default

2014-03-11 DC-DC Converter Using LT1073

Joule Thief SMPS DC-DC
on March 11th, 2014 by - 5 Comments

I bought some LT-1073 DC to DC converter chips from Digi-key for about $4.00 apiece.  I’m going to try to make something small to power 9 volt radios and other devices.  It will use a 1.5 volt AA cell for the input  battery.

The switching transistor is on the chip so the only parts that are needed are the inductor, a Schottky diode, a few capacitors and a two resistor V divider to set the output voltage.  The data sheet specifies 1 percent resistors but that kind of accuracy is not needed for a simple 9 volt power supply.  I calculated that a 1 Meg resistor and two 47k resistors in parallel should give about 9V.

The inductor is shown as 120 Microhenrys.  I’m going to use either a toroid or a pot core to minimize interference as recommended in the data sheet.  I have plenty of 180 microhenry chokes so I’m going to try one of those.  The graph in the data sheet shows a minimum value which is 120 uH, but I don’t have that value, which is hard one to find.  I may have to wind my own on a toroid core, which is easy to do.  I have some small cores that should work okay.

They talk about using low ESR capacitors or else a tantalum capacitor.  I’m going to use a regular 470uF electrolytic and another 0.1 uF capacitor across it, mainly because I found that they do a good job in my other DC to DC converters.  Maybe I should order these low ESR capacitors that they recommend, if I can find them for a reasonable price.  They’re OS-CON.  I checked the catalog and found that they’re over $1 apiece in quantities of 10 or more.  ouch.

I spent a few hours drawing a layout of the circuit on strip board.  I then mounted the parts on a piece of real strip board.  I used the 9 Volt schematic that’s in the data sheet.  The only change I made is to use a 910 K 5% resistor and 22k 5% resistor.  The 1 percent resistors they show are not necessary.  I also added a 470 Microfarad capacitor across the 1.5v battery.  I only had to make 3 cuts to the strips.  I sound a toroid core with some 24 gauge insulated wire to get 120 microhenrys.  The DC resistance of the toroid winding was very low, below a tenth of an ohm.

I added red and black wires to the input and output.  I double checked to make sure everything was wired correctly.  I hooked it up to the power and a voltmeter to the output.  It measured 9.27v for awhile.  then the voltage started dropping down to 9.2 volts.  When it got to that point it jumped back up to 9.2 7 Volts.  It continued to do this as long as there was no load.  I think the larger capacitor on the output causes this peculiarity.  If it had been smalVoltage voltage would probably not jump up and down so much.

I used a 1k resistor for a load, to make it easy to figure out the load current – 9 volts cross the resistor gives 9 milliamps load current.  As shown in the data sheet it’s capable of putting out 9.27 milliamps at 9.27 volts when the input was 1.5 volts.  But when I set the input at 1.00 volts then the output dropped to 8.7 volts.  So it will need a fresh battery often.  The input current went up to 250 milliamps with the 1k load.

I added a 9v battery snap-on connector to the output, and tomorrow I’m going to try to find a device that uses a 9 volt battery and only draws about 10 milliamps.  Most pocket radios draw 2 or 3 times that much so I don’t believe that this converter will be able to power a pocket radio.  The closest thing at hand is a digital multi-meter.  I would really like to power a regenerative receiver so I can find out how much RF interference this converter puts out.

BTW, I’m not a fan oof Digikey, mainly because they’re not the cheapest distributor.  And yesterday I got a CD in the mail – at least that’s what I thought it was.  When I opened it up, POP!  This paper cube popped out.  It has a rubber band that pulls it into a cube from its flattened state in the CD folder.  I thought to myself, that was a really big waste.  Now they’ll probably send me a full size paper catalog in the mail.  It will probably cost several dollars to print and mail.  These are just another couple reasons why I don’t like Digikey.

Note: I have been using my Android phone to enter the text in this and previous blogs.  Some text is answered – oops- entered by using the voice-to-text app.  I have to double check the text because this “feature” sometimes does very strange things, like translating certain words into complete and utter nonsense.  No matter how many times I check, the words might sneak through.  So forgive me 4 making some on the states – oops, did it again – odd mistakes occasionally.  These sometimes hilarious goofs were really its fault, not mine.

default

2014-02-28 Book Reading Light

LED Lighting
on February 28th, 2014 by - No Comments

When I go to the big box stores I usually stop look up the flashlights to see what’s new.  I often see a palm sized flashlight with 27  LED’s, 24 of them on the flat side.  The case is rounded on the corners and is typically yellow and black; the price is about $4 or $5.  A few years ago I bought one to remove the LED’s but I never got around to it.  Instead I drove a hole in the case and ran two wires out so instead of three AA cells, the light could be connected to an external power supply.  I used it for a while on a high powered Joule thief.  I disconnected it from the Joule thief and connected it to a power supply.  I liked the beam pattern because it spreads out to about the size of a page at a distance of 2 to3 feet.  I just recently made a book reading light but I thought this one would make a good one too.

My first thought was how to power this light.  3 AA cells make 4.5 volts so a good choice would be a 5 Volt power supply.  I connected the lights 2 a power supply turned it up to 4.50 volts.  The current  was about 400 milliamps, which is about 16.7 milliamps per LED.  This is a very good point to run these LEDs.

The light already has an internal  resistor so the problem is to drop the voltage from 5 volts down to 4.5 volts.  I put 1N4003 in series with the power supply but the voltage went down to 4.2 volts.  I tried a shot keep dialed shot key diode but the current was over 400 milliamps.  I tried several more diodes and finally found A 1N5403 3 amp diode that drops about .7 volts and the current is about 390 milliamps.  It got slightly warm after awhile but it’s only dissipating .3 watts.  This one looked like it will do a good job.

My next task is to find a 5 volt wall wart adapter that has the correct plug, or else cut the plug off and put something else on it.  I have some adapters have very tiny plug and I know I can’t find a jack for it.

default

2014-02-27 Adjustable Power Joule Thief

Joule Thief SMPS DC-DC
on February 27th, 2014 by - No Comments

I used an older Joule Thief that I pulled out of the pile to make an adjustable power Joule Thief.  I changed the transistor to a SS9018, put a 1 ohm resistor in series with the cathode of the white LED to allow measuring the LED current.  I left a 470 Ohm resistor between the base of the transistor and the feedback winding – it can be zero ohms.  Between the other end of the feedback winding and the positive I put a 470 K trim pot to allow adjustment of the brightness.  From this end of the feedback winding to negative I put A 150 Picofarad  capacitor.  Without this low value capacitor the LED will not light at higher trim pot settings and lower brightness levels.

The transistor is a low current preamplifier and has very high current gain, up to a thousand, so the resistance of the trim pot has to be high to allow the brightness to be adjusted to low values. Still the trim pot has point where it is very touchy and just a slight amount of movement causes a large difference in brightness.  This seems to indicate that there are two ‘modes’ of operation, and it tries to jump from one mode to the other at this touchy spot when adjusted.  This may have to do with interaction between the capacitor and coil.

The 470 K trim pot allows adjustment from about 50 milliamps battery current down to less than 1 milliamp.  Correspondingly the LED brightness goes from high to low as it’s adjusted.  If a point is found where the LED is stable, the resistance of the pot can be measured and then the pot can be removed and be replaced with a fixed resistor.

I monitored the input and output current and adjusted it for about two milliamps of input current and about 1 milliamp to the LED. The trim pot resistance was about 143k. I connected this circuit to a AAA cell on a table in my bedroom and watched it during the night. At only 1 milliamp, the LED was bright enough to serve as a good night light.  Assuming 2000 mAh for a AA cell, the JT should run for 1000 hours or about 6 weeks at 2mA.  But we all know that JTs seem to run forever as the cell is depleted and the LED brightness drops.

Update -This has been running for a few weeks with a cheap carbon Zinc battery. It makes a good night light.  I have found that these very low power JTs look brightest to the eye when the LED is an ultrabright green.  The human eye is most sensitive in the yellow to green wavelengths.  Also its most likely that a white LED is losing some efficiency when it converts blue light to white.  So the green LED is probably more efficient.

default

2014-01-31 Penny Battery Lights Red LED

Battery or Cell Free Energy Renewable Energy
on January 31st, 2014 by - No Comments

I had fairly good success with the Zinc Air (sort of) battery, so I decided to make something using pennies.  Prior to 1981, the U.S. pennies were made of solid copper, but 1981 and after the pennies have been made with zinc and a copper coating.

I got six pennies, three of the solid copper and three of the zinc, and took them outside and sanded them on the walkway cement.  I sanded the back side, to get rid of the copper coating and expose the zinc, and to get a smooth surface on the solid copper pennies.  I tried a piece of heavy sandpaper and it helped speed up the sanding.

Using some cut up paper towel, I positioned the pennies in a stack so that I had:

Copper, paper towel,  zinc, copper, paper towel,  zinc, copper, paper towel,  zinc

I used two layers of paper towel.  This gave me three cells in series.  I then clamped the stack together with a plastic spring clamp, and positioned a red LED so that the leads were clamped to the ends, with the cathode or flat spot towards the zinc penny.

I poured out a bottle of 5% white vinegar into the bottle cap and I poured some on the paper towels.  The LED lit up bright, but after ten or twenty seconds, went dim.  I thought that the spring clamp was too strong, and not enough vinegar was getting into the paper towels.  I removed the spring clamp and put a rubber band around the pennies, and poured some more vinegar onto the paper towels.  The red LED again lit up, but went dim even quicker.  The battery was working, but it just wasn’t working long enough.  I tried to get a picture of it but by the time I got the camera focused, the LED had gone dim.

I’ve read that the hydrogen bubbles build up on the plates, and that stops the current from flowing.  So I decided I needed a different battery design with the pennies dipped into jars of vinegar, which would allow the bubbles to escape.  I got three small plastic ‘custard cups’ and a plastic tube to hold them together.  I took the battery stack apart and cleaned off the pennies.  I soldered two pairs of the zinc and copper pennies together with two 3 inch lengths of solid copper wire, soldering to the edge to leave the surfaces clean.  The remaining pair of pennies each got six inch wire lengths for connecting to the LED.

I positioned the pennies in the cups to make the three cells in series.  I filled up each cup with enough vinegar to cover the pennies, and Bingo!  The red LED lit up brightly and stayed lit.  After awhile the bubbles built up on the zinc pennies, and the LED became a bit dimmer but still stayed lit.  I could shake the whole tub and get some of the bubbles off the pennies and the LED would light up brighter for a short while.

After more than thirty hours, the LED is still lit, and the zinc pennies are bubbling vigorously.  A little shake gets the LED to light up a bit brighter for awhile.  But it’s like the Eveready Bunny, it just keeps going and going.

Update Sat Feb 1 – It’s been 45 hours, and the zinc pennies are still bubbling and the LED is still lit.  The vinegar evaporates (or is being converted into hydrogen bubbles) so I added a small amount to each cup to keep the pennies under the solution.  Sort of like the lead acid batteries in vehicles: sometimes they need a bit of H2O to keep them running.  I bumped into the battery and some of the vinegar spilled out onto the counter, so I retired the experiment.

One thing I’d like to do is triple the number of pennies so that the battery will put out three times as much current.  Someone will probably recommend adding a Joule Thief, but I can’t see why I should waste half the power when I can add another cell and get higher voltage with no loss of power.

One thing I’ve noticed is that when I look through the pennies I receive, I seldom find any pennies that are solid copper.  I think the copper thieves have removed them, turned them into scrap and sold them.  I read an interesting article about how the Gramm Leach Bliley Act allowed the big banks to get into the commodities business, and how they manipulated the metals markets to raise prices.  The price of aluminum went up, and other metals did, too.  Sounded like the fox was in charge of the chicken house.  Here is one article about this.  http://www.marketoracle.co.uk/Article43786.html

default

2014-01-28 Microwave Oven Is Goofy

computers Other Off Topic
on January 28th, 2014 by - 1 Comment

When I use my Frigidaire Gallery microwave oven my mind usually restricts the numbers I input to the normal digits allowed by a clock, such as 0 through 59.  I used it this morning and meant to put in the right numbers but instead I pressed seven and zero.  But instead of giving me an error, it went right ahead and started running, counting down from 70 seconds.  HUH?  I didn’t think it was supposed to do that!

So I thought about it for a bit and decided to try it with another invalid number.  I pressed 1, 9, and 9 and hit the start.  It started counting down from 99 instead of 59!  That’s really weird.  No problem at all, it just takes whatever you press.  That’s really goofy, I thought.  But still it doesn’t really matter as long as the microwave oven actually runs for two minutes and thirty-nine seconds, which is what I input.  It just lets me decide that I want what I put in, without any prejudgments on my out-of-line input.

I have never tried this with any other microwave oven, probably because I just press the minute button twice for two minutes.  My old microwave oven (I think it was a Daewoo) didn’t have a full number keypad, it had seconds, minutes and tens of minutes buttons.  The one at work had buttons to heat a cup of coffee or cook a TV dinner; I never used the number buttons.  Next time I’m over at someone’s house I’ll have to try to trick their microwave oven into taking some invalid numbers and see what happens.

default

2014-01-27 LED Auto Dome Light Is Too Hot

LED LEDs Lighting
on January 27th, 2014 by - No Comments

P1040004S6I took apart a “Type R” LED light, which is a substitute for a wedge base auto light used in instrument panels and dome lights.  This light has an 8 mm diameter LED and a 560 ohm, 1/8 watt resistor in the wedge base.  Assuming that the LED drops 3.2 volts and the auto’s battery voltage is 13.8V, the resistor has 10.6 volts across it, and that calculates to slightly less than 20 milliamps current.  That’s fine for the LED, but if it’s rated for more than a tenth of a watt, then it could handle more current.  But the resistor isn’t so fortunate.

The formula for the dissipation of a resistance is voltage squared divided by the resistance.  In this case, we have 10.6 squared divided by 560, which gives a bit more than 200 milliwatts.  This is a 1/8 watt resistor, or 125 milliwatts, so the resistor is dissipating 60 percent or more than its maximum rated wattage.  This little resistor must be getting very hot.  And that is an invitation to catch fire, or short out, or both, in which case the LED will be subjected to too much current and also overheat and most likely fail.  And the excess heat could shorten the life or damage the LED.

And this assumes the battery voltage stays at 13.8 volts.  Problem is that the voltage during charging may go quite a bit higher, up to 15 volts.  So there are even worse conditions that the light may encounter.  Another factor is the fixture that it’s in may get warm from other things such as electronics or even other lights.  If you ask me, this does not seem to be a safe light to use in an a vehicle.

default

2014-01-25 70.7 Volt PA Transformer Impedances

Audio
on January 25th, 2014 by - No Comments

When you get a 70.7 Volt Public Address Transformer, you see that it has connections  for different speakers, such as 4, 8, and 16 ohms.  On the other winding are various taps for different power outputs such as .625, 1.25, 2.5, 5 and 10 Watts.  But there is nothing to tell what the impedance is of the winding taps.  I was curious what these impedances were, so I did some simple calculations  and found the impedance of each winding.

In a DC circuit, the power is equal to the voltage E squared divided by the resistance.  We will assume that  the reactance is zero so the answers are the same for DC and AC circuits. Algebraically  rearranging the formula, we have the resistance is equal to the voltage squared, divided by the power.

We take the 70.7 voltage and square it and we get 5000.  We divide this by the power, .625 watts.  We get 8000 ohms, which is the impedance of this winding tap. For the 1.25 way tap we get 4000 ohms.  For the 2.5 watt tap we get 2000 ohms for the answer.  For the 5 watt tap we get 1000 ohms for the answer.  And for the 10 watt tap we get 500 ohms.

So if you need an audio transformer that is 1000 ohms to 4, 8, or 16 ohms, you can use one of these and choose the 5 watt tap, for example.  You can also use one for a 5000 to 500 ohm autoformer, but since both taps are on the same winding, they are not isolated from each other.  To get isolation, you could put two transformers back-to-back, and contact the 16 ohm windings together.  You could also connect  two of them together in various series sand/or parallel combinations to get various impedance combinations.

These transformers seem to be much cheaper and more easily obtainable than an audio output transformer.  I used a 10 watt transformer in the above examples, but they come in various wattages, both higher and lower.  So next time you visit an electronics store, check these out to see what they cost.  Sometimes when buildings are demolished there are ceiling speakers left in the rubble, and even if the speaker is damaged, the transformer may have survived.  More here:  http://en.m.wikipedia.org/wiki/70_volt

default

2014-01-20 Zinc Air Battery – Sort Of

Battery or Cell Renewable Energy
on January 20th, 2014 by - 2 Comments

I saw a Youtube video on how to make a zinc-air battery using a zinc electrode and a piece of steel wool for the other electrode.  The zinc reacts with the air and electrolyte (potassium hydroxide or KOH) and the steel wool does not, it serves as the conductor.

I ordered some KOH online from a dealer, it cost a bit more than $6.00 for a two pound bottle but the shipping was nearly as much.  It is used to make soap, among other things.  Caution! This is caustic and burns skin, eyes and is poison.  The flakes get very hot when mixed with water, so you have to be careful and not let it get on you.

The video said to add 50 grams of KOH to 150 grams of water.  It did get very hot, and I put the glass jar into a pan of cool water to cool it off.  Instead of the steel wool, I used a piece of copper plate small enough to fit into the jar.  Instead of a zinc plate, I used a zinc electrical fitting.  I drilled a small hole in both of these electrodes so I could put a copper wire through and hang the electrodes in the solution by the copper wires.  The zinc electrode was completely submerged in the solution, so I don’t think it could get any air.

I measured the voltage at 1.06 volts.  I connected a germanium Joule Thief to the wires and the two yellow LEDs lit up brightly, but after ten or so seconds they began to dim.  The copper plate didn’t appear to be changed, but the zinc started bubbling and was coated with bubbles.  The LEDs continued to dim, until they went out.  The voltage dropped as the LEDs dimmed.

I disconnected he JT and the voltage recovered and climbed back up to 1.06 volts.  When I connected the JT, the LEDs flashed on brightly, but dimmed even more rapidly, going out in a few seconds.  The buildup of hydrogen bubbles and/or lack of air could be causing the fast drop in the output current.

Update Jan 21 – I thought that the solution to this dropoff in current would be to put a large capacitor across the output to store the charge as it came off the cell.  I went to the box of large capacitors and was going to take one of the 10 000 uF electrolytics, but instead I grabbed the 350 farad super capacitor.  I connected it up to the cell and of course it acted as a super large current sink.  The voltmeter barely moved as it charged.  Several hours later I checked the voltage and it had climbed to 300 millivolts.  At this rate it is going to take a lot more than a day to fully charge this capacitor.

As of 8 PM Jan 23, the meter reads 645 millivolts.  It’s still climbing slowly, a few millivolts per hour.   But now that it’s over 0.6 volts, it will run a JT with a silicon transistor for awhile, until the voltage drops down to about 0.45 volts.

As of 10 PM Jan 24, the meter reads 780 millivolts.  This is about half way to 1.6 volts.

As of 10 PM Jan 25, the meter reads a bit over 900 millivolts.

As of 10 PM Jan 26, the meter reads 965 millivolts.  I noticed that the rate of charging has gone down, probably from a number of factors.  I noticed that the zinc electrode has turned black, and the copper electrode has also tarnished.  Some of the electrolyte has evaporated.  I disconnected the super cap, and the meter read .978 volt, lower than when I started.  I took the electrodes out and washed them off and put them back, and added some water.  It looks like the current and charging rate have gone up slightly.  I’m not sure the voltage will go past 0.978 volts.  I’ll see what happens tomorrow evening.

As of 10 PM Jan 27, the meter read 1.077 volts.   I got up this morning and checked the voltage and it had passed 1 volt by a few millivolts.  The cleaning I did seemed to have sped up the charging quite a bit.

As of 10 PM Jan 30, the meter read 1.132 volts.  I have watched it for the last 24 hours, and it has maxed out and does not go any higher.  The reading varies a few millivolts between 1127 and 1132 millivolts, so there may be some change due to temperature or the meter.  I have added a bit more vinegar to the jar to make up for evaporation, but I don’t think this has affected the readings.

Feb 2 – Sometime today the copper wire holding the zinc electrical fitting gave out and the fitting dropped to the bottom of the glass jar.  I’m guessing that the localized interaction between the copper wire and zinc ate both away and the copper wire got so thin it broke.  And since  the wire and copper plate were connected by the fluid, it discharged the super capacitor.  I guess I could say that the test has ended.

default

2014-01-14 Watson’s Standardized Joule Thief

Joule Thief SMPS DC-DC Uncategorized
on January 14th, 2014 by - No Comments

DSC_0016S6I found this picture that I had posted to my late, great watsonseblog.blogspot.com.  I have added my new rustybolt.info/wordpress blog address to it.

This is a Joule Thief that uses standard parts, including a readily available toroid core.  This high permeability core can be wound with various sizes of wire, but about ten turns will give somewhere around 200 microhenrys, which is more than enough for a Joule Thief.

default

2014-01-13 Coil For FM Microphone

Radio Frequency
on January 13th, 2014 by - No Comments

I think I built one of these years ago but it must be buried in a box somewhere, probably in a pile of boxes in the garage. It’s a FM microphone transmitter, of which I have several, built from kits and from scratch. But this one was unique in that I built it with a coil of a single turn of heavy wire. I used a piece of 12 or 14 AWG bare solid house wire, the kind found inside of the walls. The insulation was stripped off and I bent it into a circle about 40 to 45 mm or 1-9/16 to 1-3/4 inches diameter. I just arbitrarily chose a length of wire and bent it, with the ends sticking out about 3/8 inch so that I could solder the other components onto it.

But this time I decided to try calculating the size of the wire from a formula given in Illustrated Handbook of Electronic Tables… by Ludwig on P. 206. I tried it with several different radiuses, but I used 0.08 inch – the diameter of the wire – for the coil length of all of them. The results I got were not close to the results I had obtained experimentally. The only thing I can think of that might be causing this is that the algebraic calculations I was doing were not the same as what is shown in the formula. It’s possible that the formula is incorrect. I’m going to have to try some other manuals to see if they agree.

I did calculations for a coil of 0.1 microhenry and a diameter of 1.8 inches which is a radius of 0.9 inches. The answer came to 1.048 turns. But I have already cut and bent a piece of wire into a circular coil with almost but not quite a single turn and a diameter of 1.8 inches and it measures 0.098 microhenry or 98 nanohenrys. The formula gives a somewhat close answer, but it is reliably high from this and a few other calculations that I did. This is acceptable as long as one knows that the calculations will be off a certain amount, so that it can be compensated.

I built the transmitter using the loop of 12 AWG bare copper wire approximately 1.95 inches in diameter. The oscillator transistor is a PN2222A, and the audio amp is also a PN2222A. The loop has a 15 pF ceramic capacitor across it to resonate at the low end of the FM band around 89 MHz. This will have to be adjusted for the various components used, especially the transistor.

The emitter resistor for the RF oscillator determines the output power. I used a 510 ohm resistor, which I a bit higher than the typical 330 to 470 ohms. The emitter current is about 2 MA, which is less than 10 milliwatts total power with fresh batteries. Less than a third of that gets out as RF, so this circuit is just ‘flea power’ compared to other wireless microphones.

The audio transistor amplifies the electret condenser microphone (round black thing at the bottom). The battery pack is three AAA cells in series for 4.5 volts. If rechargeable cells are used, there should be four, totaling 5 V.

I took the ‘bug’ out to an open field and propped it upright on a pile of dirt so the coil (or ring) was above the rest of the circuit. I used a cheap portable radio with an extra length of wire added to the whip antenna to make it about 30 inches. I could get the signal up to about 100 to 120 feet or 30 to 36 meters before it faded out. But another station was on a nearby frequency and it bled through when the bug got weak. This was around 90 MHz, at the low end of the FM band. I might have been able to receive it farther away if I had a better radio and if it had been farther away from a strong station. But in a large urban area it’s nearly impossible to find a frequency that is not already being used by a strong FM station.

© RustyBolt.Info/wordpress
CyberChimps