For an unrelated project, I bought a cheap 3Com Gigabit OfficeConnect switch ($3). I was expecting some bad caps but ALL of them were toast. Completely toast.
It had GLORIA and Wendell caps, all of the GLORIA ones were bloated, and some on the verge of releasing electrolyte. Yeah the switch didn't work.
I did however do a few modifications: I used 470uF 25v instead of 470uF 35v on the two input caps because that's all I had. I tested the voltage and it only went up to 18v. So we're within tolerance.
I replaced all 330uF 16v with the same 470uF 25v caps. 220uF 16v were replaced with 220uF 10v polymer OSCONs (all that I had on hand). Aaaand, a 47uF I just threw in 100uF. Finally, I added two more 1000uF 6.3v filter caps on two unused pads. I figure they didn't populate those just to cut costs, but I need this thing to run stably, so yeah.
Here's what it looks like AFTER:
I tilted the tall 470uF nichicons because I want to ensure they have some clearance. Also regarding the ones that have the weird leg thing going on (near the coil rectifiers), RADIAL capacitors were originally there!
I've really only seen radials on the old vintage stuff. Oh well. It looks horrible, and I generally hate making imperfect work like that, but I don't feel like hunting down some VXes just for a switch.
The small blue and small brown caps are ELNA. I have lots of ELNAs kicking around. SILMICs too-- but you wouldn't use a SILMIC on a digital circuit.
I can see why all of the original cheap capacitors were bloated, this switch gets quite warm after running for a few hours. Running 24/7 and it'll kill those cheap caps in no time (probably why not too many of these are around?)
Up and operating perfectly (just try to ignore the diskette box):
It had GLORIA and Wendell caps, all of the GLORIA ones were bloated, and some on the verge of releasing electrolyte. Yeah the switch didn't work.
I did however do a few modifications: I used 470uF 25v instead of 470uF 35v on the two input caps because that's all I had. I tested the voltage and it only went up to 18v. So we're within tolerance.
I replaced all 330uF 16v with the same 470uF 25v caps. 220uF 16v were replaced with 220uF 10v polymer OSCONs (all that I had on hand). Aaaand, a 47uF I just threw in 100uF. Finally, I added two more 1000uF 6.3v filter caps on two unused pads. I figure they didn't populate those just to cut costs, but I need this thing to run stably, so yeah.
Here's what it looks like AFTER:
I tilted the tall 470uF nichicons because I want to ensure they have some clearance. Also regarding the ones that have the weird leg thing going on (near the coil rectifiers), RADIAL capacitors were originally there!
I've really only seen radials on the old vintage stuff. Oh well. It looks horrible, and I generally hate making imperfect work like that, but I don't feel like hunting down some VXes just for a switch.
The small blue and small brown caps are ELNA. I have lots of ELNAs kicking around. SILMICs too-- but you wouldn't use a SILMIC on a digital circuit.
I can see why all of the original cheap capacitors were bloated, this switch gets quite warm after running for a few hours. Running 24/7 and it'll kill those cheap caps in no time (probably why not too many of these are around?)
Up and operating perfectly (just try to ignore the diskette box):
.
Comment