More than half of a batch of Sun Ray 1 terminals that we received in July 2000 died within two years. They first became unstable (occasional resets) and then crashed within seconds of being powered up.
A visual examination quickly revealed the cause: Four capacitors of the power supply had failed. Their top caps are bulging outside and dark electrolyte has visibly vented out of the top cap and can be seen as a small dark droplet:
As might be expected with defect smoothing capacitors, an oscilloscope shows that there is a ~1 V ripple on the the 3.3 V and 5 V power supply lines, which should be bad enough to cause the crashes.
A check in SunSolve reveals that this is a known problem (alert 24670).
The following four components are affected:
C414: 3300 µF, 10 V (Lelon RXJ 105 °C L927(M), black), h=26 mm, d=13mm, w=5mm C405: 1000 µF, 10 V (Lelon RXJ 105 °C L919(M), brown), h=20 mm, d=10mm, w=5mm C410: 1000 µF, 10 V (Lelon RXJ 105 °C L919(M), brown), h=20 mm, d=10mm, w=5mm C419: 1000 µF, 10 V (Lelon RXJ 105 °C L919(M), brown), h=20 mm, d=10mm, w=5mm
A report in IEEE Spectrum February 2003 claims that several major Taiwanese capacitor producers (including Lelon) bought electrolyte for ≥1000 µF low-ESR aluminum capacitors from a Taiwanese supplier (Lien Yan) who the report alleges to have used a stolen P-50 water-based electrolyte recipe from a Japanese competitor, but got the additives wrong. As a result, the report claims, the capacitors produce excess hydrogen and burst. I wonder whether the astonishing failure rate of our Sun Rays is related to this story. Lelon denied in a press release using Lien Yan P-50/P-51 electrolytes, however the language used leaves open whether they did in the past. (also: Dennis Zogbi, October 2002, c't, tti, more photos)
As suitable replacements, I bought from Farnell:
3 x 1000 µF, 10 V, 20x10.0 mm, Nichicon PL, Farnell 345-1100 (catalog 2-790) 1 x 3300 µF, 10 V, 25x12.5 mm, Nichicon PR, Farnell 345-1458 (catalog 2-782)
Steps for opening Sun Ray 1:
With this capacitor replacement, I was able to repair all but one of our SunRay failures so far. In this one case, the device had apparently remained powered up unattended for many months after the capacitor failure. This had led to further damage in earlier stages of the power supply. In this single device, there was clearly visible heat damage to several other components. Therefore, if you suspect an unstable supply voltage or if you can see bulging capacitor caps through the ventilation holes, remove power immediately until the device can be tested and repaired, to avoid further damage that might not be practical to repair.
Standard warning: Power supplies must be repaired only by technically qualified persons. Local regulations may require repaired devices to be tested for electrical safety (e.g., protective-earth impedance and line insulation of metal enclosure). Use information on this page at your own risk.
Thanks to Tim Harris, Ian Pratt, James Bulpin, Nick Batterham, and Paul Evans for their assistance.
created 2002-10-04 – last modified 2003-04-24 – http://www.cl.cam.ac.uk/~mgk25/sunray/caps-repair.html