0
\$\begingroup\$

I am looking for a lower power PIC18F microcontroller that supports SLEEP mode (or even DEEP SLEEP mode without RAM retention). These power saving modes are widely advertised by microchip. They write that their "XLP" technology can result in sleep current consumption in the few ten nA range.

Now my confusion starts:

I do have an PIC18F26Q84, advertised as"Low-Power, High-Performance Microcontroller with XLP Technology"

In the datasheet, page 3 (the overview section), I read:

Low Power Mode Features:
– Sleep: < 1 μA typical @ 3V

But then in the "Electrical Specification", I find under "Power Down Current":

I(PD) base @ 3V: Typ. 1.2 μA 
I(PD_WDT)    1.7 μA 

So if I understand this correctly, I end up with at least 1.2 + 1.7 = ~3 μA, given that some clock/WDT is required to wake up from sleep mode

Now we are clearly far above the few ten to hundred nA range. And I presume that are still 'best case' assumptions, with all peripherals turned off.

Do I misunderstand the specifications? Why do they seem contradictory?

Or should I use another part (I'd like to stay in the 18F family since I know these parts well).

\$\endgroup\$
5
  • \$\begingroup\$ Ten nA is going to be challenging. I haven’t looked into the PIC but the corresponding AVR got down to low hundreds of nA with wake on interrupt from external source. Have you considered one of the newer FPGAs instead? Lattice iCE40 runs on basically nothing. \$\endgroup\$
    – winny
    Jun 25, 2022 at 8:53
  • \$\begingroup\$ XLP is quite a broad term covering a wide range of performance. Look for the nanoWatt XLP devices such as the PIC24FJ64GA306 which advertises a 10nA Deep Sleep mode. \$\endgroup\$
    – Finbarr
    Jun 25, 2022 at 9:43
  • \$\begingroup\$ user5, I've used PIC18Fs for decades. Including the XLP variety and so-called nano-watt. I do love Microchip as a supplier of parts (both for hobby and for professional use) because of just how well they support a developer regardless of how small or big the problem. That said, they cannot hold a candle to some from the MSP430 family from TI (originally Germany, but that's another story.) It takes time to ratchet up a PIC18F from barely ticking away time to full speed and power is wasted. MSP430 is instantly going from zero to fast, no time wasted getting up and going. What are your goals? \$\endgroup\$
    – jonk
    Jun 25, 2022 at 16:56
  • \$\begingroup\$ First of all, I'd like to understand the specs that are given in the datasheet. They seem contradictory. Testing the device, I ended up with 30uA in sleep mode, but I have not yet turned off all peripherals. \$\endgroup\$
    – user52366
    Jul 12, 2022 at 13:33
  • \$\begingroup\$ @user52366 The 30uA sleep mode current you've got is likely given datasheet specs. Once you tweak the power control settings, you'll cut it down maybe by an order of magnitude, or a bit more. Nanoamps is maybe possible, but you'd need to buy a lot of chips, and select them. And keep them cool, too. And make sure stuff on the PCB doesn't leak. \$\endgroup\$ Aug 20, 2022 at 4:51

1 Answer 1

1
\$\begingroup\$

The simple answer is that the data sheet for the PIC18F26Q84(DS40002259B) does not contain any "official" specifications for low current behavior in sleep mode.

What Microchip has done for years is to publish data sheets that contain specifications that the part is known to meet. The power consumption specifications especially for low power operational and sleep modes depend a lot on the fabrication yields achievable in the real world. This is double speak gobbledygook that means that Microchip may publish these specifications when they have completed tweaking the process parameters to get the best yield and lowest power consumption that will result in most net profit for the device.

It's a good bet that Microchip knows the best case for power consumption but has not published it because they have low confidence that they can consistently produce parts that can meet it at an affordable cost.

So to keep you interested in designing with their controller Microchip puts a bullet item at the top of the data sheet that makes a claim for low power in sleep mode but leave out the specifications you can expect the part to meet.

You are correct to read the data sheet looking for actual specifications for parameters that are critical to your design goals. It's a big red flag when the specification for low power sleep current is hard to find or missing.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.