RF to DC Regulator Design

[florescent]

Hi. Ya'll

My chip of wireless sensors will be fed by RF signal (by inductive coils). I am now stuck at design for RF to DC regulator (2.5V VDD generation). My charge pump multiplies voltage and provides unregulated DC with small ripples. This DC varies from 0 to upto15v depending on the distance between the transponder and reader.
My CMOS only regulator is supposed to generate steady 2.5V VDD from this input voltage variation. (I have a problem to layout diode with the package I am using, so CMOS only)

Now, it's been almost a month for this design. I’ve decided no more time wasting, but ask to the experts.

Among many IEEE papers I’ve read regarding to this design, I attach this one that has given me the closest result so far.

Only problem (big problem) is the current consumption. This paper originally draws 100uA of the current. I’ve improved it to 70uA, but still too much for my biomedical application.

I am using “On Semiconductor 0.5um Tech (AMI06)” with Cadence.

Isn’t it possible to use less than 10uA with that big and old tech (0.5um)?

Would you introduce me your good ideas or IEEE papers?

 

[dick_freebird]

Since the input power is fundamentally limited I would
go with a shunt regulator, right after the rectifiers. A very
low power bandgap and shunt transistor. Or even just
a simple FET stack (say, 3*VT) maybe with a high ratio
current mirror at the bottom so you have usable drain
headroom.

Question is, how good and process-repeatable does the
"2.5V" have to be? We talking data acquisition, or just
running some low speed logic with conservative timing?

 

[florescent]

Dear dick_freebird

Thank you for your reply and great ideas

It will run quite good number of blocks (only couple of digital blocks, analog for all others)

I do have both data acquisition and slow, very slow (tens of kHz) logics. Only thing is oscillators that will generate data from sensors by load modulation. It is not necessary to have a super steady regulator. But the more steady, the better.