Overview (december 2006 - march 2007): 3 month (4 - 1 for
holidays and extensive travelling):

1. studies of properties of the actual CETTO card
2. evaluation of different active filters
3. evaluation of the PIC18F97J60 comparator circuitry:
   large thermal drift that makes it unsuitable for the
   low-level discriminator (event trigger)
4. evaluation of the LED calibration circuitry exploiting a
   principle of the controlled breakdown in the pn-junction
   (this is a standard russian method with the AL-10X LEDs -
    it doesn't work in case of modern devices)
5. design of a new ultrafast LED calibration circuitry
6. design of a new amplifier and a peak detecting circuitry
7. implementation of a sliding scale linearization algorithm
   to get red of a differential non-linearity of the ADC (it is
   obligatory for any ADC of this type because choosing of a
   better chip doesn't guaranty this parameter to be good @ all)
8. layout of the printed circuit board, assembly of components
9. investigation of component tolerances and circuit stability
10. setting up and conducting spectral measurements

 - good rate for a side job (evenings, weekends) !




Poorely-defined task:

Any of the following aspects were specified
(highly important for design optimization):

1. purpose of instrument !!!
2. temperature conditions
3. supply voltage(s)
4. power consuption
   (critical for hand-held)
5. board dimensions
6. component base
   (price sensitive)
7. peripherials
8. deadline


Motivation of the PIC18F97J60 chip for signal processing:

This IC has standard to all Microchip controllers software
environment that makes its programming easier. It has among
other devices the largest memory, it has a possibility to
attach an external memory, it supports an Ethernet protocol
for the fast data logging. 


Adjustable PZ-compensation is not needed:

Amplifier's input impedance defines a signal decay time.
It is important to keep this time short to avoid pile-up
effects. For impedances in order of kOhms the decay time
ranges from handred nanoseconds to a few microseconds.
Reducing it further makes sense for the counting systems
to provide a reasonable throughput. In spectroscopy the
limiting factor is an ADC latency: for most of successive
approximation ADCs the digitization lasts typically 20 us.

As the PMT signal shows a weak dependance of its amplitude
on the load resistance, neither variable input impedance,
nor adjustable PZ-compensation circuitry are needed.

Considering of FADC + DSP:

To keep the error less than 0.5% with respect to an ideal
CR-RC shaper there should be at least six samples between
zero and the peak with the FADC resolution more than 7-bit.
For the shaping time constant of 250 ns the 8-bit FADC
with 25 MHz sampling rate would be sufficient. The good
candidate with a reasonable power consumption is AD9280.


Work summury (april 2007 - August 2007):

1. Concept development for the MCA-1000 board (electrical
   specification and mechanical layout, component selection)
2. Design of an electrical scheme of the MCA-1000 board
   based on the PIC18F97J60 chip (basic version)
3. Layout of the printed circuit board
4. assembly of components
5. MCA-1000 tests at CETTO
6. Changes to the design concept concerning additional
   memory for the PIC processor, including extra
   functionalities into scheme:

a.) Power over ethernet: it could operate from the standard
    power supply (+45V...+60V fully IEEE802.3af compliant)
    or from your own power supply (+8...+36V fused)

b.) Two optically-decoupled input channels for interrups
    (suitable for optical sensors mounted at the gate to
    let the processor know about the incoming transport)

c.) Optically decoupled measurement of the supply battery status

d.) Two summing PMT inputs with a signal weighting for both
    (very important for the light collection from the large
    crystal - it equalizes the PMT responses if the tubes
    share the HV supply)

e.) Differential reference voltage: Vref(-) = +0.5V, Vref(+) = +2.5V
    for the ADC provides a high linearity in the whole measurement
    range

f.) The PMT signal is split and amplified separately in two channels
    It allows for covering a wide energy range 0...2.5 MeV in one
    channel while keeping a high energy resolution of 1keV in the
    other channel 0...800keV to identify isotopes from Am to Th.

              !!! This solution could be patented !!!

g.) Embedded LCD display F-51320 types (Optrex, 0...+50 C. deg.) or
    better BTHQ-128064AVE (www.data-modul.com) specified for the
    industrial -20...+70 temperature range (compatible device).

h.) 2 Mb Extended RAM

i.) 256 kB EEPROM with the write protection (for constants)

j.) MMC flash card for data

7. Redesign of the printed circuit board (the layout density
   is extremely high), assembly of components

8. Comprehensive board tests
9. Writing the board description

10. Design of a compact version of the MCA-1000 board:
    new mechanical concept, new printed circuit board.
11. Design of a new baseboard for the Hamamatsu R4607-06 PMT