LMK04610: Power consumption and Jitter in Bypass mode

Hi Kevin,

I will take this into lab and measure to see if I measure values close to yours.

Thanks,

Michael

Hi Michael, any updates on those power measurements? Thanks.

Hi Kevin,

Sorry for the delay, I have been out of office on business travel. In order to give you the most accurate readings possible, I will need to know your clock tree. Can you provide your input and output frequencies? Or could you share the configuration file you are using?

Thanks,

Michael

PLL1 and PLL2 are disabled (Bypass mode). SYSREF has a divide of 100.

Hi Kevin,

Thank you for providing these values. I will take them into lab and provide my measurements ASAP.

Thanks,

Michael

Hi Michael, any information yet? Thank you.

Hi Kevin,

I am sorry for the delay. We recently moved labs and are still locating all of our parts and components. Our logs say an LMK04610 was used recently, but I am unable to locate it. Please expect to hear by early next week.

Thanks,

Michael

Hi Kevin,

I managed to locate our LMK0461x (I am technically using a 4616, which is the same save for the number of outputs) and I was able to test your configuration. It would be a bit more precise if you were able to send me your configuration file (either as a hex file or as a .tcs file), but with the information you gave I found that the LMK04616 draws 280 mA of current. This seems like quite a large difference until you consider that this is what is being consumed by the EVM. It is difficult to count all of the resistors to GND that the device has, and those contribute to the current consumption, along with other active and passive components on board (including multiple LDOs and a VCXO). 

If we are solely referring to the LMK04616 chip, driving its outputs with no output terminations, I would say that the estimate of 84.96 mW is likely a close one. But when factoring in the entire system, the power consumption is going to be exceed 84.96 mW.

Thanks,

Michael

If you are using LMK04616EVM, I see in the user guide that the headers shown in Figure 23 could be used to source power externally thereby bypassing the onboard LDOs and Switching regulators to get accurate current measurements for the 1.8V and 3.3V rails. Nevertheless its hard to image the losses being what you say if using the onboard DC-DC converters which should be >88% efficient. If using LDOs you should be able to get at least 90% by feeding them a voltage above the dropout threshold min. All the resistors I see would still only draw hundreds of microamps. Only the LEDs might draw a few milliamps. The oscillator current could be removed from the total. Can you try and get a better measurement of the chip? I am also very interested in validating the additive jitter in Bypass mode. The datasheet only gives output jitter when using the internal VCO and assuming the OSC reference they give in the datasheet. But since I will not be using the VCO, I am only interested in additive jitter in bypass buffer mode. Thanks.

Hi Kevin,

I will see what I can do. Expect to hear by the end of the week.

Thanks,

Michael

Hi Kevin,

Just wanted to update you, I was limited on BW this week, but measuring the power consumption by the LMK04610 is the top priority for me next week. Expect to hear by the middle of next week, my apologies.

Thanks,

Michael

Thanks for taking the measurement, Michael. I was really hoping the overall power would be less. Do you happen to know what the additive Jitter was in bypass mode? 

Hi Kevin,

The additive jitter in bypass mode with the integration BW from 12 kHz to 20 MHz was measured as 85.31 fs RMS with an input of 200 MHz.

Thanks,

Michael