APC battery backup not working sometimes

[Peabody]

This is an APC Model BE550G UPS. It's just a few months old, but
has developed an intermittent problem that I'd like to fix if I
can. I know if I send it back to APC it will just happen to be
working then, and I will have accomplished nothing.

The problem is that sometimes when the UPS switches into backup
mode, no power gets to the plugged-in devices. It's apparent the
UPS thinks it's supplying backup power (4 beeps every 30 seconds,
just like the manual says), but it isn't. The battery shows as
fully charged, with 27 minutes of run time available. And when it
works, it works fine. So it's not the battery.

I first noticed this during powerup of the UPS. At first,
everything is on mains AC. Then it switches into backup mode as a
test for a few seconds, then reverts back to mains AC. What I
noticed was that my modem would turn on at first, but then turn
back off when the UPS went into test backup mode, then turn on
again at the end of the test period. Sometimes. Most days, but
not every day.

Then of course I tried a full test - unplugging the UPS from the
mains, and all the devices shut down instead of continuing to run
under backup power. But all of this is intermittent. Sometimes
when I do the test, it works fine.

I called support, and they had me reset the sensitivity of the UPS,
and of course it started working then. But I strongly suspect this
has nothing to do with sensitivity - after all, it's hard not to
notice the mains dropping out completely.

Logically, what this seems like to me is that everything is working
ok, but the relay which switches over to supplying backup power to
the devices has bad contacts that don't always make contact. I can
hear the relay click when it goes into backup, and again when it
reverts to mains, and I can hear that whether the backup power gets
through or not.

Well, I just wondered if anyone here has any experience with this
kind of thing, and might offer some advice on what to try. I have
already opened it up and made sure all the internal cable contacts
are good, even tightened a couple just to be sure, but that didn't
help.

Any advice would be appreciated.

 

[Jeff Liebermann]

On Mon, 09 Jul 2012 12:48:08 -0500, Peabody
<waybackNO746SPAM44@yahoo.com> wrote:

This is an APC Model BE550G UPS.

<http://www.apc.com/resource/include/techspec_index.cfm?base_sku=BE550G&tab=features>
Not the best UPS on the planet, but cheap at about $70.

It's just a few months old,

If "a few months" means 2 months, it should still be under the 3 year
warranty from APC.
<http://www.apc.com/site/support/US/en/product/BE550G/>

but
has developed an intermittent problem that I'd like to fix if I
can. I know if I send it back to APC it will just happen to be
working then, and I will have accomplished nothing.

How do you know? I sent a UPS back because the stupid light was
burned out. (Long story, don't ask). I got a refurbished UPS back in
the mail in 3 days. Instead of guessing, maybe reading the warranty
might help?
<http://www.apc.com/site/support/index.cfm/warranty-services/in-warranty-services/>
How Does Standard In-Warranty Service Work
Defective products may be returned to APC by the customer for
replacement or repair during the standard warranty period.
Defective products will be replaced or repaired and returned
to the customer. Customers who must have original units back
rather than replacement units due to assigned asset tags
and set depreciation schedules must declare such a need at
first contact with APC. It will be APC's standard practice
to issue factory reconditioned replacement units in exchange
for original units. The warranty on reconditioned units
delivered for in-warranty service is 90 days or the remainder
of the original warranty period, whichever is longer. Repaired
units carry the balance of their original warranty period.

So it is written, so it must be. (Charlton Heston in The 10
Commandments)

The problem is that sometimes when the UPS switches into backup
mode, no power gets to the plugged-in devices.

Been there. There's a transfer relay inside. The contacts are not
making good contact or the contacts are pitted from trying to switch a
potential overload such as a laser printer.

I called support, and they had me reset the sensitivity of the UPS,
and of course it started working then. But I strongly suspect this
has nothing to do with sensitivity - after all, it's hard not to
notice the mains dropping out completely.

Well, at least you tried calling support. Support runs on the
assumption that nothing is wrong and that you must have done something
strange to cause the failures. That's not unreasonable as 80% of the
devices returned under warranty have nothing wrong with them.

Logically, what this seems like to me is that everything is working
ok, but the relay which switches over to supplying backup power to
the devices has bad contacts that don't always make contact.

Yep, you got it.

I can
hear the relay click when it goes into backup, and again when it
reverts to mains, and I can hear that whether the backup power gets
through or not.

Well, if you're ambitious, open the box and look inside. If the relay
is an open frame type (fairly common), you should be able to inspect
the contacts. If the relay has a cover, it can usually be removed to
reveal the contacts. My guess(tm) is that you'll find pitted contacts
or a mangled relay armature.

Well, I just wondered if anyone here has any experience with this
kind of thing, and might offer some advice on what to try. I have
already opened it up and made sure all the internal cable contacts
are good, even tightened a couple just to be sure, but that didn't
help.

You opened it up and you didn't look at the contacts? Sigh. Unplug
the unit. Turn it off. Disconnect one lead from the battery. Tear
it apart again. Look at the contacts (with a magnifier). Are they
pitted? If so, clean them.

Any advice would be appreciated.

Well, if you want to risk you life, run the UPS with the covers off.
If it goes into backup mode without any output again, push on the
relay armature with an INSULATED wooden or plastic stick and see if
the output magically appears on the load.

--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com jeffl@cruzio.com
# http://www.LearnByDestroying.com AE6KS

 

[Peabody]

Jeff Liebermann says...

I can hear the relay click when it goes into backup,
and again when it reverts to mains, and I can hear that
whether the backup power gets through or not.

Well, if you're ambitious, open the box and look inside.
If the relay is an open frame type (fairly common), you
should be able to inspect the contacts. If the relay
has a cover, it can usually be removed to reveal the
contacts. My guess(tm) is that you'll find pitted
contacts or a mangled relay armature.

Thanks very much for your reply, Jeff.

The relay is a PCB style DPDT:

Goodsky
EMI-SH-212DCF
12V coil
8 amp contact rating at 250V

I found the datasheet for this series relay, but not the
exact "DCF" part, at:

http://www.goodsky.com.tw/en

searching on the part number EMI-2P over on the left.

But, I haven't been able to find anything equivalent at
Mouser.

The relay is encased in plastic, so I doubt I can get
inside, but I haven't desoldered it yet.

I found 10 with a similar number for sale on Ebay for $79,
but no singles.

This is a really dinky little relay to be rated at 8 amps,
which is probably why it is failing.

If I could find a replacement relay, it would probably cost
less to buy that than to ship the UPS back to APC - it
weighs a ton.

Of course while it would be nice, it wouldn't absolutely
have to physically fit - I could bridge from the PCB holes
to the relay with wire - but the coil and the rating would
have to be right, as well as the speed I assume.

Well, let me know if any solutions come to mind. Thanks
again for your help.

 

[Peabody]

Peabody says...

But, I haven't been able to find anything equivalent at
Mouser.

Ok, I may have lied.

On page 1896 of the pdf catalog, I find:

Omron G2RL series
8 amp rating, same size and pinout

Omron part: G2RL-24-DC12

Mouser part: 653-G2RL-24-DC12

Looks like this should work. The only question would be if
there is a speed issue - how fast it switches.

 

[Jeff Liebermann]

On Tue, 10 Jul 2012 19:19:27 -0500, Peabody
<waybackNO746SPAM44@yahoo.com> wrote:

The relay is a PCB style DPDT:

Goodsky
EMI-SH-212DCF
12V coil
8 amp contact rating at 250V

<http://www.goodsky.com.tw/general-control/52-emi-2p>
<http://www.goodsky.com.tw/files/99-101_EMI-2P.pdf>

Decoding the part number:
EMI = Appliance relay
SH = Wash tight and flux proof
2 = 2 poles
blank = Form C (DPDT)
12 = 12V
D = DC coil
C = Silver-Cadmium-Oxide burn resistant contacts
F = 250V insulation

The relay is encased in plastic, so I doubt I can get
inside, but I haven't desoldered it yet.

True. It's sealed for flux washing and isn't going to come apart.
That doesn't stop you from unsoldering the relay, and using a hack saw
blade to cut a grove around the base to remove the cover. The inside
is NOT potted.

I would just love to replace the relay with a Solid State Relay,
that's not going to work. It has to be DPDT and all SSR's are either
SPST or DPST. That would require a redesign.

Digging through Digikey[1], I find (in order of increasing price):
<http://www.digikey.com/product-detail/en/RT424012/PB967-ND/1095291>
<http://www.digikey.com/product-detail/en/RTE24012/PB988-ND/1095310>
<http://www.digikey.com/product-detail/en/RT424012F/PB968-ND/1095292>
<http://www.digikey.com/product-detail/en/RTE24012F/PB296-ND/254513>
<http://www.digikey.com/product-detail/en/G2RL-2%20DC12/Z2926-ND/368870>
<http://www.digikey.com/product-detail/en/RTE25012/RTE25012-ND/1427674>
<http://www.digikey.com/product-detail/en/G2RL-24%20DC12/Z147-ND/307667>
etc...
I haven't checked the pinout and spec sheets on all of these to make
sure they're identical. Watch out for the contact material and
switching speed. The original is 12msec max operate and 8msec max
release. The first (and cheapest) one in the list is 8msec and 6msec
respectively, which should be faster than the original. However, the
contacts are AgSnO2 which is better than AgCdO contacts of the
original.
<http://www.gljg.com.cn/en/lvc_sto.html>
Anyway, spend some time digging through the data sheet and see if I
misssed anything.

Also, if you have room, you might consider using a socket:
<http://www.digikey.com/product-detail/en/RP78602/PB525-ND/365958>

Good luck.


[1] If you're using Digikey's relay search page:
<http://www.digikey.com/product-search/en/relays/power-over-2-amps/1049447>
you can select more than on filer criteria in a given column by
holding down the shift key.


--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

 

[Peabody]

Jeff Liebermann says...

http://www.digikey.com/product-detail/en/RT424012/PB967-ND/1095291

Watch out for the contact material and switching speed.
The original is 12msec max operate and 8msec max
release. The first (and cheapest) one in the list is
8msec and 6msec respectively, which should be faster
than the original. However, the contacts are AgSnO2
which is better than AgCdO contacts of the original.

Well, that one is silver/nickel per the data sheet. The
AgSnO2 version would be RT423012, which DigiKey doesn't
stock.

But this one:

http://www.digikey.com/product-detail/en/RTE25012/RTE25012-ND/14276

74

is silver/nickel with gold plating - per the datasheet.
Would that be worth a couple extra bucks?

 

[Jeff Liebermann]

On Wed, 11 Jul 2012 14:48:14 -0500, Peabody
<waybackNO746SPAM44@yahoo.com> wrote:

Jeff Liebermann says...


http://www.digikey.com/product-detail/en/RT424012/PB967-ND/1095291

Watch out for the contact material and switching speed.
The original is 12msec max operate and 8msec max
release. The first (and cheapest) one in the list is
8msec and 6msec respectively, which should be faster
than the original. However, the contacts are AgSnO2
which is better than AgCdO contacts of the original.

Well, that one is silver/nickel per the data sheet. The
AgSnO2 version would be RT423012, which DigiKey doesn't
stock.

AgNi looks OK for contact material.
<http://www.gljg.com.cn/en/lvc_sn.html>
It has a bit higher resistance than AgSnO2 and will
therefore run warmer if you load the relay to the
maximum. However, since the UPS is rated at 550VA at
115VAC, you're limited to about 5A max on the contacts
anyway. That should work.

But this one:
http://www.digikey.com/product-detail/en/RTE25012/RTE25012-ND/1427674
is silver/nickel with gold plating - per the datasheet.
Would that be worth a couple extra bucks?

It's a bit difficult to decode what the data sheet says
about the contact material. It sorta hints that there
are several different types of contact material, but
doesn't offer any part numbers to distinguish them.
Dunno.

If it's gold, I would suggest that it's a bad idea.
Gold plating is only for a "dry" load. It's primary
purpose is to provide low contact resistance in circuits
that do not carry any current. The slightest arc across
the contacts, the few millionths of gold plating instantly
disappears, leaving a rather miserable connection to the
intermediate nickel plating between the gold plating and
copper contacts.

--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com jeffl@cruzio.com
# http://www.LearnByDestroying.com AE6KS

 

[Peabody]

Jeff Liebermann says...

http://www.digikey.com/product-detail/en/RT424012/PB967-ND/109529
1

I ordered this one today from Digikey. Came to $5.27 with
shipping. It should be here Saturday or Monday, and I'll do
the switch-out and see what happens.


I guess there's always the possibility that the problem will
turn out to be something else, but the fact that the problem
is intermittent, and that the UPS behaves normally in all
other respects, makes it seem likely that the relay contacts
are responsible. But, you know, it's only $5 either way.

I really appreciate your help with this.

 

[Peabody]

Peabody says...

Jeff Liebermann says...

http://www.digikey.com/product-detail/en/RT424012/PB967-ND/1095291

I ordered this one today from Digikey. Came to $5.27 with
shipping. It should be here Saturday or Monday, and I'll do
the switch-out and see what happens.


I guess there's always the possibility that the problem will
turn out to be something else, but the fact that the problem
is intermittent, and that the UPS behaves normally in all
other respects, makes it seem likely that the relay contacts
are responsible. But, you know, it's only $5 either way.

I really appreciate your help with this.

The relay arrived this morning, and I've installed it, and everything
works fine so far. Of course I don't know for sure that the relay was
the problem because it was an intermittent problem to begin with. But
I've gone through about a dozen tests with the new one and it hasn't
failed to work. So I'd say this looks promising.

I have to keep the old relay in case I ever need to send the unit in
under warranty. So I won't be taking it apart. But, it does rattle
when I shake it, which I think is not a good sign.

Anyway, time will tell. Thanks again for your help.

By the way, I assume the lead-free solder they're using now has a
higher melting point than the 60/40 I use - it was a bear to get the
thing unsoldered with my 30W iron. Ended up having to add some of my
solder to the pins, then braid it off, two or three times before it all
wicked up and the holes opened up.

And Digikey is amazing. I ordered late Thursday afternoon, and it
arrived in the mail Saturday morning, all for $2.20 postage. Hard to
beat that.

 

[Jeff Liebermann]

On Sat, 14 Jul 2012 12:40:32 -0500, Peabody
<waybackNO746SPAM44@yahoo.com> wrote:

The relay arrived this morning, and I've installed it, and everything
works fine so far.

Congrats, so far.

I have to keep the old relay in case I ever need to send the unit in
under warranty. So I won't be taking it apart. But, it does rattle
when I shake it, which I think is not a good sign.

It might be a spring. Some DPDT relays have two springs, one of which
might have come loose. With a spring missing, the tension on the
normally closed contacts will be reduced, possibly causing the
symptoms.

By the way, I assume the lead-free solder they're using now has a
higher melting point than the 60/40 I use - it was a bear to get the
thing unsoldered with my 30W iron. Ended up having to add some of my
solder to the pins, then braid it off, two or three times before it all
wicked up and the holes opened up.

A 30 watt iron will work, but I prefer something with a thermostatic
control and at least 60 watts. The idea is to solder hot and quickly,
so that the heat affected zone is minimal. A low wattage iron may get
up to temperature eventually, but I'm into instant gratification.

Mixing unleaded and lead-tin solder is a bad idea. However, I do it
all the time because I absolutely detest unleaded. I clean off as
much solder from the pads and pins with a solder sucker, and then
apply the lead-tin solder. If you get a dull surface finish, you
still have some residual unleaded solder in the mix. Suck and remove
it all, and try again.

I use solder braid only when I can't get a solder sucker onto the PCB.
The problem is that it takes much more heat to remove solder with
braid. If the PCB has tiny traces, or is made from phenolic, the
traces will delaminate. That can happen with a solder sucker, just
less often. If I have to unsolder many pins, I use an antique
desoldering station:
<http://802.11junk.com/jeffl/crud/pace-desoldering-station.jpg>

And Digikey is amazing. I ordered late Thursday afternoon, and it
arrived in the mail Saturday morning, all for $2.20 postage. Hard to
beat that.

Yep. Great service from Digikey. Mouser, Allied, and Newark are
other good sources of parts and pieces. I prefer Digikey because I
like their online search tool which came in handy trying to find the
relay. However, I will confess that I buy quite a few parts on eBay,
because it's much cheaper.


--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

 

[Peabody]

Jeff Liebermann says...

The relay arrived this morning, and I've installed it,
and everything works fine so far.

Congrats, so far.

I have to keep the old relay in case I ever need to
send the unit in under warranty. So I won't be taking
it apart. But, it does rattle when I shake it, which I
think is not a good sign.

It might be a spring. Some DPDT relays have two
springs, one of which might have come loose. With a
spring missing, the tension on the normally closed
contacts will be reduced, possibly causing the symptoms.

Well I'm going to declare victory and withdraw from the
field of battle.

It's been a week now, and the UPS has passed the boot test
every morning without fail. And as luck would have it, in
our 100-degree weather there have been two short-term power
outages (two that I know of because I was at the computer
each time) and the UPS switched into backup mode perfectly
each time.

So it appears the problem definitely was the relay, and my
replacement operation was successful. And all for less than
the cost of shipping the UPS back to APC. Plus, if
something goes wrong in the future, I'll know what to do.

Thanks very much for your help, Jeff.

 

[Wolfgang Allinger]

On 20 Jul 12 at group /sci/electronics/repair in article 20120720-205124.333.0@news.astraweb.com
<waybackNO746SPAM44@yahoo.com> (Peabody) wrote:

Well I'm going to declare victory and withdraw from the
field of battle.

It's been a week now, and the UPS has passed the boot test
every morning without fail.
[...]
So it appears the problem definitely was the relay, and my
replacement operation was successful.
[...]
Thanks very much for your help, Jeff.

THX for the feeback.


Saludos (an alle Vernünftigen, Rest sh. sig)
Wolfgang

--
Wolfgang Allinger, anerkannter Trollallergiker reply Adresse gesetzt!
Ich diskutiere zukünftig weniger mit Idioten, denn sie ziehen mich auf
ihr Niveau herunter und schlagen mich dort mit ihrer Erfahrung!
(lt. alter usenet Weisheit)