• State TI Thinks Resolved
  • Locked Locked
  • Replies 17 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 713 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS61291: RDS(on) of Bypass FET

Angelo Pietroforte
Prodigy 200 points
Part Number: TPS61291
Other Parts Discussed in Thread: LM3370, TPS610985

RDS(on) of Bypass Switch = 1.2 ohm.  Rather high for my liking.  Does anyone know of a part with lower RDS(on) besides the TPS6128x family?

  • 0
    TI__Mastermind 46905 points
    Hi Angelo,
    the current consumption in this device is very small in bypass mode, and normally the output current should be small in this mode.
    may I know the application condition? I would suggest a suitable solution
  • 0 in reply to Jasper Li
    Prodigy 200 points

    Nice to hear from you Jasper.

    2 AA's (Alkaline) in series is the source.  Ie an input voltage range of 2.2V-3.3V. 

    Granted yes most of the time I'm drawing 1mA or less through the bypass FET in that mode, but sometimes I can be as high as 90mA from the high of 3.3V down to 2.95V of my battery source.

    The TPS61291 powers an LM3370, and the LM3370 powers the rails of a camera.  The camera needs 2.8V and 1.4V from the LM3370.  I assume the LM3370 needs 2.85V so I don't get into dropout issues with it.

    So I need to have no lower than 2.85V coming out of the TPS61291, and with the 90mA going through that P-FET. and the P-FET's 1.2ohm, I'm dropping ~0.1V.  So 2.95V is when I have to turn to Boost mode.  I was hoping I could more efficiently let the battery's drop all the way down to 2.85V before I had to turn to Boost mode.

    Granted a lower RDS(on) would require a bigger P-FET and maybe bigger driver too, all implying a bigger IC, but I have to ask because that is what I would really want without having to go to the BGA versions in the chips that really want 2.3-2.5Vin minimum (ie TPS6128x family).  My board is nearly done being laid out, and now I'm finding the 0.1V dropout has me having to make some more final tweeks.  I guess better to know the issue now, than after the board comes out and I start scratching my head at that late hour trying to understand why I have 0.1V dropout of the PFET. 

    I'm in Milpitas CA.  Best Regards! 

    Anyway,

  • 0 in reply to Angelo Pietroforte
    TI__Mastermind 46905 points
    Hi Angelo,
    The control logic of TPS61291 is bypass mode at small loading, 1ma for example . then boost mode in high output current condition. so the p-fet should be disabled at 90mA current condition.

    because of the ESR of alkaline battery, it is hard to detect when battery voltage is lower than 2.9V. so just enable boost mode at high load condition.

    will you disable the LM3370 in standby condition? as the LM3370 consumes 30~40ua even at no load condition.
  • 0 in reply to Jasper Li
    Prodigy 200 points
    Yes, LM3370 is in Iq (SD) =3uAmax (see datasheet) 99% of the 24hr day. TPS61291 powers a micropower microController (uC) and a micropower motion detector too. So, 99% of the time the system is in a 3.0V/40uAmax kind of power state. Ie, much less than 1mA. I stated 1mA originally to allow you a quick point of reference.

    The uC has an algorithm that automatically wakes up every half hour and through use of its' internal comparator, with an external voltage divider tied to Vbatt, it compares the battery state to the comparators 1.2V reference. And, if Vbatt is at the 2.9V threshold, that I mention previously, the comparator output in uC flips, and the algorithm knows to toggle an I/O line high, ie the line tied to EN/BYP of TPS61291. Ie enter boost mode.

    Your suggestion to enable boost mode at high load sounds interesting, and I noticed the output stays stable at 3.2V/90mA (w/ Vin=3.3V) when EN/BYP goes high. And, interestingly, I see the dropout still exists in this state. I thought it would go to 3.0V in boost mode. Anyway, I see that Vo stays at 3.0V when I drop Vbatt to below 3.0V which is good. This is all good to know, but I also know that the uC comparator algorithm is NOT written robust enough to re-toggle the I/O line low again when it goes back to micropower state, ie 3.0V/40uA.

    So the question becomes is the power subsystem more efficient:
    1) being in boost mode when Vin > Vo at Vo=3.0V/90mA, or is
    2) the PFET more efficient at Vin > Vo at Vo=3.0V/90mA. PFET is just an LDO in bypass mode.
  • 0 in reply to Angelo Pietroforte
    Prodigy 200 points
    With a 40 ohm load to TPS61291 and Vin=3.30V, Iin=79mA, Vo=3.21V EN/BYP=low. EN/BYP=high, Vin=3.30V, Iin=79mA, Vo=3.21V
    With a 40 ohm load to TPS61291 and Vin=3.20V, Iin=76mA, Vo=3.11V EN/BYP=low. EN/BYP=high, Vin=3.20V, Iin=76mA, Vo=3.11V.
    With a 40 ohm load to TPS61291 and Vin=3.10V, Iin=73mA, Vo=3.02V EN/BYP=low. EN/BYP=high, Vin=3.10V, Iin=76mA, Vo=3.05V.
    With a 40 ohm load to TPS61291 and Vin=3.05V, Iin=73mA, Vo=2.97V EN/BYP=low. EN/BYP=high, Vin=3.05V, Iin=76mA, Vo=3.04V.
    With a 40 ohm load to TPS61291 and Vin=3.00V, Iin=71mA, Vo=2.92V EN/BYP=low. EN/BYP=high, Vin=3.00V, Iin=78mA, Vo=3.04V.
    With a 40 ohm load to TPS61291 and Vin =2.95V, Iin=70mA, Vo=2.87V EN/BYP=low. EN/BYP=high, Vin=2.95V, Iin=80mA, Vo=3.04V.
    With a 40 ohm load to TPS61291 and Vin =2.90V, Iin=69mA, Vo=2.82V EN/BYP=low. EN/BYP=high, Vin=2.90V, Iin=81mA, Vo=3.02V.
    With a 40 ohm load to TPS61291 and Vin =2.85V, Iin=68mA, Vo=2.77V EN/BYP=low. EN/BYP=high, Vin=2.85V, Iin=83mA, Vo=3.02V.
    With a 40 ohm load to TPS61291 and Vin =2.80V, Iin=66mA, Vo=2.72V EN/BYP=low. EN/BYP=high, Vin=2.80V, Iin=84mA, Vo=3.02V.
    From Vin=3.30V down to 3.05V I see no benefit in boost mode. I get the same or more current draw in boost mode and still dropout.
    From Vin=3.05V down to 2.95V I still see no benefit in boost mode. I can live with feeding LM3370 2.87V.

    LM3370 discusses a 100% duty cycle (low dropout control), pg 14 datasheet. I have notes that it doesn't work, but that is before I really understood that TPS61291 had a dropout, so maybe that was confusing me. I'll have to test it out further because if it really works, I can live in bypass mode down to Vin =2.90V. That's when the dropout from TPS61291 = 0.1V and then I'll be at the 2.8V rail that the camera needs. Interestingly, LM3370 does show slightly better Rds(ON) with the DSBGA package, but that requires a new layout that I don't want to do.
  • 0 in reply to Angelo Pietroforte
    Prodigy 200 points
    LM3370 pg 15 datasheet shows equation for dropout. Pg 16 doesn't show an equivalent curve for BGA package.
    I'm again thinking 2.95V into TPS61291 should be my final value before boost mode. All because with the 0.1V dropout from the TPS61291 PFET, that will get me to 2.85V input to LM3370, and then with the equation of pg15, I'll have enough dropout and not have to be so picky with respect to inductor used or worry as much about over temperature variations.
  • 0 in reply to Angelo Pietroforte
    TI__Mastermind 46905 points
    Hi Angelo,
    thanks for the detail bench test.
    in your measurement, you use a resistor to simulate the loading. however, the loading is not pure resistor. at least LM3370 is not resistance load. it is buck converter, which consumes less current if its input voltage is higher.

    I agreed with your conclusion: setting the mode transition value to 2.95V is reasonable.
  • 0 in reply to Jasper Li
    Prodigy 200 points

    Glad you concur.  I'm moving forward on the entire system layout and am about 3/4 finished.  How is someone supposed to secure 20 samples of LM3370SD-3416/NOPB when 0 stock exists with the normal distributors:  Avnet, Arrow, Digikey, Mouser?

  • 0 in reply to Angelo Pietroforte
    Prodigy 200 points
    Yes, LM3370 is not resistance load, but I had to bench test the TPS61291 the best I knew how and that is via a 40ohm load. Vo/Io=3.0V/75mA=40ohm.
  • 0 in reply to Angelo Pietroforte
    Prodigy 200 points
    Vo and Io of TPS61291 is Vin and Iin of LM3370. Vo and Io of LM3370 feeds camera image sensor. Yes buck is efficient when Vin > Vo.
    I was an analog FAE for National Semiconductor about 17 years ago, and then I became an applications engineer working on PMIC's for IDT in Si Valley about 10 years ago. Your support is still very much appreciated. You time zone is 8hr ahead. I'm guessing you provide support from England. National had a big support center in Firstenfeldbruk in Germany about 20 years ago.
  • 0 in reply to Angelo Pietroforte
    TI__Mastermind 46905 points
    Hi Angelo,
    I mainly support boost converter, so I can't reply your question about the LM3370. I also assign the post to the engineer covering buck device.
    another boost converter may be easier is TPS610985, which automatically operates at pass-through mode at VIN > VOUT and boost mode if VIN < VOUT. Giving that you have many experiences in Analog device, I'm not concern to provide more options.

    I locate in Shanghai. TI have a product line here that focus on boost converter design. Nice to know you!
  • 0 in reply to Jasper Li
    TI__Guru**** 191445 points

    Angelo,

    Please let me know what specific questions you may have about LM3370.

  • 0 in reply to JohnTucker
    Prodigy 200 points
    Hi John,
    I'm moving forward on the entire system layout and am about 3/4 finished. How is someone supposed to secure 20 samples of LM3370SD-3416/NOPB, or for purchase when 0 stock exists with the normal distributors: Avnet, Arrow, Digikey, Mouser? I used to be a National Semi FAE 17 years ago, so I apologize up front if my first question to you is non-technical. I know other flavors of LM3370 are more available, but that will require me to absolutely figure out the programming. And oddly, a few quarters ago, I lost the working recipe for those details. The -3416 allows me to bring up my system at the exact 2.8V and 1.4V that it needs.
  • 0 in reply to Angelo Pietroforte
    TI__Guru**** 191445 points

    Angelo,

    I can research this further, but right now Rochester Electronics is the only place showing stock:

    If you don't need the I2C interface, we have other (non pin2pin) options.  FYI, this is the first LM3370 inquiry I have received, or seen in the last several years.

  • 0 in reply to JohnTucker
    Prodigy 200 points
    Hello John,
    Yes, I saw www.rocelec.com/.../TEXTISLM3370SD-3416-NOPB last night, and I emailed them to see if they'd sell me 20, because the website says 329minimum. They probably don't want to break up a small reel. It always seems that the other flavors of LM3370 have some caveat. I'm trying not to rethink the power subsystem, but it seems many things are pushing me in that direction.
  • 0 in reply to Angelo Pietroforte
    TI__Mastermind 46905 points
    Hi Angelo,
    do you have further tech question about the boost converter or the LM3370 device?
  • 0 in reply to Jasper Li
    TI__Guru**** 191445 points
    I will close this thread for now. You can re-open it any time if you need further assistance.