Drawbacks of an actual CETTO board based on the PIC16F74 chip:


In general: bad layout of components on the board, usage of chip
            carriers with a high inductance.

In details:

1.  T(signal) << T(R11_C21) << T(C23_R20) - this mismatch leads to a signal
    distortion and amplitude degrading. The latter motivates a high gain and
    results in amplifying the noise (S/N worsening).

2.  Maximum rating violation for the U10A.3: negative voltage @ OPAMP's input
    with respect to ground (Vss).

3.  The Z-diode D2 doesn't protect input against overvoltage because it is
    connected to the OPAMP's virtual ground, where the AC and DC voltages
    are always zero ! Usage of Zener-diodes to shunt signals at the OPAMP
    outputs introduces a high integral non-linearity into amplification
    circuitry.

4.  Another source of a high non-linearity is a wrong operating point of
    ALL (except for the 1st one) OPAMPs - their signals start from ground
    which is a negative power rail (VSS). I made the same mistake for the
    last amplifier on the MCA-1000 prototype board

5.  10x difference in resistor values at the OPAMP inputs enhances the
    signal bias due to input offset currents. Usage of FET-input OPAMPs
    minimizes this effect, but in general this is a design shortcoming.

6.  The PMT charge integration takes place at the parasitic capacitance
    of an unbiased Z-diode D2 !

7.  Wrong concept of gain adjustment: amplifier has a high fixed gain and
    the saturated signals are being attenuated !!! It must be another way
    round: the input signal must be split and amplified. Adjustable feed-
    back circuitry would be even better: no signal and power losses @ all.

8.  Resistor in series with a sample-and-hold switch results in a variable
    baseline depending on a signal rate.

9.  The comparator based on the OPAMP U7B has no input-referred hysteresis
    while discriminating slowly rising signals - may lead to noisy switching.
    The width of the output pulses is not standardized.

10. Maximum rating violation: the peak detecting circuitry may drive +12V
    which are divided in proportion 3:4 = 9V and fed onto PIC16F74's pin
    AN1, whose input voltage should not exceed +5 Volts.

11. The common to all circuits 4x-potentiometer U6 makes routing of critical
    analog signals very long. The chip itself is unreliable (has a limited
    lifetime).

12. Maximum rating violation: the +5V linear voltage regulator U2 is too weak
    to supply the entire circuit. The dissipated power = (12 - 5) V x 60 mA
    is very close (may exceed) to the absolute limit of 500 mW. !!! Loosing
    of 500 mW just because !!!

13. the Z-diode D1 parallel to the input power rails has no current limiting,
    in case of breakdown it will shunt the circuit forever.

14. The HV voltage source has a L7_C12-filter whose resonance frequency is
    equal to the operating frequency of the DC-DC converter U8, it results
    in high oscillations on the PMT's output. !!! Any inductance needs
    a current through to work as a filter !!!

15. The ADC of the PIC16F74 microcontroller shows a very high (50% and
    above) differential non-linearity thus making this device unsuitable
    for spectroscopic measurements without special correction scheme.

16. The PMT baseboard supplies the anode with unfiltered HV.

17. Maximum rating violation: the 1kV decoupling capacitor between
    anode and load resistor on the baseboard does not comply with
    1.5kV PMT supply voltage.

18. COMPLETELY wrong grounding scheme: tying to ground an input power
    rail which must stay floating (it carries in the pick-up noise), no
    ground connection to the main board (electrical potential mismatch),
    inappropriate usage of filtering chokes: the L2 enhances common mode
    noise, the L6 enhances power surges of the DC-DC converter that are
    immediately visible with LED calibration pulses.


  ... and if it wasn't bad enough:


     -------------
    |             |
    |        (I/O)| -------100 uH choke (separate)
    |             |     |
    |    MAX485   |    100 nF                         2 x 2 = Yellow ?!
    |             |     |
    |        (I/O)| -------100 uH choke (separate)
    |_____________|


                            *     *     *

!!! It is simply too much for the board consisting of 3 OPAMPs and 1 PIC
device... How was it certified ?