MSP430 vs MICROCHIP XLP

Hi Alessandro,

well, I would answer: write an efficient software and stay in low-power modes as long/often as you can.

Pls keep in mind: Every silicon manufacturer has the best products when compared to others!

Rgds
aBUGSworstnightmare 

Thanks aBUGSworstnightmare,

I know..you are right. It was just to make the best choice at the beginning, at least from a theoretical point of view.

Alessandro

Alessandro Giuliano said:

PIC MCUs with XLP have "Higher performance across 1.8-3V than MSP430

power vs. performance is not a linear relation, and just looking at the top pont is nice for advertizing, but completely misses the demands of really low-power projects.

Example: on 100mA, a cheap standard silicon 4148 diode will have almost the same forward voltage drop as a schottky BAT42, which is way more expensive. One could advertise the 4148 as way cheaper at the same performance. But if you don't need 100mA and go down to 1 or 0.1mA, the BAT42 wil only have a fraction of the 4148s forward voltage drop and fully justifies the increased costs.

BTW, iI don't like the PICs processor structure. It's a real mess. At least the 12/14 bit PICs I worked with.

That's true.

I was not considering that. Do you think I can get these performance info on different levels just  checking deeply on datasheets?

Alessandro

Alessandro Giuliano said:

Do you think I can get these performance info on different levels just  checking deeply on datasheets?

At least the MSP datasheets list different power consumption levels for different core clocking. I don't know for other device types. Sometimes the marketing department has demanded to get rid of this information as it would render their marketing strategies futile.
It's like with cars. On every advertising you'll see teh (maximum) horsepower, but not the torque or information on which rpm this applies and how it changes relative to rpm. It's nice to have a car with 500PS, but it is useless if these 500PS are at 10.000RPM and it doesn't have the torque to accelerate from 0.
I experienced this with out companies BMW. It has lots of PS, but only above 2000RPM. So if you start from a traffic light, it comes slowly, and then you'll suddenly get a kick in the ass (and possibly in the rear end of the car before you). Nothing about this in the datasheet (or at least deeply hidden and only indirectly to determine).

Back to MSP430 power efficiency: while computing power is linear to the clock, power consumption is reduced squared or even further when you clock the device down. This is why the LPM features, which completely stop the core, bring so much benefit even if you work with maximum speed the rest of the time.

To quote Winston Churchill: 'I don't trust  statistics I didn't forge myself". This is even more true for advertisings and marketing statements. If you don't have the information about the underlaying data and the selection criteria, you can state almost everything and prove it later (you know, legally an advertising statement has to be true, but this truth does not need to have any sane meaning)

To get back to the datasheet of the MSP: the 54xx datasheet has two tables about this. One lists the core current when operating on 3V for execution from flash ro ram (usually, your code will execute from flash, but it's possible to copy it to ram and execute it there to save more power if LPM isn't an option e.g. due to latency considerations) on 1,4,8 and 16MHz.
The other table lists different typical currents for different LPMs (tha tmeans core off, oscillators on/of depending on LPM). (independently of the core frequency).
The resulting efficiency can be calculated by multiplying these values with the proper duty cycle.
So if you properly use LPMs when you do not need the processor power, but then run the MSP at max when you need it, you can still be way more efficient that any competitor, even if the energy/MIPS ratio looks works at first.

Hi,

I'm working on MSP430F47187 for a year. I have used PIC24FJ256GA106 on a big project. I'm not sure for XLP but I can say PIC24 series will provide more rich peripherals to you, you can develop more robust hardware and cheaper comparing to similar MSP430. I didn't test XLP series. For power consumption in low power mode, MSP430 is better than normal PIC series, but thats it.

BP.

I have no experience with the PIC24, only the older ones. And it had a reason why we switched from them to the (also old) MSP430F1232. The PICS weren't more robust (in fact I toasted several, while I hadn't that many deaths with the MSPs, neither in absolute numbers, nor in relation to the number of devices built (way more with MSP now).

Also, I really disliked the PIC processor philosophy (engine, memory organization etc.) and much more. The only advantage was that they were cheap.

But neither price nor power consumption was a criteria when switching. Maybe the newer families have been designed with a better concept. But then, who wants to use a processor family where you have to start learning from scratch for the next bigger/better version? I prefer the MSP approach of widely being backwards compatible with the concepts. It really saves money if you don't have to re-invent the wheel anew each time.

Hi friend, I Have used PIC and both MSP. PIC is not robust. I have burnt many PIC pins. And I have many times connected VCC to GND and GND to VCC to MSP for half a minute, The controller Heats up, But for my experience, it never dies. It works properly again. I have connected more voltages than controller working on to MSP but it port pins never fails in my experience. In my opinion, MSP is more robust than PIC. And when the issue of documentation and well-written datasheets comes Microchips never stands in front of TI. TI has very high quality of documentation and user guides, which microchip don't have.