SOLVED Deep Sleep comsumption - Sparkfun Pro Mini Vs Clone Pro Mini


I have two arduino pro mini (3.3V - 8MHz) :

  • A sparkfun original. I have remove LDO & pwr LED
  • A chinese clone. I have also remove LDO & pwr LED

I’m trying to understand the difference in consumption between my two modules in deep sleep mode :

  • SF : 4.5 µA
  • Chinese : 170 µA

To measure the current I am using a tinyCurrent module with a stabilized power supply (TTI QL355P).

My code :

#include <LowPower.h>
void setup()

void loop() 
  LowPower.powerDown(SLEEP_FOREVER, ADC_OFF, BOD_OFF);

My question :
In your opinion what are the differences between the two modules that can cause this difference in consumption?



The Chinese clones are often built with cheaper components and NOT to the original specification.
Those cost cutting items can often lead to odd results in some instances and you may have found one of them.

I carried out a new test: I unsoldered all the components on the card except: the microcontroller, the quartz and the capacity in series on DTR.

Result: 165µA.

Can the microcontroller be a counterfeit?

If the markings on the chip match the ones on a real AVR/
It may indeed be a clone of a clone but there is no sure fire way to check that.

The Atmega328P microcontroller on the Chinese clone is a counterfeit. I had the same problem, and asked Kevin Darrah to help figure out what's going on. He has released three videos on the subject, the second of which presents the results of a forensic analysis of the suspect chip with x-ray photos and decapping the chip to reveal the die. Here are the videos:

We don't know whether the counterfeit chips don't meet datasheet specs in other ways than sleep current. I think this is potentially a big deal. It's not just Pro Mini clones. Uno and Nano clones also use the 328P, so they could have counterfeits too.

The third video has what appears to be a way to determine whether your 328P is genuine or a counterfeit. This can be done by running a sketch Kevin wrote, and doesn't require testing for sleep current.

Thanks for the links.

Indeed, my microntroller seems to be a counterfeit.

I performed the test suggested by Kevin Darrah (The sketch is available here: Arduino atmega328p unique id/serial number · GitHub )

Results :
Genuine Atmega :

1E 9C 95 FF F C6 FF 26 
FF 9 FF 17 FF FF 55 34 
37 31 35 30 FF 4 8 20 
17 1 12 6 13 6 FF FF

Clone :

1E BA 95 FF F FF FF 26 

This shows that it is surely a counterfeit.

Bottom line: if you want to work on low power, don't use a Chinese clone.

Note: I bought this clone on Aliexpress. The seller is "win win", the product page: Carte Pro Mini 328 ATMEGA328 ATMEGA328P AU pour Arduino, 3,3 V/8 MHz, 5 V/16 MHz, | AliExpress

Naming the seller is pointless as there are so many.

The lesson you really learned is that if you want genuine you know where to go the first time. 8)

The sad thing is that it seems for all these years we've been able to buy clones with genuine Atmel processors, or at least it has seemed that way, but now that may no longer be true. The regulators may not have been all that great, but you can deal with that. I guess if you're handy at rework, you could still buy clone Unos, Nanos and Pro Minis, and buy genuine chips from Digikey, and replace the processors on the clones if they test as counterfeits. There might even be a market for MCU-less clone boards.

Anyway, Arduino made their designs open source, so there's nothing illegal about the clones, and they sure do save a bunch of money. But the appearance of counterfeit processors definitely complicates things. They may work well enough for most things, but if something doesn't work right, is it your fault, or the chip's fault?

So far, only chips with date codes in 2019 have been found to be counterfeits. So new old stock would be a good thing to find. I even bought a couple used Pro Minis on eBay with 2015 date codes. They have genuine Atmel 328Ps. But I don't think we know about 2020 date codes yet.

Guillaume, what's the date code on yours? It would be the bottom line of markings, the first 4 digits. "1942" would be the 42nd week of 2019.

Date code on counterfeit Atmega :

  • 1902
  • 1903
  • 1947