strokes. But then there are "different strokes for different folks". If you've been using one for so long, I'm sure you know all about its capabilities. Back in the day working in civil engineering I used an HP67.

 

Best Regards,
Russ


  • tim53 likes this

#7 Oregon-raybender

Oregon-raybender

    Optical Research Engineer

  • -----
  • Vendors
  • Posts: 493
  • Joined: 13 May 2010
  • Loc: Oregon, South Western Coast

Posted 13 March 2019 - 02:14 PM

All my optical design work was done on HP equipment, 11c, 12c, 19c, 21, 45, HP 67, HP 85, 9815, 9825, 320. I have a collection of HP cals ( all work great) which I find at second hand stores. I found a HP musuem in Hillsboro, OR, they carry the old (all) HP manuals on stick.

 

Starry Nightswaytogo.gif


  • Rustler46 likes this

#8 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 13 March 2019 - 05:46 PM

Here's a report on the DIY kit's performance. With a "dead" AAA battery the light remained lit (though dim) after two days. It started off quite bright and hard to look directly at. I'll give a test on a new battery later. Here's how it makes a handy little light when attached to single AAA battery holder.

 

Joule-thief-8.jpg

 

Joule-thief-9.jpg

 

My 30+ year old red LEDs are just too dim. So in a few weeks when I get some bright red LEDs from China, I'll make something similar with home made coil and perf-board circuit. I hope to have come up with a variable brightness control. 


Edited by Rustler46, 13 March 2019 - 06:16 PM.


#9 Tenacious

Tenacious

    Viking 1

  • *****
  • Posts: 629
  • Joined: 19 Oct 2015
  • Loc: Northern KY

Posted 13 March 2019 - 08:06 PM

I love stuff like this!  We go thru batteries like food (large family).  Now, I have a reason to salvage some compact fluorescent, too.  It seems to me that normal 2-cell flashlights (even a Mini Maglight) would make a good home for your circuit.      I wonder how long a fresh D cell would last.

 

Yeah, I'm an Hp collector, too.


  • Rustler46 likes this

#10 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 14 March 2019 - 12:19 AM

I love stuff like this!  We go thru batteries like food (large family).  Now, I have a reason to salvage some compact fluorescent, too.  It seems to me that normal 2-cell flashlights (even a Mini Maglight) would make a good home for your circuit.      I wonder how long a fresh D cell would last.

Yeah I find this sort of stuff quite interesting. Nothing really implicated, but seems to produce amazing results. I wonder why they don't built a circuit like this into a single-cell flashlight. But just going by the miil-amp hour capacities of alkaline batteries a fresh D-Cell would last a long time. Typical milli-amp hour capacity of various sizes of alkaline-manganese dioxide batteries is given in the following chart.

 

Screen Shot 2019-03-13 at 9.28.11 PM.png

 

With a used (but not dead) AAA battery the little circuit draws 38 milli-amps, with the LED light too bright to stare at. The amp-hour capacity would be a bit lower with the higher current draw (versus 10 mill-amps in the chart). But say it still had 900 milli-amp hours, then it would last 24 hours (900 ÷ 38). It's kind of a rough calculation, since the milli-amp draw will drop as the voltage drops. But it's a ball park estimate. Starting with a "dead" AAA battery it was still putting out some light at 48 hours.

 

As for a fresh D-cell we can be more confident, since the load is so low for that size battery (38 milli-amps versus 200 milli-amps in the chart). But it would likely be in excess of 14 days (13,000 ÷ 38 = 342 hrs. = 14 days). 

 

I don't know if the circuit would work with a 2-cell battery. But I can test it out with my DC power supply to run the voltage up to 3 volts.. The worst that could happen is I fry an inexpensive transistor. 


  • Tenacious likes this

#11 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 14 March 2019 - 05:01 AM

I found another very interesting option for a Joule-thief type circuit. This one uses a small 4-pin chip, one small inductor, battery and LEDs. 

 

Screen Shot 2019-03-14 at 1.55.23 AM.png

 

The chip is for a solar powered garden light using rechargeable battery. But it will work fine without the solar attached, using either alkaline or rechargeable battery. It has one drawback - it shuts off when the battery voltage dips down to 0.9 volts. This was to protect the solar powered rechargeable battery from being discharged too deeply. But for our purposes there's not much life in an alkaline battery when it reaches that low a voltage. If using a rechargeable you just recharge.

 

There seems to be some variance in the recommended inductor value - 220 μHenry or 330 μHenry. So I searched on EBay and found 20 chips and inductors for around $4.50. I'll order a range of inductor values to see which one works best, and let you know which one to get. Of course you'll still need battery holder, switch and other parts. But the electronics (chip and inductor) are less than $0.23 per light in lots of 20.

 

This circuit has a lot going for it:

  • Few parts to assemble
  • Low cost
  • Good Joule-thief type battery use. 

 

The 0.9 volt cutoff isn't a deal-breaker in my view. When my parts arrive from China I'll share the results. It seems like a simple red light like this could be used to mark tripod legs in the dark to reduce chance of someone bumping a leg. For whatever use, here are the Ebay sources for parts:

A club or group of amateurs could go together to cut costs for electronics and other needed parts. Here are some more parts sources I found:


  • Tenacious likes this

#12 Tenacious

Tenacious

    Viking 1

  • *****
  • Posts: 629
  • Joined: 19 Oct 2015
  • Loc: Northern KY

Posted 14 March 2019 - 05:29 PM

Thanks for sharing the circuits and your ideas.   Your posts bring up several questions.

 

   Did your super-bright LEDs come in, what is the light output rating (Usually given in mcd, milli-candles.  Values can range from 1000 to 10,000 or more)?

 

   I like your idea of an adjustable brightness (post #4).  I wonder if the super-brights can produce the desired illumination at 5 or 10 mA, increasing efficiency, especially for a red astro light.

 

   The circuit in your last post shows the white LEDs connected in parallel, which should require 2x more current than a single LED.  Will the 5252F chip work with a single LED?

 

 

Around here, I have toyed with the idea of house hold emergency white lights wired into the walls and utilizing spare automotive lead-acid batteries (I have a few of these).  Until your post, the idea was to use pure DC excitation.  My strategy was to increase voltage (two 12 volt batteries in series ~ 24 Vdc) and make a longish loop of series white LEDs with 1 current limiting resistor (say 20 mA).   The forward voltage on my white LEDs is around 4 volts, IIRC.  That would allow 5 or 6 LEDs in a loop all drawing 20 mA.  Naturally, I could switch-in more loops as desired.  At current consumption in this range, lead-acid batteries would last a very long time!  Also, it doesn't take a lot of solar panels to recharge a system like this.    The downside of the series loop, of course, is the old Christmas light string scenario - "..if one goes out, they all go out..".   This is not a problem for a guy with a VOM.

 

 

I'm very interested to see how your projects turn out.  Please keep us up to date.



#13 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 15 March 2019 - 04:59 AM

Thanks for sharing the circuits and your ideas.   Your posts bring up several questions.

 

   Did your super-bright LEDs come in, what is the light output rating (Usually given in mcd, milli-candles.  Values can range from 1000 to 10,000 or more)?

 

   I like your idea of an adjustable brightness (post #4).  I wonder if the super-brights can produce the desired illumination at 5 or 10 mA, increasing efficiency, especially for a red astro light.

 

   The circuit in your last post shows the white LEDs connected in parallel, which should require 2x more current than a single LED.  Will the 5252F chip work with a single LED?

 

 

Around here, I have toyed with the idea of house hold emergency white lights wired into the walls and utilizing spare automotive lead-acid batteries (I have a few of these).  Until your post, the idea was to use pure DC excitation.  My strategy was to increase voltage (two 12 volt batteries in series ~ 24 Vdc) and make a longish loop of series white LEDs with 1 current limiting resistor (say 20 mA).   The forward voltage on my white LEDs is around 4 volts, IIRC.  That would allow 5 or 6 LEDs in a loop all drawing 20 mA.  Naturally, I could switch-in more loops as desired.  At current consumption in this range, lead-acid batteries would last a very long time!  Also, it doesn't take a lot of solar panels to recharge a system like this.    The downside of the series loop, of course, is the old Christmas light string scenario - "..if one goes out, they all go out..".   This is not a problem for a guy with a VOM.

 

 

I'm very interested to see how your projects turn out.  Please keep us up to date.

I'll try answering your questions, Tenacious.

 

None of the parts I've ordered have come in except for the DIY kit, assembled with a single AAA battery. That one only had to come from Fremont, CA to Coos Bay, OR, thus the quick transport. For the other parts the "slow boat from China" will take longer - typically 2-3 weeks, depending on the vendor.

 

Using the DIY kit with a NiMH rechargeable AAA works quite well. That cell has a valuable characteristic of a flatter discharge curve (1.5 volts - fully charged, 1.25 volts - 50% charged, 1.0 volts - almost fully discharged). The downside is this makes it almost impossible to know the amount of capacity left based on its voltage alone. It stays at or above 1.1 volts until it is almost completely discharged. But it will hold a higher LED brightness until it gives out. It is not a good idea to completely run down a NiMH battery. So don't leave a light on only to come back later to a dead and maybe damaged battery. If the light goes out under use just pop a spare in while the exhausted one is being charged. There's always alkaline standby. For NiMH batteries the 0.9 volt shutoff of the QX5252 chips is a definite plus. At that voltage there is very little energy left in the cell. Discharge to 0.8 volts is where damage starts to occur. So it seems like the wise course is:

  • Use a traditional Joule-thief circuit for alkaline batteries for maximum harvest of energy.
  • Use the QX5252 for rechargeable NiMH batteries. Shutoff at 0.9 volts protects battery.
    Can also be used with Alkaline batteries. With early shutoff (0.9 versus 0.5 volts) the last 5% of battery capacity is not harvested. Easier to build compared to Joule-thief circuit.

 

NiMH discharge Curve.jpg

 

The discharge curve for an alkaline battery is a constantly dropping voltage as capacity is used. 

 

Screen Shot 2019-03-15 at 2.56.08 AM.png

 

As for the super bright LEDs on order, this is what the vendor states:

  • Size: 5mm
  • Angle of light 120 °
  • STYLE: round top led
  • Voltage: 1.8 - 2.2 V
  • Power:0.06W

The voltage is a bit puzzling since most white or blue LEDs require 3.2 - 3.8 volts. The reds use 1.8 volt. But I guess specifications can vary. As for light output of these LEDs, I don't know if there's a correspondence between watts and lumens. One source gives 80-100 lumens per watt. So 0.06 watts X 90 lumens/watt = 5.4 lumens. I don't know how to covert that to candles.

 

The circuit shown in post #3 draws 24.7 milli-amps with 1.25 volts supply. The LED is passing 8.4 mA, appearing quite bright looking straight at it. But I'm sure it could be passing a lot more current - many can handle 20 mA. With 1.57 volts supply (fresh alkaline battery) the LED is drawing 12.7 mA and is brighter, while the entire circuit is drawing 31.8 mA.

 

As for the QX5252 circuit I think they can produce significant current, since many of the garden lights have multiple white LEDs, assumedly in parallel. The reference in this link gives current capacity of 300mA. But I assume it will work with a single LED, providing the current being draw by the load attached.

 

I experimented a bit with changing the resistor feeding the base of the transistor. That does change the brightness of the LED with corresponding change in current draw of the entire circuit and of the LED. I only have one reasonably bright white LED (an extra in the DIY kit). All of my red LEDs are quite dim. So I'll just wait 'til the super bright ones arrive to determine what's needed to make a bright enough red light.

 

I don't claim to be an electronics wizard. But for your household emergency light, I suggest using one 12 volt battery powering a string of parallel wired LEDs, This  would be safer than 24 volts in my opinion, and easier to implement. Having the LEDs wired in parallel is better than a series arrangement. If one dies, the rest remain lit. Just include a suitable current limiting resistor in series with each LED. For a battery under charge the voltage might rise to nearly 14 volts. If we limit the current to something safe for the LED, say 15 milli-amps then a 650 ohm, 1/4 watt resistor would work - 14 volts minus 4 volts (LED voltage drop) ÷ 0.015 amps = 667 ohms. When not being charged the battery might be around 12.6 volts. The 650 ohm resistor will then limit the current to about 13 mA. Be sure to put a suitable fuse on the positive line to avoid any chance of a fire. If you use reasonable sized wire the voltage drop should not be excessive. With only 15 mA per LED they would all be seeing about the same voltage. If the LEDs at the ends of your string are too dim, just use a smaller resistor to increase the current a bit. If your LED voltage drop is different than the 4 volts used above, then an adjustment to the resistor might be necessary.

 

I don't know if this setup will meet your needs. It will require a bit more assembly effort since each LED will need a resistor soldered on. But it adds the important advantage of one bad LED not spoiling the string. In any case that would be my suggestion. Your mileage will depend on the road you take. But I hope may suggestions will be of use.

 

Best Regards,

Russ


Edited by Rustler46, 15 March 2019 - 04:53 PM.

  • Tenacious likes this

#14 Tenacious

Tenacious

    Viking 1

  • *****
  • Posts: 629
  • Joined: 19 Oct 2015
  • Loc: Northern KY

Posted 15 March 2019 - 08:12 AM

Good Morning Russ

 

I think you are getting a great price on those items that you ordered from China, especially in large quantities.  I find it tough to wait when excited about a new project.   With the loss of Radio Shack, most areas of the country have no local source for electronic parts in small quantities.  I'm lucky where I live to have a few remaining outlets (Micro Center, Debco, and a handful of small businesses).  Does the retail chain Frys exist in Oregon?  Just wondering.

 

I agree about the forward voltages required by LEDs of different colors, white usually near 4 Vdc.  The single color LEDs (red, green, etc) usually require half of that.

 

Like you, I think the power spec is the volt/current dissipation, not light producing efficiency.  This efficiency improves all the time and probably varies with manufacturer and process.



#15 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 15 March 2019 - 02:23 PM

Sadly our local Radio Shack closed last year. I believe there is a Fry's in Portland. So that is an option. For most of what I use it's either rooting through my spare parts cache or waiting for the China shipment. Ebay is great now that I've learned to use the Advanced Search feature. One option is to sort in ascending order by price + shipping cost. They still list some sponsored items first. But it reduces how many listings one has to look at in finding what you want.

 

So the wait is on. Time to switch gears, clean up my electronics mess until parts start arriving. 



#16 Bob4BVM

Bob4BVM

    Surveyor 1

  • *****
  • Posts: 1763
  • Joined: 23 Mar 2015
  • Loc: W. Oregon

Posted 15 March 2019 - 02:59 PM

Nice work there Russ !

Another use for your ckt would be powering the LED in a Telrad on  a single AA

To adjust brightness of you LED, just stick a pot in series (wiper + one end) with either LED lead to adjust the current. A low value pot like 50-150 ohms would be best if you have one.

 

BTW have you used your bino-chair lately ? I am in the process of motorizing mine.

 

CS

Bob


  • Rustler46 likes this

#17 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 15 March 2019 - 04:49 PM

Nice work there Russ !

Another use for your ckt would be powering the LED in a Telrad on  a single AA

To adjust brightness of you LED, just stick a pot in series (wiper + one end) with either LED lead to adjust the current. A low value pot like 50-150 ohms would be best if you have one.

 

BTW have you used your bino-chair lately ? I am in the process of motorizing mine.

 

CS

Bob

Nice to hear from you, Bob.

 

Thanks for the information. That could be an option for powering the Telrad. I've considered using the circuit in an Orion variable red flashlight. But since it is a $30 item, it makes more sense to leave it alone and make my own red light.

 

For one of my Telrads I made a dew heater following directions on the DewBuster website. It keeps dew at bay when run at medium heat. I've found the aftermarket dew-shield (the one that flips up to cover the window) to be useless.

 

As for the Binoculars-chair, I haven't used it much since discovering my 15 X 80 binoculars have gotten way out of collimation. They used to be just slightly off, well within the range of one's eyes to correct. But I'll need to get brave and open up the binos to see what is amiss. Maybe a prism has become unglued from its mounting hardware. So many interesting DIY projects for cloudy nights. 



#18 BGRE

BGRE

    Surveyor 1

  • *****
  • Posts: 1943
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 15 March 2019 - 05:11 PM

Nice work there Russ !

Another use for your ckt would be powering the LED in a Telrad on  a single AA

To adjust brightness of you LED, just stick a pot in series (wiper + one end) with either LED lead to adjust the current. A low value pot like 50-150 ohms would be best if you have one.

 

BTW have you used your bino-chair lately ? I am in the process of motorizing mine.

 

CS

Bob

You obviously don't understand how the circuit works. Installing a sufficiently large resistor in  series with the LED can lead to destruction of the transistor. 

Installing the potresistor between the battery and the circuit is more effective and avoids transistor breakdown. Try simulating the effect of modifications to see what actually happens. Alternatively use a scope to monitor the diode current and collector emitter voltages.


  • Rustler46 likes this

#19 Geo.

Geo.

    Vendor - Nexstar Parts

  • *****
  • Vendors
  • Posts: 4991
  • Joined: 01 Oct 2008
  • Loc: Upstate NY

Posted 19 March 2019 - 08:40 AM

Sadly our local Radio Shack closed last year. I believe there is a Fry's in Portland. So that is an option. For most of what I use it's either rooting through my spare parts cache or waiting for the China shipment. Ebay is great now that I've learned to use the Advanced Search feature. One option is to sort in ascending order by price + shipping cost. They still list some sponsored items first. But it reduces how many listings one has to look at in finding what you want.

 

So the wait is on. Time to switch gears, clean up my electronics mess until parts start arriving. 

I've been buying discrete components from eBay sellers for a while. Way better prices than you ever saw at RS. 

 

Recent purchases:

 

80 x 0.1uF 50V ceramic disc caps 4 days delivery - US seller  - $3

100 x 1/6W 2KΩ Metal Film Resistor Tolerance ±1% - $1

50 x 4.7KΩ 1/4W 0.25W 5% Carbon Film Resistors - $2

100 x RJ9 4P4C modular jacks - $19. DigiKey gets $4-5 a piece plus shipping for modular jacks.

 

Yes, I have plans for these components. 


  • Rustler46 likes this

#20 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 22 March 2019 - 02:27 AM

Here's an update on my Joule-thief project. None of the discrete components ordered have arrived from China. But a couple of rolls of 30 gauge wire for the coils have arrived from a domestic source. There are enough dead fluorescent lights in my recycle box to provide some toroid coil cores. So tomorrow I'll get to work and wind some coils with available supplies. 

 

My home-made variable linear DC power supply has failed (the one shown in some of the photos). So I ordered a nice replacement (switching) power supply - 0-30 volts DC, up to 5 amps. This has arrived and is very handy for testing variations in the circuit on the bread board test setup. It displays current draw and voltage using 4 digits for each display. Voltage and current limits can be set. Once the current limit is reached, voltage is automatically adjusted to stay within current limit. This is very handy for testing LEDs. I just set the current limit for 15 ma and hook the lamp directly to the supply without series resistor. This protects the LED and verifies it is working.

 

For testing an overall circuit one can set the current limit to be above actual current draw. The supply then displays what the circuit draws with variations in supply voltage and in some component values. This bench supply seems to be a good buy at $49 post-paid from a US source. Though I almost never buy the add-on warranty, this time I added an extra full coverage, 3-year policy for $7. While it has a manufacturer's warranty, it's only 30 days and I'm responsible for shipping. The price was so good for the supply, adding a little insurance felt like a good deal. If it doesn't die during the first 3 years, it likely will stay the course. I must have gotten one of the last ones, since it no longer shows up on Ebay. Here's what it looks like:

 

Screen Shot 2019-03-21 at 9.52.46 PM.png

The supply is quite compact, just 6 inches in height. 

 

I used my multimeter to measure the oscillation frequency of the Joule-thief circuits. They are as follows:

  1. Single AAA-cell with DIY Kit - 63,000 Hertz (cycles/second)
  2. Breadboard circuit with home-made coil - 107,000 Hertz

The second circuit has a smaller toroidal core, though both circuits have nearly the same number of windings. Perhaps the smaller coil is contributing to the higher oscillation frequency. 

 

No wonder the light appears to be continuously lit. It's flashing much too fast for your eye to notice the flicker. I have no way to measure the duty-cycle, which is the percentage of time the light is on during each oscillation cycle. But these circuits typically have a 25% duty cycle. This is part of the efficiency of the Joule-thief. Even though the light is only illuminated 25% of the time, it appears to be continuously lit - a real savings in battery usage.

 

I'm looking forward to changing circuit components to see how some variation in LED brightness can be enabled, particularly as the voltage drops down in the vicinity of 0.5 volts. I've found there is some light output at 0.42 volts. But it takes around 0.60 volts for the circuit to turn on.

 

Without being burdened by knowing exactly what I'm doing electronically, it will be necessary to keep track of current being passed through the transistor. Seems like lowering the base drive resistor under low voltage conditions might help brighten then LED as it draws that last Joule of energy from the AAA-cell. My multimeter has a milli-amp measuring capability which will be useful to measure current in different legs of the circuit. The power supply just measures overall circuit draw. When the overall circuit is drawing 30 ma, the LED is passing around 7 ma.

 

The new bench supply along with my multimeter has made testing much easier. Since I have a supply of small toroid cores, it might be fun to experiment with different number of windings or even change the turns ratio from the present 1:1. What better thing to do on a cloudy night.

 

When I get all supplies, I'll make some dim red lights for marking tripod legs and GEM counterweight. I don't want to run into these in the dark. 

 

This looks like a potential source for electronics:


Edited by Rustler46, 22 March 2019 - 05:06 AM.

  • Tenacious likes this

#21 Jon Isaacs

Jon Isaacs

    ISS

  • *****
  • Posts: 76151
  • Joined: 16 Jun 2004
  • Loc: San Diego and Boulevard, CA

Posted 22 March 2019 - 03:49 PM

I believe there is a Fry's in Portland.

 

 

A friend and visited the Fry's in San Diego last month.  There were very few customers, the shelves were nearly empty, it looked like bankruptcy is just around the corner..

 

Jon



#22 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 22 March 2019 - 07:33 PM

Today I've extracted a selection of small toroidal ferrite cores from deceased fluorescent lights. They are in various sizes:

  • 6 mm O.D. 
  • 9 mm
  • 10 mm

Also they come in various colors:

  • Black
  • Grey 
  • Green
  • Blue
  • Salmon

I assume the greater total volume of the core material along with number of turns of wires on the core is related to the total energy that can be stored in the magnetic field. This is the energy that will be dumped into the LED as a higher voltage spike.

 

Not being burdened by knowing what I'm doing, I have some questions for those knowledgeable in things electronic:

  1. What do the colors of the toroidal cores signify?
  2. Is this something to be concerned about with a Joule-thief circuit?
  3. Does the frequency of oscillation relate in some simple way to the duty cycle of the LED?
  4. Is there any reason to vary the turns ratio from the default 1:1?

Since I don't have an oscilloscope, waveform cannot be examined. I can measure current, resistance, capacitance and frequency with the tools available. In any case I'll keep up experimenting - it is quite interesting. With the smallest core, I can easily get 10-15 turns of red/green 30 AWG wires. The larger cores will allow for more turns of wire.

 

I suspect that I can't do too much damage, no matter the choices made. The supply is short-circuit protected with voltage and current set limits. No big deal if I fry a cheap 2N2222 transistor. I'll be careful not to poke my eye with any wires. If I burn myself with the soldering iron, it won't be the first time.

 

Any suggestions for this old experimenter? 



#23 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 22 March 2019 - 07:33 PM

A friend and visited the Fry's in San Diego last month.  There were very few customers, the shelves were nearly empty, it looked like bankruptcy is just around the corner..

 

Jon

frown.gif



#24 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 27 March 2019 - 01:55 AM

While waiting for supplies to arrive from China, I've done some more online research and some experimenting with my test-bed circuit. There is a tremendous amount of information on the web about the Joule-thief circuit. Here's some of what I found:

  • Drawing an alkaline battery down to 0.8 volts or below makes it much more likely to leak. So be aware of that possibility and not leave a dead battery in its holder for a long period of time. This is just good procedure even for new batteries, since they can all leak.
  • The simplest Joule-thief circuit is not very efficient. With a more complex circuit providing brightness control, this can be improved.
  • My simple Joule-thief circuit still provides useful light down to about 0.45 volt. But it is drawing 5-10 milli-amps.
  • Adding some additional components that provide brightness control, it provides useful light down to 0.55 volts and below. 
  • The improved Joule-thief is much more efficient, drawing less than 1 milli-amp at low, but useful brightness. At 2 milli-amps it is quite bright. At full brightness (uncomfortable to look at) it is only drawing around 7 milli-amps at 1.50 volts input.

    Edit:
    The modification is quite simple. The output from the Joule-thief oscillator is rectified with a small signal diode S1 and is used to charge the 0.1 uF capacitor, C2. The variable resistor VR1 is used to pick off a portion of that DC voltage and send it to the base of the regulating transistor Q2. Its collector is attached to the base of the oscillator transistor Q1. When enough positive voltage turns on the regulator transistor, the base of the oscillator transistor begins to be grounded. This affects the oscillator in a way that varies the brightness of the LED fed by the Joule-thief circuit. Capacitor C1 is not needed with a battery powered circuit.

    In use with a fresh AAA battery the light output can be reduced to what is needed, which will conserve battery energy. As the battery nears the end, the light will need to be adjusted to maximum brightness to continue functioning. 

The improved Joule-thief circuit is shown below:

 

Improved Joule-thief.jpg

 

The unlabelled resistor on top-left is 10 K-ohms. Input power (battery or power supply) is along the right side. I used a 475 K-ohm variable resistor (VR1) rather than the indicated 220 K-ohm. It worked fine with that change. Another change necessitated by available supplies was that I used 2N2222 transistors, which also worked quite well.

 

This circuit came from the discussion in the following link:

While there is an increase in the number of components, the benefits of brightness control and improved efficiency are worth the added complexity. Here are a couple of photos showing the test-bed circuit:

 

Improved Joule-thief-01297.jpg

 

The multimeter gives a more accurate current measurement of 2.1 milli-amps compared to the power supply's reading of 1 milli-amp.

 

Improved Joule-thief-01298.jpg

 

The big orange 0.1 micro-farad capacitor is the only one I had. But a physically smaller wafer capacitor would work as well. I'm hoping to mount all components on a small perf-board that fits into the extra space in a battery holder. I'll be posting what results as soon as the "boat from China" arrives.

 

Edit:

 

Some supplies came in today:

  • Battery holder for single cell AA or AAA battery with switch. This has enough space inside for the electrical components (coil, transistors, resistors, capacitors).
  • Super bright LEDs of white plus 4 other colors (20 each) - this will give an assortment of colors to choose from, mostly just white and red.

I'll drill a hole for the LED in the bottom of the extra space in the battery holder. Then most components will be soldered to the leads for the LED as shown on my original post. The coil will set in the space separately. When all is attached and tested I'll insert the LED in its hole and glue into place. The coil will be anchored by some glue as well. 

 

The variable resistor will be a problem - finding one small enough to fit inside. I might just substitute suitable fixed value resistors to provide maximum light and forgo the adjustability. The circuit would still be much more efficient than the straight Joule-thief circuit.

 

When all is according to my liking I'll make some more of these with red lights for use at the telescope. One could be used to illuminate my sketching pad. Others could be attached to the mount's counterweight and tripod legs to avoid bumping these in the dark


Edited by Rustler46, 28 March 2019 - 01:06 AM.

  • Bob4BVM, Tenacious and drneilmb like this

#25 Rustler46

Rustler46

    Viking 1

  • -----
  • topic starter
  • Posts: 686
  • Joined: 10 Feb 2018
  • Loc: Coos Bay, Oregon

Posted 28 March 2019 - 11:53 PM

I've assembled simple a Joule-thief flashlight. 

 

Joule-thief Flashlight-01300.jpg

 

This test circuit is the lower efficiency one with just 4 components - an LED, transistor, resistor & coil. The most expensive component is the battery case at $0.99. The rest of the components add up to just pennies. This circuit operates at 90,000 cycles per second. Here's a close-up of the innards:

 

Joule-thief Flashlight-01303.jpg

 

With the smallest coil (seen in the above photo) the rest of the components for the more efficient version should fit into the case. This will considerably lessen the draw on the battery. This improved version will add 6 components to the count - a diode, capacitor, transistor & 3 resistors. While it would be non-adjustable as for brightness, it would be much more efficient. Current draw from the battery would drop from 24 milli-amps to around 8 milli-amps with a fresh battery at 1.5 volts. 

 

I'll need to order some more parts if the circuit is assembled according to plans. My test circuit was assembled from available resistors. To get a single 220 k-ohm resistor for VR1, it was necessary to string together 3 lessor value resistors. This would be hard to fit in the available space. But not wanting to wait for weeks to get the needed components, I decided to experiment with what I had on hand.

 

This turned out to be a success. I used fixed resistors to replace VR1 with a voltage divider. Resistors of 120 k-ohms and 18 ohms in a divider provided voltage to the base of the regulator transistor. Also the 0.1 mfd capacitor C2  was replaced with a much smaller one of 0.0002 mfd. Oscillating frequency is around 150,000 cycles per second. The current draw with this version is 8.5 milli-amps at maximum battery - down from 24 milli-amps. At low but useful brightness (at 0.60 volts) the current is less than 2 milli-amps. At lowest light, just before cutoff (at 0.53 volts) the current is less than 1 milli-amp. So the improved efficiency circuit made with available parts uses about 1/3 the power compared to the simple Joule-thief circuit. I'll make this one up tomorrow.

 

Once both versions of the light are finished, I can compare the longevity of each. I expect the more efficient light to last for days with a fresh AA cell. Great fun on cloudy nights!

 

 


  • siriusandthepup and Tenacious like this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics