LPC4337 Power Consumption Problem (more than 200mA)

[lpcware]

Content originally posted in LPCWare by razor9119 on Thu Jan 08 13:25:16 MST 2015
I have both ngx expresso and keil mcb board. Also i made custom board with their reference circuitry. I tried all same code on these boards. But whatever i do, they have more than 200mA power consumption at 5v. So they have far more than 200mA consumption at 3.3v. I used both keil and ngx libraries. But i couldnt figure out why.

In the datasheet, power consumption is ~80mA at full speed.

Any suggestion in both software and hardware is ok for me. Core m0 is in reset while boot. If you have lowest possbile power cons. software project, please send that project. I want to try it. I need to figure out what is the problem ? software or hardware ?


*I made this custom board with keil mcb4300 and ngx expresso reference circuity.

[lpcware]

Content originally posted in LPCWare by bavarian on Fri Jul 10 02:17:05 MST 2015
Can you be more precise with regards to the measurement setup?
On the Keil board at least you have the jumper block J1 near to the loudspeaker which allows you to measure the current for 3.3V Analog/Regulator/IO.
However, I don't think that the numbers you see are absurd.
"Having no code" means the chip at least started up in a specific boot mode and hangs up there. Let's say it's starting in UART0 bootmode, then at the end of the bootcode it sits in a loop waiting for a '?' character. This is a well defined situation, the chip runs on 96MHz, M0 core is in reset and peripherals are enabled.
According to the diagrams in the data sheet this will give you a base floor of 40mA for the ARM and the bus system, on top of that you will get about 50mA for all the enabled peripherals. consume.
If you start with a bootmode which boots from memory but there is only random data in there, the result can be anywhere because you don't know what the ARM is doing with it.

If you drive the system to the maximum clock of 204MHz, then of course the power consumption will be higher. How high? what do you think?
The power consumption of a clocked piece of silicon goes linear with the clock. If you look for the M0 number in the table in the data sheet, 3.3mA @48MHz and 6.6mA @96 Mhz then you might get a feeling what it will be at 192MHz. Right, 13.2mA. The whole system will have double the power consumption if I double the clock.

What can be done to get it down:
[list]
  [*]  disable branch clock for peripherals which are not used
  [*]  decrease the input clock for peripherals (where it is possible) with a pre-divider
  [*]  supply the chip with a lower voltage if feasable
[/list]

Regards,
NXP Technical Support Team

[lpcware]

Content originally posted in LPCWare by smckown on Tue Jul 07 16:29:37 MST 2015
Seeing a similar problem on some custom LPC4337 boards.  ~ 90 mA with no code on board, almost 200 mA with code running on board.  Initial development using the NGX dev board, which also acted as a hardware design reference.  All IO pins appear to be configured correctly and board is operating properly.  Looking at current consumption numbers from the data sheet, I'm thinking ~ 100 mA is closer for the peripherals in use.

Razor9119, did you make any headway with your power consumption issue?

[lpcware]

Content originally posted in LPCWare by razor9119 on Fri Jan 09 08:59:32 MST 2015
But Total periph consumption isnt MORE than 50mA even all is open.
İnternal Flash use more 100mA???

[lpcware]

Content originally posted in LPCWare by embd02161991 on Thu Jan 08 18:44:14 MST 2015
Hi,
The code has to run from RAM and also the peripherals need to be disabled to reduce power.

Thanks,
NXP Technical Support