PDA

View Full Version : Low Power, High Speed Stamp



Loopy Byteloose
01-28-2006, 01:04 PM
A few [or several] months ago, there was discussion of the possiblity of making a 3.4 volt version of the BasicStamp that would be both high speed and lower power.

Obviously, this would allow a lower voltage regulator threshold too.и Four rechargible NiMH or NiCads could be used and recharged by solar panel.

Am I just hoping, or is this being actually considered?

I do understand that thereиmay have to be an obvious move to 3.4volt logic or some adaption to jump over to 5.0volt logic.и

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"When all think alike, no one is thinking very much.' - Walter Lippmann (1889-1974)

ииииииииииииииииииииииииииииииииииииииииииииииииии ииииииииииииииииииии Warm regards,иииии G. Herzog [и黃鶴 ]иin Taiwan

KenM
01-28-2006, 10:02 PM
Isn't power inversly proportional to speed?
If one goes up, the other goes down?
Of course, the terms "high speed" and "low power" are relative terms..

Kramer said...
version of the BasicStamp that would be both high speed and lower power.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Ken

Loopy Byteloose
01-29-2006, 12:05 AM
Well the BasicStamp uses a SX-48 and as the clock speed goes up it uses more power.
When you use more power, you get more heat.
Also, when you you more power you have more EMI problems to handle as they use more power too.

So, the answer to your question is 'no'.

On the one hand, you can have the SX-48 operate at 50mhz at 5 volts and something like 75ma
Or you can have the SX-48 operate at 50mhz at 3.4 volts and something like 40-50 ma. [Sorry, but I am not looking at Guenther's data].

Since Amps X Volts = Power [in Watts], you begin to see that the 3.4 volt solution is significantly less in Power Consumption, in Heat, and in EMI problems.

You can still clock at 50mhz [or faster if that is called for]. The microprocessor just runs cooler and cleaner. I suppose one might argue that 3.4 volt logic is a bit more tricky as the range between 0 and 1 is narrower, but the newer 74HCXXs seem to handle it well.

Presently, the BasicStamp is setup to automatically reset with a brown-out fuse set at about 4.5volts, so you cannot just bypass its regulator and provide 3.4volts.и Parallax has to actually program the units to have a brown-out fuse which is set at about 3.2volts and make sure the EEPROMs can accept the lower voltage too.

Having done that, it seems that switching the voltage regulator to the lower voltage would be optimal.и If you wanted a 5volt unit, you could bypass the regulator and put regulated 5volts on the direct voltage supply pin.и Of course you would be back to higher power consumption, more heat, and more EMI.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"When all think alike, no one is thinking very much.' - Walter Lippmann (1889-1974)

ииииииииииииииииииииииииииииииииииииииииииииииииии ииииииииииииииииииии Warm regards,иииии G. Herzog [и黃鶴 ]иin Taiwan

Post Edited (Kramer) : 1/28/2006 5:11:50 PM GMT

Paul Baker
01-29-2006, 12:41 AM
KenM said...
Isn't power inversly proportional to speed?
If one goes up, the other goes down?
Of course, the terms "high speed" and "low power" are relative terms..

Power consumed by a digital silicon device is proportinal to cv2 where c is the switching speed and v is the operating voltage. Since its v squared, reducing the operating voltage permits a squared of speed increase for the same power, so doubling the speed and halving the voltage produces a system that runs at twice the speed and half the power. This is why they keep dropping the operating voltage on subsequent versions of Pentiums, so they can run faster while keeping the heat generated within acceptable limits.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
и1+1=10

Chris Savage
01-29-2006, 01:08 AM
To add to what Paul was saying (and for Kramer), running the SX48 at 5V/50 MHz will generate a lot of heat!и However running it at 3.3V/50 MHz makes it a whole lot more acceptable (generates less heat).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com (mailto:csavage@parallax.com)

Beau Schwabe (Parallax)
01-29-2006, 03:31 AM
Just playing with some numbers here...

Suppose at 5V/50MHz you are drawing 20mA.... This is effectively a 250 Ohm load.

R=V/I
R = 5V / 20mA
R = 250 Ohms

So now take your 250 Ohm load across 3.4V and you get...

I = V/R
I = 3.4V / 250 Ohms
I = 13.6mA

13.6mA vs. 20mA is a 32% reduction in current consumption

So To figure Power.... P=I*V ...

5V/50MHz = 100mW ( 5V * 20mA)

3.4V/50MHz = 46mW (3.4 * 13.6mA)

46mW vs. 100mW is a 54% reduction in power consumption.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe (mailto:bschwabe@parallax.com)

IC Layout Engineer
Parallax, Inc.

Post Edited (Beau Schwabe (Parallax)) : 1/28/2006 11:30:00 PM GMT

Loopy Byteloose
01-29-2006, 07:10 PM
So, is there even going to be a 3.3 volt Stamp?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"When all think alike, no one is thinking very much.' - Walter Lippmann (1889-1974)

ииииииииииииииииииииииииииииииииииииииииииииииииии ииииииииииииииииииии Warm regards,иииии G. Herzog [и黃鶴 ]иin Taiwan