SX28 current draw with and without "TURBO"

Bean

I read somewhere that when you don't specify TURBO the SX28 divides the clock by 4.
If that is so, can I use 32KHz internal RC clock without turbo to get even lower current draw ?

Has anyone ever measured current with and without TURBO ?
Bean.
·

Paul Baker

I believe all that non-TURBO does is unstage the pipeline (Table 2.7 in the User's Manual), meaning each instruction is serialized, and the the next instruction isn't executed (starting with the fetch cycle) until the write phase of the previous is completed. Since each stage is active only every fourth cycle and the quiescent current is much lower for CMOS than the active current, you should see a reduction in power consumption. But depending on your application the amount of current saving you'll see will vary. If you are using all of your pins as outputs, the current savings will likely be quite small percentage wise. But if your SX has no interface with the external world the difference could be considerable, but this will depend highly on which subsystems are used and some subsystems will not be affected in non-TURBO (for example the RTC subsystem should still increment every cycle). Even in the ideal situation I would be somewhat suprised if the current draw came close to 50% (I would stab a guess at 60-75%) Of course the only way to know is to try it. Perhaps you can experiment and provide use some numbers like Tracy Allen did with the comparator function. I'm sure the group would appreciate it.

-Paul

Post Edited (Paul Baker) : 1/19/2005 4:51:53 AM GMT

Bean

I was about to do that, but I thought I would check to see it anyone else had already done it.

I'll try a couple experiments...

Bean.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Cheap used 4-digit LED display with driver IC·www.hc4led.com

Low power SD Data Logger www.sddatalogger.com
SX-Video Display Modules www.sxvm.com
Stuff I'm selling on ebay http://search.ebay.com/_W0QQsassZhittconsultingQQhtZ-1

"USA Today has come out with a new survey - apparently, three out of every four people make up 75% of the population." - David Letterman

Post Edited (Bean (Hitt Consulting)) : 1/24/2007 1:32:20 PM GMT

Paul Baker

It would be great to get a set of numbers for each subsystem so those on a power budget can make accurate predictions for battery life. Maybe Ill attempt this in March (after the contest). -Paul

Bean

Well, it seems the non-turbo mode does NOT divide the clock speed. At the slower clock rates it really doesn't save much current at all.

Here is the test program:
DEVICE SX28,STACKX,OPTIONX,OSC128KHZ
FREQ 128_000

PROGRAM Start NOSTARTUP

Start:
TRIS_A=0
TRIS_B=0
TRIS_C=0
Main:
TOGGLE RB.5
PAUSE 5000
GOTO Start

Here is what I measured:

32KHz Turbo = 1.32mA
32KHz Non-Turbo = 1.30mA

128KHz Turbo = 1.46mA
128KHz Non-Turbo = 1.40mA

1MHz Turbo = 3.1mA
1MHz Non-Turbo = 2.4mA

4MHz Turbo = 7.7mA
4MHz Non-Turbo = 5.2mA

Considering that 1MHz Turbo (3.1mA) has the same performance as 4MHz Non-Turbo (5.2mA). It is better to use the turbo mode if you are looking at MIP per mA.

Overall I was surprised by how low the current was in general.

I want to try repeating the test at lower supply voltages.

Terry

Stephen

The clock speed will always stay the same. Non-turbo mode simply adds a few cycles delay to emulate the instruction timing of a PIC. It's purpose was simply to allow PIC firmware to be directly loaded to the SX and still be able to run. Nobody uses the SX this way anyway, as all you really gain is Flash programming - the power and cost will always be higher. Everyone runs from turbo mode to take advantage of the speed anyway. The faster design simply has a much higher overhead... You'll notice that the ramp in clock freq. vs current consumption is not exactly linear.

Stephen

In the past I have measured as low as 100uA running at 2.5V at 32KHz.