|Re: RUBBER KEYPAD under SPICE calculators and rebuild notes|
Message #9 Posted by Ellis Easley on 18 June 2003, 7:38 p.m.,
in response to message #1 by Norm
I opened up my 34C's last weekend to lubricate the switches and generally clean up the one that I never opened before. I used the dental floss for the first time and it works great! I used a double loop (a single loop broke the first time I tried it) and tied the loop to a screwdriver handle with a triple half-hitch (I remembered a "dumb criminal" story about a couple of convicts who made a rope out of dental floss and cut their hands up escaping from an upper-story prison window.)
My original unit is the soldered type. I had opened it before with a modified crack-of-doom and twisting operation which resulted in breaking the narrow lip along the front edge, which I had glued back as well as I could. The dental floss re-cracked the end of my glue joint but otherwise didn't affect the front edge. I hadn't remembered some details, such as that the main assembly is a regular fiberglass PCB (its S/N is 2051S)- with so much discussion about the flexible PCB in the solderless design, I though the only difference was some solder! Then I opened the other unit and saw the solderless design up close for the first time (its S/N is 1934A).
My newer, soldered unit has metal disks under clear sticky tape for the keyboard click, like the Voyagers. It also has the thick rubber sheet between the keys and the tape over the disks. The older, solderless unit has a sheet of plastic over the keyboard with printed traces and the clicky domes molded into it - actually it looks vacuum formed. There is a third layer of plastic between the sheet with domes and the main flexible PCB, with holes under the domes where the column lines make contact with the row lines. There are also five holes where the column lines on the outer layer are "spot soldered" to the main flexible PCB. Now I see how much greater is the travel of the older keyboard. When switching back and forth between them, the the older, long travel keyboard seems tiring, but after using just it for a while, it seems normal.
The sheet of rubber in the newer, soldered unit isn't just flat on my unit. It has disc-shaped regions molded in over each keyboard disk.
My older unit had corrosion on the power switch but was otherwise pretty clean. I removed the corrosion and cleaned the rest of the flexible PCB pads - and the IC and display and P.S. assemply pins - with a "Pink Pearl" pencil eraser.
It worked when I lubed the switches and put it back together except some segments of the LED are open. I found that my multimeter diode test range is good for testing the LED display - it puts out just enough current to light up the segments, and the voltage readout should show up segments that are marginal - in fact I thought the minus sign segment was somehow different because it registered a higher voltage when I first lit it up (all other segments had registered about 1.5V), then I found the other terminal that connects directly to the minus sign cathode and determined that the display assembly has a regular diode connected between the two outer terminals - what's that for? It's between the power supply positive (after the power switch) and negative inputs - like a shunt reverse polarity protection diode, except without a fuse - but why is it part of the display?
I traced out a schematic of the solderless unit including the power supply and compared the two units. I noticed two differences in the power supplies: The newer unit has a 10K resistor instead of a diode to provide the standby power from the battery - I imagine that keeps the memory alive with a lower battery voltage - also, the capacitor that holds some charge for the standby power when the battery is out is 22 mf in the older unit and 33 mf in the newer one.
My newer, soldered unit doesn't "flash" (at least not as much) and the older one does, but I don't believe it has anything to do with deterioration of the IC's - they have different IC's. The units have two IC's in common - the two 8 pin packages that are not connected to standby power: 1MA1-14 and 1MA1-15 (-0015 on the older unit but I imagine it is the same as -15). On the older, flashing unit the other 8 pin is 1LB5-002; on the newer unit it is 1LB9-02. The 40 pin IC on the older unit is 1820-2162-A, on the newer unit is is 1820-2162 (no A). The 40 pin IC's are made by AMI and the datecodes are 7906 and 8047 respectively. Who knows if the A means they are different?
Regarding flashing, the difference between my units is this: when executing a function like Sine, the older unit flashes random commas and the minus sign; the newer unit is completely dark until the result is displayed. Both units flash random numbers, commas, periods and the minus sign when a user program is running.
BTW, both units return the same number for the Calculator Forensics test as the 41CX and the 15C - does this mean that the Spice models contain some kind of Nut CPU and firmware?
I also made some current measurements. The battery charging current is 150 mA. This would be the .3C "quick charge" 5 hour rate. With the battery almost discharged (just before the low battery indication starts) the unit draws 150 mA with the display set to FIX 9 just after power on (all zeros) and 190 mA after STO ENTER (all segments lit). With a freshly charged battery, the unit draws 125 mA with all zeros and 160 mA after STO ENTER. Also with a freshly charged battery, it draws 65 mA displaying "1." and from 60 to 75 mA executing the Integration function - the higher current occurs during evaluation of the function in the user program, when random numbers are displayed.
One final thing about my older, solderless unit: it has a mistake in the printing on the keyboard bezel - above the "A" key, both "FIX" and "DEG" are printed in gold! Also, the blue arrows in "->D->R" and "->H.MS->H" have a bit of gold at the very end of the tails.