I used the hexar on my vacation in the US and it performed flawlessly there despite 50 C heat in Las Vegas and Hoover Dam. This is a camera that is easy to get spoilt by. It is small, quick, exposes perfectly, is sharp and is silent. No wonder I have come to enjoy it!
But… would you believe it! When I grabbed the Hexar for a walk the other day it refused to work! It did not react to anything! At all! I have read about this dreaded symptom. The prospects were grim!
But since “its only electronics these days” I had a peek inside it and found the culprit! It appears that the vias on the flexible circuit board apparently can develop breaks. In my case a via for the positive supply did not work.
The circuit board is populated with surface mount devices, SMDs, and since my nearsightedness has not left me yet, I have no problems soldering the little buggers even without a microscope. In this case it was a simpler case of following the traces on the double sided board but still my eyes and an Armytek head lamp helped a lot. And the problem was not SMD-related at all.
An overview of the bottom PCB with the camera’s bottom plate removed (in fact I removed MUCH more just do analyze the problem, but this is where the problem was). The battery compartment has a rechargeable battery for testing here (remove when soldering:). The battery current is fed directly to a filter capacitor (the rightmost above) before going to the motor drive ICs to the right. The LB8620 contains two motor drives. There are two of these chips, the second is just discernible to the far right. The two pins in each corner of the chip are ground and Vcc is fed on pin 3 (3rd from above left).
Here is a detail of the PCB. The two very small cisterns are the two 100uF/10V capacitors. The leftmost vertical trace is the positive supply running on the upper side of the PCB to the right capacitor’s positive pin, situated to the right of the blue capacitor on the PCB’s under side, not visible in the picture above. From this pin the supply changes side to the lower side of the PCB and runs to the motor drive circuitry, among other things. It was this through-hole that did not work.
With a temp controlled solder iron I removed the right cap altogether and found some green corrosion on the pins and some smudge on the PCB itself. Isopropanol cleansed the areas but a simple resoldering did not alleviate the problem. Perhaps the cap is giving up slowly? Resistance measurement of it showed a increasing resistance as it should.
One could hope that solder and reheat on the pin would connect the upper and lower side of the PCB. Not so in this case. So I decided to do a jumper instead of relying on the PCB. The jumper is the overly red wire soldered to the SOT-89 type IC (marked OK14). From here it runs to the under side of the PCB just beside the plastic little “screw tower”, it is then soldered to the right capacitor’s positive pin on the underside of the PCB. From there on it is on its own and the PCB seems to be fine.
I used just over 300 degrees (Celcius of course, we’re civilized) on the solder iron but also tested 350 degrees on a bit of the unused flexible PCB. It turns out that the PCB plastic is quite tolerant to heat so that no extra precautions seemed to be necessary. Of course one should not be overly slow when heating the capacitor pin but there is no need for split second timing either. In my case a 2-3 second heating was enough.
Looking up the various SMDs, and specially the CPU chip, I found no indication that 6.5 V would be dangerous, most components have an “Absolute maximum rating” of 7 V. Note that the data sheet “Absolute maximum”-figures indicate where no damage will happen to the component – no actual function is promised though. I have previously noted that 6.4-6.5 V is some kind of maximum voltage where the camera will run, above this the LCD acts up. Better not charge the battery with more than 6 volts then…
After this modification the camera reacts normally in all aspects and sounds OK too. I will have to do a short test roll just to check the the focussing still is calibrated.