Standlight circuits (Will this circuit work?)

[FrontRanger]

Re: Will this circuit work?

FrontRanger
It is years since I understood poles and zeros!
I think it is the impedance of the capacitor at potential oscillation frequencies that is important.

Thanks for the link to the data sheet. From what's written there, the 0.1 uF input cap is not for stabilizing the regulator loop but rather for bypassing noise on the input side of the regulator. (It's not even part of the feedback loop.) So if desired, you should be able to use a "smoothing cap" with this circuit.

 

[Alex Wetmore]

Re: Will this circuit work?

Cool! I've tried to work your mods into the original drawing. Please tell me if this drawing is accurate:

I'm assuming that this standlight charging circuit could just be dropped in Martin's circuit 8 in between C1 and the LEDs in his design? If the goldcap were a pair of 1.5F supercaps in series instead of a single, how do the internal resistances of the caps in series fit into the calculation for the discharge rate resistor? Do you just subtract the sum of both internal resistances? I guess it might also be necessary to change the values of the control resistors for the LM317L so that it sets a different max voltage for the supercaps in series?

Thanks again!

I built this circuit and it is working nicely. I just have it on a breadboard now, I haven't finished my headlight.

My only concern is getting better life out of the supercap. On my Lumotec IQ Fly headlight I get about 5 minutes of life at what appears to be half brightness. With this circuit and no discharge rate resistor I get about a minute at more like 1/4 brightness. 50ohm resistor increases the life, but makes it even dimmer.

I'm wondering if I bought the right supercaps. The better of the ones that I have is a NEC/Tokin FYH series. 5.5V rating and 1.0F.

What is the best mechanism to measure internal resistance of the capacitor? Is ESR on the data sheet (20 ohms) the internal resistance? That actually sounds quite reasonable, but I don't think it is accurate. I'm seeing it deliver 100ma for a very short period, and then it drops down to about 25ma quickly before levelling off a bit.

What are the best goldcaps/supercaps for this type of project? I see that Panasonic makes a HW series goldcap designed for high ma projects, but they are limited to 2.1 volts...not enough.

 

[FrontRanger]

Re: Will this circuit work?

My only concern is getting better life out of the supercap. On my Lumotec IQ Fly headlight I get about 5 minutes of life at what appears to be half brightness. With this circuit and no discharge rate resistor I get about a minute at more like 1/4 brightness. 50ohm resistor increases the life, but makes it even dimmer.

I understand that the IQ Fly also uses an XR-E. Perhaps not the R2 bin, but other than that, a fair comparison to your light. Why the difference? It may simply store more charge on its cap, as you mention. Alternatively, the Fly may use something fancier than a resistor (more efficient) for the standlight. With that said, the simplest fix may well be to increase the resistance and tolerate the reduced brightness (doesn't take much to be seen) or increase the capacitance, if you have room for it.

 

[syc]

Re: Will this circuit work?

I understand that the IQ Fly also uses an XR-E. Perhaps not the R2 bin, but other than that, a fair comparison to your light. Why the difference? It may simply store more charge on its cap, as you mention. Alternatively, the Fly may use something fancier than a resistor (more efficient) for the standlight. With that said, the simplest fix may well be to increase the resistance and tolerate the reduced brightness (doesn't take much to be seen) or increase the capacitance, if you have room for it.

FWIW, I picked up a pair of 1.5F 5.5V Panasonic Gold Caps spec'd at 30 ohms (P/N P11343-ND at Digi-Key) for the simpler circuit. I was only planning on using one at a time for my projects, but you were saying that putting them in parallel was an option on your smart standlight circuit, right?

 

[syc]

Re: Will this circuit work?

My only concern is getting better life out of the supercap. On my Lumotec IQ Fly headlight I get about 5 minutes of life at what appears to be half brightness. With this circuit and no discharge rate resistor I get about a minute at more like 1/4 brightness.

ps.
What's the difference in performance when you wire it up using the simple supercap charging circuit? How much of a difference in light intensity do you notice during the charging phase, and any difference in standlight duration?

 

[FrontRanger]

Re: Will this circuit work?

... you were saying that putting them in parallel was an option on your smart standlight circuit, right?

syc, you must be thinking of another poster, as I have not smart standlight circuit. I was just replying to Alex that increasing the capacitance would help his standlight runtime. One way to do that is put more of the same capacitor element in parallel.

 

[Alex Wetmore]

Re: Will this circuit work?

ps.
What's the difference in performance when you wire it up using the simple supercap charging circuit? How much of a difference in light intensity do you notice during the charging phase, and any difference in standlight duration?

With the simplest circuit (just a supercap in parallel with one of the LEDs, no resistor) it takes a couple of seconds to charge the capacitor and the LED is ramping up in brightness as that occurs. The capacitor charges very quickly. With the "smart" circuit the primary LED doesn't dim noticably as the capacitor charges (so it is instantly bright), but it takes more like 20-30 seconds to charge the capacitor fully. Discharge seems similar in both cases...which makes sense since the LM317 is out of the way during discharge.

alex

 

[Alex Wetmore]

Re: Will this circuit work?

I built this circuit and it is working nicely. I just have it on a breadboard now, I haven't finished my headlight.

My only concern is getting better life out of the supercap. On my Lumotec IQ Fly headlight I get about 5 minutes of life at what appears to be half brightness. With this circuit and no discharge rate resistor I get about a minute at more like 1/4 brightness. 50ohm resistor increases the life, but makes it even dimmer.

I think I've figured out one of the major problems with this circuit.

The LM317 is setup to deliver around 6 volts (actually too much for the supercap). However the LED is in parallel with the supercap and seems to prevent more than 3.5 volts or so from being dropped across it. If I remove the LED from being in series to charge the supercap, then unplug my power supply and plug in the standlight LED I get both a brighter LED and a longer lifetime.

Since the supercap never seems more than 3.5 volts it isn't charging fully. It is wasting 66% of the useful charge because the LED stops having useful light below about 2.5V. I'm getting 1V of useful potential out of the supercap instead of the ideal 3V of useful potential (3 + 2.5 = 5.5, the capacitor's maximum rating).

I tested this by disconnecting the second LED while charging the capacitor, then unplugging my power source and plugging in the second LED. It had a longer life and was brighter than in the normal circuit.

Any ideas on how to address this? Would it be easier to have a 3rd LED that only came on as a standlight?

alex

 

[syc]

Re: Will this circuit work?

Sorry, my mistake (shouldn't rush off posts just before bed). It was Bandgap's design.

syc, you must be thinking of another poster, as I have not smart standlight circuit. I was just replying to Alex that increasing the capacitance would help his standlight runtime. One way to do that is put more of the same capacitor element in parallel.

 

[Steve K]

Re: Will this circuit work?

hi Alex,

You've stumbled across the fun and annoying aspect of standlight circuits..... the simple ones don't work that well.

The one suggested by Bandgap relies on some isolation between the led and the supercap in order to fully charge the supercap. A 500 ohm resistor is a partial way to achieve this, but involves a lot of I x R losses.

A current source could also work, but involves more parts and will have some headroom requirements (i.e. will require some voltge to just run the circuit).

Something that I'm planning to use in an upcoming standlight circuit is to use a small boost converter to drive the led. The Zetex ZXSC310 can run from voltages from 0.8v to 8v, so it could use quite a bit of the charge in the supercap.

Plenty of options, so you have an excuse to buy parts and play around.

Steve K.
(known on the i-bob list simply as Steve Kurt)

 

[Alex Wetmore]

Re: Will this circuit work?

Something that I'm planning to use in an upcoming standlight circuit is to use a small boost converter to drive the led. The Zetex ZXSC310 can run from voltages from 0.8v to 8v, so it could use quite a bit of the charge in the supercap.

That looks like it'll do the job, but surface mount gives me a headache and inductors give me an even larger headache (unless I don't need to think about how they are working). Any ideas on similar devices that are designed for through soldering?

I guess I also don't understand how you charge the capacitor to ~5 volts without a large resistor. The ZXSC310 will help you make use of the 2.5 down to .8 volt range of the capacitor, but it won't let you charge it over the 3.5 volt range clamped by the LED. Is that correct? If you put a large resistor between the capacitor and ZXSC310 then it will still restrict the current.

Figuring this out and dealing with optics really makes me appreciate the IQ Fly that much more. It is expensive but a well dialed in solution. It's fun to build these things, but not as simple as it first appears.

alex

 

[FrontRanger]

Re: Will this circuit work?

I think I've figured out one of the major problems with this circuit.

The LM317 is setup to deliver around 6 volts (actually too much for the supercap). However the LED is in parallel with the supercap and seems to prevent more than 3.5 volts or so from being dropped across it. If I remove the LED from being in series to charge the supercap, then unplug my power supply and plug in the standlight LED I get both a brighter LED and a longer lifetime.

Since the supercap never seems more than 3.5 volts it isn't charging fully. It is wasting 66% of the useful charge because the LED stops having useful light below about 2.5V. I'm getting 1V of useful potential out of the supercap instead of the ideal 3V of useful potential (3 + 2.5 = 5.5, the capacitor's maximum rating).

...

Any ideas on how to address this? Would it be easier to have a 3rd LED that only came on as a standlight?

Hi Alex, your analysis looks correct to me. As you point out, during discharge, your capacitor is limited from discharging much below 2.5 V, because the LED is nearly off at that point. And during charging, the same LED clamps it to around 3.5 V or so. (It will actually be a bit higher than the LED voltage because of the resistor between the cap and the LED.)

As Steve K. says below, the simple standlights don't work that well. Let's say that "simple" means only passive elements and diodes here. With that restriction, I don't think you can do much about the first problem (the 2.5 V min voltage). The boost regulator that Steve K. mentions sounds promising for extracting as much charge as possible out of that cap, but that's a bit more complicated. If you want to keep it simple, you might be able to attack the second problem (3.5 V max voltage). To maximize the amount of charge stored on the cap, I scribbled out this little circuit:




If each LED is 3.6 V @ 500 mA, you'll get 7.2 V at the top. An ideal clamp would charge your cap to 5.5 V (assuming no margin). If standard Si diodes drop 0.7 V, and Schottky diodes drop 0.4 V, then two of the former plus one of the latter will clamp the cap to 7.2 - 1.8 = 5.4 V, just under the 5.5 V max. The resistor must be sized so that only a small current goes through it during normal steady-state operation. Let's say we choose 10 mA for that current. Then the resistor should be (5.4-3.6 V)/10mA = 180 ohms. The values of the diode drops are approximate, so you'd have to play with the resistor value a bit, or measure the diodes accurately and calculate exactly.

I don't have the parts laying around to try this myself. This won't be as good as the boost regulator, but will nearly triple the usable stored charge if it works as planned (3.5-2.5) becomes (5.4-2.5). Hope it helps.

 

[Alex Wetmore]

Re: Will this circuit work?

I didn't get all that the discharge rate resistor was doing for me. It isn't only slowing down the capacitor discharge, but it also allows the capacitor to charge to a higher voltage.

With a 150 ohm discharge rate resistor I have a 10ma current flowing through it when power is applied to the circuit. The voltage drop across the resistor is 1.4 volts. That allows my capacitor to charge an extra 1.4 volts above the 3.6 volts, giving me 5 volts at the capacitor.

During discharge I start with around 110ma going through the circuit at first, and it drops down to around 80ma where it stays for a long time. That works out pretty well.

I had made the mistake of trying to maximize discharge rate, without realizing that it was also hurting my maximum possible charge with this circuit.

 

[Steve K]

Re: Will this circuit work?

I didn't get all that the discharge rate resistor was doing for me. It isn't only slowing down the capacitor discharge, but it also allows the capacitor to charge to a higher voltage.

With a 150 ohm discharge rate resistor I have a 10ma current flowing through it when power is applied to the circuit. The voltage drop across the resistor is 1.4 volts. That allows my capacitor to charge an extra 1.4 volts above the 3.6 volts, giving me 5 volts at the capacitor.

During discharge I start with around 110ma going through the circuit at first, and it drops down to around 80ma where it stays for a long time. That works out pretty well.

I had made the mistake of trying to maximize discharge rate, without realizing that it was also hurting my maximum possible charge with this circuit.

hi Alex,

The resistor does serve two functions, and optimizing one tends to degrade the other.

I've got an idea that may help, but of course, it adds parts and eats up a little voltage. Not much tho.....
The idea is to add a transistor switch between the LED and the cap/resistor. The transistor only turns on when the dynamo is stopped. The means that the original resistor only has to control the discharge current.

I've left it to the student to calculate the resistor values.

The only thing that requires attention is the base resistor value. Set the base current to be about 0.03 of the collector current (roughly). Make sure that you do this calculation for the lowest desired operating voltage of the cap.

I certainly like most surface mount parts, and it can work even with crude tools. I just hack up a copper clad board with a dremel tool to make most circuits. This is an example of the Zetex boost circuit as used in a rear standlight.

That ought to provide some options.

Steve K.

 

[syc]

Re: Will this circuit work?

The idea is to add a transistor switch between the LED and the cap/resistor. The transistor only turns on when the dynamo is stopped. The means that the original resistor only has to control the discharge current.

Cool - I was wondering how you'd use a transistor switch to isolate the supercap during the charging - I was going to ask a friend with an ECE background about it, but looks like CPF came through first.

By the way, my ECE friend also pointed out the wasted charge in the supercap, and said that he could sketch out a simple current pump circuit for me that could squeeze out more of the charge. Then he wandered down a path with adding a pic microcontroller...
I'll try to get the current pump out of him before feature bloat takes over, and post some updated diagrams.

 

[syc]

Re: Will this circuit work?

As Steve K. says below, the simple standlights don't work that well. Let's say that "simple" means only passive elements and diodes here. With that restriction, I don't think you can do much about the first problem (the 2.5 V min voltage). The boost regulator that Steve K. mentions sounds promising for extracting as much charge as possible out of that cap, but that's a bit more complicated. If you want to keep it simple, you might be able to attack the second problem (3.5 V max voltage). To maximize the amount of charge stored on the cap, I scribbled out this little circuit...

I was wondering if you could simplify things a little more with this circuit?

(sorry if its nonsensical - I've been reading up a little on electronics, but have been slow about it)

I wasn't crazy about the supercap soaking up current from both LED's while charging, so went back to the supercap parallel to only a single LED. What I'm trying to get is the benefit of the 150 ohm resistor that Alex reported, but on a very simple circuit. Would this circuit allow the supercap to charge up to around 5V, as well as throttle the discharge rate?

 

[Steve K]

Re: Will this circuit work?

I was wondering if you could simplify things a little more with this circuit?
<image removed... see above>

(sorry if its nonsensical - I've been reading up a little on electronics, but have been slow about it)

I wasn't crazy about the supercap soaking up current from both LED's while charging, so went back to the supercap parallel to only a single LED. What I'm trying to get is the benefit of the 150 ohm resistor that Alex reported, but on a very simple circuit. Would this circuit allow the supercap to charge up to around 5V, as well as throttle the discharge rate?

the circuit does simplify things, but the performance is greatly reduced. The supercap will only charge up to the LED's Vf, which means you won't get much use of the stored energy.
[to review: energy = 0.5 x C x V^2]

The use of the diode and resistor does allow the supercap to charge up faster while producing a dimmer standlight with a longer run time.

The circuit that uses the voltage regulator does divert current from the LEDs, but only while charging. Since it charges to a higher voltage, it stores more energy, but does reduce light output longer than when the supercap is only charged to Vf. This is a normal trade-off, and up to the user to decide if it is appropriate.

Everyone has to decide what factors are important for their use. The factors include size, cost, complexity, efficiency, ease of assembly, reliability, run time, brightness, and how it will be used. (I may have missed some factors...)

To me, the voltage regulator and boost regulator are pretty simple, and the supercap is expensive and big and limited in ability to store energy. As such, I tend to use a AA nicad with a boost regulator and am pretty happy. There are times when the supercap will be better than the nicad (i.e. if you expect that there won't be much time to charge, and the nicad or cap will completely discharge). You'll need to pick the best solution for your use.

regards,
Steve K.

 

[FrontRanger]

Re: Will this circuit work?

I was wondering if you could simplify things a little more with this circuit?

(sorry if its nonsensical - I've been reading up a little on electronics, but have been slow about it)

I wasn't crazy about the supercap soaking up current from both LED's while charging, so went back to the supercap parallel to only a single LED. What I'm trying to get is the benefit of the 150 ohm resistor that Alex reported, but on a very simple circuit. Would this circuit allow the supercap to charge up to around 5V, as well as throttle the discharge rate?

the circuit does simplify things, but the performance is greatly reduced.

Yes, and...

The supercap will only charge up to the LED's Vf

... it's not even that good. The cap will only charge up to one diode-drop below the LED's steady-state Vf. As Steve K showed that's a quadratic degradation in stored energy.

The circuit I posted in #52 dramatically improves this while not electrically overstressing your cap. Increasing the stored energy while restricting design to only two-terminal devices was the idea there. The circuit in post #56 works on the same clamping principle as the one in #52, except that the clamp is referenced to only one Vf above ground instead of two, and the clamp voltage is smaller.

 

[Calina]

Re: Will this circuit work?

This is likely the most interesting thread that I've seen on CPFabout stand lights and super caps.

I'm not sure it deserve a sticky status since stand lights are unfortunately not a very popular subject but it should at least be included in a "thread of interest" on the first page.

Thanks to all for this very informative thread.

 

[Bandgap]

Re: Will this circuit work?

I think I've figured out one of the major problems with this circuit.

The LM317 is setup to deliver around 6 volts (actually too much for the supercap). However the LED is in parallel with the supercap and seems to prevent more than 3.5 volts or so from being dropped across it. If I remove the LED from being in series to charge the supercap, then unplug my power supply and plug in the standlight LED I get both a brighter LED and a longer lifetime.
alex

Ooops.
I have been off-line for a few days.

Circuit charging at 6V - which is indeed too much for the supercap

My calculations show 5.3V for 240 and 820 Ohms.
Did you use exactly those?
If so, the 820 need to be a little lower.

Also - if the capacitor voltage never gets to 5V under charge, too much current is being drawn by the led, the current limiting resistor needs to be higher.
This is a simple circuit - only really good for 20 or 40mA discharge.

EDIT
Just noticed the the wise Steve K has already said this: The one suggested by Bandgap relies on some isolation between the led and the supercap in order to fully charge the supercap. A 500 ohm resistor is a partial way to achieve this, but involves a lot of I x R losses.


With a 30 Ohm internal impedance, the supercap output voltage will immediately drop as soon as you draw current because of this resistance.

Hope I am not missing the point.

By the way: If you want to have a high-current (greater than 50mA ) standby, supercaps are the wrong technology in my view.

Steve