Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

"The Scavenger" A Joule Thief inspired Boost regulator. (unfinished)

Status
Not open for further replies.
Now that I see the schematic; I've the following questions/suggestions on the "finished product" circuit.

1) What core did you use?
2) Do you have a scope or multimeter which measures the frequency it is switching?
3) Wouldn't it be better that ZD1's cathode be connected AFTER R5, to compensate for the voltage drop under load (at 100mA, it will drop a full volt).
4) If you will be powering logic circuits with this supply, wouldn't it be better if the zener is instead 4.3 or 4.7 volts? The zener + Q2be voltages could exceed the max 5.5 volts of most microprocessors.

Other than that, a single word: BRILLIANT!
Congrats!
 
Hi schmitt. Let me answer some of those questions for you.

1) What core did you use?
The primary inductor core is not critical, use whatever you have around, it's just like the Joule Thief in that regard. More generally though, higher inductance's will mean lower switching frequency, so more ripple and more noise(sound). But lower frequency has more efficiency do to switching less often and more cleanly. Ideally you would probably want to aim for over 2X max human hearing range to get a good compromise. One should only go higher if you need smaller caps and/or better ripple figures.

2) Do you have a scope or multimeter which measures the frequency it is switching?
Sadly no, I have nothing that can measure frequency, especially not a scope. Honestly, the scope thing has been a thorn in my side for a long time. The cure is spendy though as I don't want to settle for a crappy tool. Unfortunately, I also don't have a lot of money, being a NEET. So spending any money on what is, for me, basically just "toys" is hard to justify. I fret over spending $20 on my hobby honestly. But if you know of a super good quality scope I can get for under (LOL) $50 US, let me know ;-) I could certainly churn out better quality stuff if I had one of those.

3) Wouldn't it be better that ZD1's cathode be connected AFTER R5, to compensate for the voltage drop under load (at 100mA, it will drop a full volt).
Perhaps, you do make an excellent point that I hadn't considered. R5 was intended to form an output filter with Cout to lower ripple current, making the circuit better for logic use. But in all truth it's really an optional part that could be removed. As for your idea, I believed that putting ZD1 on the outside of that resistor/filter would probably compromise cycle by cycle regulation. Don't forget that with little to no capacitance on the inside of R5, ZD1 will see the voltage directly through it as if it was not even there, but it will see coil voltage much much better. However, if it's sense from the outside of this loop, cycle by cycle regulation will be largely effected by the slow response of the resistor+capacitor. A better idea would be to use some other filter topology, but doing that efficiently involves adding MORE inductors. If you don't need the ripple removal, take out R5. Increasing Cout capacitance a whole lot also helps.

4) If you will be powering logic circuits with this supply, wouldn't it be better if the zener is instead 4.3 or 4.7 volts? The zener + Q2be voltages could exceed the max 5.5 volts of most microprocessors.
I know right? I expected as much myself, but in the sim and in my test circuits the only way I could get a specific voltage out was if the zener was that same exact voltage. Best I figure is because there is added voltage drop in Q1 and D1... or something along those lines. But maybe it was actually the R5 voltage drop like you were saying? Q2 really should be biased such that the instant current flows through ZD1 it engages regulation. But with too much biasing the tinny leakage current increase just before the sharp ZD1 break down causes Q2 to start to go liner, which starts to bias Q1 linear too. Not good. This problem as a whole is why C2 and R3 exists. Clean, square switching of Q1 is absolutely essential to efficiency.


Finally, thanks for the kudos. Hope I prove to deserve it :)
 
Thanks for your reply.
I'm indeed familiar with the joule thief being almost foolproof, and with almost any core being capable of creating a satisfactory circuit.
But the point I was attempting to make is that it is sometimes wise to experiment with a couple of different cores, as the efficiency may increase substantially......but this is only the icing on the cake. :)

Now, to your scope question...I see that you post a lot here, and many of your posts are clever and passionate.
When one is passionate about a "hobby", one must not fret in obtaining the means and tools to make the hobby better and more rewarding.
I can tell you from first-hand experience, that getting yourself a scope will open widely your understanding and enjoyment of electronic circuits.
I sincerely encourage you to save a litlle at the time, be on the lookout at Ebay and Craiglist, and you may find a perfectly good and affordable scope.
 
I can second that. My scope got broken on the move here and I was lost. I finally found one locally on ebay for $50. (Tek 465) Drove over and had a look and brought it home. The voltage switches need a good cleaning now but with spice I don't build as much just to see how it works. :rolleyes:
 
Ron;
you are totally correct, LTspice has allowed us to "build" circuits in a computer, and test ideas for feasibility.

But that is like a first down. It gets you into an advancing position, but one has not scored yet. The "score" is to have an actual working circuit.

To illustrate my point about the core: I once built a joule thief for an attic light. A friend saw that and asked me (actually paid) for me to build him a copy.
I used another core I had, everything else being identical, and was surprised at the improved brightness. Swapping some components between the circuits, I finally determined it was the inductor. In both instances the wire gage and number of turns was identical, but not the core.
Monitoring the switching waveforms I noticed that the high efficiency core produced sharper transitions. That meant better energy transfer.

Bottom line, although I would not have required a scope to determine which core was better, having one allowed me to understand why it is so.
Later, further reading allowed me to learn that there are different core materials with different permeabilities and core losses.

I see ()blivion as a talented individual, and would certainly like him to further improve his capabilities.
 
Sure, I think both have merit and are complements. I have built a few things that didn't last long enough to scope, but when run in spice the problem could be seen without smoke. ;)
 
Oh, hey. I didn't see all the info krb686 posted in post #35 until just now.

...Reads...

Good stuff, you really do your homework. That's excellent.

I'm not trying to bad mouth, only making observations, but for Dan M to claim that simulators should not be used because they are inaccurate really shows his age, stubbornness, and doesn't do simulators the justice they deserve. Simulators have been bad in the past, but have made great strides in recent years. No real pro circuit designer doesn't at least have some simulator in his arsenal these days. Even if he was right though, and they were all unreliable, one would be paused to remember that simulators are rough drafting tools designed to flesh out a project quickly. Prototyping boards are great and definitely not going anywhere, but they are not nearly as fast as a good simulator is for getting the budding ideas out of your head. Doing things "out loud" does wonders for train of thought. You can usually spot a great deal of preliminary problems right away when you rough draft them a simulator. This lets you quickly figure out what the key parts of your project are. Though do keep in mind, in the end, one must always build a real circuit before being able to call it a working design. Because, as useful as simulators are, they certainly are not perfect.

krb686 said:
Are the simulators just not accurately programmed or are they missing something?

The main thing about simulators is that they can only ever do what you tell them too, and that usually means that they don't take into account every aspect of a real life circuit. The falstad sim is a great example. When you add a part, it will not include stray values, and wires are always 0 Ohms, even if these details might be critical to the real circuits performance. This is most apparent with the falstad sim when you do something like add a capacitor in series with power and ground without any resistance. Because a capacitor looks short to changing voltages, and the sim says there is absolutely no other resistance, the current can get up to a ridiculous level, something like a billion giga amps. The the sim usually kills it's self when this happens with a convergence error or something similar. In any case, just like you can model stray component values, you can add them to a simulator manually as phantom components. I usually do this in falstad sim for things that have the biggest tendency to cause problems, mostly capacitors and ESR.

Anyway, moving on fro the simulator topic, Dan seems to also only want to consider the main transformer as acting strictly as a transformer. While it's true that it is technically a transformer, it's primarily acting as an inductor, just with an added feedback winding. The effect the secondary has on power and efficiency of the circuit is really only a secondary effect of what kind of feedback it is giving to the control circuitry. It's really just a sensor winding, and is not otherwise taking any part in the power conversion. But I think you already get that...

krb686 said:
...when Q1 rapidly cuts off, the inductor voltage spikes to whatever level, in that case 10.3V, necessary to push current through the next availabe path...

Yup, as far as I know this understanding is exactly right. If it was primarily working as a transformer, then a 1:1 ratio wouldn't produce more voltage than the input, but proof from real circuits clearly would show otherwise. What's really happening is the inductor is gaining energy from the current charging up the inductor "flywheel", then when the switch turns of, the inductors forward EMF is added to the input voltage, then pumped into the output capacitor. It's typical boost conversion at it's finest. With this in mind, I personally only play with the primary coil if I think the inductance is not right for some reason or another, mainly frequency related. I rarely would change it because of it's effect on the secondary.

That all being said, the ratio of the secondary to the primary *IS* important, as it controls how the feedback works. So in theory it can be taken advantage of. Though I may not necessarily 100% correct, I tend to think of it about like this.

Case 1: With higher base turns, you get stronger feedback while the device is oscillating, but because you tend to increase R1 as well, you get less drive during startup. Stronger feedback means cleaner switching, and thus better efficiency.

Case 2: With lower base turns, you get weaker feedback while the device is oscillating, but any transformer action allows for more current on the secondary. And you tend to have to lower R1 to get better feedback, which also leads to better low voltage starting.

Case 1 probably has the best applications for efficiency and power handling. Case 2 seems to be better for widening the operating range. With this in mind, I think Case 1 serves for better circuit improvement. In truth, there is not actually much energy left in cell below 1V when compared to fully charged, so aiming for that is less useful. The real point of the circuit is efficiency.

krb686 said:
I'm curious why reality departs from theory in that aspect.

I'm not sure myself. All I know is my testing and simulating did not strongly indicate that having an odd turn ratio was any better performing than just 1:1. But my gut tells me otherwise, as recently indicated above.


So krb686, if you do decide to play around with it more, Let me know if it does anything for you.
 
All of the links to your schematics are not working, and I would like to try a crack at this design. Could you repost the final version?
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top