One of our CCR1036 began acting funny. After a few hours of head scratching looking into the software while taking angry customer calls, the unit was replaced.
On visual inspection of the faulty unit, we discovered these bulged capacitors. They seem to be there to stabilize power to the Ethernet chips. Lack of physical stability of the Ethernet ports was consistent with the symptoms observed. Therefore, we believe that these capacitors are to blame.
Now for the next problem. Which is to find replacement capacitors. You would think that the markings on the capacitor body would be enough, right? Well, no. Markings are:
MG 105ºC
1209(M)
Now, the part about the temperature rating is pretty clear. But, what does 1209(M) mean?
A usual convention uses 3 digits (XYZ), meaning the capacity is XY10^Z pF. But in this case, that would work out to 1210^9 pF, or 12 mF (milliFarad, not microFarad), which is huge for an electrolytic of that size and presumed voltage rating.
I have tried searching and asking around, and nobody seems to know what is the capacity of those things. Does anybody have an idea?
P/S: Mikrotik, listen, we will pay you 20 euros more for each CCR and you put quality capacitors, ok?
Also remember the value of such decoupling caps is not at all critical.
Check what voltage is across them (5v, 12v, 24v) and get a capacitor with a suitable voltage rating and same physical
size, and you will be OK. But the value is likely at the other side.
So the value is 680 µF, 6.3 V.
The manufacturer for these is Su’scon. I have ordered for replacements Würth Elektroniks of the same value, rated for 125°C instead of the 105°C of the failed ones. At the princely cost of 0.818 euros apiece in RS, I hope them to last for a lifetime.
Also check the caps in the power supply.
Such electrolytic caps are the first components to fail in modern electronics.
There are several reasons for that:
temperature of operation
relatively large AC current passing through capacitor e.g. due to switch mode power supply
cost cutting on components
planned longevity
All capacitors have a series resistance, not a physical resistor but inherent due to the design, and the high
AC current dissipates power in that resistance and heats up the component from inside. Any outside heating
(in switch mode supplies often because of the rectifier diodes immediately adjacent to the caps) is added to
that. As the component has operated on high temperature for a while, the series resistance (ESR) increases,
which sends the whole thing into a death spiral: increased ESR, increased dissipation, increased temperature,
even more increased ESR, etc. The top of the cap bulges until the security seal bursts and the remaining
electrolyte evaporates, killing the component and your device.
Usually any modern device that has gradually started to act funny after a couple of years of operation can be repaired
by replacing caps.
Thanks for the input, especially the interesting explanation on how capacitors die. In this particular unit, the power supply had already been replaced. Actually this is one of the units I fitted with dual power supplies.
Now hoping the remainder capacitors remain good for a few more years!
