3A LED driver powered by 2 NiMh batteries

Discussion in 'The Projects Forum' started by toms.699, Jan 13, 2011.

  1. toms.699

    Thread Starter New Member

    Jan 13, 2011
    7
    0
    Hello

    After trying to find answers on my own I give up. Maybe someone here can help.

    I’m wondering what is a maximum current that boost driver can deliver when powered from 2 NiMh batteries. D size batteries should deliver required input current.

    Assuming that physical size and components price is not a big constraint what would be a limitation? Is it possible to get in 3A area? What would be required?

    Looks like inductor would need to be rated about 8-9A, 10-47uH seems reasonable since bigger inductors usually have lower current rating and also higher ESR. This means that frequency needs to be pretty high, seems like 300kH or more should get inductor ripple current in reasonable area.

    Very low Rds(on) N-MOSFET as a switch (3-5 mOhm, Vg(th) about 0.8-1.0V) seems like only option.

    How to drive it? Is there any driver available on market (I can not find anything)? I tried to push ZXSC300 but all I can get is about 1.1A. Can it be pushed further, another alternative maybe? Could 555 timer + comparator be used for that (2V minimum power supply may be a problem).

    Thank you.
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    What are the specs on the LED?

    Post your circuit as it is now. .png format is preferred. Use the "Go Advanced", "Manage Attachments" buttons to upload them to the forums.
     
  3. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    I have never seen a D-size Ni-MH cell. Energizer's datasheet for theirs shows the same spec's as a little AA cell so maybe there is a little AA cell inside the big nearly empty can.
     
  4. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,784
    945
  5. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
  6. Markd77

    Senior Member

    Sep 7, 2009
    2,803
    594
  7. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    A battery holder will melt with a current of 20A or more.
    There are many no-name brand battery cells made in ***** that are no good.
     
  8. toms.699

    Thread Starter New Member

    Jan 13, 2011
    7
    0
    For battery, I have AccuEvolution LSD cells. For 10Ah D size cells required current is below 1C. Would be also possible to use 18650 NiMh (or NiCd for that matter) cells that are commonly used in power tools. They can deliver huge currents. Input current should not be above 8-9A (may be higher on startup but only for few microseconds).

    Anyway, power issues are not my concern at this point.

    Question is if circuit with input voltage in 1.8-2.5V range can deliver output 3.6V with 3A current?

    About a LED I’m thinking about Cree XM-L but my question is general. It could be Luminous SST-50 under driven (it is rated for 5A). My assumption is that I need boost only since any white power LED will have Vf above 3V at currents higher then 1A (with lower output current Vf will drop and at some point driver will go direct drive effectively prohibiting going lower with output current).

    I’m assuming that LED Vf will be in 3.0-3.6V range for given current (if such driver is possible I will try to have 2 modes – high 3A, about 3.6V Vf and “low” 1A with Vf about 3.3V).

    I tried to build circuit using ZXSC300, one like on page 6 of spec (attached picture). Instead of transistor I have MOSFET and I tried it with extra resistor between gate and ground (470k) but it does not change anything.

    Below is list what I replaced:

    Transistor -> N-MOSFET NXP PH2520U,115 Rds(on) = 2.1mOhm, Vg(th) = 0.7V

    Diode -> tried two SBR10U40CT and 95SQ015

    Inductor -> tried three PM2120-470K-RC (47uH, 8.2A, 19mOhm), 2300LL-820-V-RC (82uH, 8.1A, 25mOhm), 1130-101K-RC (100uH, 9A, 34mOhm)

    For sense resistor I have few in 10, 15, and 20 mOhm, I tried few combination even with three of them in parallel for just few mOhm but I can’t get more than about 1.1A. I measured current with multimeter but also measured brightness of LED with lux meter to eliminate error that multimeter may introduce.
     
  9. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    The "threshold" voltage of a Mosfet is when it barely conducts and is almost turned off.
    The Mosfet you used is guaranteed to turn on very well when its gate voltage is +4.5V and to turn on fairly when its gate voltage is +2.5V.
     
  10. Markd77

    Senior Member

    Sep 7, 2009
    2,803
    594
    The zxsc300 is designed to use a <ed> bipolar </ed> transistor for switching, not a MOSFET. It just isn't going to work well with a MOSFET, they have different switching needs.
     
    Last edited: Jan 14, 2011
  11. toms.699

    Thread Starter New Member

    Jan 13, 2011
    7
    0
    Hmm, one of my ideas was that MOSFET may not be fully turned on. But current flowing thru it will be less than 10A and looking at Fig. 5 in spec even line with 1.3V at gate looks reasonable, 1.6V and 1.8V should be more than enough opening.

    Am I wrong? Do I need better MOSFET? Maybe MOSFET driver? Is there a chance that better MOSFET will give higher output current?

    Are inductors I’m using reasonable? From my test it looks like it gives highest output with smallest inductor, I guess that expected (I don’t really have knowledge about why but I think bigger inductance kind of “chokes” current) but small inductance will give bigger ripple current.
     
  12. toms.699

    Thread Starter New Member

    Jan 13, 2011
    7
    0
    Can you elaborate a little why is that?

    ZXSC300 outputs 0-0.2A low level when is off, and (Vin - 0.3)V when is on (will be more than 1.5V in this case). Frequency will be probably about 300kHz (1.7us fixed off time and about the same on time assuming ~0.5 duty cycle). Drive current does not matter (it is 3mA). I tried ground MOSFET gate with 470k resistor to discharge it when off but it seems make no difference.

    LED lights and current thru it is in 1A range, so ZXSC300 is working but I can't get it go further. May be a limit but why? I may need different drive or it may be something else.

    There are even some design papers with ZXSC300 and ZXSC310 using MOSFET.
     
  13. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    The curves shown on the datasheet of a Mosfet are for a "typical" one. But you can't buy a "typical" one, you get one that might have the minimum or maximum spec that is printed in text on the datasheet.

    The Mosfet you used might need a Vgs of 4.5V to completely turn on.
     
    toms.699 likes this.
  14. Markd77

    Senior Member

    Sep 7, 2009
    2,803
    594
    I would have expected the current limited drive to turn on the MOSFET too slowly because of it's gate capacitance. Clearly it does work though. Are there any boost converters in the ap notes that use MOSFETs? The ones you linked to in post 12 are both buck converters.
    How does the gate capacitance of your mosfet compare to the ones in the links?
     
    toms.699 likes this.
  15. toms.699

    Thread Starter New Member

    Jan 13, 2011
    7
    0
    One is using ZXMN6A07F, other ZXMN2A01F. Indeed they have much lower gate charge 3nC vs 78nC in my MOSFET (total gate charge). Qgd = 18nC in my MOSFET. With 3mA drive, looks like max switching frequency is 166kHz, this may be lower than what I need.

    Do you think this is a problem and if driven with higher current it should work?

    I didn’t see any boost converters using MOSFET. Looks like maybe because it is impossible to find MOSFET for high currents required with low enough gate charge.

    I don’t think I can use ZXSC300 with BJT, hFE would need to be in 3000 range (switching 9A with 3mA at gate), right? Darlington would require twice voltage and probably would not work either.

    Different drive then.

    Do you think if MOSFET drive would be sufficient it is possible to get 3A?
     
  16. toms.699

    Thread Starter New Member

    Jan 13, 2011
    7
    0
    I looked little bit more into things, especially gate charge and drive voltage.

    Do I get it correctly that theoretically typical boost driver should work and problem is getting real parts that will provide required parameters?

    Seems like the main issue is switch. Is MOSFET a best choice? Would using gate driver help? I can't find gate driver with such low supply voltage. Can I just use 2 MOSFETS (N & P) to drive gate of main MOSFET? This should help with gate capacitance but voltage would be still low.

    Any thoughts about having second boost driver to deliver 5V constant voltage to power gate driver. It would solve voltage and gate capacitance issue. Second inductor and other parts would be needed. Is this crazy idea or it should work?
     
  17. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    It's not silly to consider a boost circuit.

    You might be able to use something like an industry standard ICL7660S, a switched-capacitor booster, to double your input voltage efficiently and not require another inductor - just a few capacitors and perhaps a diode (to prevent latchup).
     
  18. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Not silly at all. There are ICs specifically for this, search for "High Side Driver". Typically used to increase Vgs for turn on in H-Bridge circuits, they use a switched capacitor (charge pump) to increase voltage to what is needed.

    With the prevalence of low voltage devices, I'm sure there are a few solutions for what you are trying to do, though I don't have a part number off the top of my head.
     
  19. toms.699

    Thread Starter New Member

    Jan 13, 2011
    7
    0
    I looked for drivers with bootstrap and there are many of them but seems like all need at least 4V supply. Maybe I need search a bit more.


    This ICL7660S looks very interesting. I wonder if it will give enough current.


    I'm not sure from where MOSFET gate driver gets current on output. Many drivers can deliver 2A or more. This is in short pulses and I'm assuming there will be always some capacitor to store energy. Am I right?
    How to compute requirements for driver supply current?
    Assuming 1Mhz switching frequency and 10nC gate charge, seems like 10mA is needed (constant current). Is that right?
     
Loading...