LMK05318: LMK05318: 1PPS from gps for create a 640 Hz clock output

An UPDATE:

i can lock a 10Mhz out and 1Khz out if i use a REF1 at 100Hz.

Why not work at 1Hz ref1 also?

Hi Luca,

The 1pps input won't work with the default 48.0048MHz XO on the board. You'll need to change the XO to TCXO or OCXO (for example, < 4.6ppm for 12.8 MHz XO). Note that this 4.6ppm is with regards to the DPLL reference. So your DPLL reference must also be accurate.

For sanity check, you can first frequency synchronize the XO and the reference (use external XO port), see if everything locks. After that, you can use some real TCXO or OCXOs.

Another thing is that we've developed a much cleaner wizard for the LMK05318B. You can download the latest Ticspro, find LMK05318B, and in the first wizard page choose "backward compatible" and use that wizard to generate LMK05318 configurations.

Regards,
Hao

In fact, I abandoned the lock tests with the pps.
In no way, even with an ocxo (very precise) at 10Mhz did I manage to make the dev board work.

At the moment I'm trying to generate a square wave at 1250 Hz, using the board's default xo at 48.0048MHz.
The out is OUT7P, in cmos + -.
The reference is REF1 isGPS_neo7_48Mhz_500Khz_1250Hz.tcs configured in single ended, with a 500Kz square wave coming from a GPS (about 20nsec of jitter).
The board lock the reference, but often it loses synchronism and it is necessary to reset the board with the button.
I attach the best configuration to the moment I used, but I'm still not satisfied with the result.
Could someone please send me a working configuration file for the dev-board?
1) xo 48.0048MHz
2) REF1 500khz from 20nsec (peak to peak) jitter gps
2) OUT7 at 1250 Hz cmos + -

I attach the photo showing the lock between the 500Khz (red trace) and the 1250 Hz (yellow trace).
The problem that the configuration is not stable.
Synchronism is often lost and hardware reset is required.

Hi Hao!

The file you sent me doesn't work in the dev-board.
During this time I made many tests and in the end I managed to get what I wanted.
I have 1 GPS with 100pps output, and I get a synchronized 400Hz output on out7.
1) 100HZ lvttl (0V -2.7V on 50 ohm) on REF1
2) Default Xo 48.000xx Mhz (Xo set as CMOS)
3) OUT7 400Hz cmos + - aligned with ref1 (rising edge)

I attach the file that works well for me if someone needs it.

I noticed an anomaly on the software though:
The value of the time calculated in the delay window by the gui is twice that of what I measure with an oscilloscope.
If the gui tells me 300nec I measure 150nsec

GPS_neo7_48Mhz_100hzpps_400Hz.tcs

yes, I confirm that I used the revB of the gui.
Have you tried the configuration you sent me on the dev board and confirm that it works?

Hi Luca,

No I didn't try it on a board. Let me know if you still need me to do so.

Regards,
Hao

No it is not neccessary.

have you news for the delay calculation?

Hi Luca,

Sorry I haven't been able to look into this. Let me try and check it by next Tuesday.

By the way, I did find that LMK05318B's "backward compatible" mode has a software bug so that DPLL_LOFL cannot be cleared. We'll fix that asap. For now just use the LMK05318 GUI.

Regards,
Hao