LMK04828: PLLatinum Sim component blank issue

Hi Tony,

Generally, if PLLatinum Sim calculates a value without indicating "design failure" then it can be considered a valid calculation. Forced components are usually reserved for cases where the device includes integrated loop filter components. The LMK04828 is a 4th order device with integrated 3rd and 4th order components, so it doesn't make sense to calculate results for VCO0/1 with anything less than 4th order. To modify the 1st and 2nd order components, use the Loop Filter Components box in the upper left panel, and do not use the forced component values. To modify the 3rd and 4th order components, use the Forced Component Values and recalculate the loop filter. You can technically use the Loop Filter Components box, and it should update immediately; but if you respecify design targets in Filter Parameters and click Calculate Loop Filter, it will overwrite the loop filter with the Forced Component Values.

As for "blanking out" forced component values... it seems that null user inputs are "corrected" to 0, despite null input being initially valid and being handled differently than 0. I'll make sure that the next version handles the Forced Component Values (and their interaction with integrated components in the Loop Filter Components box) a little more intelligently.

For now, a workaround in case you "soft-lock" PLLatinum Sim by tweaking a "forbidden" forced component value: go to Select Device, select a different device, then reselect the original device and reload; this should reinitialize the forced component values, including the nulls.

Regards,

Tony,

The device does have integrated components which should be programmed - see registers 16C and 16D. In PLLatinum Sim, the programmable values are already pre-loaded into the Forced Component Values for C3, C4, R3, and R4.

The loop filter merges C1 with C3 when R3 = 0, and C1 with both C3 and C4 when R3 = 0 and R4 = 0. This effectively renders the loop filter second order. However, 0Ω is not a valid value for the loop filter on PLL2 of LMK04828, which is why I said earlier that all calculations should be 4th order.

The integration of the LMK products into PLLatinum Sim is a recent development; most prior products were LMX devices with more flexible loop filter order, so it made sense to allow simulation of every filter order. To prevent confusion about loop filter design for LMK devices, our next update should remove invalid filter orders for parts with integrated components, and deactivate user input on integrated components in the Loop Filter Components box.

Regards,

Tony,

If I load the device and so as you say, I get the same warnings.  So if C3_Fixed and C4_Fixed are there, but R3 and R4 are not, then there is an issue.

Maybe in the next release, we can try to address this.

The forced components restricting loop bandwidth warning comes up and some ways to deal with this are (1) increase charge pump gain, (2) increase phase detector frequency, (3) reduce integrated component values, (4) reduce phase margin, (5) increase gamma optimization parameter.

The optimizer uses the Hook-Jeeves method and you are seeing something converge very close.  You can sometimes see small difference if you press the button repeatably.   In this situation, a wider loop bandwidth is needed, but it is being restricted by the forced components.  So the optimal solution is just at the point where the loop bandwidth is restricted by the forced components.  So it makes sense why sometimes you see the warning and sometimes not.

Regards,

Dean