https://www.youtube.com/watch?v=uSVLemZFAbs
The LED is a 3.2V @20ma 16000 mcd variety and the video shows some results I need help with. The biggest thing I need help with is how to accurately measure efficiency with a scope and a DMM because that is all I have.
Here is the schematic diagram as I can't find the link on his very discombobulated blog.
The one I replicated is the top one and no matter what I changed the frequency would not budge.
Hello,
The new blogsite is http://rustybolt.info/wordpress/?page_id=560 (http://rustybolt.info/wordpress/?page_id=560)
I tried to search for the circuit you refer to above but has not found. I found this though:
http://rustybolt.info/wordpress/?p=221 (http://rustybolt.info/wordpress/?p=221)
By the way he is also a member here if I recall correctly: http://www.overunity.com/13175/25mv-joule-thief-powered-by-peltier-merely-using-our-body-heat-free-energy-247/msg351059/#msg351059
Gyula
Quote from: Legalizeshemp420 on October 11, 2013, 08:46:23 AM
Here is the schematic diagram as I can't find the link on his very discombobulated blog.
The one I replicated is the top one and no matter what I changed the frequency would not budge.
What do you mean? There are many things that should change the frequency of any JT circuit. Change just about anything and the frequency will change. The frequency will change even as the supply battery voltage drops, as the transistor warms and cools, as you change winding numbers or even the position of the transformer wrt the other components.
Are you sure, with your new scope, that you are actually displaying the frequency of the JT's oscillations and not something else? I remember we had some trouble with that when you were using the simulator. Several times you had displays that didn't actually show what the JT was doing but were missing the JT oscs or were showing probe and instrument noise only.
ETA: How many turns on the little toroid? (I had to take apart 3 CFLs to find a 3/8" toroid.)
All I know is that it is hooked up on the anode and cathode of the LED and that 400khz is what I am getting to it.
For those who may have missed the image.
Quote from: Legalizeshemp420 on October 11, 2013, 05:48:17 PM
All I know is that it is hooked up on the anode and cathode of the LED and that 400khz is what I am getting to it.
For those who may have missed the image.
Did you wind the toroid?
Well, that certainly looks like a JT waveform. What's the timebase setting? What happens when you bring a strong magnet over to the toroid?
Frequency changes.
The toroid I did not wind but performs exceptionally well in my standard JT so I am using it for this purpose but 400khz?
Timebase setting is .5us (lowest this 20mhz scope will go) and I counted about .5 for a full wave which is .25us for the period.
Silly me, I put two layers of #27 on my little 3/8 inch toroid from a CFL before I measured it. It was 1.5 milliHenry! Now I've stripped off all but 14 turns (each winding) and it still measures 500 microHenry.
How many turns on the toroid?
18 on each side of 20 or 22ga wire or possibly 24ga but no smaller wire.
I can't measure the inductance value of it myself though.
Quote from: Legalizeshemp420 on October 11, 2013, 06:20:02 PM
Frequency changes.
The toroid I did not wind but performs exceptionally well in my standard JT so I am using it for this purpose but 400khz?
Timebase setting is .5us (lowest this 20mhz scope will go) and I counted about .5 for a full wave which is .25us for the period.
There are 9 minor divisions across the shot you show. That makes 9 x 0.0000005 sec = 0.0000045 sec or 4.5 microseconds for the full screen. There are 17 full periods shown. 17/0.0000045 = 3.78 MegaHz.
It's easy to misplace decimal points, for sure, I do it all the time, and maybe I did again. But whenever possible on an analog scope, use as many screen divisions and peaks as possible to do your frequency calculations. Accuracy improves with more cycles and more divisions.
Quote from: TinselKoala on October 11, 2013, 06:29:33 PM
There are 9 minor divisions across the shot you show. That makes 9 x 0.0000005 sec = 0.0000045 sec or 4.5 microseconds for the full screen. There are 17 full periods shown. 17/0.0000045 = 3.78 MegaHz.
It's easy to misplace decimal points, for sure, I do it all the time, and maybe I did again. But whenever possible on an analog scope, use as many screen divisions and peaks as possible to do your frequency calculations. Accuracy improves with more cycles and more divisions.
You lost me.
In that shot I was counting the middle reticule until I saw the same pattern with the next wave. I counted that at 2,4,half way between, 6,8,div. So, .5?
Quote from: Legalizeshemp420 on October 11, 2013, 06:23:25 PM
18 on each side of 20 or 22ga wire or possibly 24ga but no smaller wire.
I can't measure the inductance value of it myself though.
Sure you can. You have a scope, right? Set up a tank circuit with a known capacitance, poke it with a little jolt from a power supply and measure the ring frequency. Then calculate the inductance based on the capacitor value and the measured frequency. That's how inductance meters do it.
I know what I meant was I don't have a way to constantly poke it since I can't freeze a capture like I could in a DSO. Because I can't do that I would need something that repeats. I saw the circuits but do not have the parts here to replicate.
Quote from: Legalizeshemp420 on October 11, 2013, 06:32:25 PM
You lost me.
In that shot I was counting the middle reticule until I saw the same pattern with the next wave. I counted that at 2,4,half way between, 6,8,div. So, .5?
Look at the screen graticule markers. There are nine full divisions across the scopeshot you showed. That means there are 9 x 0.5 microseconds across the whole screen. Right?
How many peaks are there across the whole screen? Since the first one is aligned with the first leftmost graticule line, it is the "zeroeth" one and doesn't count. SO there are 17 peaks across the screen until you get to the rightmost graticle marker shown. You have 17 full cycles in 4.5 microseconds.
Your own value, 1 cycle in 0.25 microseconds, also yields a value of 4 MHz, not 400 kHz.
Quote from: TinselKoala on October 11, 2013, 06:35:57 PM
Look at the screen graticule markers. There are nine full divisions across the scopeshot you showed. That means there are 9 x 0.5 microseconds across the whole screen. Right?
How many peaks are there across the whole screen? Since the first one is aligned with the first leftmost graticule line, it is the "zeroeth" one and doesn't count. SO there are 17 peaks across the screen until you get to the rightmost graticle marker shown. You have 17 full cycles in 4.5 microseconds.
Your own value, 1 cycle in 0.25 microseconds, also yields a value of 4 MHz, not 400 kHz.
Ahhhh, we came to the same result only via a different path PLUS that damn misplaced decimal point, lol. btw, 4mhz makes more sense than seeing that many patterns on my 20mhz scope for a 400khz signal.
Now, UGH 4mhz? WaTaH?!? lol
Quote from: Legalizeshemp420 on October 11, 2013, 06:35:42 PM
I know what I meant was I don't have a way to constantly poke it since I can't freeze a capture like I could in a DSO.
Ah, you have no signal generator, I take it. Well, your scope does have a single-shot trigger mode, and you do have a camera....
So rest the camera on something steady, set the scope to single-shot, open the shutter, poke the tank with a wire to get a ringdown, then close the camera shutter. You'll have what the designers of the first digital scope had to use: a photograph of a single sweep on an analog scope.
Or you could build a simple 555 squarewave pulse generator in half an hour, and use that.
Or, you could use a 30-dollar Arduino and make your own selfcontained inductance meter that does its own poking and measuring. I see that Radio Shack is now stocking Arduino! Or you can order them for half that price from Singapore or China if you want to wait a couple of weeks.
Quote from: TinselKoala on October 11, 2013, 06:39:39 PM
Ah, you have no signal generator, I take it. Well, your scope does have a single-shot trigger mode, and you do have a camera....
So rest the camera on something steady, set the scope to single-shot, open the shutter, poke the tank with a wire to get a ringdown, then close the camera shutter. You'll have what the designers of the first digital scope had to use: a photograph of a single sweep on an analog scope.
Or you could build a simple 555 squarewave pulse generator in half an hour, and use that.
Well, I do have an ancient 555 around here I found in my antiquated parts pile (what was left after the stupid Oct 28, 2006 flood/fire). Most of my parts date from 1985-1992 and my cans of old IC's I miss the most but something had to go to make room years ago after the fire so I tossed them. About 50 dollars worth so not that great of a loss but a lot were ceramic.
Now if only I could manage to have been allowed to keep those Germanium trans I had but no go as I never got those back after the cleanup that took 3 months from hell.
Hmm... OK, I have 14+14 turns of # 27 with short leadin wires to my little toroid, and each winding measures right about 400 microHenry. But the one on the schematic says 38 microHenry. I wonder if there is another decimal point issue here.
He says he uses a really high perm toroid so less winding for a greater inductance but 38uH is awfully low.
Well, I dunno. I just got this one running, by taking off more and more wire from my toroid, it finally started oscillating. Now it only has 7+7 turns on the toroid, and it still measures 85 microHenry or so. But my frequency isn't anywhere near as high as yours. I'll take off a couple more turns in a few minutes, need to eat some supper first.
I built it onto a clothespin so I could grasp a battery easier. Here it is running brightly on a AG1/LR621 button cell. I'm using the exact cap values as in the schematic and a BC337-25 transistor, a 3/8" toroid from a CFL....
OK, I just looked at your video again, and based on my findings and your toroid, I think you have _way_ too much inductance, just like I had at first.
TK:
I am pretty sure you know this already but, those toroids from CFL's are powdered iron and not ferrite hence the low permeability. They still work as you have seen but, you have to play around with them. (I don't like them at all) Of course, they did do the job they were engineered to do in the first place but hey, can't these people design electronics with better components that we can salvage? I mean, come on, we need good stuff, ha ha.
Bill
Quote from: Pirate88179 on October 11, 2013, 10:52:18 PM
TK:
I am pretty sure you know this already but, those toroids from CFL's are powdered iron and not ferrite hence the low permeability. They still work as you have seen but, you have to play around with them. (I don't like them at all) Of course, they did do the job they were engineered to do in the first place but hey, can't these people design electronics with better components that we can salvage? I mean, come on, we need good stuff, ha ha.
Bill
Yes, my magnet sucks right to them and they are pretty heavy for their sizes. Since the shipping on toroids is more than the price of the toroid I use what I can salvage. I bought the one in the video as well as a surprise pack (both from Elec. Goldmine) and they are all iron which is a bummer. :(
btw, when I just used my toroid I wound with 28ga 2 turns on a toroid from a CFL my frequency went way up. It was up so much I could no longer zoom in to see what frequency it went to.
Heh.... Well, I left mine running on the tiny AG1 cell, from the time I took that picture above until a few minutes ago, and the light was out. I took off the toroid and took some wire off it. With 5 turns remaining on the toroid I get about 45 microHenry on the meter. If I go to 4 turns I get less than the spec of 38 uH, so I went with the 5 turns.
Now... I put the same AG1 cell back in... and it lights up, but doesn't run very long. Then it shuts off.... and in a few seconds, lights up again, dims, shuts off... and repeats. IOW it seems to be doing the same thing that the DALM thing is doing (yes it's still doing it, but much longer between osc bursts, still on same battery) but with light in the LED. I made a little video showing the cycling, it will be processed and up for viewing shortly. I haven't scoped this version yet to see if the spikes are narrower than they were before when I had too much inductance.
http://www.youtube.com/watch?v=mTOuUMsivWA
For my purposes this will not work since it will not even operate much above 20-30ma whereas the DALM will.
How 'Super Efficient' is this?
I dunno, I'm not the one making the claim. My build doesn't seem all that "super efficient" to me, but it is running on batteries that won't run (to light) some of my other JTs.
I like the "gasping for power" effect though.
MileHigh has shown how to make an easy analog power measurement computer out of a single TL082 dual opamp chip costing under a dollar, with a few resistors and a capacitor, that produces an output that can be read on an ordinary DMM to give power. I don't know if it will work on a JT signal but I'm going to give it a try, later on. I also have my light intensity meter using Arduino that I can use to measure the light output of its "standard" white LED. I'm not sure how accurate this is when looking at a pulsed light source, but I will be finding out.
I found a way to test for inductance but I need a sine wave source of 1hz to 2mhz that is super easy but I must be able to manually sweep it.
Btw, on this BK2120 it came set for 110Vac but has 110vac/120vac/220vac/240ac. I haven't ever, in 30 years of probing the AC outlets with a meter, seen 110vac unless something was wrong it was always 120-122vac. Should I leave it at 110vac or switch it to 120vac?
OK, I've at least measured the battery voltage during the gasping performance.
This one needs around 715 mV to turn on, and turns off at around 635 mV. This is higher by about a tenth of a volt than the performance of the DALM.
Quote from: Legalizeshemp420 on October 12, 2013, 11:20:13 AM
I found a way to test for inductance but I need a sine wave source of 1hz to 2mhz that is super easy but I must be able to manually sweep it.
Btw, on this BK2120 it came set for 110Vac but has 110vac/120vac/220vac/240ac. I haven't ever, in 30 years of probing the AC outlets with a meter, seen 110vac unless something was wrong it was always 120-122vac. Should I leave it at 110vac or switch it to 120vac?
I'd move it over to the 120 VAC setting. It probably won't make any difference. If you were in Japan you would be better off using the 110 position.
You really really do need a signal generator. I have one of these and there is nothing wrong with it for the purpose you need.
http://www.ebay.com/itm/1MHz-Function-Generator-Elenco-FG-500K-Kit-Waveform-/331042803024 (http://www.ebay.com/itm/1MHz-Function-Generator-Elenco-FG-500K-Kit-Waveform-/331042803024)
For your inductance measurement you will be sweeping a tank circuit and looking for resonance, so you just need to change the value of the capacitor in your tank in order to get the resonance frequency into the range of your instrumentation.
If you are setting up an electronics workbench you might like to budget for something a bit better. This looks really cool to me, I wish I could afford one right now.
http://www.ebay.com/itm/AE20125-10-MHz-Sweep-DDS-Function-Generator-Kit-with-USB-and-Modulation-/251169054479?pt=LH_DefaultDomain_0&hash=item3a7ad79f0f (http://www.ebay.com/itm/AE20125-10-MHz-Sweep-DDS-Function-Generator-Kit-with-USB-and-Modulation-/251169054479?pt=LH_DefaultDomain_0&hash=item3a7ad79f0f)
Or, for a little less, you can get an Arduino and learn how to use it. They are amazing critters, very useful for all kinds of things.
http://www.youtube.com/watch?v=S6N8ys8FiA4
DALM is a better circuit from my testing here as well. I am not sure what AcmeFixer was on about and I can't tell either since he has not appeared on this thread YET his name, and circuit, was mentioned in the title of this thread.
I guess we can safely say it is pretty much rubbish and if you notice he supposedly made it even better in the second version but it did get more complicated as well. I have my suspicions about that circuit as well.
Well, time to stop what I am doing until I can measure inductance. :(
We seemed to have posted at the same time.
I am loving the looks of that second one and 10mhz would be perfect for measuring down to the sub nH level but at 120 delivered I would just save for a RIGOL or something. More expensive but way better imo.
http://www.amazon.com/Waveform-Function-Generator-Frequency-Counter/dp/B00D55WR0S/ref=sr_1_60?ie=UTF8&qid=1381596371&sr=8-60&keywords=function+generator
That one is way more expensive but it includes a frequency counter (I have one that I hardly ever used and has all of the original materials and box). Model DFC-1000 that I used less times than the fingers on your hand (singular).
I can only purchase one spool of magnet wire so which is a good size to buy for winding, or rather what is mostly used in hand wound toroids?
Well, I'd like to be driving around in a Lexus... but in the meantime my 20 year old Chevy gets me where I need to go.
People wind coils with all kinds of wire. Years ago I got a good deal on a big spool of #27.... so all the coils I wind are made from that. I still have about 3 pounds on the reel. I've made a dozen or so JTs, a few Tesla coil secondaries, a Slayer Exciter, lots of electromagnets of various sorts... even some 14-strand Litz-like wire from doubling it back and forth across the back yard and twisting it up with a drill. Nine pounds of wire goes a long way.
If you can find an old CRT TV or monitor that hasn't been scavenged yet, there is a lot of good fine wire, usually in two gauges, on the ferrite deflection yoke around the neck of the CRT. Around here, the professional scavengers open up the cases and just break the CRT necks to get the yoke off, so you have to be quick and pounce on the discarded set before they get to it.
This DFC-1000? That's a nice instrument, you should unbox it and use it. I consider my Philips PM6676 universal counter to be the Master Instrument in my lab, from which I can calibrate all the others.
http://www.bonanza.com/items/like/123005221
Yes, that is the one only mine is in pristine condition. I originally bought it for 50-80 dollars when it was brand new (I forgot the exact price) back in the late 1980's I think it was. I used it to measure the frequency of some negative ion generators I was making (hand wound ferrite E cores from Coilcraft). Goes up to 1ghz I think it is (1000mhz).
My car I am doing my back brakes on today and in the morning and it is a 1998 Ford Taurus so I know what you mean.
I'm happy to report that the high-voltage transistors 13003 that were in the CFL that I took apart to get the 3/8" toroid... they work fine in the JT circuit!
I have not yet done any power comparisons, but the battery in the photo below is the same AG1 button cell that I used in the "burst oscillation" video. It shines brightly with the 13003 transistor, but has now gone out after five minutes or so while I'm typing and uploading the photo. I'm waiting to see if it will do the "burst oscillations" on this battery like the BC337-25 did.
There are several different versions of this transistor as I have found. The pinouts could be either BCE or ECB , left to right, looking at the numbered side of the transistor. These I have are BCE. I had to solder some wires to the legs to make them long enough for comfortable installation in the clothespin.
This is a highvoltage transistor! It might be a good alternative to the 2n3055 in HV JTs. I'll be trying that, later on.
I have two of those exact same transistors in one of my CFL bulbs. Was a 100 watt CFL as the 60 and 75 use smaller parts and I have a 150 watt that once it goes I wanna see what is in it. Been a few years since I bought it because I don't use it much but the 150 watt light equiv CFL cost me almost 12 dollars to buy it back then. It was not cheap but I have noticed that CFLs are getting rarer to find and their prices have sky rocketed in 2013 as I had to replace some and their prices tippled.