After I’ve now had the second device with some design flaw that causes it to fail just slightly outside of the warranty period, I wanted to write down my experiences.
Synology DS415+ and the Intel Atom C2000 clock signal issue
Back in January 2015, I’ve bought a Synology DiskStation DS415+. It came with 2 years of warranty (later extended to 3 after Synology heard of the problem but didn’t see any cause for action). It happily purred along 24/7 until it didn’t. In late August 2018 – 6 months outside of their warranty period – I started to get random lockups and a few days later it didn’t boot at all. Instead, it showed all HDD-LEDs as yellow with the status LED blinking blue for a few minutes until it shut down on its own again.
Turns out there was a problem with the implementation of the clock signal for the Intel Atom C2000 CPU. It appears as if it wasn’t properly grounded and over the years didn’t make the CPU “tick” anymore. There’s also a huge thread in the German (inofficial) Synology-forum.de.
While Netgear – who had some devices that suffered from the same issue – did a recall and fixed all devices for free, the German Synology support was like “Meh, it’s outside of warranty. Nothing we can do about this. But we’ll happily assist you in selecting a new Synology device to purchase.” – totally useless. Even after hinting that other Synology users (in other countries, though) got their units fixed even outside of the warranty period. Luckily, these users then checked the returned devices for modifications and found this:
After pointing out to the Synology support that there seems to be an easy fix by soldering a small resistor to the PCB, they said something along the lines of “Yes, we know. But this isn’t public information and thus we’re not telling customers.”
This was the moment when it became clear to me, that my next NAS won’t be a Synology device.
However, after adding the resistor to my device, it has been running fine ever since. And it’s still running now in May 2025 and just refuses to die. Still, the way their support has handled this left a very bad taste in my mouth.
AVM Fritz!Box 7590 and 2.4 GHz WiFi
In Germany, we have the so-called “Routerfreiheit” (freedom of router choice) – meaning you don’t have to use the router from your Internet provider but can use any other instead. And your provider must provide the necessary details to be able to connect that 3rd party router to their network.
AVM is a Berlin-based company building and selling routers that do a lot more stuff than just routing. Their devices usually come with support for telephony (POTS, ISDN and VoIP – all with support for DECT handsets, too), VPN, media server and nowadays even home automation. All packaged up in one neat device.
So, I’ve bought their Fritzbox 7590 back in January 2018 (upgrading from their 7390 model). These come with a 5 year warranty. The Fritzbox also worked as expected. In early 2023, we moved and upgraded our main router to AVM’s Fibre-capable model. However, the new flat is very long and has lots of metal parts in the walls, so we’ve “downgraded” the older model to repeating duty.
There’s one device permanently in the area of that 7590 repeater – an old Raspberry Pi 2 receiving ADS-B signals from aeroplanes in the area. Due to the walls, 5 GHz has a hard time and that Pi usually used 2.4 GHz connections.
For several months now, my Zabbix shows that this Pi has lots of issues with its WiFi connection. And just by accident I’ve found this German blog post about an issue with the 2.4 GHz WiFi module. I’ve dug out my old Android phone and used the WiFi monitor feature of Network Analyzer Pro to check the signals. And lo and behold: Nothing on 2.4 GHz where the Raspberry Pi is located. Only when about 10cm away from the Fritzbox, Network Analyzer showed its WiFi signal.
It seems like there are some MP1477 step-down converters and a too small capacitor causes the bootstrap voltage to spike up to 6V when that pin is only designed for up to 5V. This causes the MP1477 to die over time, raising its output voltage. (Details are explained in this recent blog post.) If it’s not fixed in time, it’ll also kill the components after that, namely a component marked 055F which seems to be a Qualcomm QFE-1922-0-24MQFN – the WiFi frontend. That’s when the 2.4 GHz WiFi will fail (or only be receivable when being very close to the Fritzbox). And if still left untreated, it can even raise the voltage so far that it’ll cause a resistor to burn through the PCB.
A news article published in May 2024 dug up this thread where users report that they’ve seen the issue arise as early as about slightly over 5 years from purchase. Looks like I’m “lucky” it only happened after 7 years with my device.
Reposts