> Hi Ruben,
>=20
>=20
> Nice circuit, but a little complicated. It seems that your circuit needs
> a lot of different power domains.
>

Only 3.3V and 5V and some circuits can be switched on/off by the CPU to kee=
p=20
current consumption down when those curcuits are not used.
=20
> Are you stepping-up the battery voltage to 5V and then stepping it down
> again to 3.3V? What is the overall efficiency of the process?
> I was thinking about doing that, but I'm not sure it will have good
> efficiency.

Yes, since I need 5V anyway which draws the most current (LCD backlight) an=
d=20
doing like this I do get 3.3V without using a buck/boost regulator and can=
=20
still operate the circuit down to about 3.0V battery voltage. Not sure abou=
t=20
the overall efficiency but it is good enough for my needs. I did get around=
 3-4=20
hours continuously operating time, with backlight on (I think it was a 1000=
mAH=20
battery but I'm not sure now). Normal operation is on for around a minute a=
t a=20
time for 25 to 50 cycles per day and then downloading data via USB which=20
charges the battery.

>=20
> Wouldn't it be better if you connect the resistors of your P-FETs
> differently so you don't have a voltage divider in the gate? Which
> values do you use for the Source-to-Gate resistors and Gate-to-PIC
> resistors?
>=20

Source to Gate 1M and looking at the BOM I have 10K Gate to PIC which seems=
=20
very high. Can't remember why now (I did the circuit around 3 yers ago) and=
 the=20
BOM I have here may be wrong (copying and pasting components in the cad). I=
=20
usually put a low ohm resistor (100 ohm perhaps) between the PIC pin and th=
e=20
Gate for low frequency switching to limit short current surges in the PIC p=
in.=20
It also helps when doing the layout on a 2 side board. This only got to the=
=20
prototype stage for circuit evaluation.

/Ruben


>=20
> On 23/01/2015 21:11, Ruben J=F6nsson wrote:
> > Attached is what I came up with when doing something similar. Circuit i=
s
> > powered either from battery or USB and charged when attached to USB. Ch=
arger
> > circuit is MCP73833. When battery operated (USB detached), circuit is p=
owered
> > up with a keypress (CN10 pin 1/T2) and then self held until powered off=
 by
> > software (T4/T3). When USB is attached, circuit is always powered from =
USB (T5
> > not conducting). IC9 is a reset circuit that cuts power from the batter=
y
> > (except VBUP to RTC) when voltage is too low in order to prevent the ba=
ttery
> > from discharging too much.
> >
> > There is a 5V, TS1935 boost regulator and a 3.3V, MCP1603 buck regulato=
r that
> > has VIN from the 5V regulator in order to get the circuit operating at =
battery
> > voltage down to about 3.0V. The battery voltage is also measured with a=
n AD in
> > the CPU. Voltage to the AD is cut off with a P Channel MOS FET (T16) wh=
en not
> > measured.
> >
> > Charging is activated by the CPU after USB current has been negotiated =
to
> > 500mA or never leaves POWERED state (powered from USB charger with powe=
r
> > only).
> >
> > /Ruben
> >
> >
> >> Dear All,
> >>
> >>
> >> I need to design a mobile equipment operated by an 1-cell Li-Ion batte=
ry. I
> >> want the equipment to be able to charge and operate simultaneously whe=
n
> >> connected to a power supply and operate on battery power when not conn=
ected.
> >>
> >> Some restrictions:
> >> 1) I cannot afford the Vf drop of a diode from the battery to the circ=
uit; 2)
> >> When the circuit is charging and power is suddenly removed, it must co=
ntinue
> >> operating without any interruption;
> >>
> >> I devised the attached circuit (PDF). I hope it performs as expected.
> >>
> >> - Transistors T1 and T2 are P-channel MOSFETs with very low Rdson. Wit=
h
> >> both transistors off, current cannot flow in any direction because of
> >> the face-to-face transistors' body diodes.
> >> - When not connected to a power supply and not powered on, resistor R1
> >> keeps both transistor's gates at source potential and both transistors
> >> are off.
> >> - When connected to a power supply, diode D3 keeps both gates at sourc=
e
> >> potential and both transistors are off (even with transistor T3 being)=
..
> >> - When not connected to a power supply and with transistor T3 on (or S=
W1
> >> pressed), both gates will be at around 10% of Vbat and the MOSFETs wil=
l be
> >> on.
> >>
> >> Possible problem: Power supply voltage decreases slowly, preventing th=
e
> >> MOSFETs from turning on and keeping the circuit in brown-out condition=
..
> >> Possible solution: Use a PNP transistor controlled by a TL431 before
> >> diode D3.
> >> Con: More parts.
> >>
> >> Does anybody have experience with this kind of circuit? Is my analysis
> >> correct? Are there any obvious circuits better (simpler, cheaper) than=
 this?
> >> Comments and suggestions are welcome.
> >>
> >>
> >> Best regards,
> >>
> >> Isaac
> >>
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2015.0.5645 / Virus Database: 4273/8986 - Release Date: 01/24/15
>=20
>=20


--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.