What's the smallest input value with the vth=0.25vref in sub-ADC?

[lhlbluesky]

my comparator is the one that has built-in threshold,it have two differential input and two differential reference ,if my adc is 1.5 bit per stage,and total bits are 10,vdd=1.8,then when my vth=0.25vref,how is my vref,1v or 500mv(my input 0.4-1.4,vcm=0.9)?and what's the smallest input value with the vth=0.25vref?

1mv or 125mv or more?

i'm confused about this ,please help me .

thanks !!!

 

[Btrend]

Re: about sub-adc

usually the notation are :
Full Scale (FS)=2Vref, Vth=Vref/4.
but in fully differential implementation, all quantites should be halved.
1V=2*(2Vref) ==> Vref=0.25V , so ur refp=Vcm+0.25, refn=Vcm-0.25
Vth=0.25/4 ==> Vth =0.0625V
also,
Vcm=(0.4+1.4)/2=0.9V

what's the smallest input value with the vth=0.25vref?

it is 1LSB=FS/2^10, but again if u want to know the amplitude of this input , it should be halved in differential implementation too, i.e. the smallest amplitude of each branch of differential input is FS/2^11=4.9mV

 

[n1cm0c]

Re: about sub-adc

I usually see Full Scale (FS) equal to the absolute value of Vref...
In fully differential the sign inversion is trivial, just exchange pos and neg inputs,
but the signal range is always the same, equal to Vref (not twice Vref).
In a balanced fully differential signal the FS signal is defined py pos=Vref/2+vcm
and neg=-Vref/2+Vcm (for a +Vref input, pos-neg=+Vref).

a +Vref/4 input is then pos=Vcm+Vref/8 and neg=Vcm-Vref/8 .
Reverse pos and neg and you'll have the -Vref/4 diff input signal

Added after 2 hours 15 minutes:

In your case Vref=1V , Vcm is your choice (usually mid way between supply rails, so 0.9V here), you say 10-bit resolution (9-bit magnitude plus sign bit?) .
your LSB is either Vref/2^9 or Vref/2^10, depending on what you mean by 10-bit, and the signal itself is pos=Vcm+LSB/2 neg=Vcm-LSB/2 for a +LSB input.

 

[Btrend]

Re: about sub-adc

In a balanced fully differential signal the FS signal is defined py pos=Vref/2+vcm
and neg=-Vref/2+Vcm (for a +Vref input, pos-neg=+Vref).

how about neg-pos=-Vref ?
in fully differential implementation, the FS=[neg-pos, pos-neg]=[-Vref,+Vref] =2Vref ?

 

[n1cm0c]

Re: about sub-adc

You can define FS as the maximum absolute excursion (Vref) or the maximum excursion (as you do), 2*Vref. I prefer the max abs swing (Vref), since the circuit effectively handles the same amplitude signals, no matter if you are amplifying
+Vref and -Vref (the positive and negativ signs are only polarity inversions).

Because of my definition (absolute voltage swing) I exclude the sign bit from the N when I compute the LSB value, quantization noise, etc. It's a matter of taste, really.