CW305 Artix7 xca100 Power Consumption


#1

Dear All:
Hope you are doing well and in good health.

I have the CW305 Artix7 xca100, with 0.5 Ohm shunt resistor.

I had this board because I am intrested to measure the power consumption of the digital design implemented on the FPGA.

My question is: first of all I am calculating the power consumed by the Artix7 using This formula:
Vshunt=(Vsupply-Vcore)=Voltage difference accross the shunt resistor.
Ishunt=Vshunt/Rshunt=Vshunt/0.5 ==> Power=Vcore*Ishunt. Noting that Vcore is just the terminal connected to the FPGA and Vsupply is the terminal connected to the power supply. Is this the correct way to calculate the power consumption?

In addition May I know please the exact amount of the Capacitance of the FPGA to get the energy using this formula E=0.5C*(Vsupply-Vcore)^2.???

Many Thanks,
Yehya.


#2

Hi Yehya,

Are you looking to measure the overall power consumption (including the static power consumption), or just the dynamic power consumption (for CPA attack purposes)? If it’s the second, you don’t really need to factor Vsupply or the shunt resistance, since the voltage on the high side of the shunt will be effectively constant if there’s enough capacitance there. That means you can consider the power to be proportional to Vcore^2. Otherwise, your formula should be correct.

Regarding your second question, what are you trying to measure the energy of? If you’re trying to measure how much energy the FPGA has consumed, you’re better off sampling the total power consumed and numerically integrating it.

Alex


#3

Hi Alex:

Thanks for your reply:

I am confused concerning the power consumption measurement on the CW305, XCA100T with shunt resistor 0.5Ohm.

In fact What I am trying to do is charecterizing the power consumption of a digital design implemented on the Artix7.

I send data and clock to the digital design. and I connect X2(Vsupply)and X3(Vcore) accross the shunt probes to measure the voltage and to do the difference to get the power consumption as I explained before: Ishunt=Vshunt/Rshunt=Vshunt/0.5 ==> Power=Vcore*Ishunt.

But in fact the total power consumption is not changing, so for example static power consumption is arround 40mW
and this value we got it when data=off and clock=off so the total power consumption is equale to the static power consumption.
and when we start sending data and clock to let the design on the FPGA the total power consumption (static+dynamic)~static so dynamic is too low and this is not logical and is not normal!
Static power consumption is arround 40mW.
What do ypu think?

and If I want to use the LNA onboard on X4 which is the Vcore if I am measuring this using oschilloscope normally I am measuring the Vcore so how can I get the power consumption or the dynamic power consumption from the measured Vcore after the +20db LNA?

Noting that I using Data aquisition board to measure the voltage difference accross the shunt resistor.

Many Thanks,
Yehya.


#4

Hi Yehya,

I’m a little confused about your setup. What do you mean by Data acquisition board? Are you using something like an OpenADC to measure the power consumption?

Alex


#5

I am using Labview with MyDAQ. Yes it is an external solution to measure the voltage difference accross the shunt.


#6

Hello Alex:

and If I want to use the LNA onboard on X4 which is the Vcore if I am measuring this using oschilloscope normally I am measuring the Vcore so how can I get the power consumption or the dynamic power consumption from the measured Vcore after the +20db LNA?


#7

Hi Yehya,

I haven’t used MyDAQ before, but I’ll do my best to help with that. Looking at the user guide, I’d guess you’ll want to have the AGND pin from the screw terminal connected to GND on the board and use AI0 for Vcore and AI1 for Vsupply? Most of the stuff measuring consumption only seems to use AC stuff for ramping between low power and high power states, at least for microcontrollers. In any case, you should be seeing much higher power consumption when getting a clock signal vs not getting one. Are you able to verify that the device isn’t being clocked when you think it isn’t?

Regarding the LNA, keep in mind it’s AC coupled, so it’ll only be useful for measuring higher frequency stuff off of Vcore. I’m not sure what the low frequency cutoff is, but if it’s anything like the OpenADC’s, MyDAQ probably won’t get much useful stuff off of it. I doubt you’ll get much useful information off of stuff at that frequency anyways.

I haven’t used MyDAQ before, so forgive me if any of the above stuff is obvious.

Alex