Hi !
that's what I'm looking for ; actually I use Lion accu and HT7333 to feed my sleeping ATMEGA's and the whole card with NRF24L01 is under 3,5uA. Main part of power is for regulator (about 2uA) ...
I cant use one CR2032 or so : NRF needs at least 3V (and not more than 3,3V) to keep some power to transmit.
does anyone know a more sober regulator ?
Thx
2uA for a regulator quiescent current consumption is not at all bad.
That is 1mAh in 20 days.
Can you switch off the power to the card with NRF24L01 using a mosfet during quiescent periods ?
IIRC "bare" nRF works down to 2.0V. Maybe even less. If you don't use the PA+LNA version just remove the regulator and use single CR2032. And I would try single unregulated (non-rechargeable) coin cell even with PA+LNA. Add a HUGE cap to cover power spikes and it may work well.
6v6gt:
2uA for a regulator quiescent current consumption is not at all bad.
That is 1mAh in 20 days.
Can you switch off the power to the card with NRF24L01 using a mosfet during quiescent periods ?
Sure, 2uA is not that bad... 1 should be better 
usage of NRF is irregular.. some times 1 hour sleeping sometimes 2 times in the same minute. Power on is greedy so sleep is ok with less than 1 uA.
With 120mAH Lipo I get about 18 months... I hope to reach more than 2 years like my 800MHz intrusion detectors.
Smajdalf:
IIRC "bare" nRF works down to 2.0V. Maybe even less. If you don't use the PA+LNA version just remove the regulator and use single CR2032. And I would try single unregulated (non-rechargeable) coin cell even with PA+LNA. Add a HUGE cap to cover power spikes and it may work well.
Thats what I use (may be I will integrate the chip on my pcb, I'm working on small dipole wire antenna to improve comm range)
NRF24L01 works down to 2V... but with 3V it can communicate at max 8 meters inside a builbing and that's not a lot. At 2V noway... A cap for more than 5 mS should be VERY HUGE in fact and VERY expensive and cumbersome. CR2032 has about 10 Ohms of IR ; better to use it with less than 10mA charges.
AFAIK the nRF has a LDO inside that powers it with 1.8V. I am not sure if it is for the whole chip or the logic only. Are you sure the range is influenced by supply voltage? I get decent range with nRF transmitter powered by 2 alkaline cells (about 2.4V currently) - the receiver is mains powered module with PA+LNA. With 1000uF cap the transmitter is able to send many messages before it dies.
Smajdalf:
AFAIK the nRF has a LDO inside that powers it with 1.8V. I am not sure if it is for the whole chip or the logic only. Are you sure the range is influenced by supply voltage? I get decent range with nRF transmitter powered by 2 alkaline cells (about 2.4V currently) - the receiver is mains powered module with PA+LNA. With 1000uF cap the transmitter is able to send many messages before it dies.
Interesting... what length ? in which space ? do you know how long that take to transmit with repetitions
CR2032 and other lithium coin cells are able to provide only a few mA of current, if you need larger current
pulses some other battery is needed. Alkaline 1.5V button cells can produce far larger currents.
Anyway 2uA draw is less than the self-leakage/self-discharge of pretty much every battery type out there, its good enough.
MarkT:
Anyway 2uA draw is less than the self-leakage/self-discharge of pretty much every battery type out there, its good enough.
"good enougth" is a very personal advice
For information I've done precise measurement of the power consumption of this ATMEGA card working at 3.3V -
1 day average with 120mA lipo accu (more than 500 days autonomy)
@ awake period 2sec with single level detection ; 800uS data acquisition every 30sec ; nrf24L01 transmit 12/hour :
3,5mS delay from interruption to vector fetch 17% aver. 1,2mA
sleep 37% aver. 3,5uA
accu leakage 35%
other(working+ transmit time) 11%
So 1uA of sleep consumption saved is 10% more time... not negligible
In the same spirit, Is it possible to choice/modifiy the fuses configuration when burning MiniCore ?
I wish to try with shorter cristal start time.
Thx
MCP1700 1.6uA quiescent current typical
MCP1711 0.6uA quiescent current typical
The ATmega328p can tolerate a voltage range between 1.8 and 5.5 volts assuming the clock speed is 4MHz or less, and the brown out fuse is appropriately set.
If the regulator is needed only for the radio part, then power the MCU directly from the lipo cell and switch a mosfet during the time the radio is required. The mosfet controls power to the radio part and the voltage regulator so the quiescent current is no longer an issue.
max777:
"good enougth" is a very personal adviceFor information I've done precise measurement of the power consumption of this ATMEGA card working at 3.3V -
1 day average with 120mA lipo accu (more than 500 days autonomy)
@ awake period 2sec with single level detection ; 800uS data acquisition every 30sec ; nrf24L01 transmit 12/hour :3,5mS delay from interruption to vector fetch 17% aver. 1,2mA
sleep 37% aver. 3,5uA
accu leakage 35%
other(working+ transmit time) 11%So 1uA of sleep consumption saved is 10% more time... not negligible
In the same spirit, Is it possible to choice/modifiy the fuses configuration when burning MiniCore ?
I wish to try with shorter cristal start time.Thx
Very few battery chemistries have a self-discharge time constant much longer than 500 days. Yearly battery replacement/recharge is a good target to go for - less often means you simply forget to do it, once a year can go in the diary.
Self-discharge accelerates with higher temperature too, so its rather variable. For long battery life you are better looking at specialist batteries like large primary lithium (shelf-life 10 years), or other primary cells with low-self-discharge. LiIon isn't the best choice.
srnet:
MCP1700 1.6uA quiescent current typical
MCP1711 0.6uA quiescent current typical
thank you ; mcp1711 interesting ... unfortunately I have to modifiy pcb for 5 leads sot23
(MCP1700 more expensive than HT7333 and not very different)
6v6gt:
The ATmega328p can tolerate a voltage range between 1.8 and 5.5 volts assuming the clock speed is 4MHz or less, and the brown out fuse is appropriately set.
If the regulator is needed only for the radio part, then power the MCU directly from the lipo cell and switch a mosfet during the time the radio is required. The mosfet controls power to the radio part and the voltage regulator so the quiescent current is no longer an issue.
My bad, thx a lot... I'll test that
MarkT:
...Yearly battery replacement/recharge is a good target to go for...For long battery life you are better looking at specialist batteries like large primary lithium (shelf-life 10 years), or other primary cells with low-self-discharge. LiIon isn't the best choice.
Bad access to the devices so we try to have the longest life for accus ... and low price too, thx
max777:
My bad, thx a lot... I'll test that
I could have added:
A P-channel mosfet e.g. Si2333DDS ( good for Vgs down to 1.5 volts ) or just a PNP transistor like 2n3906.
I have, incidentally, an application for which the remote part has been running just under 2 years on a pair of alkaline AA batteries. In that time the battery voltage has dropped from 3.295 volts to 2.757 volts. It is actually a mail box notifier system. The remote transmitter part consists of a barebones ATMega328p running at 4MHz, an NRF24L01 with PA + LNA and an IR/ LED sensor array and a 3.3 volt boost converter. It wakes up every 32 seconds, scans the mailbox, and sends a report including the mailbox status and the battery voltage. I'll be changing the batteries soon, not because of the reduced battery voltage - it has been tested down to 2.2 volts, but because of the risk of damage resulting from battery corrosion, leakage etc.
Hi
6v6gt:
I could have added:
A P-channel mosfet e.g. Si2333DDS ( good for Vgs down to 1.5 volts ) or just a PNP transistor like 2n3906.I have, incidentally, an application for which the remote part has been running just under 2 years on a pair of alkaline AA batteries. In that time the battery voltage has dropped from 3.295 volts to 2.757 volts. It is actually a mail box notifier system. The remote transmitter part consists of a barebones ATMega328p running at 4MHz, an NRF24L01 with PA + LNA and an IR/ LED sensor array and a 3.3 volt boost converter. It wakes up every 32 seconds, scans the mailbox, and sends a report including the mailbox status and the battery voltage. I'll be changing the batteries soon, not because of the reduced battery voltage - it has been tested down to 2.2 volts, but because of the risk of damage resulting from battery corrosion, leakage etc.
Hi !
I usualy use DMG2305 for that kind of purpose (less than 100milliOhms @ 1.6V for Vgs) ; it seems cheaper 
2xAA batteries with more than 3.2V ! good stuff ! I need very small battery for my small hardware (here 1.0 release) 2xAA bigger than the box. I have no more room for power off circuitry and just go to unlimited sleep when battery voltage goes under 3.2V. With your solution, sleep current is under 0,5uA : maybe 6 more months !
Actually I'm bored with fake LM335/TMP36 from china... do you know some cheap and fast to92 thermo ?
