WIP: Reverse engineering the programming protocol for SiTime SiT8008

[lty1993]

Hello,

I got a SiTime Time Machine 2 and few blank SiT8008BI chips few days ago.

Since I cannot find any information about programming protocol online, I would like to reverse it myself.

It is been busy recently, the progress maybe slow.

==============================================================================
[Dec 8, 2018]

I managed to capture the programming waveform using a scope. (The model number I am programming to the blank device is SiT8008BI-21-XXN-8.000000)



The channel 1 is OE/ST/NC, channel 2 is OUT, channel 3 is VDD.

Here is hdf5 file I captured:
CH1.h5
CH2.h5
CH3.h5

==============================================================================

[Rerouter]

It really looks like you have a serious amount of crosstalk.

at a glance it looks almost like CH3 is used as a chip select,
CH2 / CH1 as data pair, possibly SPI,

Any chance you can clean up your crosstalk issue, and zoom in a bit on the waveform to get an idea of how the 3 pairs are transistioning,

My best guess is the OE is used as MOSI, and the OUT as MISO,

[lty1993]

It is really hard to probe the signal right without making a custom jig.



Anyway, I do not think the cross talk will affect much on interpreting the signal.

The OE is a three level signal as shown in next screenshot.



I posted the hdf5 files for each channel in the original post. You can use matlab, infiniium offline, or something similar to analyze them.

[Rerouter]

The image is a much easier thing to analyse than 1GB of files that I cannot open without specialized software,

So far, CH1, is a clock, don't know what the positive bits are, but clearly 800,000 Hz

To me its exactly like SPI, maybe a 1 wire variation on it, looks like update on falling clock, and read on rising.

[amyk]

I suspect that it's not just a frequency that's directly programmed into the device, but dividers, drive strengths, and other parameters that the programmer tunes in order to get the right frequency output.

[Bassman59]

Hello,

I got a SiTime Time Machine 2 and few blank SiT8008BI chips few days ago.

Since I cannot find any information about programming protocol online, I would like to reverse it myself.

Or you could see what Silicon Labs has to offer. Their timing products are well-documented and well-supported.

[lty1993]

Silicon Labs provides either fixed factory programmed oscillator or serial user programmable parts. However, they do not have any one time user programmable oscillator.

It is great to get few one time user programmable oscillator lay around for prototyping, and order factory pre-programmed parts in same form factor for production.