Author Topic: coin cell powered R4  (Read 11682 times)

joelucid

  • Hero Member
  • *****
  • Posts: 869
Re: coin cell powered R4
« Reply #30 on: May 02, 2016, 05:12:06 AM »
Quote
They can be, but you need to follow careful guidelines when laying them out e.g. reducing the two current loops, minimizing stray inductance

No kidding - mine only tend to work well if I let Tom put the finishing touches on my design  :)

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: coin cell powered R4
« Reply #31 on: May 02, 2016, 05:16:56 AM »
BTW Linear do some nice ultra-low EMI boost regulators. Their app notes include designs for post and pre filtering on the power supply to reduce conducted noise and have PCB layout recommendations. If you're unsure what you're doing then follow the design layout guidlines to the letter, do exactly as they say and you should get a quiet design.
Mark.
« Last Edit: May 02, 2016, 06:39:13 AM by perky »

joelucid

  • Hero Member
  • *****
  • Posts: 869
Re: coin cell powered R4
« Reply #32 on: May 02, 2016, 05:30:49 AM »
Thanks. I recently read up a little on the subject of noise filtering (e.g. http://www.murata.com/~/media/webrenewal/support/library/catalog/products/emc/emifil/c39e.ashx is pretty thorough). So I hope I know a little more now.

I'm getting the first rev of my noise filtered espgw board today. It uses a mic5219 ultra low noise regulator, common mode choke, 3 terminal filter cap, ferrite bead and several decoupling caps to filter the power supply and a MAX14850 as digital isolator on the SPI / DIO0 lines. Can't wait to see how it'll do on a noisy power supply.

davegravy

  • Jr. Member
  • **
  • Posts: 56
  • Country: ca
Re: coin cell powered R4
« Reply #33 on: May 02, 2016, 05:09:24 PM »
I'm going to push forward a bit further with the boost converter...  Pulled the assembly off my solderless breadboard and direct soldered my components to the DIP adapter to keep leads as short as possible.  See attached picture.

Works now... at least, it gives me expected output voltage. Not sure about EMI, ripple current, efficiency or anything else yet.  But if I can get the whole project working with it, then things should only get better when it gets PCB-ized, right?.


TomWS

  • Hero Member
  • *****
  • Posts: 1892
Re: coin cell powered R4
« Reply #34 on: May 02, 2016, 05:25:43 PM »
I'm going to push forward a bit further with the boost converter...  Pulled the assembly off my solderless breadboard and direct soldered my components to the DIP adapter to keep leads as short as possible.  See attached picture.

Works now... at least, it gives me expected output voltage. Not sure about EMI, ripple current, efficiency or anything else yet.  But if I can get the whole project working with it, then things should only get better when it gets PCB-ized, right?.
Gee, I don't know... I think your implementation has that sort of je ne sais quoi that is so appealing!   8)

Rule #1 for hobbyist implementation: If it works, what the hey?  Go for it.  Let us know how the entire scheme works!

Tom

davegravy

  • Jr. Member
  • **
  • Posts: 56
  • Country: ca
Re: coin cell powered R4
« Reply #35 on: May 04, 2016, 10:38:34 PM »
Drawing 9uA 6.2uA now in listenmode.

Steps I've taken:

* Pulled the regulator
* switched to internal oscillator @ 8Mhz
* reduced BOD fuse to 1.8V

Code: [Select]

for (uint8_t i=0; i<=A5; i++)
{
   pinMode(i, OUTPUT);
   digitalWrite(i, LOW);
}

flash.initialize()
flash.sleep()
radio.listenModeStart();
LowPower.powerDown(SLEEP_FOREVER, ADC_OFF, BOD_OFF);

I recall reading that ~2uA is possible in listenmode. I'd rather keep the flash soldered, so I realize I might not get all the way there, but I should be able to do better right?. I haven't experimented with dropping the clock rate further, but what else can I do?

I tried
Code: [Select]
 disable_wdt(); 
but it had no effect. Guessing that gets called already within the powerDown function.
« Last Edit: May 05, 2016, 10:26:34 AM by davegravy »

davegravy

  • Jr. Member
  • **
  • Posts: 56
  • Country: ca
Re: coin cell powered R4
« Reply #36 on: May 05, 2016, 10:25:25 AM »
Scope calibration was off, actual consumption is 6.2uA, not 9.0

For comparison's sake I loaded the deepsleep example sketch and I'm getting 4.8uA.

At this point I think idle power draw is acceptable to meet my stated goals, and I've just developed an irrational obsession with cutting power. You guys gave me some kind of bug  :o
« Last Edit: May 05, 2016, 10:29:29 AM by davegravy »

joelucid

  • Hero Member
  • *****
  • Posts: 869
Re: coin cell powered R4
« Reply #37 on: May 05, 2016, 10:48:57 AM »
Quote
You guys gave me some kind of bug  :o

The lowest power mode with stock moteinos is to use the radio as timer by using listen mode with a RSSI thresh of 255. That gets you down to around 1.2uA. I ended up with a combination of listen mode and this low power setup. Essentially I switch to listen mode when I need interactivity and otherwise use the timer.

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 5996
  • Country: us
    • LowPowerLab
Re: coin cell powered R4
« Reply #38 on: May 05, 2016, 11:20:03 AM »
The lowest power mode with stock moteinos is to use the radio as timer by using listen mode with a RSSI thresh of 255. That gets you down to around 1.2uA. I ended up with a combination of listen mode and this low power setup. Essentially I switch to listen mode when I need interactivity and otherwise use the timer.

What timer do you refer to? Some kind of interrupt in listen mode?

WhiteHare

  • Hero Member
  • *****
  • Posts: 1298
  • Country: us
Re: coin cell powered R4
« Reply #39 on: May 05, 2016, 11:41:21 AM »

davegravy

  • Jr. Member
  • **
  • Posts: 56
  • Country: ca
Re: coin cell powered R4
« Reply #40 on: May 05, 2016, 11:42:53 AM »
The lowest power mode with stock moteinos is to use the radio as timer by using listen mode with a RSSI thresh of 255. That gets you down to around 1.2uA. I ended up with a combination of listen mode and this low power setup. Essentially I switch to listen mode when I need interactivity and otherwise use the timer.

What timer do you refer to? Some kind of interrupt in listen mode?

Think he's referring to https://lowpowerlab.com/forum/index.php/topic,1617.0.html and https://lowpowerlab.com/forum/index.php/topic,1325.0.html... but a lot of this is still over my head so I could be misunderstanding.
« Last Edit: May 05, 2016, 11:47:41 AM by davegravy »

davegravy

  • Jr. Member
  • **
  • Posts: 56
  • Country: ca
Re: coin cell powered R4
« Reply #41 on: May 05, 2016, 03:31:45 PM »
Quote
You guys gave me some kind of bug  :o

The lowest power mode with stock moteinos is to use the radio as timer by using listen mode with a RSSI thresh of 255. That gets you down to around 1.2uA. I ended up with a combination of listen mode and this low power setup. Essentially I switch to listen mode when I need interactivity and otherwise use the timer.

What I want to confirm is that 6uA is about the best you can do in an (interactive) listen mode without further amputating the moteino hardware (in my case, removing the flash chip which is according to datasheet is good for another ~1uA).  There's no other software optimizations I should be researching that you know of?


davegravy

  • Jr. Member
  • **
  • Posts: 56
  • Country: ca
Re: coin cell powered R4
« Reply #42 on: May 05, 2016, 05:09:47 PM »
No.  Re-read reply #37.

Maybe I'm misunderstanding it, but Joe's using a combo of two different listenmode methods

1) with the RSSI_threshold=255, uses 1.2uA but is non-responsive to other nodes (in simple terms it's a less expensive sleep mode than
Code: [Select]
Lowpower.powerdown(SLEEP_8S, BOD_OFF, ADC_OFF)
)

2) responsive listenmode... as per the listenmode_node and gateway examples. i.e you can actually wake the moteino with a transmission burst. This is more expensive. Within this state I'm getting 6uA, and thought I read elsewhere that Joe was reporting 2uA within this mode, so I just wanted to confirm that 6uA is reasonable and I'm not missing some optimizations within this.


TrendSetter

  • Newbie
  • *
  • Posts: 7
Re: coin cell powered R4
« Reply #43 on: June 20, 2016, 03:55:50 PM »
Does anyone have any thoughts on the LTC3525 in relation to the other boost regulators mentioned?  I have used them with success and learned of them from the jeenode micros.
I have opted for them over direct battery because of them being able to pull power from a battery that has almost no voltage.  Does that approach have merit?

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 5996
  • Country: us
    • LowPowerLab
Re: coin cell powered R4
« Reply #44 on: June 20, 2016, 04:17:02 PM »
I really like the LTC3525-5v too, I use them on the PowerShield. They just work and are very low quiescent draw.