Author Topic: LTC3525-3.3  (Read 4576 times)

markus1234

  • NewMember
  • *
  • Posts: 14
  • Country: at
LTC3525-3.3
« on: April 07, 2017, 04:02:31 AM »
Hi,
has anyone experience with the LTC3525-3.3?

My atmega328 uses (unregulated - directly) 4.6 uA from 3.3v.

With the regulator, wired like in the image below, the current drw is more like 38 uA. Is it really that bad?



best regards

ChemE

  • Sr. Member
  • ****
  • Posts: 419
  • Country: us
Re: LTC3525-3.3
« Reply #1 on: April 07, 2017, 06:41:05 AM »
38uA at what input voltage?

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: LTC3525-3.3
« Reply #2 on: April 07, 2017, 07:28:23 AM »
I  agree with ChemE's question.  A lot of LDO's (maybe all of them?) current consumption goes berserk if the input voltage is below the regulated voltage. 

Another thing to also try: have you tried measuring the quiescent current?  Datasheet says it's a max of 15ua.  If it's not, then you've got work to do figuring out why.
« Last Edit: April 07, 2017, 10:48:45 AM by WhiteHare »

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 6866
  • Country: us
    • LowPowerLab
Re: LTC3525-3.3
« Reply #3 on: April 07, 2017, 08:10:47 AM »
Guys the LTC3525 is a booster, not an LDO. So obviously there will be more quiescent draw than say the MCP1703 (just ~2uA stepdown). The DS has the graphs but the higher the gap from the battery to the output voltage the more energy is spent boosting. From my own experience the LTC3525 chip has been fantastic and I use the 5V version on the Moteino PowerShield.

ChemE

  • Sr. Member
  • ****
  • Posts: 419
  • Country: us
Re: LTC3525-3.3
« Reply #4 on: April 07, 2017, 08:51:54 AM »
Exactly my thought Felix.  If it is drawing 38uA from a 0.9V dead alkaline battery then that might not be too bad at all.  38uA from a 3.6V LiPo doesn't sound too appealing.

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 6866
  • Country: us
    • LowPowerLab
Re: LTC3525-3.3
« Reply #5 on: April 07, 2017, 10:06:06 AM »
If you find a booster that has an output of 400mA (not from the lowest voltage) and is such low quiescent, i want to hear about it :)
I did a lot of research before I decided on this one, and I tried a few which were flaky or not performant. This is the best I could find. LT parts are truly great in my own practice.

markus1234

  • NewMember
  • *
  • Posts: 14
  • Country: at
Re: LTC3525-3.3
« Reply #6 on: April 07, 2017, 01:03:57 PM »
My current results are the following:

"Real" current usage after  the LTC3525: 4.9uA

current usage before the LTC3525:

Vin: 5V - 42 uA
Vin: 3.3V - 38 uA
Vin 1.5V - 36 uA

Is this reasonable?

Thanks for all the answers

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 6866
  • Country: us
    • LowPowerLab
Re: LTC3525-3.3
« Reply #7 on: April 07, 2017, 03:16:38 PM »
Yes that looks quite reasonable.

markus1234

  • NewMember
  • *
  • Posts: 14
  • Country: at
Re: LTC3525-3.3
« Reply #8 on: April 07, 2017, 03:33:26 PM »
Wow, this is not the 70 - 85% efficiency I expected from the datasheet  ;D

It seems to be really hard to power something from a single AA with that.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: LTC3525-3.3
« Reply #9 on: April 07, 2017, 04:49:03 PM »
If you're worried about the magnitude of the quiescent current, you could use your boost converter to charge a capacitor.  Run your Moteino from the capacitor and have your Moteino disable the booster after the capacitor is charged.  According to the DS, when disabled, the LTC3525 draws <1ua ("typically" just 0.1ua).  When the voltage drops low enough, turn it on again.  Repeat the cycle indefinitely. 

To go even further, perhaps you could also optimize to run the LTC3525 only during its phase of maximum efficiency.   That's probably your best chance at powering your Moteino at something near the claimed efficiency...   :)
« Last Edit: April 07, 2017, 05:28:25 PM by WhiteHare »

markus1234

  • NewMember
  • *
  • Posts: 14
  • Country: at
Re: LTC3525-3.3
« Reply #10 on: April 07, 2017, 05:37:25 PM »
I guess this would be something like: https://lowpowerlab.com/forum/low-power-techniques/a-new-sleep-mode-using-225-na/ right?

Honestly, I wanted to keep it simple, but now.... I guess I need to grab some mosfets :)
(Or just eliminating it completely and use  2 AAs instead of 1)
« Last Edit: April 07, 2017, 05:39:56 PM by markus1234 »

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: LTC3525-3.3
« Reply #11 on: April 07, 2017, 06:24:39 PM »
Yeah, if you want efficient AND simple, it's hard to beat two AA's.

markus1234

  • NewMember
  • *
  • Posts: 14
  • Country: at
Re: LTC3525-3.3
« Reply #12 on: April 08, 2017, 02:53:33 AM »
Hm, can you tell me,  according to http://oregonembedded.com/batterycalc.htm I should expect a battery life of 4 years on an alkaline AA battery (calculated with 2000mAh). Isnt the self discharge in 4 years even higher than my usage (with the ltc3525) ?

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: LTC3525-3.3
« Reply #13 on: April 08, 2017, 04:48:26 AM »
Hm, can you tell me,  according to http://oregonembedded.com/batterycalc.htm I should expect a battery life of 4 years on an alkaline AA battery (calculated with 2000mAh). Isnt the self discharge in 4 years even higher than my usage (with the ltc3525) ?
How did you arrive at that conclusion?

markus1234

  • NewMember
  • *
  • Posts: 14
  • Country: at
Re: LTC3525-3.3
« Reply #14 on: April 08, 2017, 08:54:03 AM »
For example http://gammon.com.au/power states 2% - 80uA continous for alkaline AA. Google picture graphs also seem to settle on this 2% / month value. But I'm not 100% sure. Last week I dug up a multimeter which I used in 2005 and  the 4 AA batteries in it are still at 1.1 V under load.   :o