I am having some problems getting the LTC3884 to measure the output current accurately. I am using the part to provide power to the core voltage of a FPGA and it is connected very similar to the DC2084A demo board but with no LTC3874 for extra phases. I changed out the inductor to a PA3790.201 and ran the LT Power CAD to adjust the series R on the Isense+ pin to be 634ohms. I used the LT Power play with the DC1613A adaptor to change the registers settings in the LTC3884 and LT Power play is also used to display the output current. After making all of the adjustments, i get a much higher current reading than expected on each of the 2 phases on the LTC3884. In my test setup, i have the LTC3884 powering a resistor and the output current should be a total of 14A to this resistive load. In LT Power play, it is showing me a measured current of 11A on Phase 0 and 12A on phase 1. Is there any other register settings that i need to make adjustments to?
I am currently changing the following registers:
Iout_Cal_Gain = 0.29mohm (DCR of my inductor)
Mfr_Temp_1_Gain = 0.991 (As per the datasheet's recommendation)
I have also placed this resistive load on the demo board and its results are in line with the expected value. Also, i noticed something interesting with the Demo board's settings. The DCR of the inductor used on the demo board is 0.32mohm (which is the default value in the LTC3884) but when LT Power play is loaded up, i read a value of 0.34mohm in the Iout_Cal_Gain register. I'm not sure why a higher DCR value is used in the register settings for the demo board.
Please make sure the MFR_IOUT_CAL_GAIN_TC is 3900ppm/C in LTpowerPlay. Most likely the issue is caused by the pcb layout of the DCR sensing and/or inductor soldering. Kelvin sensing to inductor pads is…
Did you verify that this section of the datasheet remains true?
If the RC values are chosen such that the RC time constant matches the inductor time constant (L/DCR, where DCR is the inductor series resistance), the resultant voltage appearing across the capacitor equals the voltage across the inductor series resistance (VDCR) and thus represents the current flowing through the inductor.
Yes i did verify this portion of the datasheet and adjustet the R to 634ohms.
Please make sure the MFR_IOUT_CAL_GAIN_TC is 3900ppm/C in LTpowerPlay. Most likely the issue is caused by the pcb layout of the DCR sensing and/or inductor soldering. Kelvin sensing to inductor pads is required. To verify this, you can use a voltage meter to measure the voltage across the cap of RC for the DCR sensing. If this voltage divided by 0.29mohm DCR gives similar current (11-12A) as LTpowerPlay reading which proves my analysis.
On our standard demo board DC2084A, we use 0.34mohm (0.32mohm DCR inductor) to account DCR plus soldering resistance (~0.2mohm).
Thanks for the suggestion, i am out of town right now and will test this next week when i get back. I will keep you posted on what i find,
I checked the voltage across the capacitor and it seems ok. I have attached a picture of my layout where these traces are being routed, can you have a look and let me know if this looks ok? I have circled the current sense lines from each phase in green. Thanks