Forums

LPC2124 power consumption

Started by Unknown March 2, 2004
I have looked at the data sheets for several of the LPC processors and
they all list 30 mA as the "typ" power consumption with the CPU
running a NOP loop and no peripherals. Obviously this is not a very
useful number without info on the peripherals. Anyone measured the
core and IO currents while running a app that uses some portion of the
IO while accessing both the internal RAM and Flash? Even a combined
total current number would be more useful than the data sheet.



An Engineer's Guide to the LPC2100 Series

Hi,

power consumption has to be specified in an enviroment that everybody
can reconstruct. As soon as an application is involved, this does not
hold true any more. The power consumption of the peripherals is rather
low compared to the CPU and the memory itself.

I/O currents are completely indenpendent from the chip itself, they
are solely depending on your (or somebody else's) application and are
not representative for the micro.

Regards, Robert --- In , redsp@y... wrote:
> I have looked at the data sheets for several of the LPC processors and
> they all list 30 mA as the "typ" power consumption with the CPU
> running a NOP loop and no peripherals. Obviously this is not a very
> useful number without info on the peripherals. Anyone measured the
> core and IO currents while running a app that uses some portion of the
> IO while accessing both the internal RAM and Flash? Even a combined
> total current number would be more useful than the data sheet.





--- In , "philips_apps" <philips_apps@y...> wrote:
> Hi,
>
> power consumption has to be specified in an enviroment that everybody
> can reconstruct. As soon as an application is involved, this does not
> hold true any more. The power consumption of the peripherals is rather
> low compared to the CPU and the memory itself.
>
> I/O currents are completely indenpendent from the chip itself, they
> are solely depending on your (or somebody else's) application and are
> not representative for the micro.

That is where we disagree. If I could "reproduce" the measurement,
then I would just make the measurement myself and I would not need the
data from you.

I need a reasonable approximation for a realistic power consumption.
I think everyone understands that power consumption is a function of
what software/hardware is running at the time. The number in the data
sheet is meaningless (literally since I don't even know if it is
accessing Flash or RAM) and has no value in estimating requirements
for a power source. I don't know if I need to provide 50 mA or 500 mA.

If I can't come up with some sort of number to allow me to size a
power supply, I'll just use another MCU that provides better data,
like the OKI parts.




OK,

here we go, let's go OKI. Let's just assume our device follows the
ratio between current used for core and current used for I/O as the
OKI device, let's compare.

The OKI 674001 series is specified with typical 40 mA running @ 33 MHz
while I/O current is typical 18 mA @ 33 MHz. So I would consider the
core current to be MUCH more significant than the I/O current.

May be I missed something in the OKI Manual but what is running while
the "typical" power consumption for I/O is measured? One timer, all
timers, one serial interface, all serial interfaces...???

My point being, the typical power comsumption of the OKI device at 33
MHz is higher than the LPC2106 at 60 MHz and whether it is specified
in more detail than the LPC2000 devices, granted we have work to do
but to make a qualified decision which one is better in power
comsumption the data is there.

The typical for the 675001 at 60 MHz is much higher (75 mA) than that
of the LPC2124 @ 60 MHz.

And we are only talking mA not even Volts. Taking the Volts into
account, 2.5V * 75 mA typical versus 1.8V * 50 mA typical makes a nice
factor of 2. Approx. 190 mW versus 90 mW

If you think you need to go for the great OKI device because of a more
detailled (but much higher) power spec, I won't be able to stop you.

Sorry for being a little defensive here, Robert

Hint: The power supply for your board is highly depending on your I/O
drive requirements. Add to those worst case I/O requirements 50 mA for
the micro at full speed and you will do fine.If you want to make it
bullet proof at -40C, and running the micro at 2V insteadof 1.8V, add
70 mA instead of 50 mA

--- In , redsp@y... wrote:
> --- In , "philips_apps" <philips_apps@y...>
wrote:
> > Hi,
> >
> > power consumption has to be specified in an enviroment that everybody
> > can reconstruct. As soon as an application is involved, this does not
> > hold true any more. The power consumption of the peripherals is rather
> > low compared to the CPU and the memory itself.
> >
> > I/O currents are completely indenpendent from the chip itself, they
> > are solely depending on your (or somebody else's) application and are
> > not representative for the micro.
>
> That is where we disagree. If I could "reproduce" the measurement,
> then I would just make the measurement myself and I would not need the
> data from you.
>
> I need a reasonable approximation for a realistic power consumption.
> I think everyone understands that power consumption is a function of
> what software/hardware is running at the time. The number in the data
> sheet is meaningless (literally since I don't even know if it is
> accessing Flash or RAM) and has no value in estimating requirements
> for a power source. I don't know if I need to provide 50 mA or 500
mA.
>
> If I can't come up with some sort of number to allow me to size a
> power supply, I'll just use another MCU that provides better data,
> like the OKI parts.





I am not trying to pick a fight. I am pointing out that your
information in the data sheet is of no value to an engineer. I have
adequate information from OKI which you may not have access to since
OKI is unlikely to provide you with support.

My main concern is that when I asked *you* for more information,
instead of getting that information, I got noise. I do not understand
what you are saying below, since in part, it *is* mainly defensive and
not responsive. Regardless of what is in the OKI data sheet or anyone
else's data sheet, I am looking for information on the Philips parts.

At the end of your message you mentioned several current figures, but
I don't know what these currents are. If your MCU is running a
realistic program that is hitting both Flash and RAM and lets just say
the UART is running at 38.4 kbps both xmit an rcv, can you give a typ
and max current figure for the CPU core 1.8 volt supply?

BTW, in this case Watts are irrelevant because I am trying to size my
power supply which will work equally well at 1.8 volts or 2.5 volts.
The total power consumption may be a factor in battery life, but this
is pretty much swamped out by the DSP on our board. I am just trying
to figure out if I need a 100 mA, 250mA or 500 mA regulator. Size is
important, so I don't want to to use a part any larger than I have
to.
--- In , "philips_apps" <philips_apps@y...>
wrote:
> OK,
>
> here we go, let's go OKI. Let's just assume our device follows the
> ratio between current used for core and current used for I/O as the
> OKI device, let's compare.
>
> The OKI 674001 series is specified with typical 40 mA running @ 33
MHz
> while I/O current is typical 18 mA @ 33 MHz. So I would consider the
> core current to be MUCH more significant than the I/O current.
>
> May be I missed something in the OKI Manual but what is running
while
> the "typical" power consumption for I/O is measured? One timer, all
> timers, one serial interface, all serial interfaces...???
>
> My point being, the typical power comsumption of the OKI device at
33
> MHz is higher than the LPC2106 at 60 MHz and whether it is specified
> in more detail than the LPC2000 devices, granted we have work to do
> but to make a qualified decision which one is better in power
> comsumption the data is there.
>
> The typical for the 675001 at 60 MHz is much higher (75 mA) than
that
> of the LPC2124 @ 60 MHz.
>
> And we are only talking mA not even Volts. Taking the Volts into
> account, 2.5V * 75 mA typical versus 1.8V * 50 mA typical makes a
nice
> factor of 2. Approx. 190 mW versus 90 mW
>
> If you think you need to go for the great OKI device because of a
more
> detailled (but much higher) power spec, I won't be able to stop
you.
>
> Sorry for being a little defensive here, Robert
>
> Hint: The power supply for your board is highly depending on your
I/O
> drive requirements. Add to those worst case I/O requirements 50 mA
for
> the micro at full speed and you will do fine.If you want to make it
> bullet proof at -40C, and running the micro at 2V insteadof 1.8V,
add
> 70 mA instead of 50 mA