Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

[Per Hansson]

I've got a 100/100Mbps connection in my apartment, been real satisfied with it.
It's connected to a Linux firewall acting as a router and WiFi access point.
This Sunday morning it abruptly stopped working though, firewall reported that the NIC dropped the link.

Only positive thing with the power supply was that it was held together with a screw.
It has no S marking and not even a faked CE marking, only the CCC marking but that's only for China. It also has a SA mark but I think it might be fake.
Especially seeing how they even managed to make the polarity drawing incorrectly, I mean really the whole PSU is a big short circuit!?
Main cap had failed, managing to kill the fuse in the unit.
After finding a new cap I went looking for a fuse but found a much nicer D-Link PSU with twice the capacity and all Japanese UCC caps.
So I moved the cable over to that and installed it, now the fiber converter worked great again!

[momaka]

Re: Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

Badcaps strikes again . And that is one shitty-looking power adapter! I'll give the manufacturer some credit, though - at least it has an input common-mode RF choke.

[Per Hansson]

Re: Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

Yea it really is crap, desoldered the secondary caps now too, they had started to fail as well, 0.15 vs 0.25 ESR in them...
It's almost amazing how cheaply built it is, yet they still went to the trouble with some safety caps and filtering, it's puzzling!
What's more puzzling is an ISP using this shit in their equipment, think how many unnecessary service calls and service visits it creates!
I guess "a cent saved but a dollar wasted" is a logic they live by....

[mariushm]

Re: Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

The company making the fiber converter doesn't make the power adapter, they just buy it from a third party because they don't want to bother with certifications, approvals, ul listings etc, they're not a power supply company.
In theory, they would pack the converter with a quality power adapter but some cheapen out.

A few years ago when I moved into a new apartment, I was lucky to have an installer from the ISP that actually lived in the apartment building my apartment was in, so he was in a talking mood.

I asked about the cable modem quality and what fails most often.. with those modems, he said the power adapters were the main cause of failure. But, with every box of 20 cable modems they bought, they received 24 or 25 power adapters in the package so they didn't care. If user was complaining, they came with new router and power adapter and replaced them on the spot, a 5 minute job, so it didn't cost the company much. At their service center, the old modems were tested and if good, put back in service to other customers.

The new cable modems come with serious power adapters, they don't get hot so I think they'll last quite a lot.

[momaka]

Re: Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

It's almost amazing how cheaply built it is, yet they still went to the trouble with some safety caps and filtering, it's puzzling!

^
That is quite funny indeed.

Definitely a candidate for the the gutless, bloated, and fried power supply hall of shame thread.

What's more puzzling is an ISP using this shit in their equipment, think how many unnecessary service calls and service visits it creates!
I guess "a cent saved but a dollar wasted" is a logic they live by....

Probably a decision made by the a manager or some bean counter rather than the engineering department at the ISP company. I've seen this practice a lot in IT as well. A power supply or battery for something dies, and the usual solution is to go online and buy the cheapest item that comes up first.

[Per Hansson]

Re: Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

So I borrowed a ladder from a neighbor to look at the fiber converter which the landlord had conveniently installed in the attic...
All I could see was that mine was dead but my neighbors was working.

What's more puzzling is an ISP using this shit in their equipment, think how many unnecessary service calls and service visits it creates!
I guess "a cent saved but a dollar wasted" is a logic they live by....

Heard some noises from my attic yesterday, turns out my neighbors fiber converter had finally failed three weeks ago and a local electrical contractor was here replacing it
Spoke to the guys that were here, they had a refurbished one from the ISP with them...
They gave me the old failed one to have a look at, same 400v "XJ" bulk capacitor and same failure mode, who could have guessed
They came back only an hour or so after having replaced it, apparently the "refurbished" one was not working.
So they gave me that one too and guess what? Yea same cap failed again! Third's time the charm right?
I took a look at some more of those he had in his car, at least a few had a redesigned PSU with more safety markings.
They had removed the choke though so the credit momaka gave them needs to be retracted!

The company making the fiber converter doesn't make the power adapter, they just buy it from a third party because they don't want to bother with certifications, approvals, ul listings etc, they're not a power supply company.

I actually tried to locate the manufacturer of the power supply, on the PCB is written: 中山泽龙电子
But turns out that is just a company that you can custom order PCB's from, i.e. not the real manufacturer...

This time I also took a look at the PCB inside the fiber converter, it uses the same Chengx capacitors with no model marking rated for 105°C, 4x 100uF 16v 5x12mm and 1x 1000uF 16v 10x13.5mm.
The small caps ranged in ESR between 0.35 > 0.80 while the large one in each converter was 0.1 vs 0.6 respectively, so one converters caps had degraded significantly more than the others.
So we can conclude that at least low ESR caps should be used, but good luck finding an actual datasheet for these caps

[momaka]

Re: Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

So they gave me that one too and guess what? Yea same cap failed again! Third's time the charm right?

They had removed the choke though so the credit momaka gave them needs to be retracted!

Well, I can't edit my post anymore, but I do (verbally) take all I said back .

[mariushm]

Re: Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

Well, the fiber converter looks OK.

You have the input which is 5v, the voltage goes through D1 as a basic reverse voltage protection so the whole fiber converter will already see only about 4.3-4.6 volts (due to voltage drop on the diode), then you have the L1 inductor doing some small filtering and you have the big capacitor doing bulk energy storage and then the power goes into the linear regulator AIM 1084 something.

That's an adjustable linear regulator, you can figure out the voltage based on the two small resistors above it (RP1 and RP2 in the picture).. basically the output voltage will be 1.25 * ( 1 + R1/R2).. My guess the output voltage for that linear regulator will be 3.3v or lower, because the linear regulator has an internal voltage drop of about 1-1.1v so at best the output voltage will be 4.5v - 1v = 3.5v

Anyway, the point is that only that big 1000uF capacitor kinda has to have low esr as the energy comes from a switching power supply (the power adapter). The designers chose well a capacitor rated for 16v even though the connector is supposed to take in just 5v (so in theory you could use a 6.3v or 10v capacitor).

If there isn't any other circuit using the raw voltage of 5v (you could check by following the traces from the connector) then in theory you could have powered the fiber converter even from a more common 7.5v DC power adapter but that has the potential to overheat the 1084 linear regulator.

The linear regulator would have a relatively clean output voltage so there's no need for low esr capacitors after it, basically it makes no difference what you use after. In fact, some linear regulators actually need some ESR on the output, some regulators specify in datasheets that capacitors should have esr above 0.1 ohm or something like that.
So basically don't rush to change those small capacitors on the fiber converter, the esr is normal for such small value capacitors and for how they're used in their circuit (to the best of my knowledge and limited analysis from the single picture posted)

[Per Hansson]

Re: Fiber Converter NMN-M19-CN100-S20SC Power Supply Dead

Thanks for your insight mariushm, I should clarify better what I meant in the conclusion of my last post.
One of the fiber converters had degraded more than the other, so what I meant was that I think the lower of the ESR readings I got;
0.1 for the big cap and 0.35 for the small caps is in the ballpark of what they should be...

I attached a new top-down picture of the unit, the linear regulator is an AIC 1084PM RP1 & RP2 are 100Ω each.
The converters PCB is designed to accept both SMD or through hole caps, and since the latter was used it's very easy to check the voltages without having to turn the PCB around
The large bulk cap has 5v across it.
CP3 & CP4 which are the two small caps near the input have 2.5v across them.
The other two small caps that are near the fiber connector have 5v across them.
The controller is a IC+ IP113C LF and it needs 2.5vdc