Hi,
Someone close to the Netherlands that can supply a CCR1036 power supply, urgently?
Just had ANOTHER CCR starting to fall over due to a suspected failed power supply, shrugs… 
And there she goes…
Found a replacement at a online shop in Germany that luckily has stock. Hopefully I’ll receive it tomorrow!
MT - you REALLY need to work on the reliability of your hardware 
Hi Chris,
Let us know if the C10 capacitor is leaking, and if it has the black plastic around it.
Thanks,
Andrew
Will ask the DC engineers to take some pics when they replace it. 4 CCR’s purchased (over a period of 2 years), this is now the 3rd one failing
I guess we found the issue. New PSU is in, 23.7V on the dot just like it used to be.
Hey, I’d be grateful if you could name the shop you got a replacement power supply from as I need to replace two of them myself and my distributor knows of no such thing.
http://www.meconet.de/ in Germany, ~14 in stock if I remember correctly. I also saw a few US companies with stock (but shipping & customs would have taken too long in my case).
My first unit came from MT directly, but I think that was on a special case only.
BTW @nz_monkey…
These power supplies with the better, green caps.. Aren’t these supposed to be the NEW improved power supplies?
And yes, this is in a huge DC, so a very stable supply on the AC side, on UPS.
Hi Chris,
Yep the PSU with the green capacitors are the revised ones.
You have me really concerned now
I wonder if the cap failures are due to low quality capacitors, or poor circuit design..
Currently described problem seems to be specific power supply issue.
For the last 4 months none of RMA CCRs was returned due to problems with power supply, neither we have received any other request about faulty supply.
I have two routers CCR1036-8G-2S+ and both have failed psu.
And yes, the C10 capacitor fails.
Link to g.drive if photo can’t load
I can’t do RMA, becouse warranty was over a year ago.
For one router, I changed the power supply from the LED strip. It has a power of 100Watt.
Link to g.drive if photo can’t load
For the second (backup) router, I replaced both capacitors (C10 and C11).
It’s nearly always about heat. Heat has an extreme aging effect on electrolytic capacitors. When a particular value capacitor in a device fails it’s mostly the ones closest to the heat.
Fig 2.15 shows the extreme reduction in life caused be elevated temperatures.
http://www.nichicon.co.jp/english/products/pdf/aluminum.pdf
I don’t disagree. The PSU in the CCR1009-1036 receives no airflow so if anything was going to overheat, it would be the PSU. Coincidentally the cap that seems to fail is C10 which is closest to the heatsink.
Our units that have failed are in Tier3+ facilities with clean power and excellent cooling, so it does not seem to be directly related to external environmental factors.
I am seeing wider reports of CCR power supply problems now.
Mikrotik, maybe you are not seeing these problems via RMA as the units are 12+ months old, or people are just replacing the PSU and not returning the routers.
Well the COST associated with a RMA is a factor. The DISTRIBUTORS just pull up their shoulders (frankly, I was forced to purchase my first replacement PSU from MT directly, as the DISTRIBUTOR insisted that replacement PSUs doesn’t exist and can’t be replaced. My second PSU was imported, for the same reason). Talk about DISTRIBUTORS being uneducated.
If it’s THAT big of an issue to JUST get a replacement PSU, just THINK about what the issues would be with a RMA though the DISTRIBUTOR. Seeing the DISTRIBUTOR doesn’t want to assist, it becomes financially unfeasible to bother to do a RMA to begin with. Why would I spend how many hundreds of $ on courier charges to send a faulty unit back to MT, if I can just go and purchase a replacement PSU for $15 odd? Having a CCR fail in a production network, there’s the matter that it is a time critical failure too. You want to get back up and running ASAP. You don’t have the TIME to spend weeks and weeks on RMA procedures and wait while distributors and/or MT plays games and sit and think about what could possibly have caused the failure.
MT is sorely unaware of these issues because of the lack of RMAs received (due to their distribution channels), the problem is FAR greater than what MT would like to accept. MT’s distributors (in SA at least), is a joke, to say the least.
Matters like this, faulty hardware, unreliable hardware, distributors pulling up their shoulders not assisting clients… These are all SERIOUS concerns, which for me personally, is making me think very long and very hard whether I want MT in core production networks.
Just my 2c.
As far as I know, the last 6-8 months we are shipping units with an improved PSU, the C10 has been changed to a better one. Some parts have been changed to better handle the heating.
We also had 3 failed CCR PSU within the last 12 months. All Units are located in a datacenter with controlled temperature without overheating. We replaced the PSU with spare parts and never notified MT about it as the CCR were out of support.
The third PSU died last friday and it seems like it killed the CCR as well as since then VRRP with AH starts to fail 1-6 hours after activating it (the message is saying that the AH encryption is wrong).
The VRRP config was running fine for 2 years without changes and this error occurs randomly after enabling the ccr.
VRRP is behaving normally, both nodes see each others and failover can be done by changing priority without problems, until suddenly, between 1 and 6 hours after enabling, the ccr with the failed PSU messes up the vrrp ah config and both units decides to become master.
Its really bad that a failed PSU can damage the CCR and makes him act that bad. We disabled VRRP for now and will remove all CCR from our core network and use them only in non critical environments. We are really sad about that as we really like our CCR but it looks like they are not ready for critical environments yet.
We started widely deploying 1036s about 3 years ago and less than 2 years in we started seeing some fail. Now after 3 years with a couple hundred deployed its become a big issue for us. In fact I’m waiting for a tech to run out and replace a power supply of a down CCR right now. We have a handful of these power supplies on hand but the outages caused by them are killing us. The one I’m replacing right now is 1y 6m old and was installed in a data center. Monitoring shows the 1036 has been running at a nice even temperature. The max it ever hit was 59C on the CPU and 55C on the chassis. With the average being only 43C on the chassis. MikroTik says its tested up to 60C.
I’m glad to hear they’ve at least made some changes to the newer power supplies but I’m still nervous. Has anyone tried the 1009s with passive cooling and external power supply?
Thanks
Same there, cooled DC, 1036-8G-2S+, 1y6m, C10 capacitor in PSU is blown up. Voltage - 23.9 without load, step down to 23.3 when starting. Switching CPU freq to 800 (within bootloader) helps starting it up. On full freq - cyclic reboot.
Same situation here in Belgium.
A few CCR1036 in production (bought from Interprojekt).
C10 blown after a bit more than 1 year of 5 of them.
Distributor has changed the power supplies with better ones for 4 of them.
The 5th one was blown last week. We’ll change the power supply on all the remaining devices.
I would be glad to know if other models have the same power supply, to change them all.
When it is so clear what is wrong (as usually with SMPS), why don’t you replace only the caps instead of the whole PS?
On this type of board and with this kind of cap that should be easy to do for any electronics repair person.
One year is too short a lifetime, but those caps are going to be the first thing that fails in most of today’s equipment.