6 posts / 0 new
Last post
ggkang
Offline
Last seen: 5 years 7 months ago
Joined: 2015-03-05 07:31
Question about SPI operation

Hi. i would like to design a simple remote control, which can detect just one signal sent by transmitter.

I have one question about SPI interface.

In Figure 5 in the document 'UM-B-012', U can see point A,B,C and D and i wonder the coded Flash memory can be activated at the point D after VDD is powered.

I think the waiting time (or delay) is so long from A to D and this time can waste an energy before DA14580 send a signal.

JE_Dialog
Offline
Last seen: 7 hours 40 min ago
Staff
Joined: 2013-12-05 14:02
Hello ggkang,

Hello ggkang,

if you want a simple remote control, the simplest way is to use the keyboard project as a template. Section 10 of the user manual shows how to customise the config.h to the different options of keyboard mapping - here you could have as simple as 1 button remote (we have customers already using tihs appraoch).

Also, fig 8.3 in the bootloader doc describes the compelte boot time : when run in extended sleep , then the start-up time is few mS from sleep to connection, not 100's of ms. In extended sleep, RAM is all retained, so no need to re-load from FLASH., therefore much quicker and it is what we use on keyboard applications.

BR JE_Dialog.

 

ggkang
Offline
Last seen: 5 years 7 months ago
Joined: 2015-03-05 07:31
The current challenge that we

The current challenge that we face is that the remote control has no battery (it uses a piezoelectric device)

Therefore every time the button is pressed, the RF block has to start from power-up, which means the data must be stored in flash and RAM is always reset.

In this particular case, does the chip consume 100's ms to load data from flash? (or, does it consume a few mA during boot-up as shown in the figure?)
This is a quite crucial question in selecting our RF terminal, so I would appreciate your reply.

ggkang.

JE_Dialog
Offline
Last seen: 7 hours 40 min ago
Staff
Joined: 2013-12-05 14:02
Hello ggkang, yes, each time

Hello ggkang, yes, each time you load from FLASH on  power reset, it will take time to read contents from FLASH to SRAM.  It can be reduced and speeded up using a secondary bootloader (in OTP) as in table below  This could be reduced depending upon the exact flash device.

Please see table below : it shows the energy consumption to load memory into SRAM :

If you can use OTP only : then loading SRAM from OTP memory on boot is much less power than FLASH  - and also much quicker... (one of the advantages of OTP).

If you can use OTP, i would suggest to look at the DA14581 as it has a faster boot time (its optimised for A4WP which has as requirement a 60mS max overall system boot time).

BR JE_Dialog

ggkang
Offline
Last seen: 5 years 7 months ago
Joined: 2015-03-05 07:31
according to the data it

according to the data it seems using an OTP it takes about 4 mA for 30ms. which means it will consume at least >100uJ for startup (while transmission seems to consume only a few 10s of uJ)..

Our energy budget goal is around 50-uJ at maximum, so if there is no way other than using an OTP, I think I can say this chip is beyond our energy budget.. :(
If you can provide any other way to reduce the bootup time or bootup energy, please let me know.

ggkang.

JE_Dialog
Offline
Last seen: 7 hours 40 min ago
Staff
Joined: 2013-12-05 14:02
Hello ggkang,

Hello ggkang,

below please find a cold boot plot of a 14581 development board... Ignore the high current spikes at the start of the test : this is just inrush current to the development board PCB on power up.. (didn't have a spare battery powered 14581 board available to measure).

The total energy needed by the DA14581 running proximity app, at 3V, booting from OTP is 38uC (120uJ).... and takes 30mS..with peak current of 5mA.

BR JE_Dialog