Using a $2 DS3231 RTC & AT24C32 EEprom from eBay


From the comments on our ‘Using DS1307 and DS3231 real-time clock modules with Arduino‘ post, Edward Mallon writes:

I’ve been putting the $1 DS3231 boards through their paces for quite a while now…and they have been performing well.

More details at Edward Mallon’s blog.

Via the comments.

Join the Conversation


  1. Ewww… a cheap Maxim IC part, something of an oxymoron. Ha ha ha… personally I couldn’t get myself to buy these dodgy stuff.

    A lot of issues seemed to be due to using modules, modules, modules… Note that it’s easy to hit microamp load targets for hobby circuits if you design them yourself.

    The data platform looks nice… and that’s my problem with it, uh, it’s round. A lot of times with such battery-powered IoT things, instead of sticking to CR2032s, using 2xAA batteries is a great choice. All your battery load, power and capacity headaches will go away. Often, when size is not a critical issue, just use a box sized for 2xAA plus all the boards and sensors.

    I would have used normal RTC and some means of adding GPS time stamps to the log. For examplw, with 2xAA, just add some cheap wireless communication. Only when data is high value, e.g. science data, then a better RTC may be desirable, then you’d need calibration procedures too, and of course one would never use dodgy chips to generate serious science data.

  2. Yeah, all those ueber complex modules, I hate them, too. They have all this stuff on them, like, say, resistors or what, like, capacitors! Or even those other things, like, what, LED or such.

    Look at the picture. I already count about 7 or such pieces I never know what they are and such hardship to understand this stuff. Where do all the wires go? And blue!

    And all these letters. What could 32k mean? Bra size? Sometimes there are even magnets involved. Who knows how magnets work? Those clowns even don’t, like me?

    But anyway, I’m off to get my amplifier working to boost the GPS signal down in my cave and under water. You know, related to that article and such like what?

  3. Quick Thought…

    Note: In this post I have NO evidence of fake DS3231 chips at all. But I do raise a question…

    Q. Is it possible that these cheap DS3231 RTC modules (BoB’s) from China have Fake DS3231 chips on them?

    Someone (not me right now, I don’t have one) needs to test these Chinese DS3231 RTC Break-out-Boards (BoB’s) against known real DS3231 parts. Maybe someone on the Time-Nuts Group has done this already? Drill here:

    Keep in-mind the Real Maxim/Dallas 32XX RTC parts have some sort of on-die temperature-compensated reference (MEMS?) So if the chips on the Chinese BoB’s are fake, the difference in time-keeping should show up pretty soon.

    Check this thread (not conclusive, but worth a read):

    1. It has a 32.768KHz crystal, it’s the oscillator circuitry that gives it the accuracy. The datasheet talks about it, nothing magical, it checks temp regularly and uses a lookup table to adjust parameters.
      Could be the Maxim originals are expensive due to calibration costs (multiple point calibration?). Could be the clones good around 25C only, so lab tests will be fine. But the boards are so cheap that it’s easier to believe the board makers got them for next to nothing…
      Pity I could not find definitive marking documentation for Maxim chips. All MCHP datasheets have ’em. These dodgy parts are dirt cheap, so possibly they are not going to duplicate the lasered marking 100% exactly. Or see if the chip really does temp calibration every 64 seconds and is accurate enough within the spec’ed temp range…

    2. About calibration, probably calibration is done after encapsulation with the crystal, it wouldn’t make sense otherwise. So we may have rejects that are fully functional but out-of-spec over some of the temp range. They may be in spec at the usual 25C min error design point. Someone should decap one to see once and for all… clone or not?

  4. I have no illusions about these cheap RTCs, and up to this point I have been doing drift checks on a significant number of units every four months or so. After correcting for the compile & upload time, they generally seem to creep forward about 15 seconds over a typical deployment in reasonably stable temp environments at about 25C. But even the units that get subjected to high temps (up to 60C) in the tropical sun have not suffered much more drift that those in the caves. Up to this point code updates forced the four month cycle, but I deliberately left the RTCs alone on the last round, and some will be running till mid 2016, giving me a chance to do a 9 month drift check on many different builds with these DS3231s.

    @ KH: I understand the concerns people have about using cheap modules from eBay, but many seem to be missing a key point of my project. I am trying to see if I can reach the ‘good enough’ point using cheap components & housing materials because that enables many things that just can not be done with professional equipment (which we also use). We would never risk deploying a Hydrolab somewhere that it would be damaged by a storm, or more often, simply be stolen because of it looks like it has value. Most of the units that I build are for these kind of exploratory tattle-tale deployments, with no pretense of being commercial instruments, although sometimes we come surprisingly close. Not to mention the undue attention you get trying to travel with equipment worth thousands of dollars, whereas the last time we had a suitcase full of my little DIY units inspected, it was easy to convince them that the whole lot had no commercial value, because they honestly don’t.

    So it actually helps us that the Pearls look like a high school science project made from modules and jumper wires, and I think this rough appearance has also encouraged others to start building them on their own. As soon as you go to a custom PCB, class A parts, etc., you cross a line that discourages beginners (…like me…) from trying their hand.

    1. No worries, just shooting the breeze. :-) I’m a dinosaur, when folks talk about how they did all the contortions to fit this or that module into their projects, hmmm, it’s just good eye-rolling exercise for me…
      I choose not to enrich makers of _dodgy_ boards, however little they earn.
      Plenty of ways to skin the cat. Dinosaurs like me are often cheapskates, custom PCBs or class A parts never crossed my mind.

Leave a comment

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.