PFD delay for eliminating the charge pump dead zone

[Analog_starter]

pfd dead zone

Hi all,

How long is it proper for the delay time? And how to simulate the charge pump dead zone?

Thanks & Best Regards
Analog_starter

 

[huanchou]

.option post=1

Try to simulate the PFD+Charge pump then plot the pahse error (X-axis) Vs. (delta charge; ΔQ of charge pump output (Y-axis)). According to the plot to add the delay until the dead zone is eliminated.

 

[Analog_starter]

zero deadzone pfd

Hi huanchou,

The question is how I can get the phase error as X-axis? Which simulator do you use? Spectre or Hspice?

Thanks & Best Regards
Analog_starter

 

[huanchou]

the function of delay charge

Hi, I mean the HSPICE simulator.
Try to use the two pulse function as the PFD input.
Let one of the pulse function has the delay parameter say 'delay+phase error'
For example:

V1 in1 ground pulse( HI LO delay rising falling duty period)
V2 in2 ground pulse( HI LO 'delay+phase error' rising falling duty period)

Then run the tran. simulation and sweep the "phase error" this parameter.

 

[Analog_starter]

charge pump dead zone

Hi huanchou,

Thanks a lot!

Thanks & Best Regards
Analog_starter

Added after 7 minutes:

BTW, which parameter of charge pump stand for the delta charge (Y-axis)? How can get it?

Thanks & Best Regards
Analog_starter

 

[huanchou]

spectre special fuction pfd deadzone

Try to use the LX0() to plot the charge in the cap.
Don't forgot to put the cap in the charge pump output.

 

[Analog_starter]

what is a pfd pump

Hi huanchou,

I have tried it, but can not get the .tran sweep plot waveform that "LX0(CCtest)" as the Y-axis VS. "p_error" as the X-axis. Could you show me the error?

And I think it should be the ΔQ VS. p_error around the phase error time piont, not the "charge stored in capacitor (LX0)", right? How can you catch the ΔQ value in a transient simulation around the phase error? I am really puzzled by it.

Thanks a lot !!!

My sim file as below:

***module for testing dead zone***
.inc 'PFD_CP.cdl'
.param ups=1.8
.param p_error=14n
.TEMP 25.0000
vdd vdda 0 ups
vss gnda 0 0
vdd2 vdd 0 ups
vss2 gnd 0 0
vsleep sleep 0 0
I_CPN I_CPN 0 -50u
I_CPP I_CPP 0 50u
vref ref 0 pulse ups 0 10n 0.1n 0.1n 25n 50n
vfb fb 0 pulse ups 0 p_error 0.1n 0.1n 25n 50n
************************************************************
.OPTION post=1 PROBE=0 accurate=1
.lib 'XXXX.lib' tt
************************************************************
.tran 0.1n 100n
+sweep p_error 9n 11n 0.1n
.print tran LX0(CCtest)
.end



Thanks & Best Regards
Analog_starter

 

[Analog_starter]

charge pump deadzone

Can anybody give me some suggestions?

Thanks & Best Regards
Analog_starter

 

[cirand]

dead zone pfd

just measure the voltage on the capacitance and multiplied by C, that is the Q .
you should add the .meas statemant in hspice list:
.mear tran vcpo avg v(o) from=XXn to=XXn.
.meas charge param='vcpo*C'
where v(o) is the voltage on the capacitance .
the plot is "charge stored in capacitor VS phase error" ,not "ΔQ VS. p_error " you can the value of ΔQ / p_error by measureing the curve.

 

[dumbfrog2]

dead zone pfd

also do a post layout extracted simulation