1792 réponses à ce sujet

[Yannovsky]

I got the charging problem only once, never drain the battery off second time, so I'm unable to confirm the issue yet. But I think the problem was more due to BMS issue. I'll try to do some tests later. Already disconnected the charger from my robot.

[Serveurperso]

sur la PI t'as un jumper qui l'empêche de démarrer et la bloque en consolation de quelques milliamperes. c'est mieux qu'un interrupteurs général bouïné

[firened]

hey that sounds great. I'll take a look at that

[Mike118]

In any case, the switch is something that have been asked by three people now... Yes we made it to be H24 switch on ... But some people want to be able to easily switch it off ... So if it can solve both thing at the same time, adding a switch is not a bad solution... But I am still open to any other practical suggestion =).

[firened]

it's pretty hard to find information on it..

the PEN or RUN pin could be used to achieve the same as the switch. to keep a pi shut off (except some mA), and to get it out of a discharged state.

raspberry pi 2 B+ and newer have a RUN pin.
raspberry pi 3 B+ and newer have RUN and PEN.

I don't know whether the PEN or RUN pin uses less power..
they are both pulled high. if we connect a switch from one of those pins to GND it will disable the pi.

all in all, cutting a trace on the UPS would not be necessary this way, but it still uses a tiny bit of power (6Ah charge last 24 days @ 10mA).

and it's still a manual solution for when the battery is empty.

[Serveurperso]

RUN is a processor reset pin (processor disable)
PEN is general power disable
RUN and PEN are pull-up to VCC
RUN GND = 100mA
PEN GND = 10mA

[firened]

pulled down RUN on a zero is only 3mA.
https://raspberrypi....-zero-long-term

[Serveurperso]

yes but zero is useless for vigibot
remaining 10mA is for the red leds

[firened]

yeah. i couldn't find power consumption for any other board.

[Serveurperso]

RUN to VCC = CPU only disabled = 100mA
A pulse to RUN is like Reset on any microcontroller
PEN is like a hardware power off
without current requierment on a bigger switch

[firened]

where did you find information on current consumption?

[Serveurperso]

and bad contact risk of a bad switch
everywhere
https://www.google.c...Q4dUDCAY&uact=5
https://www.raspberr...ic.php?t=207961
interesting. a real poweroff from software
GPIO to PEN
linux halt = full poweroff
dtoverlay=gpio-poweroff,active_low=y
which will safely shutdown raspbian, then finally bring GPIO26 / PEN low to shut off power.
11mA measured
and poweron pulse button
this is the missing featue on our PI
a real poweroff and pulse to power on
now an UPS can auto poweron @ 10% battery and no more problem
and serious design
like any ATX computer
a power switch on low voltage is a bad desig
n
The power LED will probably be consuming about 4mA, the PEN pull-up about 2.5mA, the Transorb over-voltage protector, the PMIC and high-side switch for HDMI 5V should not be drawing much. So there's about 4-5mA unaccounted for.
the other solution is small solid state relay
https://fr.farnell.c...s-de-relais=cms
A good one = https://fr.farnell.c...t-no/dp/1608203
4.5A 0.1ohm

[firened]

interesting information

[Mike118]

So this is what I am studying to use to solve the problem : https://www.pololu.com/product/2808

[firened]

-i just remembered the low voltage alarm pin on the CW2015, maybe it can be connected to a switch and FET to solve both requirements!

-the pololu pushbutton switch looks like a possibility too. CTRL pin probably wants clear signals, so maybe also the alarm pin from the CW2015?
from the pololu pushbutton spec:
"Switch can lose its state when power is disconnected (the switch typically defaults to being off when power is first applied, but it can default to being on near the upper end of the operating range)"
would that be a problem or how many days will the battery last after the pololu switch turned off?

[Mike118]

there is also the same product with a 2 position switch

[firened]

ah nice
@Pascal do you have a discharged battery (at around 2.4V) that you can test with your 4 diodes?

with 4 x 0.6V diodes the voltage will get stuck at 2.6V and will never reach 3.0V to actually start charging the battery.

either way. I'm not a fan of this, i think it's dangerous and advise against using it.
this is the correct group here to discuss this i think.

[Mike118]

yes best place here to discuss about how to improve the UPS

[Serveurperso]

yes
0.75V because load
yes 4 diode is too much !!!
3 diode look perfect
but i need to measure the CC part @ the begin of charge
remember two 18650 can handle a charge up to max 10A without problem !!!
5 better
we are limited with the charger, the low voltage, and wire vdrop !!!
and current is BMS limited also !

[Mike118]

if someone plug the charger without batterie connected all the power will be provided trought the diode ...