Notes on the Analog Board for the Height Sensor in the Advanced Dicing Technology 7122 Dicing Saw --------------------------------------------------- Rev. Date: 30-Apr-2021 These notes are from the repair of the Dicing Saw in April 2021. The problem was that the machine failed during the Height Test, i.e. the measurement of the blade diameter. As found there were 3 Analog Boards for the Height Sensor and they all had the same problem, i.e. the +12V and -12V power rails on the card appeared to be shorted together by about 150 Ohms. The 100 mA to 200 mA load from this short circuit caused the +-12V power supply to collapse and that blocked the dicing saw from working at all. On these broken Analog Boards there was no unusual connection between the +-12V power rails and ground. On all 3 boards, the short between the +12V and -12V power rails was traced to the first OpAmp on the card - IC1. Refer to the schematics of the Analog Board in the dicing saw Parts Book. Document page 118 (pdf viewer page 125) shows the original version of the Analog Board which includes a 3 pole low pass filter. The "new" version of the Analog Board shown on document page 119 (pdf viewer page 126) is the same except that the low pass filter has been disabled by not installing 3 capacitors on the printed circuit board. It appears that the copper connections on the original and new versions of the circuit board are the same. In the new version of the Analog Board, which is used in our saw, the circuit consists of (input to output): IC1 - Current to Voltage Converter with an adjustable gain of 1 uA --> 47 mV minimum gain up to 1 uA --> 547 mV maximum gain controlled by VR1 and an adjustable offset zero provided by VR2. The 100 pFd capacitor in the feedback provides a roll-off in the range 2.9 kHz to 34 kHz depending on the setting of the gain control. IC2A - Inverting voltage gain stage with a gain of 20. IC2B - Does nothing - voltage follower IC3A - Is a strange absolute value circuit IC3B - Is a voltage follower with a diode in its output so that the analog output signal from this board can only pull-up. There is no source of pull-down current. Test Point 1 is the output signal from the board. Test Point 2 is the output of the voltage gain stage IC2A. Test Point 3 is the Ground of the Analog Board. I do not know why they included an absolute value circuit. The input to the Current to Voltage converter is the Anode of the Photo-Diode in the Height Sensor. Details of the Height Sensor are shown on document page 64 pdf page 71. The Analog Board includes a current source to drive the LED in the height sensor. Its current output is: (control signal input voltage - 0.7 V) divided by 249 Ohms. So for a 5V control input it makes about 17.3 mA output. Note that in the schematic for both the original and new version of the Analog Board that IC1 is listed as an LF356 and that the center wiper contact of the offset zero pot for this stage is connected to the positive power rail +12V. This connection the center wiper on the offset zero pot matches what is shown in the datasheet for the LF356 on page 31. The full schematic of the LF356 is shown on page 15 of its datasheet. Why is there a Vcc to Vee power rail short inside IC1 on all 3 Analog Boards that we have ? - It could be caused by a spike on the +-12V power supply provided to the Analog Board but I don't think this is likely because: the Analog Board includes a fancy CLC input filter on its power rails, and because IC2 and IC3 (which use the same power rails as IC1) do not appear to be damaged on any of the 3 Analog Boards that we have. - Perhaps there was a big spike on the input signal to the Analog Board which runs directly to pin 2 the inverting input to OpAmp IC1. I don't this this is likely because when checked with an Ohm meter the gates of the input P channel junctions FETs on IC1 look intact. - On all 3 of the Analog Board that we have, the actual part installed at IC1 is a TL071 not a LF356. A TL071 is a more modern OpAmp than the old LF356 which may not even be in production any more. The datasheet for the TL071 shows its offset zero circuit on page 37. Note that the center wiper contact is connected to the negative Vee power rail -12V. The full schematic of the TL071 is shown on page 38 of its datasheet. From that schematic it is clear that if the offset zero control circuit for a LF356 is used with a TL071 OpAmp (as is the case on the Analog Boards that we have) then turning the offset zero control pot to either its CW or CCW limit will destroy the TL071. Destruction of the TL071 happens because the full 24V power supply is put across one of the 1080 Ohm emitter resistors in the OpAmp's input circuit. This results in over 500 mW dissipation in that resistor. I replaced the shorted TL071 at location IC1 of the Analog Board with an LF411 (the only similar OpAmp that I had on hand). At the same time I had to connect the center wiper contact of the offset zero control to the negative Vee -12V power rail to match the zero control circuit needed by a LF411 (or by a TL071). See page 1 of the LF411 datasheet for its offset zero circuit and page 11 for its full circuit. The Dicing Saw would now run, i.e. its +-12 Volt power rails were within limits so the machine did not automatically shut down, but it would not complete a Height Test without errors. I believe that it pasted the Height Test the first time that we powered it up and after that the saw reliably failed the Height test. I assume that the analog output signal from the Analog Board goes to an ADC someplace in the controller for the dicing saw. I assume that the blade diameter measuring routine must consist of: - With the LED OFF and the blade out of the optical path measure the output of the photo-diode. This is expected to be some low value - perhaps 200 mV output from the Analog Board or something like that. - With the LED ON and the blade out of the optical path measure the output of the photo-diode. This is expected to be some high value near full scale - perhaps an 8 Volts output from the Analog Board or something like that. - Take the center of the difference between these two readings, in this case 3.9 Volts, and that is the target to seek while now slowly lowering the blade into the optical path. When the output of the Analog Board is 3.9 Volts, the bottom of the blade is in the middle of the optical path. - Note that this target value is re-measured and calculated for each cycle of the blade diameter measuring procedure. In that way blade diameter measurement errors, due to drifts in the zero or gain of the analog electronics, are minimized. A problem with this is that the output stage of the Analog Board IC3B can not supply any pull-down current. Thus if there is any bias current trying to pull the input of the ADC in a positive direction - it will just go up. It could be that ADT does have a pull-down resistor at the input of the ADC that measures the Analog Board output (on whatever board the ADC is located in their controller setup) and that this resistors is just disconnected or damaged in our machine. In any case on the Analog Board I added a 10k Ohm pull-down from the card's output to the card's ground plane and then the dicing saw was able to complete its Height Tests. The OpAmp at IC3B can easily drive the 1 mA current pulled by this resistor at the full +10 Volt output, the highest output voltage that could be expected from the Analog Board. A smarter setup might be a 20k to 50k pull-down to -12V so that the class B output stage of IC3B is always turned On in one direction. With help from ADT support we also had to adjust the gain of the Current to Voltage Converter stage to match the LED and the Photo-Diode in the Height Sensor. Note that this gain adjustment is not to compensate for anything on the Analog Board itself - rather the gain adjustment is required to match the characteristics of the LED and Photo-Diode in the Height Sensor on a particular machine. On the work bench, measuring the gain of the Analog Board that was adjusted on our saw with ADT's help we find that 1 uAmp of current input causes a 1.8 Volt output at TP1 or TP2. This is linear up to about a 5 uAmp input (about a 9 Volt output) which is near the limit of the board when running from +-12V supplies. Note that this is near the low end of the range of gain available from the Current to Voltage Converter.