circuit questions???

[Robocop]

While doing a few of my own run time tests I noticed one of my batteries had ran down to 0.37v and the light was still going (moon mode) I tried the dead battery in every other AAA light I have and it would not power up. I placed it again in the tested light and it powered up and ran for about another hour.

What component or feature is different with this AAA circuit when compared to my others? Is it simply a combination of just the right LED and just the right components that make this happen? I do not keep up with circuit stuff that much and as such know nothing at all about their workings. I am wondering if this ability to power from a dead cell has any effect on other things.....you know as in if this is possible with this light why do others not have similiar circuits?....what is the trade off in other words?

 

[MrAl]

Hi there Robo,

Well, it's mostly the technology used, as in what boost chip
it is using. Some boost chips are made to work down to
very low voltages (and at least put out some power) and
others are not.
Part of what helps determine this is what kind of transistors
are used in the chip, and if it is using any form of bootstrapping.
Bootstrapping is a technique where you run the boosted output
voltage back into the circuit (partially) so that the output stage
can continue to operate down to low voltages.
Obviously boostrapping is desirable if you are concerned about
running on very low, run down batteries.

 

[Steve K]

to expand a little on the issue... is there a list of good boost converters that can run on insanely low voltages?? I need to rig up another standlight circuit for a new bike, and this capability allows for better use of the stored energy.

I've been using the trusty Zetex boost converters... ZXSC310, for instance. Works great running from a single nicad cell. I wouldn't mind finding a converter that can run with low input voltages like the 310 does, but could convert a little more power.

thanks!

Steve K.


p.s. "standlight" is a term used for a light driven by a bicycle dynamo that stays powered when the bike is stopped. Typically, it's not very bright, and doesn't run more than a few minutes. Most use a supercap for energy storage and minimal circuitry to drive a single 5mm led.

 

[Mr Happy]

I think it is very unusual to find a boost converter that will work with input voltages below about 0.6 V. This is related to a natural limitation of silicon based circuits. I suspect that creating a circuit able to overcome these limitations and work with really low voltages may lead to compromises in other areas.

Another point worth noting is that while an alkaline cell still has some power left when down to 0.37 V, a NiMH cell is for all intents and purposes completely discharged at that voltage. It has probably given up over 99% of its stored energy by then.

 

[MrAl]

Hi again,


Some good points Steve and Mr Happy.

There are ways to overcome voltage limitations, namely, as
mentioned, bootstrapping, however the 'starting' voltage
still has to be enough to start the circuit (0.7 or more).
Lucky, many cells voltage climbs up a little once the light
is turned off, allowing it to start up again.

I use mostly rechargeables so the MrHappy 99 percent rule
applies to almost everything i use too.

Steve, i worked on one of the original Zetex 300 circuits and
if i remember right the main limitation is the output drive
current, which limits what kind of transistor you can use for
the main driver. It may be possible to boost this with a
transistor and apply that to another transistor which would
allow a higher output current. Of course speed is an issue too,
so maybe a MOSFET for the second transistor.

 

[Robocop]

Thanks for the information and this stuff is mostly way over my head anyway. I am curious now that I know these circuits can be made to work from very low voltages why do all of todays lights not use them? It must be some negative involved with the circuit.....does it cost more to make a circuit that runs from such a low power?

Also if you do have a circuit that powers from such a low voltage does this only apply to 5mm LEDs?

 

[Curious_character]

The energy in an alkaline cell is nearly all gone when the cell voltage reaches about 0.8 - 0.9 volt. So you're wasting very little by throwing it away when it reaches this value. Circuit designers know, or should know, this, so reasonably don't put forth much effort or expense to make circuits run on very low cell voltages.

I'm not certain it's true, but believe that cell leakage is more likely to occur when the cell is completely discharged. If so, this would be another good incentive to limit the minimum functioning voltage.

Finally, NiMH cells, which are very nearly out of energy at 1.0 volt, don't take kindly to extremely deep discharges below this. So devices which users might put NiMH cells into would be better off quitting at a reasonable voltage.

c_c

 

[Steve K]

Hi again,

Steve, i worked on one of the original Zetex 300 circuits and
if i remember right the main limitation is the output drive
current, which limits what kind of transistor you can use for
the main driver. It may be possible to boost this with a
transistor and apply that to another transistor which would
allow a higher output current. Of course speed is an issue too,
so maybe a MOSFET for the second transistor.

hi Al,

Adding a transistor to the normal FMMT614 output seems possible, but would also give up the ability to run over the normal voltage range of a nicad cell (right?). Seems like you'd need 1.3v to get two BJT's turned on, and perhaps more if the FMMT614 is driving a logic level N mosfet.

.... or is there an app note that I missed? ....

I'm presently using the Zetex boost converter to drive 4 yellow 5mm LEDs in series. I'd love to use it to drive one of the 4 dies in the upcoming Cree MR-E. That would simplify the packing to be able to eliminate the 5mm LEDs from the design. I'm sure the Zetex converter could push current through the MR-E die, but I'd like to get a higher current level.

I'd also like to use a supercap instead of the nicad cell, hence the desire to find a converter that uses all of the energy in the supercap. Unfortunately, the capacity of a nicad is much larger than that of a similarly sized supercap, so I'll be sticking with the nicad. I've had very good luck with the nicad in my current designs, so can't complain. Still, it would be nice to not have to ponder the condition of the nicad.

hey Al, any comments on the intended application for the zxsc310? Strictly for single-cell LED flashlights, or was there something else in mind? A power supply for little mp3 players, maybe?

regards,
Steve K.

 

[MrAl]

hi Al,

Adding a transistor to the normal FMMT614 output seems possible, but would also give up the ability to run over the normal voltage range of a nicad cell (right?). Seems like you'd need 1.3v to get two BJT's turned on, and perhaps more if the FMMT614 is driving a logic level N mosfet.

.... or is there an app note that I missed? ....

I'm presently using the Zetex boost converter to drive 4 yellow 5mm LEDs in series. I'd love to use it to drive one of the 4 dies in the upcoming Cree MR-E. That would simplify the packing to be able to eliminate the 5mm LEDs from the design. I'm sure the Zetex converter could push current through the MR-E die, but I'd like to get a higher current level.

I'd also like to use a supercap instead of the nicad cell, hence the desire to find a converter that uses all of the energy in the supercap. Unfortunately, the capacity of a nicad is much larger than that of a similarly sized supercap, so I'll be sticking with the nicad. I've had very good luck with the nicad in my current designs, so can't complain. Still, it would be nice to not have to ponder the condition of the nicad.

hey Al, any comments on the intended application for the zxsc310? Strictly for single-cell LED flashlights, or was there something else in mind? A power supply for little mp3 players, maybe?

regards,
Steve K.

Hi Steve,

I dont think you missed any app notes but to turn on a bipolar you need about 0.7v
and that's about it because some modern bipolars can saturate down to 30mv or so,
so all you need is the right configuration of a PNP and an NPN and that might do it.
I say 'might' because theoretically it should work, but i've never tried it at so low
a voltage before. Speed is also an issue, so the only thing i can say is if you wanted
to try this as an experiment maybe i can come up with a circuit. Of course this is also
dependent on finding a transistor that can saturate to a low voltage at a high enough
current, which means i would probably look at the Zetex line again.
One thing you might want to keep in mind is that the only thing that prevents a driver
from working at low voltage is the drive voltage available, which doesnt have to be
able to deliver the full output current but only the driver current. This means if
circuit complexity is not an issue (sometimes it is though) a tiny low current boost
circuit could be used just to develop the drive voltage, then go from there. Of course
that's the worst case situation.
From what i see about the 310 chip is it does already work down to 0.8 volts, so
that's not bad i guess. With some current boost on the output maybe it would be
possible to get a lot more drive current. i cant say how the efficiency would be though.

Note that the configuration might have to change a bit to accommodate two transistors
instead of one. Might be a little bit of a trick too so the output doesnt get inverted.
Maybe three transistors? Maybe a high speed inverter?
Are you into experimenting with these kinds of ideas for higher output current?

The 310 was originally designed to work in LED backlighting applications, so it was
bound for LED driving from the start. I dont think this prevents one from using it
in another app that needs voltage boosting however.

 

[Steve K]

hi Al,

thanks for taking the time to post a reply. My first assumption is that two transistors will need to be added to avoid inverting the signal, but this will indeed slow things down. Getting BJT's to shut off quickly is a bit of an art, especially when significant current is involved. Might be worth tinkering with, though. Any recommendations for the final power BJT?

For now, this is on the back burner. It's gotta wait until the fancy new Cree is actually available, and I can find time to work on it! In the meantime, I'm busy putting together my dynamo system on the new recumbent bike. Need to get it done before the days get too short!

I do appreciate learning that the Zetex part is still the best for this sort of application, though. Now I can order a few more parts and not feel like I'm wasting money.

regards,
Steve

 

[Robocop]

Holy smoke this quickly turned into a thread way over my head.....I wish I understood all the technical terms however I do better understand the concept now. My main reason for asking about this was I could not understand as to why this low voltage ability was not present on all circuits.

If it does not cost much more and is a simple addition then what would it hurt for others to include this ability....even if most people do not run such low cells it still seems as if it would be another plus for already good circuits.

 

[TorchBoy]

Re: circuit questions?

... what would it hurt for others to include this ability.

I think someone mentioned that you don't really want to run NiMH cells too flat, so having the driver stop at a slightly higher figure is a good thing in that respect.

 

[chimo]

I have found that the dirver in the Fenix E01 will start on a voltage as low as 0.3V and, once running, will operate as low as 0.1V (of course - out of regulation). It's a voltage boost circuit. Info here.

Paul

 

[Robocop]

Thanks TorchBoy as I missed that information before.....kind of makes sense to me now as many these days are using rechargeables.