Clock Chip Comparison

[NZ0I]

Real-Time Clock Chips

The Rev X1 PCBs included two different real-time clock chips: the pricey Maxim DS3231 (~$8), and the NXP PCF2129 (~$3). Both devices include built-in high-accuracy temperature-compensated oscillators, and have similar features. Both devices come factory-calibrated, and include mechanisms for tweaking that calibration to improve clock accuracy.

The idea was to test the two clock devices side by side and see how they compare. At this point in time, no rigorous testing has been conducted, but both devices have been operating at room temperature, on two separate boards a few inches apart. Both had their times synchronized to NST time on April 28. On that date both clocks were within about 100 milliseconds of each other, and of NIST time. Now, 25 days later, the Maxim DS3231 maintains a time that is indistinguishable from NIST. The NXP PCF2129, by contrast, has drifted by about 2.5 seconds.

While I suspect that the cheaper NXP PCF2129 would benefit from being manually calibrated, that would add an additional calibration step that most builders would probably be willing to pay $5 to avoid. I had commented in a previous post the PCF2129 registers behave differently from what the NXP documentation indicates that they should. And the third strike: the pinout of the PCF2129 does not support selecting to have either a periodic clock signal, or an IRQ signal, come out of the same pin, as does the DS3231 - requiring that we use an additional logic chip if we wanted to implement that behavior in our application. (And I do want that behavior.)

At last report, Patrick indicated that his Maxim DS3231 experience has been similarly positive at room temperature, over a period of several months.

Based on this limited amount of testing and experience, I'm inclined to select the more expensive Maxim DS3231 over the NXP part. The slightly less expensive Maxim MEMS device (DS3231M) is probably worth examining as well, in order to shave a few $$ off the BOM cost. But for now I think we should target the DS3231 in our Digital Interface board design and transmitters, and use it in the first prototypes.

[Gerald Boyd]

Looked at the data sheet and the DS3231 looks like a good choice to me.

FYI just ordered a new scope for the home shop. It's a two channel 200 MHz 1 gig sample unit. When it comes will try some measurements of the drive levels to the mixer board. The scope can also decode serial interfaces like spi i2c can and rs232.
Will experiment with looking at the clock gen serial input and see how good the scope is at decoding. There was an advertisement in the recent QST on page 131 for a siglent model 1000x-e. I ordered the model 1202.

Sent from my iPad

> On May 23, 2017, at 8:07 AM, NZ0I <charles....@gmail.com> wrote:
>
> DS3231

[Charles Scharlau]

Hi Jerry,

Are you stateside now? If so, welcome home!

I've been focusing on everything but the Receiver board lately. My goal is to get everything set up so that you (and anyone else that is interested) can easily work with the RDP experimental and prototype boards using just an FTDI cable, a computer, a Digital Interface board, and a Receiver board.

You will be able to control the boards by typing commands into a terminal emulator, like Hyperterminal, or something more modern.

You will be able to bootload new software into the receiver processor from a PC running AVRDude.

You could build your own software if you want, or I can email updates to you to fix bugs and add features.

This is all mostly working right now. I've just got a few details to iron out with using the FTDI cable to automatically reset the processor when bootloading, and to implement a keyboard-friendly interface and command set.

Once it all seems to be rock solid, I would like to send you a working Digital Interface board with bootloader installed, and a partially-populated Receiver board. You should then be able to work with it to debug the hardware using your PC, and have no need for the Control Head.

How does all that sound?

-Charles

[Gerald Boyd]

Charles,

Yes I am now back in the US.

I would like to help test the board set. What is the model number for the FTDI cable? I can get one on order so it's ready. Also have putty on the PC. 

Should also have some experience using the new scope by then.

Jerry

Sent from my iPad

[Charles Scharlau]

I would suggest this cable:

https://www.adafruit.com/product/70

SparkFun sells a nearly identical product, at the same price, but theirs lacks the LEDs indicating data line transitions.

The prototype PCB boards will include the option of installing the FTDI FT232RL chip on the processor board, allowing you to simply plug the receiver into a PC using a standard USB-to-microUSB cable like what comes with most cell phones nowadays.

-Charles