Author Topic: Brymen BM235 10A fuse silently destroyed by a switched off UPS leakage current..  (Read 6952 times)

0 Members and 1 Guest are viewing this topic.

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr

. A good UPS should not even have an on/off switch, since that leaves open the possibility of accidentally interrupting the power to the protected devices.

Exactly, especially when it has one and does nothing...
It gives you a false sense of security when it just isolates between live and neutral.
They should have at least put a huge sticker on it with a warning.

I've just checked my UPS on my main computer which is a much more expensive unit
with a pure sinewave output and when you switch it off you have proper isolation
from the mains.  Its off.  You have to press the power off button for at least 20 seconds
in order to switch it off, so you can not turn it off by accident. 
 

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr
Johnny B Good: " Bearing the fact of the single pole changeover switching in the UPS, it's extremely important that you don't reverse the Live and Neutral connection to the mains supply. The better Line Interactive UPSes actually incorporate mains polarity connection fault detection LEDs to warn of such problems. If you have the Live/neutral reversed, this could readily account for the winking LED via the Y cap in the RPis's smpsu wallwart."

viperidae : "Could be the ups is only switching the neutral line. That would explain the voltages."

YES, YES!
That was the problem.
I reversed the mains plug on the wall and everything works fine now.
No mains output when switched off and no blinking LED on the Raspberry Pi.

Thank you!!

They should specifically mention that this UPS needs a proper live/neutral wiring!

« Last Edit: March 06, 2019, 04:51:12 am by hgg »
 

Offline malagas_on_fire

  • Frequent Contributor
  • **
  • Posts: 591
  • Country: pt
  • Kernel Panic
    • Malagas Lair
Normally that feature is on the manual of the UPS or in the low end model of some know brands, eg, APC surge outlets with small batteries:

https://www.apc.com/kh/en/faqs/FA158817/

If one can make knowledge flow than it will go from negative to positve , for real
 

Offline johnkenyon

  • Regular Contributor
  • *
  • Posts: 123
  • Country: gb

 Bearing the fact of the single pole changeover switching in the UPS, it's extremely important that you don't reverse the Live and Neutral connection to the mains supply. The better Line Interactive UPSes actually incorporate mains polarity connection fault detection LEDs to warn of such problems.

Easily retrofittable - wire a red mains neon indicator (with integral resistor) between the incoming neutral and ground., and mount it on the rear or front panel, labelled "L-N reversed" or "P-N reversed"
« Last Edit: March 06, 2019, 05:36:29 pm by johnkenyon »
 
The following users thanked this post: hgg

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr
Not a bad idea at all. 
I think I will do it.
Problem Solved.

One question remains though!

When I connected the 40 watt light bulb from live to ground the RCD tripped immediately.
When I connected the BM235 the fuse blew before the RCD had the chance to trip.
The RCD needs less than 30mA to trip, usually 25 to 28 if its in a good condition.

So the fuse blew with 230V and less than 30mA for a few milliseconds.
Does this make any sense?
Was the fuse faulty or is there something I do not understand?  (probably..   :) )


 

Offline IanB

  • Super Contributor
  • ***
  • Posts: 12537
  • Country: us
It was a fast blow fuse. Maybe the fuse blew before the RCD had time to act?
 

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr
Fast blow fuse, but with how much energy?
From its data sheet : "Intended to carry 100% of rated current indefinitely." which is 11A.

The RCD tripped with less than 30mA @ 230V in a couple of milliseconds.

How much energy did the fuse receive in order to blow before the RCD?
Its obviously less than the energy received by the RCD.

Somebody to do the math?    ???
 

Offline malagas_on_fire

  • Frequent Contributor
  • **
  • Posts: 591
  • Country: pt
  • Kernel Panic
    • Malagas Lair
Look at page 2 of the datasheet , the graph and watch the time vs possible Amps on the fuse and compare to the RCD one  :

http://www.cooperindustries.com/content/dam/public/bussmann/Electrical/Resources/product-datasheets-a/Bus_Ele_DS_1047_DMM-B.pdf

Enough energy to prevent further damaged to meter or user
If one can make knowledge flow than it will go from negative to positve , for real
 

Offline IanB

  • Super Contributor
  • ***
  • Posts: 12537
  • Country: us
The RCD tripped with less than 30mA @ 230V in a couple of milliseconds.

The RCD will also trip with 30 A in a couple of milliseconds. It will trip with 300 A in a couple of milliseconds. How do you know how much current actually was present when the RCD tripped and when the fuse blew? (You said a 40 W bulb tripped the RCD, so that was 40 W / 230 V = 170 mA. Already much more than 30 mA.)

Bear in mind that based on the previous comments in this thread where you had the mains polarity reversed, it appears you had the full power of the mains live conductor available to supply current. If you connect that to earth/ground you could expect to see hundreds of amps flowing in cases where the earth conductor is bonded to neutral at the supply.

Since you blew a 10 A fuse, you can deduce that much more than 10 A flowed through the fuse at the time it opened.
 

Offline malagas_on_fire

  • Frequent Contributor
  • **
  • Posts: 591
  • Country: pt
  • Kernel Panic
    • Malagas Lair
The BM235 has low-z mode but be carefull AutoV (LoZ). Check page 7 of its manual .
If one can make knowledge flow than it will go from negative to positve , for real
 

Offline Electro Detective

  • Super Contributor
  • ***
  • Posts: 2715
  • Country: au
Low Z mode is great for stray leakage and confirming capacitive ghost voltages, but will trip the RCD every time, unless DUT is isolated from the mains.
FWIW: some meter brands need to cool down (up to one hour) after using Low Z mode on mains voltages,
as resistance readings in Ohms mode may not be stable or accurate with the user believing he/she has a lemon meter   :-\
 

I've played with UPS units in the past and noticed on some that the input Live and Neutral can be wrongly wired by the manufacturer,
especially "international market wiring colour scheme" snafus > black and white vs brown and blue vs red and black
vs drag revue pink and purple   :palm:

AND I've observed Live and Neutral flipped/reversed on the UPS output during a power outage as referenced to the input sine wave 

Some can be wired properly so that input and output POLARITY is correct, with some expected voltage/current phase shift difference expected

By saying 'Some can wired be properly' I mean if the rewire does not affect the battery charging circuit and battery if there are common connections.

If the UPS output runs through an isolation transformer, like some APC SmartUPS and similar Compaq/HP IIRC, those are easy to correct -if- they are out of wack.
Yes, some are, most not, luck of the draw.
Still, their transformer output is isolated (many times via two 120v legs plus CT feeding 240v to L and N output) with earth/ground still connected at the wall socket,
so the connected load/equipment won't care



All my blab above assumes correctly wired wall socket/s, IEC kettle cord leads, decent power strip boards,
with no moisture or short causing dirt/metal particle ingress anywhere.
and non flipped or poorly repaired DIY extension cords (classic widow makers)   :scared:


@ hgg: we may never know if the meter fuse blew because of what happened somehow with the UPS, or from a previous test, a severe meter drop/vibration, 
or the fuse was a rare dud,
or someone borrowed the meter and zapped it with voltage on the current setting, whatever... 

The good news is that it got you pumped to find out what the go was, and corrected a far bigger problem/issue
...than worrying about the euros you will blow on another fuse     :'(
   
YES, they are WAY-Y-Y overpriced for what they are,
another necessary credit card rape item
but still a bit cheaper than buying a replacement meter and or blowing money at a hospital/funeral

...and the sellers know it   8)


The sneaky cheaper knockoff fuse floggers aren't doing the electronics community any favours either  **** them    :--
« Last Edit: March 06, 2019, 10:39:17 pm by Electro Detective »
 

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr
If I read the fuse datasheet correctly, it can hold 125 Amps for 40ms.
The 30mA RCD can hold for the same amount of time 150mA (500% of its rated trip current)

The 40W incandescent lightbult had an initial cold resistance of 97Ω, which means that
it drew 2.37Amps when it was connected between live and ground.

When the fuse was connected from live to ground it was a dead short (almost)
How can we calculate the amperage at the time of the short?
The device was connected to a 10A fuse which did not blow.

(I am trying to understand because there was not the tiniest spark when I connected the
multimeter probes and one probe was barely hanging by itself on the ground terminal.)
 

Offline Cnoob

  • Regular Contributor
  • *
  • Posts: 132
  • Country: gb
Quote
(I am trying to understand because there was not the tiniest spark when I connected the
multimeter probes and one probe was barely hanging by itself on the ground terminal.)

That could suggest the fuse was already broken.
And why the RCD didn't trip it could be that when you shorted it out there was already another current path to complete the circuit.

You tripped the RCD with a light bulb, you may have blown a expensive fuse by shorting mains level voltages and you wired up your UPS
incorrectly.

To me that suggests you should leave mains and mains level voltages alone.
 

Offline Jeroen3

  • Super Contributor
  • ***
  • Posts: 4209
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
If I switch off my UPS, it goes into bypass mode. Maybe only interrupt neutral indicates a problem inside the UPS?
If I want to power down anything behind my UPS, I have to switch if off, and unplug it.
 

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
RCDs have TWO parameters, a rated trip current, often 30mA (actually this is a 'must trip' level, they generally go at half of this), AND a trip time....
Common examples have a must trip time of 30ms, and will generally go in 20ms.

These two parameters mean that there is ample time for a stiff mains supply to eat a 10A fuse  and clear the fault before the RCD trips. Yea, sure, 1kA or so of PSC (Could be much higher if close to the incomer and using heavy cables) is way more then 30mA, but if the fuse opens and quenches the arc in under 10ms the RCD will probably not trip because the fault has not been sustained for long enough. 

You can actually buy time delay RCDs which are usefully deployed upstream of the usual sort as they provide substantial fire risk reduction while still providing the discrimination to avoid tripping the whole installation due to some stupid downstream fault that should just trip the local one. Standard sizes for these are (100mA, 100ms), (300mA, 300ms) and the biggie for large three phase jobs, (3A, 3 seconds).

Note that a 30ms delay 100mA RCD does not provide useful discrimination with a 30mA, 30ms RCD because  it is a crap shoot as to which one will trip first on a earth fault of more then 100mA.

The fuse did its job, and THIS is why meters have expensive fuses on the high current ranges. 

Regards, Dan.
 
The following users thanked this post: hgg, Electro Detective

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr
"1kA or so of PSC" and not even the smallest spark on the multimeter leads?
 

Offline Cnoob

  • Regular Contributor
  • *
  • Posts: 132
  • Country: gb
I look at Bussmann's graph for that particular fuse which is call a clearance graph of time v current.
Assuming I reading the graph right.

I'm reading that graph as the time taken for the fuse to clear a fault condition and as this was a short circuit near instantaneous currents are involved. How high ? it depends on your impedance of your voltage source. I think it's safe to assume it's a high current and on the graph. A near instantaneous current of 150 Amps the fuse will act in 15mS. I suspect the fuse could act quicker but Bussmann are erring on caution.

So it is quite possible for that fuse to blow before the RCD can trip.
« Last Edit: March 07, 2019, 01:55:17 pm by Cnoob »
 
The following users thanked this post: hgg

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Yep, RCDs are not (Unless RCBOs) overload sensitive, they will pass a lot of current and merely get hot, overload protection is a different type of breaker.

I would expect a LOT more then 150A in a dead short to 'earth' situation assuming TN-C-S or TN-S supplies are in use (TT, the fuse might not even have blown in a week! This is NOT a good thing!).

Like I say, THIS is why good multimeters have expensive fuses, quite possible a kA or so flowed, but probably the only part that failed was the fuse, and it kept all the damage contained inside the fuse housing, quite impressive actually.

Regards, Dan.
 
The following users thanked this post: hgg

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr
Do you think that not seeing any sparks was expected ?
 

Offline IanB

  • Super Contributor
  • ***
  • Posts: 12537
  • Country: us
"1kA or so of PSC" and not even the smallest spark on the multimeter leads?

Do you think that not seeing any sparks was expected ?

You tend to get more of a spark when breaking a circuit than when making it. I have shorted 240 V mains a couple of times in the past, and it was not very dramatic. I hardly noticed I'd done it until afterwards, when the fuse had blown or the breaker had tripped. Each time I had just barely touched a live wire against a grounded terminal and that was enough to trip the circuit protection.
 
The following users thanked this post: hgg

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr
So it looks like that all this shows the safety of a good multimeter.

I bought 2 of the ebay fuses but I am 99% sure they are fake.
I just want to open them and compare them with the original.
The original costs a Chinese multimeter..   :)
 

Offline Electro Detective

  • Super Contributor
  • ***
  • Posts: 2715
  • Country: au

Test them to failure (SAFELY) and take notes,

then open them up and see how well the fuse housings and materials fared.

If 'BS' spec 10 amp fake cheapie fuses blow at 8 amps instead of 10 amps reasonably fast, at a fraction of the cost,
they are cheap and sort of do the job for the price, and a better deal than no fuse (or galvanised fencing nail  :palm: ) hey what do you want?  :D

If they are labelled 10 amps and blow at 12 or 15 amps, or your meter smokes first,  :-\
don't waste your cash..and meter   :phew:

 
 

Offline rsjsouza

  • Super Contributor
  • ***
  • Posts: 6106
  • Country: us
  • Eternally curious
    • Vbe - vídeo blog eletrônico
So it looks like that all this shows the safety of a good multimeter.
I think that is the most interesting conclusion of all this. Lots of folks argue how idiotic or exaggerated our claims are for meter safety around this forum, but this scenario is a perfect example of how even a CAT I system can be quite powerful when doing a simple measurement.

Would an el-cheapo meter or a fake HRC fuse kill you in the same scenario? Probably not, but it would give you a scare at a minimum and (big maybe) injure you if the case couldn't contain the energy.

Regarding arc flash on regular outlets, I have had my share of very impressive events on both the typical 15A circuit and some higher ones (40, 80A).
Vbe - vídeo blog eletrônico http://videos.vbeletronico.com

Oh, the "whys" of the datasheets... The information is there not to be an axiomatic truth, but instead each speck of data must be slowly inhaled while carefully performing a deep search inside oneself to find the true metaphysical sense...
 

Offline malagas_on_fire

  • Frequent Contributor
  • **
  • Posts: 591
  • Country: pt
  • Kernel Panic
    • Malagas Lair
I had one cheapo meter exploded while probing mains in Volts a long time ago, which gave a good jump scare  (PS this are the older remains stored in garage so there might be some spider webs) : :

[Edit here they are details :P Don't blame the spiders it happened well before they nest ...]

https://www.eevblog.com/forum/testgear/new-aneng-q1-9999-counts/?action=dlattach;attach=458422

https://www.eevblog.com/forum/testgear/new-aneng-q1-9999-counts/?action=dlattach;attach=458389

https://www.eevblog.com/forum/testgear/new-aneng-q1-9999-counts/?action=dlattach;attach=458395
« Last Edit: March 11, 2019, 08:29:49 pm by malagas_on_fire »
If one can make knowledge flow than it will go from negative to positve , for real
 

Offline hggTopic starter

  • Frequent Contributor
  • **
  • Posts: 960
  • Country: gr
-  "Regarding arc flash on regular outlets, I have had my share of very impressive events on both the typical 15A circuit and some higher ones (40, 80A)."

Do you mind sharing the details?   :)


-  "Test them to failure (SAFELY) and take notes,
then open them up and see how well the fuse housings and materials fared."

Will do.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf