Hi,

the last two months i had 5 rb450 with the same problem. It stops to work , and them the eth4 and eth5 leds turns on a very weak light. All these RB450 had been working for about 8 to 10 months.. when opened, i found ttwo capacitors leaked and blowed ( su'scom 560uf 6.3 volts..) I just bought about 50 rb450g and found out that they have the same capacitors... should i be worried?? is this a known issue? the problem is the lost of reliability .


There's anyway to fix the broken boards?

Welcome to the world of Mikrotik.

I've had 20 RB450s with blown capacitors. The reseller has been successful in replacing some, the others I've just gone out and bought the 6.3v capacitor (I upgraded to 680uf) and soldered them in myself. You'll need 3 per board.

thanks!!

any problem with the 450G version?


i have some rb150 that's been working for years.. with out problems..

I'll have to try to find a board that I haven't repaired yet. However I can point out the capacitors I had to replace in order to make the board operational again.

Thanks

I purchased 20 RB450s in 2008 from Tom at Roc-Noc. During 2009, each of the 20 failed. Tom mentioned that Mikrotik was aware of the capacitor issue and would offer a replacement of the 20 RB450s. In that time I've received a partial replacement. Sadly, due to the substantial delay in receiving the remainder of the RB450 replacements I've had to implement my own repairs, replacing the 3 capacitors in the picture below.

If your RB450 fails, you will see 3 bulging and brown capacitors in the areas I've pointed out below. Replace those and your RB450 will be functional again as well as not having the issue where a rapid power outage will cause the board not to reboot.

The three brown capacitors in the picture I uploaded are the replacement caps I purchased locally. Tom at Roc Noc may be able to take a picture as he has several blown RB450s on his hands.

I've got the same problem, here's some photos of the bad caps

My RB450 has been in continuous use for about 1.5 years now. It was also probably running a bit hotter than it should have been (sitting on top of a Linksys AP that throws of a lot of heat).

I'll probably just resolder some new caps, as I always like an opportunity to bust out my Weller.

Now I'm curious.
We are using a RB450G as our main router.
Can we have an official statement on whether we should get it replaced by another model?

I found this post: http://www.badcaps.net/forum/showthread ... =su%27scon

Best regards

Max

We have seen only the Green one damaged. It would be great to see actual damaged ones ...

I have about 20 of these back from customers. Here is one of them. Notice the leaking electrolyte and the three bulging electrolytic caps.

If you can make it fit, try to get caps for a higher voltage instead of higher capacity. It could be that the original ones are a bad batch but so many?? may be at a higher voltage they never fail again.

The trouble is that for a higher voltage they are bigger, so take the board to the electronics shop to see if they fit before you buy.

EDIT: just to be sure, measure the voltage over the capacitors while the board is turned on... you should see a DC voltage. Check if the polarity is right and how high that voltage is. Now you know if the caps have a too low voltage rating and if they are mounted at the right polarity!!

cheers,
Nick.

any problem with the 450G version?

I'd also like to know if the "G" version has the same caps and potential for problems, or if this is limited to the plain 450.

We have also seen this.
See http://forum.mikrotik.com/viewtopic.php?f=2&t=35400...

any problem with the 450G version?

I'd also like to know if the "G" version has the same caps and potential for problems, or if this is limited to the plain 450.

So far I have only seen this on the RB450. The RB450G appears to be quite different in design.

Tom

any problem with the 450G version?

I'd also like to know if the "G" version has the same caps and potential for problems, or if this is limited to the plain 450.

So far I have only seen this on the RB450. The RB450G appears to be quite different in design.

Tom

I'll second this. We have only seen RB450 units with this problem at this point.

Scott @ WISP-Router

Seen that problen MANY times on PC hardware..

Dell was FULL of them for a while...

They were swapping mail boards like crazy...
I had a pair of 270 small form factor units affected...

Intel, acer, and just about everyone has had there share of them...

We call them " Jiffypop " caps...

Swap them out with higher qualtiy caps and all is well.. ( If the unit hasent been run too long and damaged the power circuit..)

Keep the values the same ( Voltage can be a bit higher )..
Cap "values" "CAN" be critical... if it is a passive filter cap then not so much,,
If it is part of a "charge pump" circuit then it can be very critical... (Can affect output voltage or load on charge pump voltage controler)

Change out sould only take about 3 - 5 Min each... An experianced tech with a bit of practice can do them all in about 5 Min..

I just had this happen to a RB450G I own. Here are the pics:
I will have had the unit for a year next week. I bought it from rocnoc.com, so hopefully they can repair/replace it for me. Otherwise I'm going to have to learn to do some soldering I guess. I'm not very happy about this

Wonder when MT will use solid capacitors just like Gigabyte on their motherboards. RouterBOARDs work 24/7 and they should have them. Maybe not in 750/G model (SOHO).

I really hope I dont start seeing this problem. I have about 80 450Gs I rolled out in the last 6 months.

I really hope I dont start seeing this problem. I have about 80 450Gs I rolled out in the last 6 months.

I don't think you will. This is the first RB450G that we have seen with bulging caps. I am hoping this is an anomaly.

The 450G power circuit design is much different from the older RB450 power circuit where we have seen all the cap problems.

Tom

I really hope I dont start seeing this problem. I have about 80 450Gs I rolled out in the last 6 months.

I don't think you will. This is the first RB450G that we have seen with bulging caps. I am hoping this is an anomaly.

The 450G power circuit design is much different from the older RB450 power circuit where we have seen all the cap problems.

Tom

Normis wrote that remedy was to change the suplier indicating parts quality as the issue - not the design. Therefore it remains to be seen how many boards were made with low quality parts before problem was identified.

Why dont you start using solid state capacitors? Are they that much more expensive? I guess that for router HW, which can be easily in operation for more than 10 years customers would wish to see capacitors that will last that long.

+1

I also have the problem with a RB450G I replaced with another 630uF 16V and works fine yet..

an another things Im seeing the same capacito 6.3V 560uF with the same color, manufactor on RB750 and RB750G






please normis can you confirm that these lots of RB750/RB750G doesnt have that problem

thanks

I ran through all my faulty RB450's, and they all seem to originate from the same three shipments received in 2008. It would seem that the issue is not present in newer batches of RB450.

My Main router which is an RB450 seems to be having the same problem. I opened up the case and noticed "brown" stuff leaking out of the 3 capacitors previously noted in previous posts.

I'm going down to Fry's to look for replacement capacitors.

What information should I give the guys at Fry's so that I get the right replacement capacitors?

I can't recall the exact component name, but write to support, they happily gave that information to me.

Would any of these capacitors work: http://www.alliedelec.com/passive-compo ... 7106&No=20

Additionally, why am I forced to buy capacitors if this was a known "bad" batch of capacitors which is causing this problem.

I called the Distributor who I bought my rb450 from, and they told me that I can't RMA the box.

We've had some RB450G's hit by power surges previously that I've fixed myself.

Had no problems replacing the 6.3v 560uF original capacitors with *similar* ones.
In this case the replacements are 10v 470uF and the 3 replaced are the ones in a line in the middle of the board.
Gives me a chance to practice my soldering skills and teaches people why they should invest in surge protectors hehe.

yes, that's what we did

Put on solid state caps?
Got a picture of one of the newer boards?

I'm interested to see as if the mounts for the solid state caps are the ones I think they are I might replace the next one that comes across my desk with those instead.

Also tempted to try building my own RB compatible device using IC's from a broken one at some point.. just as a hobby

In this case the replacements are 10v 470uF and the 3 replaced are the ones in a line in the middle of the board.

Only 10 volt? How about this one?

400 Volt
-40/+105 °C

Lead spacing and package is too big. Use 470uF - 680uF and 6,3V - 16V capacitors with 3,5mm lead spacing.It's posible use 5mm lead spacing but need aditional lead forming. Don't forget about polarity.

I know that lead spacing is too wide, the capacitor is 25mm in diameter. The type of capacitor in the picture is probably one of the best you can find, it'll last forever. I think the capacitors life time is 18 000 hours at the specified temperature (105°C).

No, just 4000h at 105C

Standards IEC 60384-4 Long Life Grade 40/105/56,
in accordance with CECC 30 301-809
Capacitance range 150–22000 μF
Capacitance tolerance –20 to +20%
Rated voltage 35–450 VDC
Temperature range –40 to +105°C
Leakage current, IRL (μA) 0.003 (μA) x CR (μF) x UR (V)
Operational life time 4000 hours at +105°C
Shelf life 4 years at 0 Volt and +40°C
Diameter range 25–35 mm

You are right, I posted wrong picture, but 4000 hours at that temperature is not bad.

Found the Capacitors at Fry's. Replaced them on the board, and now I'm operational again.

Thanks to all for your assistance !!!

Have 4 x 450G's Anyone experienced the issue on the 450G's to any extent. I am sending some off to a site in China, so need to know asap, so I can swap the caps out now if need be.

Have 4 x 450G's Anyone experienced the issue on the 450G's to any extent. I am sending some off to a site in China, so need to know asap, so I can swap the caps out now if need be.

Power them with a 12vdc supply instead of the normal 24vdc supply. They run much cooler and the caps won't have to work as hard (should last much longer).

Tom

I just finished replacing the caps on 5 RB450's. Still no offers of compensation from Mikrotik or the reseller.

I just finished replacing the caps on 5 RB450's. Still no offers of compensation from Mikrotik or the reseller.

I wonder how MT warranties their products in Europe. Europe has mandatory warranty period I believe.

I just finished replacing the caps on 5 RB450's. Still no offers of compensation from Mikrotik or the reseller.

I wonder how MT warranties their products in Europe. Europe has mandatory warranty period I believe.

Mikrotik usually handles the warranties by repairing the equipment themselves, not by reimbursing someone else to repair the equipment. Not sure if that is allowed under the different EU countries laws, though.

I wonder how MT warranties their products in Europe. Europe has mandatory warranty period I believe.

Effective from 1st January 2002, the EU Directive to extend warranty on consumer products from 1- to 2- years, came into effect. This directive only impacts EU member state countries and will be phased in as the EU member states implement the directive. As of 1st January 2002 the following EU member states have implemented this EU Directive: Austria, Belgium, Denmark, Finland, Germany The Netherlands, Norway, Italy and Sweden.

So the country you bought it in is very important, since a directive is merely an instruction to each government to implement a law, and has no meaning until they do.

Directive 1999/44/EC of the European Parliament and of the Council of 25 May 1999 on certain aspects of the sale of consumer goods and associated guarantees

http://eur-lex.europa.eu/LexUriServ/Lex ... 44:en:HTML

Not in the home of Mikrotik. 1 year, bye-bye.
Offices answer on the phone was - "buy a new one!"
Yes I will, but why would I buy another Mikrotik?
Of course will try to solder new capacitators, but if that won't help - any working OpenWRT/DD-WRT solution or will use my Linux webserver as a router.

Hi,

I'm sorry to report that my RB450G died today - looks like faulty capacitors
This unit was purchased on March 2009.
--
Morten

Hi,

I'm sorry to report that my RB450G died today - looks like faulty capacitors
This unit was purchased on March 2009.

rb450g_capacitor.JPG

--
Morten

Who did you buy your RB450 from?

Hi,

I'm sorry to report that my RB450G died today - looks like faulty capacitors
This unit was purchased on March 2009.

rb450g_capacitor.JPG

--
Morten

Who did you buy your RB450 from?

roc-noc.com, this is the RB450G though.

--
Morten

Took 1000 microfarads from dead motherboard, soldered in. Works fine.
Another problem is the heat. Found 2 HDD fans, they fit fine as 1 part of the 450s box, but there is no place to connect them on the board! Some old boards have 3V pins for fans, this one doesn't
Now i need another socket for another power supply (7,5V from dead D-Links will do fine) for the fan... f0ck!nk shait

Hi,

I'm sorry to report that my RB450G died today - looks like faulty capacitors
This unit was purchased on March 2009.

rb450g_capacitor.JPG

--
Morten

Who did you buy your RB450 from?

roc-noc.com, this is the RB450G though.

--
Morten

Roc-noc is where I received my bad 20+ RB450s. I wonder if they received most of the bad shipment. Unfortunately Tom has 6 of my defective RB450s and wont return them either fixed or for me to solder myself. Hopefully Tom can advocate for us with Mikrotik since he's been left holding the bag.

We've also have about 50 RB450 and changed nearly all of them.
At Some I must resolder new Capacitors (I've took 1000uF 16V, the last one who fits into), cause of unavailability for spare Boards and the RB750 I've ordered instead are still not arrived (our internal Spare Parts already used at customer side cause we never had so much Spare Units for so much dead ones)

Also someone wrote to Power just with 12V instead the default MT PS with 24V cause then the are cooler.

So the Question is, does it also make sense to set the RB cpu-frequency and cpu-mode? so instead of the default 300Mhz just 200, 150 or 100Mhz?

Of course CPU freq. is directly related with CPU heat disapation but in this case use 12V PSU make greater sense.

I've tried this one day but couldn't say that there's much difference. It's all time hot just in a RB450Case on the desk.
Let's look how long it would work with new capacitors.

Roc-noc is where I received my bad 20+ RB450s. I wonder if they received most of the bad shipment. Unfortunately Tom has 6 of my defective RB450s and wont return them either fixed or for me to solder myself. Hopefully Tom can advocate for us with Mikrotik since he's been left holding the bag.

Let me know off-line if you have not received them yet. We worked like crazy through a bunch of RB450 cap repairs over the past few weeks and finally got caught up. They should have all been returned by now. I strongly recommend 12vdc on all RB450 and RB450G.

We have sold many RB450s but I don't believe we received all the bad ones. I know this because we have fixed some that were sold by other distributors.

Tom

HI to everyone,

I had the same problem with the caps.

I changed the 3 green one caps with other, but with different value: 470uF 16V

Now the Router boots, with green led blinking and beep, but the 5 ports are already down.
I have the led on the ports lighted very weakly, in decrescent order from tha fifth to the first.

If I connect a computer at any port, the criteria doesn't come up.

Any idea about the causes?

Thank's in advance

470uF@16V is OK. Check again your soldering carefully that you not make any shorts. Else try replace that one cap. what located nearly from swich chip.

Effectively, after the replacement,the RB worked fine.
All orange led in the port socket blinked and connected with the WinBox.

But now, after about 2 weeks from replacement where the RB was powered off, the problem with the ethernet port is yet the same.

About the soldering, I had some difficult to remove the old ones. The last problem was cleaning up the holes.
At the end I used the SMD plots for solder the folded reophores of new caps

At the moment I'm not able to try to change the caps near switch chip, because I'm out for job.
I will do it when back at home.

Many thanks for support.

Cheerio, Sergio.

We've also changed the 3 capacitors at about 10 boards last 2 weeks, hoping the'll work fine

Capacitors I used is some posts above, changing was not the easiest, but also not a too hard work, we made this in field in the car.

I think some capacitors may have a little bit thicker pin, and therefore it's hard to clean the hole good enough.
Unsoldering with a very thin soldering iron at 450C worked fine.

Hi to all,

during a series of voltage measurements, perahps i touched more pins together.
The RB turned off. I unpluged and plugged the power socket, and the RB this time reboot fully. From this time the RB works fine.
I will be out for job for a week. During this time the RB will be powered off.
When I will be back at home, I will try to turn on, and hope that it work fine again.

Cheerio.

We just replaced bad caps on 3 RB450's. Found a bad motherboard with comparable caps that we could re-use.
It wasn't to bad to do just a PITA that we have to do it.... Seems as though there was some power hack for the RB100's as well?
I really wish the RB line was a bit more hardened for longer MTBF.

-Louis

I use Pace desoldering and soldering tools...

http://www.pacedirect.com/product.php?id=543

Makes life MUCH better !!

Makes quick work of bad caps...

Thanks for the posts on this subject.....

We had placed (4) RB450's into seemingly light duty on a shelf, in a rack with other computer equipment, in an air conditioned area.

In about 18 months we had 3 out of the 4 RB450's fail..... Not good.



QUESTION for Mikrotik.... has the reliability issue been resolved in new production??


.

Same problem on two green capacitors of RB450G bought about a year ago from Lin-ITX.

Same here!

Two bulging green capacitors on a couple of (now dead) RB450G bought from Duxtel in Australia, October last year.

Cheers

Chris

yes, that's what we did

Put on solid state caps?
Got a picture of one of the newer boards?

Are the newer boards made with solid state caps?
I have a 450 in front of me that came from Streakwave last week and it clearly has rolled electrolytics.
Su'scon MG series.

2 450's are on their way to me. Both will not power up. I suspect this problem.

There are 2 things in electronics you don't try to pinch pennies with: electrolytics and magnetics.

It takes a strong engineer to stand up to the purchasing department.

Is there a date code or serial number list for product affected by the faulty caps?
I would much rather gear up a field retrofit than deal with outages.

Has distribution been instructed to purge stock and return any product affected?

In addition to the RB450 mention above, I also have an RB450G in my hand that was purchased about 5 months ago and it has 2 of the green SJ series caps C826 & C830.
How could that be if MT changed components?
Is it possible it came from the affected lot and why was sold?

This is starting to worry me.
I'd rather know where the time bombs are than watch them go off one by one.

Oh great.

Just had my first 450 die with this problem.

Am now waiting for more to die and more unhappy customers as a result.

Great.

I just sent an email to MT support asking clarification on warranty period.

One of the RB450's that failed for this capacitor problem is "out of warranty".

According to Streakwave, the warranty is 15 months and it starts from the time THEY receive the product, not when I receive it.

In the case of the RB450, it's a difference of 7 months.

So, effectively, an 8 month warranty.

This is like a Cracker Jack box.

You don't know the prize until you open it - you don't know the warranty until you RMA it.

I cannot imagine MT intended it this way.

OK, I understand the intent and I actually did something similar in a previous life.

However, we forced the distributor to warrant for 12 months.
It then fell on the distributor to rotate and move product.

As it stands, I, the end user, have no idea what my warranty is when I receive MT product!!

That is completely unacceptable in any business.

In reference to my earlier post, has all stock been purged of affected product?

If not, why?

In the UK, the Sale of Goods Act pretty much covers you (see http://www.berr.gov.uk/whatwedo/consume ... 38311.html for a reasonable summary). The problem as I see it is that it's the reseller which is going to take the hit for this and since MT's liability has ended.

It seems to me that MT have admitted that there was a manufacturing defect and that MT should, in some way, compensate their users for this.

Mikrotik sells the device to the distributors. We have informed them what do do if the devices are defective. We can't make them check the devices sooner, or make them sell the devices sooner.

Surely there are things you can do, if you want to to help the customer. Few things come to mind:

Publish when warranty expires for each serial number.
Put manufacturing date on the product.
Put a sticker "warranty expires on xx/xx/xxxx", or freshness guaranteed until XX/XX/XXXX...


Btw, it looks like it's much better to purchase from balticnetworks than streakwave:

Streakwave:
"Warranty returns are subject to the individual manufacturers warranty return policy." [which means 15mths from manufacture date]

Balticnetworks:
"MikroTik, MaxxWave, Ubiquiti, Laird/PacWireless and Kam-Fab equipment is warranted for 1 year from the date of purchase."

Streakwave is terrible. Terrible sales, terrible service and an attitude to go with it.

I once bought a PoE adapter from them and it was DOA. They wouldn't take it back!

Thank goodness VISA allows disputing charges.

Sounds like Baltic is a good alternative. I'll tell my buyer.

I second the motion for some kind of datecode. I don't trust the distributor as it is in their best interest not to handle a return.

Is there a way to know the mfg date by serial number?

======================================
Also, an earlier question was never answered.

Are newer production units built with solid state caps?
======================================

Another 450 dead of capacitor attack today.

I really, really, really wish MT would publish a serial number or mac list of affected production so I/we can be proactive about fixing this.

This is ridiculous.

MT cannot grow up until it figures out how to respond to issues like this.

You cannot simply hide and ignore it.

Step up, take your licks and get on with it!

I have some news. The first three RB450s that we changed caps in, around Oct. 2009, died this week, almost at the same time. They were running with different power supplies (I do not have the type or voltage with me now), and with the same replacement caps, which we bought at the local hobby electronic store (not the cheapest ones of course).

Because they were the first ones, we used them for some testing, and then deployed to some VIPs...

The first one that bit the dust was at my colleague's friend. On the same day my colleague's own one died, with not just the three changed caps bulging, but the fourth original one also. These two boards were running on UPS. Today my girlfriend's parents complained about their router hanging every two or three hours, and after getting them to open the case, they told me that one of the small black items inside was leaking some stuff.

My colleague was planning to replace the board to a 493G in a few weeks, but not anymore.

As we have several repaired items in production, we are getting ready for another capacitor change wave... Getting soldering irons started...

dont solder in just any old cap...

buy a good quality one.

I can recommend either some Rubycon Nichicon Chemi-Con.

you can buy from digikey, or badcaps.net or most electronics retailers.

No, all caps are not created equal.

dont solder in just any old cap...

buy a good quality one.

I can recommend either some Rubycon Nichicon Chemi-Con.

you can buy from digikey, or badcaps.net or most electronics retailers.

The board should not be putting so much demand on the capacitors that only the very best top of the line caps survive.
This kind of behavior seems to suggest a design flaw.

dont solder in just any old cap...

buy a good quality one.

We are using YAGEO, because they are easily available locally. I don't know if they are really so bad.
However, in one of the mentioned units, the only unchanged SU'SCON bulged, and our YAGEOs were intact.
Nota bene: next time changing all the SU'SCONs will be the better decision.
I can't understand why MikroTik is using such crap quality parts for RouterBoards.

Not only RB450 have this problem. Today I replaced two RB433 with bad capacitors ... Normis - do you want photo?

Electrolytic Capacitors in general are just crap

high temperatures will kill them fast as well..

many manufacturers of parts have problems with caps.... very common.

dont solder in just any old cap...

buy a good quality one.

I can recommend either some Rubycon Nichicon Chemi-Con.

you can buy from digikey, or badcaps.net or most electronics retailers.

The board should not be putting so much demand on the capacitors that only the very best top of the line caps survive.
This kind of behavior seems to suggest a design flaw.

Yes! And that is why we switched from 24vdc to 12vdc power supplies for indoor RouterBoards.

Tom

Electrolytic Capacitors in general are just crap

high temperatures will kill them fast as well..

many manufacturers of parts have problems with caps.... very common.

Conspiration theories rock http://en.wikipedia.org/wiki/Capacitor_plague

Electrolytic Capacitors in general are just crap

high temperatures will kill them fast as well..

many manufacturers of parts have problems with caps.... very common.

Conspiration theories rock http://en.wikipedia.org/wiki/Capacitor_plague

So we're going to wikipedia now are we?

Here is some good reads that is directly applicable to this thread.

http://en.wikipedia.org/wiki/Electrolytic_capacitor Esp. last paragraph.

and

http://en.wikipedia.org/wiki/Arrhenius_equation

I've got my RB450G second handly. The one from I bought it said, this one is brand new, he just got this as a replacement, because his unit has failed, probably with this bad capacitor problem.

This was two mounts ago. From this and this post for me it seems, that in production the capacitors had been changed, with a layout modification of course. But the replacement unit is still not equipped with solid state capacitors, and these posts are more then 1 year olds.

What is the current situation then?

the problem was not with capacitors in general, but with a specific bad batch. we have changed the supplier, and there is no more problem. they look very similar to the previous ones, I think only the color and manufacturer info is different

I think it is important to present accurate and complete information.
We have seen written here multiple times by MT that the "supplier" has been changed.
e.g.

...we have changed the supplier...

new batches are already made with different parts and will not have this issue. as soon as we got first problems, we switched supplier immediately

A very small number, we found the problem and changed supplier immediately

The "new supplier" is the same supplier: Su'scon

The "new" capacitors are brown. The old capacitors are green. But they both have "Su'scon" printed on them.
The new capacitors are MG series, 680uF/6.3V; the old capacitors are SJ series, 560uF/6.3V.
Both are aluminum electrolytic. MG series is rated longer at high temperature.
http://www.su-scon.com.tw/products_li.html

I, for one, never said the specs of the capacitor where to blame. I simply pointed out that not only did the manufacturer not change, the part number and series did change. Perhaps that was to make it easy to identify new vs. old.

So now, many months and posts later, we have what appears to be an accurate and complete story:

There is no design problem.
A bad batch of caps - from a previous supplier - not a bad batch from the manufacturer.
DOA, defective and fake caps were mixed with good caps.
The supplier has been changed, the manufacturer has not been changed.
The value, series (and color) and part number has been changed.
The previous supplier has been nailed to a tree. (That part is fiction, yes?)

If someone wants to repair a unit, it would be better to match the specs of the new parts, no?

Replace With os-con's like the new rb units

solid state = much better

I really hope I dont start seeing this problem. I have about 80 450Gs I rolled out in the last 6 months.

I don't think you will. This is the first RB450G that we have seen with bulging caps. I am hoping this is an anomaly.

The 450G power circuit design is much different from the older RB450 power circuit where we have seen all the cap problems.

Tom

Well. I've seen it. About 160 rb450gs in field, 15 493ahs, and a handful of 433ahs. Maybe 50 rb411s. At least dozen 450gs have failed as well as a few 411ars and 3 433ahs. The 493ahs were under warranty still. The others are just out.

We have approx 60 450Gs in operation. Those purchased after January, 2011 are blowing the 6.3v 560uF CAPs. So far, 4 450Gs blown. We're replacing the caps with:

Mouser #: 647-UPW0J561MPD
Mfr. #: UPW0J561MPD
Manufacturer: Nichicon
Desc.: Aluminum Electrolytic Capacitors - Leaded 6.3volts 560uF 8x11.5 20% 3.5LS
RoHS: RoHS Compliant

0.177 each/100 Mouser Electronics.
GAF
AireBeam Broadband

Mikrotik: How about posting a MAC LIST of all of the 450Gs with this batch of Jiffypop Caps!!!
GAF

I really hope I dont start seeing this problem. I have about 80 450Gs I rolled out in the last 6 months.

I don't think you will. This is the first RB450G that we have seen with bulging caps. I am hoping this is an anomaly.

The 450G power circuit design is much different from the older RB450 power circuit where we have seen all the cap problems.

Tom

Well. I've seen it. About 160 rb450gs in field, 15 493ahs, and a handful of 433ahs. Maybe 50 rb411s. At least dozen 450gs have failed as well as a few 411ars and 3 433ahs. The 493ahs were under warranty still. The others are just out.

And what power supplies were you running on these? 24vdc or 12vdc?

We switched to 12vdc for most of the RB4xx family over a year ago and have seen much less cap failure.

Tom

Just had the same thing happen to one of my 450g's this morning. Made it all through the heat of summer and died on a cool morning at 3:50AM

It was powered with 24v. Going to drop to 12v now.

Made for an early start to the day and some angry customers. Not sure when I got this one. It's prob year and a half old. Gonna replace the caps and dump it in a less critical part of the network that has redundancy. So the green capacitors are the older versions?

Caps replaced board lives again Even bough extras in case other fail

Hi ...

Does MT have something like board serial numbers range that was filled with this bad batch? Not only 450s but 411ARs, 433UAHs, 433AHs & 433?

Regards;

I'm sorry but it's not possible to tell an exact range.

Hi.

I see. Well, as soon as I receive a recently ordered 433UAH I'll inspect a production one that since last november is showing some strange behaviour. Once or twice a day it reboots and alarm says "cause 1".

I already swaped power supplies (UBNT 24V 1A, a 19V 5A notebook power supply, POE x power jack, used both ways isolated or at the same time, rebuilt cables, etc).

PCI cards are R52nM (2) and a XR9.

It's not related with CPU load nor online users (peak 70 CPEs divided by 2 R52nM) / heavy traffic.

A RB450 that shares the same nobreak shows an uptime of 144 days.

Regards;

cause1 sounds like a kernel panic, send to support a supout.rif file, so we can see what caused it.

Is autosupout.rif ok? Every reboot like that there is a new one of course.

Regars;

Done.

Anyway, the new 433UAH will arrive next week. Actual uses a plastic MT box and the new one will be assembled in an aluminum weatherproof box. RF poluted area (cell tower 300m away, airplanes crossing above this place around 1000 feet with their transponders spittin' some Kw at 1050MHz besides the ground radar etc).

Regards;

you can also check power budget of board and combined power consumption of the board (on routerboard.com for max power output to extension cards) as that also can be a problem, that with worse conditions, cards go into lower data rates for more TX power - result, not enough power to power card and crash.

you can also check power budget of board and combined power consumption of the board

Well:

R52nM = 1,95W (2W)
R52n = 2,40W (2,5W)
XR9 = 3.3V @ 1.2A = 3,96W (4W)

2+2,5+4 = 8.5W

RB433UAH = 16W to cards.

This is valid to a healthy onboard power supply. But if some bad electrolitic capacitors are there, who knows?

Anyway, a 433AH in a couple of days will replace this 433UAH. Then I'll have chance to search for hw failures at the bench.

Regards;

load seems to be reasonable, check PSU and caps, maybe there is the problem.

MikroTik use capacitor with low peak voltage.
Is crazy to use 6,3v in line of 5v.
I've benn replaced all the capacitors in my RB in high tower with 560uF 10v (doubled than 5v)
and all works ok from 12v to 48v of power supply.
My RB damaged is: RB800, RB450, RB433AH and many RB411AH supplyed at 24v or 48v.

I hope MikroTik engineers change the capacitors and increase flash disk in future RB's (see RB1100AHx2)

OrCAD (master designers)

My RB450G failed on Sunday. I purchased it in July of 2011. No longer under warranty.
The capacitors are bad. The Nichicon UPW series or the new Polymer capacitors are my two choices.
The board made a sizzling noise before it failed.
Mike

My RB450G failed on Sunday. I purchased it in July of 2011. No longer under warranty.
The capacitors are bad. The Nichicon UPW series or the new Polymer capacitors are my two choices.
The board made a sizzling noise before it failed.
Mike

you like these? http://www.mouser.com/ProductDetail/Nic ... 9k%252bUHZ

actually, scratch that, I think those are too long, going back through thread to find exact model microtik uses so I can see length/width

the suscon looks to be 8x12mm and earlier the post said 3.5 lead spacing
http://www.su-scon.com.tw/en/product/Lo ... 5C/MG.html
there isn't going to be a lot of head room in standard case so going over 12mm probably not good

Also, there is a super tiny 10 ohm 1/100 watt resistor that feeds 24vdc to pin 2 of the power supply control chip.
It read 4300 ohms. I removed one from a control pcb from an old broken hard drive and tacked it into the circuit.
It boots up and works fine now.
This resistor is near the power supply control chip, right near pin 2, connects to a 1uf chip cap right there too.
Hope this helps anyone who has a power supply problem that is not bad caps. I would think that this
tiny tiny resistor is too small of wattage value for the application.
MIke

Not sure exactly when I bought this RB750, but it's the second one to fail (at around 6am yesterday) in the exact same place in the network.
On reading this thread, I've swapped out the 24V adaptor that was powering it to a 12V for the replacement RB750. Fun start to a Monday morning...

RB450G just stopped working. Bought on Jul 9, 2010 from rOc-nOc.com. 560uf caps poped and leaked.

try instlall solid state caps instead.
preferably - not oxide caps with polymer dielectric, but tantal, palladium caps and etc kinda expensive stuff.
but thats about [amount of]money[you want invest into improving/modding you unit/router].

Unfortunately reported two RB450G with 2 capacitors burned

Buy 3 in November 2010, 2 were burned in 15 days, as a precaution I'll change the capacitors that still works.

Old thread, but thought I'd report that my trusty RB450G just bit the dust after ~10 years of continuous service in my home.

Opened it up, and the two 6.3v, 630uF caps were bulging a bit.

Replaced them with a couple scavenged 1000uF caps and its back to life!

Hopefully for another 10 years...