Opinions on adding fuses to power amp

[N Cook]

400 watt RMS amp for use on 240 V mains.
Destructed due to metal dropping in and shorting the amp .
The only fuse is on the mains rated at 8 amp, maker's design rating, which
shows no sag , discolour or anything like that, ie untouched.
Shorted power trannies, burnt low power trannies and even a piece of 3mm
trace from the main bridge rectifier burnt through and that had 2 runs of
added solder over the track for current carrying, the other polarity trace
overheated but not ruptured
I'm thinking of bridging that gap with a fuse and another on the other rail
after cutting it.
The mains transformer is rated at 2x 47V,5 amp.
So drop the 8A fuse (for 2 KWatt !) to what value ? and the 2 added fuses of
what rating ?

 

[Eeyore]

400 watt RMS amp for use on 240 V mains.
Destructed due to metal dropping in and shorting the amp .

How did that happen ? The 'cabinet' is supposed to prevent such things, it's an
IEC safety requirement.

The only fuse is on the mains rated at 8 amp, maker's design rating, which
shows no sag , discolour or anything like that, ie untouched.

Sounds like an incompetent manufacturer to me. 8 amps will provide a continuous
2kW almost ! Given typical fusing characteristics, consider that up to 4kW for
maybe 10 seconds.

Shorted power trannies, burnt low power trannies and even a piece of 3mm
trace from the main bridge rectifier burnt through

IEC 60065 failure very likely.

and that had 2 runs of
added solder over the track for current carrying, the other polarity trace
overheated but not ruptured
I'm thinking of bridging that gap with a fuse and another on the other rail
after cutting it.
The mains transformer is rated at 2x 47V,5 amp.
So drop the 8A fuse (for 2 KWatt !) to what value ? and the 2 added fuses of
what rating ?

Probably something like 3.15A but you may need to add a surge-gard type NTC in
the mains input.

What make / model was it ?

Graham

 

[Tim Shoppa]

400 watt RMS amp for use on 240 V mains.
Destructed due to metal dropping in and shorting the amp .
The only fuse is on the mains rated at 8 amp, maker's design rating, which
shows no sag , discolour or anything like that, ie untouched.
Shorted power trannies, burnt low power trannies and even a piece of 3mm
trace from the main bridge rectifier burnt through and that had 2 runs of
added solder over the track for current carrying, the other polarity trace
overheated but not ruptured
I'm thinking of bridging that gap with a fuse and another on the other rail
after cutting it.
The mains transformer is rated at 2x 47V,5 amp.
So drop the 8A fuse (for 2 KWatt !) to what value ? and the 2 added fuses of
what rating ?

The power supply to your amp probably has massively oversize filter
capacitors.

Average current draw at 400W out might be only two amps, but because
of the rectifier-filter construction it doesn't draw that two amps
continuously. It draws it in very brief gulps 120 times a second.
Typical conduction angle might be 20 percent in a well designed
supply; with massively oversize filter caps it could be as small as 5
percent. Massively oversized filter caps have been all the rage in
consumer audio for too many decades now.

The size of fuse you have to put in depends not on simple average
current (2 amps) but the root-mean-square (RMS) current (which could
indeed be 8 amps at full load).

And remember, fuses aren't really there to protect your electronics,
they're in there to prevent fires.

If you want a supply that doesn't have such a small conduction angle,
you go to choke-input supplies, which for some bizarre reason aren't
nearly as popular as they were 50 or 70 years ago .

Tim.

 

[Meat Plow]

400 watt RMS amp for use on 240 V mains.
Destructed due to metal dropping in and shorting the amp .
The only fuse is on the mains rated at 8 amp, maker's design rating, which
shows no sag , discolour or anything like that, ie untouched.
Shorted power trannies, burnt low power trannies and even a piece of 3mm
trace from the main bridge rectifier burnt through and that had 2 runs of
added solder over the track for current carrying, the other polarity trace
overheated but not ruptured
I'm thinking of bridging that gap with a fuse and another on the other rail
after cutting it.
The mains transformer is rated at 2x 47V,5 amp.
So drop the 8A fuse (for 2 KWatt !) to what value ? and the 2 added fuses of
what rating ?

So what you are saying is that this catastrophic failure of a 400 watt amp
didn't blow an 8 amp fuse? I find that really hard to believe not that you
are making this up mind you.

 

[N Cook]

The power supply to your amp probably has massively oversize filter
capacitors.

Average current draw at 400W out might be only two amps, but because
of the rectifier-filter construction it doesn't draw that two amps
continuously. It draws it in very brief gulps 120 times a second.
Typical conduction angle might be 20 percent in a well designed
supply; with massively oversize filter caps it could be as small as 5
percent. Massively oversized filter caps have been all the rage in
consumer audio for too many decades now.

The size of fuse you have to put in depends not on simple average
current (2 amps) but the root-mean-square (RMS) current (which could
indeed be 8 amps at full load).

And remember, fuses aren't really there to protect your electronics,
they're in there to prevent fires.

If you want a supply that doesn't have such a small conduction angle,
you go to choke-input supplies, which for some bizarre reason aren't
nearly as popular as they were 50 or 70 years ago .

Tim.

The 2 reservoir caps are 6800uF, 80V rating with presumably about 65V on
each , when all is in working order.

 

[N Cook]

So what you are saying is that this catastrophic failure of a 400 watt amp
didn't blow an 8 amp fuse? I find that really hard to believe not that you
are making this up mind you.

Well semi-catastrophic, the positive side in a sorry state but the negative
rail side cold-tests ok so far. I was wrong when I said about the overheated
trace it was the ground return not the negative rail, the + trace burned
through as a "fuse", no deliberate necking at those points.

 

[Meat Plow]

Well semi-catastrophic, the positive side in a sorry state but the negative
rail side cold-tests ok so far. I was wrong when I said about the overheated
trace it was the ground return not the negative rail, the + trace burned
through as a "fuse", no deliberate necking at those points.

Hmmmm well I suppose anything is possible. I haven't gotten into these new
fangled Chinese amps though. My experience was with good old Crown,
Soundcraftsman (yuk) Peavey, BGW and all those that had a dozen or so
devices like MJ-15024/25 pairs on each rail. I have a 20 year old long
defunk SCS (Sound Code Systems) 750 WRMS MOSFET amp down in my basement
that still works great after I repaired it. I used to use it for bass
guitar in the bridge mode. Has an infinite variable speed fan that adjusts
with the heat sink temp. Uses now obsolete J-50's and K-134's. I bought up
a bunch of them years ago when I was luck enough to find some so I would
have spares if the amp decided to toss an output which it hasn't in the
15 years since I repaired it. I use the amp now for my PA in the basement
where I practice.

Anyway, I think fusing the rails is feasible.

 

[Eeyore]

The size of fuse you have to put in depends not on simple average
current (2 amps) but the root-mean-square (RMS) current (which could
indeed be 8 amps at full load).

Absolute utter complete and total rubbish.

'Oversized' caps will only slightly change the conduction angle. It's readily
shown by calculation. Certainly no way whatever will the rms current be 8A !

Graham

 

[Eeyore]

Anyway, I think fusing the rails is feasible.

It's good practice IMHO.

Unfortunately it's unlike to *prevent* further 'burn ups' although it might help
mitigate them slightly.

To actually prevent burn-ups you need to design for that and the small Chinese
companies I've come across who offer cheap OEM products don't have that level of
design skill. With bipolar output amplifiers you'll normally need to use a number
of fusible resistors (or else rate the output devices more conservatively - which
cheap designs never do of course).


Graham

 

[Eeyore]

The 2 reservoir caps are 6800uF, 80V rating with presumably about 65V on
each , when all is in working order.

That's a reasonable size. They won't be causing excessive switch on current.

Is the power transformer toroidal or EI ? That's where the big difference is.
Toroids will take a considerable switch-on current (it's largely due to their
magnetic characteristics).

Graham

 

[CJT]

The power supply to your amp probably has massively oversize filter
capacitors.

Average current draw at 400W out might be only two amps, but because
of the rectifier-filter construction it doesn't draw that two amps
continuously. It draws it in very brief gulps 120 times a second.
Typical conduction angle might be 20 percent in a well designed
supply; with massively oversize filter caps it could be as small as 5
percent. Massively oversized filter caps have been all the rage in
consumer audio for too many decades now.

The size of fuse you have to put in depends not on simple average
current (2 amps) but the root-mean-square (RMS) current (which could
indeed be 8 amps at full load).

That shouldn't be terribly hard to calculate, assuming a conduction
angle of your choice. So how about adding some support for the
assertion?

 

[N Cook]

That's a reasonable size. They won't be causing excessive switch on current.

Is the power transformer toroidal or EI ? That's where the big difference is.
Toroids will take a considerable switch-on current (it's largely due to their
magnetic characteristics).

Graham

Toroidal , not E-I lamination transformer
I'm coming around to thinking 5 amp anti-surge in the mains fuse-holder and
a 5 amp fuse shunted in the + and - rail traces from the DC side of the
bridge rectifier before the reservoir caps, but undecided whether quick-blow
or anti-surge ones there.

 

[Meat Plow]

Toroidal , not E-I lamination transformer
I'm coming around to thinking 5 amp anti-surge in the mains fuse-holder and
a 5 amp fuse shunted in the + and - rail traces from the DC side of the
bridge rectifier before the reservoir caps, but undecided whether quick-blow
or anti-surge ones there.

Fix the amp, put the fuses of your choice in and put it through its paces.
You'll probably want to run some music through it at high volume
preferably some Hip Hop <barf>. See which fuses blow and which don't.

 

[Tim Shoppa]

That shouldn't be terribly hard to calculate, assuming a conduction
angle of your choice. So how about adding some support for the
assertion?

10% conduction angle gives peak current of 20A, RMS current of 6.3A,
not too unreasonable to choose an 8A fuse in that case.

6.25% conduction angle gives 8A RMS, but that'd blow the fuse way too
often.

It takes big capacitors and toroidal transformers with really stiff
windings but many folks are building audio power supplies that way. I
personally don't like razor-thin conduction angles like 10% but that's
the style of other folks, not me!

Tim.

 

[Phil Allison]

"N Cook"

400 watt RMS amp for use on 240 V mains.

** Is that 400 watts per channel or 200 per channel ?

Destructed due to metal dropping in and shorting the amp .
The only fuse is on the mains rated at 8 amp, maker's design rating, which
shows no sag , discolour or anything like that, ie untouched.

** Then the AC supply current did not exceed 8 amps by much for more than
a fraction of a second.

Shorted power trannies, burnt low power trannies

** Err - " trannies " = transistors ??

The mains transformer is rated at 2x 47V,5 amp.

** Say it is rated at 500 VA, for simplicity.

So drop the 8A fuse (for 2 KWatt !) to what value ? and the 2 added fuses
of
what rating ?

** From another post I see it is a toroidal type.

Couple of facts:

1. With secondary shorted ( before or after the bridge) primary current will
rise to circa 42 amps rms - so bye bye to any 8 amp fuse real quick.

2. At switch on, the peak supply current will regularly exceed +/- 100
amps for the first half cycle - diminishing to the idle value over the
next 10 - 15 cycles or so.

3. At 400 watts output, the AC current draw will be around 4 amps rms -
assuming this is a typical, low bias, class AB amplifier design. With hard
clipping the figure will rise to over 6 amps rms.

So - the maker's choice of an 8 amp AC fuse is not unreasonable, given the
above facts.

The knee jerk reaction of matching the fuse rating to the VA rating of the
AC tranny does not work in practice - just try using a 2 amp fuse if in
doubt.

Maybe try a 6.3 amp " anti-surge " fuse - the kind with spiral wound fuse
wire.

Keep a few spares handy.



....... Phil

 

[Phil Allison]

"N Cook"

So drop the 8A fuse (for 2 KWatt !) to what value ? and the 2 added fuses
of
what rating ?

** Be very wary of adding +/- DC rail fues to any power amp designed without
them.

Very likely if one or other DC fuse blows or is removed, the amp's output
will swing fully to the rail with its fuse still intact.

Recipe for fried speakers.

You have been warned.




........ Phil

 

[msg]

Phil Allison wrote:

** Be very wary of adding +/- DC rail fues to any power amp designed without
them.

Very likely if one or other DC fuse blows or is removed, the amp's output
will swing fully to the rail with its fuse still intact.

<snip>

One could add overcurrent shutdown circuitry on each rail designed to shut
down the entire supply, or for antique simplicity use 'indicator fuses' which
have a spring-loaded plunger which makes contact with an endstop terminal
when they blow -- the endstop is often connected to a crowbar which blows
main fuses upstream. These are small cartridge fuses with the same form
factor as standard varieties; I had to replace lots of them in old line
printers where they were used in the hammer drivers -- in this
application they did not crowbar but instead were wired to status circuits
that reported the failure.

Regards,

Michael

 

[Phil Allison]

"msg"

<snip>

One could add overcurrent shutdown circuitry on each rail designed to shut
down the entire supply,

** How idiotic !

or for antique simplicity use 'indicator fuses' which
have a spring-loaded plunger which makes contact with an endstop terminal
when they blow -- the endstop is often connected to a crowbar which blows
main fuses upstream.

** Or simply use no DC rail fuses and let the AC supply fuse blow when
overloaded - as originally intended by the designer.

Those power amps that DO use +/-DC rail fuses, almost invariably have them
fitted in line with the output devices but NOT in line with low power,
voltage amplification circuitry.

It must be possible to remove either DC ( or both) rail fuse with the amp
under load and drive and have no circuit damage occur. Any competent amp
designer can arrange things so this is the case - but not likely the OP.





........ Phil

 

[msg]

"msg"




** How idiotic !






** Or simply use no DC rail fuses and let the AC supply fuse blow when
overloaded - as originally intended by the designer.

Evidently the designer did not account for the sort of catastrophic
failure described by the OP. The large capacitive reservoir in the supply
is very much like that in the line printers that employed the sorts of
protection I described, including the "idiotic" independent supply rail
overcurrent protection scheme with very fast response times (faster than
any fusible link).

Those power amps that DO use +/-DC rail fuses, almost invariably have them
fitted in line with the output devices but NOT in line with low power,
voltage amplification circuitry.

Indeed that was implicit in my response; the rails in question are the
power amp rails, _not_ low-power rails.

It must be possible to remove either DC ( or both) rail fuse with the amp
under load and drive and have no circuit damage occur. Any competent amp
designer can arrange things so this is the case - but not likely the OP.

Yes, but the OP did not design the amp and short of re-engineering it he
would need workable solutions.

Regards,

Michael

 

[Phil Allison]

"msg"
Phil Allison

Evidently the designer did not account for the sort of catastrophic
failure described by the OP.

** Nonsense.

Output stage S/C failure is intended to blow the AC supply fuse.

The circuit damage is already done when it failed.

( snip drivel)

Indeed that was implicit in my response; the rails in question are the
power amp rails, _not_ low-power rails.

** No - you have got it WRONG again !!!

It is the same two DC rails, split to service the output stage devices via
fuses.

YOU have NO clue whatever about power amplifier circuitry.

Better you learned to shut the **** up.

Yes, but the OP did not design the amp and short of re-engineering it he
would need workable solutions.

** Well, he ain't gonna get any from a posturing ASS like you then.

The only *problem* the OP ever had was due to dropping a metal object
inside the amp - to which the solution is damn obvious.




....... Phil