Maker Pro
Maker Pro

Opinions on adding fuses to power amp

M

msg

Jan 1, 1970
0
Phil said:
"msg"
Phil Allison




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

Huh? I was agreeing with you and pointing out that my previous post
was based on that assumption...

YOU have NO clue whatever about power amplifier circuitry.

I will withdraw from this discussion at this point; I have designed
my share of power amps, mostly involving large computer peripherals
and have experience with big servo amps and shaker tables as well.

Regards,

Michael
 
P

Phil Allison

Jan 1, 1970
0
"msg"
Phil Allison


** Piss off - you snipping IMBECILE.

I was agreeing with you and pointing out that my previous post
was based on that assumption...


** Total bollocks.


I will withdraw from this discussion at this point; I have designed
my share of power amps, mostly involving large computer peripherals
and have experience with big servo amps and shaker tables as well.


** But obviously YOU have never studied how typical audio power amps are
made or have the tiniest a clue why.

PISS OFF

- BLOODY FOOL




......... Phil
 
J

John Tserkezis

Jan 1, 1970
0
** Total bollocks.

Wonderful. You're arguing with people who are agreeing with you.

Quite entertaining.

You're not trying to approach the notoriety of Rod Speed are you? He might
be different, but he still makes sense.

You on the other hand make no sense whatsoever. Still entertaining though...
 
P

Phil Allison

Jan 1, 1970
0
"John Tjerkshis Fuckwit "

( snip autistic, verbal drool )


** Those with the IQ of a donkey only understand donkey speak.


Hee - haw, hee - hawww......





...... Phil
 
N

N Cook

Jan 1, 1970
0
Phil Allison said:
"N Cook"



** 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

I'm just trying to engineer a more respectable equivalent of fuses for trace
rupture.
Eventually one DC rail ruptured as a fuse. I've now decided to do some
calculation to determine what the rupture current for that trace was.
Then as the common return almost fused and is the same dimension but in the
even worse case then carries twice the current , halving this trace rupture
current to use as the fuse rupture current and reduce that for current
carying capacity , by half again ?
If all output trannies fused short circuit all round so loading both DC
rails the combined common return would have ruptured at whatever that
rupture current is for the trace that did burn up.
Anyone know the term for laying double traces of solder along copper traces
to increase current carrying , so I can research it ?
I'm intrigued how you lay quite accurate molten solder runs and parallel
over and with existing traces.

Anyway measurements:
Copper trace 3.5 x .02mm and as 2 half approximated elipses of solder then
area of 1 elipse of tin+lead which is Pi x a x b , a and b minor and major
axes of .15mm and .8mm.
From tables of fusing currents for copper, also tin and lead (NB in the form
of circular wires ) for different diameters.

Copper fusing current of the trace = 12 amps
Lead+Tin elipse then 6 amps (not as much as I would have intuitively
thought)
Total 18 amps so 18/4 = 4.5 amps conventional fuse rating.
Presumably the circular to sheet allowance would up this 4.5 amp figure ,
but by how much ?
What sort of correction factor for thin sheet/non-circular heating then
rupture allowance?
 
R

Ron(UK)

Jan 1, 1970
0
N said:
I'm just trying to engineer a more respectable equivalent of fuses for trace
rupture.
Eventually one DC rail ruptured as a fuse. I've now decided to do some
calculation to determine what the rupture current for that trace was.
Then as the common return almost fused and is the same dimension but in the
even worse case then carries twice the current , halving this trace rupture
current to use as the fuse rupture current and reduce that for current
carying capacity , by half again ?
If all output trannies fused short circuit all round so loading both DC
rails the combined common return would have ruptured at whatever that
rupture current is for the trace that did burn up.
Anyone know the term for laying double traces of solder along copper traces
to increase current carrying , so I can research it ?

Beefing up!
I'm intrigued how you lay quite accurate molten solder runs and parallel
over and with existing traces.

When we used to build classic 'Hitachi' mosfet amps, we would lay a
thickness of solder along the DC rails to the outputs transistors and to
the output terminal. It`s a bit like welding, once you get the hang of
it, you can lay down a respectable run of solder that looks quite
professional. I spose you could beef up with tinned copper wire, but it
wouldn`t look so neat. Some amps have hard wired connections directly to
the output trannies, or, the DC feed from the PSU go directly to the pcb
lands upon which the outputs are soldered/bolted.
Anyway measurements:
Copper trace 3.5 x .02mm and as 2 half approximated elipses of solder then
area of 1 elipse of tin+lead which is Pi x a x b , a and b minor and major
axes of .15mm and .8mm.
From tables of fusing currents for copper, also tin and lead (NB in the form
of circular wires ) for different diameters.

Copper fusing current of the trace = 12 amps
Lead+Tin elipse then 6 amps (not as much as I would have intuitively
thought)
Total 18 amps so 18/4 = 4.5 amps conventional fuse rating.
Presumably the circular to sheet allowance would up this 4.5 amp figure ,
but by how much ?
What sort of correction factor for thin sheet/non-circular heating then
rupture allowance?

Did you ever tell us what make and model of amplifier this is?

Is it really worth all this effort to redesign? You're never going to be
able to protect against stray metal object clattering around inside, and
I`m sure the original designers never gave it a thought.


Ron(UK)
 
N

N Cook

Jan 1, 1970
0
Ron(UK) said:
Beefing up!


When we used to build classic 'Hitachi' mosfet amps, we would lay a
thickness of solder along the DC rails to the outputs transistors and to
the output terminal. It`s a bit like welding, once you get the hang of
it, you can lay down a respectable run of solder that looks quite
professional. I spose you could beef up with tinned copper wire, but it
wouldn`t look so neat. Some amps have hard wired connections directly to
the output trannies, or, the DC feed from the PSU go directly to the pcb
lands upon which the outputs are soldered/bolted.


Did you ever tell us what make and model of amplifier this is?

Is it really worth all this effort to redesign? You're never going to be
able to protect against stray metal object clattering around inside, and
I`m sure the original designers never gave it a thought.


Ron(UK)

These beefed-up traces are far too regular to be done by hand, also precise
45 degree X+Y plot lines so some sort of plotter type delivery system
following the traces/etch artwork co-ordinates. But its not just the solder
lines but the amount of solder delivered that is so uniform.
 
R

Ron(UK)

Jan 1, 1970
0
N said:
These beefed-up traces are far too regular to be done by hand, also precise
45 degree X+Y plot lines so some sort of plotter type delivery system
following the traces/etch artwork co-ordinates. But its not just the solder
lines but the amount of solder delivered that is so uniform.

Ah you're talking abput the hatched (scotched) type thing, I`m sure
that`s solder paste screen printed on during manufacture of the pcb.

Ron(UK)
 
N

N Cook

Jan 1, 1970
0
Ron(UK) said:
Ah you're talking abput the hatched (scotched) type thing, I`m sure
that`s solder paste screen printed on during manufacture of the pcb.

Ron(UK)

None of that on this one but I know what you mean. But some extrusion of
solder paste process via a vinyl sign cutter type plotter mechanism in place
of the cutter would be more likely than molten solder. Then fused onto the
board before populating with components but would that take solder bath
soldering of the components afterwards ?
The "beefing" is not applied after the main component soldering operation.
 
Toroidal , not E-I lamination transformer
I'm coming around to thinking 5 amp anti-surge in the mains fuse-holder and

whatever fuse the toroidal gets needs to be antisurge. Toroids are
known for their sometimes heavy inrush.

This will give no protection to the output devices of course, it
really will be nothing more than basic fire protection.

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.

You'd get much better fuse response if you put them after the
reservoirs, because they dont then need to handle high peak cap
charging currents.

I'm just trying to engineer a more respectable equivalent of fuses for trace
rupture.
Eventually one DC rail ruptured as a fuse. I've now decided to do some
calculation to determine what the rupture current for that trace was.
Then as the common return almost fused and is the same dimension but in the
even worse case then carries twice the current

worst case is 1x not 2x. The common psu rail carries near zero if
similar fault currents flow in both + and - rails.

, halving this trace rupture
current to use as the fuse rupture current and reduce that for current
carying capacity , by half again ?
If all output trannies fused short circuit all round so loading both DC
rails the combined common return would have ruptured at whatever that
rupture current is for the trace that did burn up.

then there would be near zero current in the common rail.
Anyone know the term for laying double traces of solder along copper traces
to increase current carrying , so I can research it ?
I'm intrigued how you lay quite accurate molten solder runs and parallel
over and with existing traces.

I dont think I've ever seen any attempt to do it to accurate
dimensions. Neither flow soldering nor hand soldering work like that.

Copper fusing current of the trace = 12 amps
Lead+Tin elipse then 6 amps (not as much as I would have intuitively
thought)
Total 18 amps

No, because cu and slobber blow at very different temps. 18A probably
wont even be close.


I'm left wondering precisely what youre trying to achieve. If youre
trying to protect output devices, none of the above will do it. If
youre trying to protect speakers, the above are not whats wanted. If
all you want is to prevent charring of pcb due to repeated blowing,
its going to be cheaper in the end to prevent the repeat blowing.
Really whats propsed so far doesnt seem to solve any real world issue.
The amp sounds roughly designed, intended for a limited service life,
and to change that you'd need to add a full set of current and safe
area protection for the output devices, almost none of which can be
achieved with fuses.

What are you really looking to achieve?


NT
 
N

N Cook

Jan 1, 1970
0
I'm left wondering precisely what youre trying to achieve. If youre
trying to protect output devices, none of the above will do it. If
youre trying to protect speakers, the above are not whats wanted. If
all you want is to prevent charring of pcb due to repeated blowing,
its going to be cheaper in the end to prevent the repeat blowing.
Really whats propsed so far doesnt seem to solve any real world issue.
The amp sounds roughly designed, intended for a limited service life,
and to change that you'd need to add a full set of current and safe
area protection for the output devices, almost none of which can be
achieved with fuses.

What are you really looking to achieve?


NT

I'll repeat it without the clutter.

I'm just trying to fashion a more respectable equivalent of fuses instead of
trace burning and rupture.
 
E

Eeyore

Jan 1, 1970
0
msg said:
One could add overcurrent shutdown circuitry on each rail designed to shut
down the entire supply,

He's *repairing* an amplifier not redesigning it.

A decent amplifier has some element of overcurrent protection built in anyway.

Graham
 
E

Eeyore

Jan 1, 1970
0
msg said:
Evidently the designer did not account for the sort of catastrophic
failure described by the OP.

You don't have much experience of high power audio amplifiers do you ?

Graham
 
E

Eeyore

Jan 1, 1970
0
N said:
Eventually one DC rail ruptured as a fuse. I've now decided to do some
calculation to determine what the rupture current for that trace was.

You'll find it was probably getting on for 100 amps or so.

Graham
 
E

Eeyore

Jan 1, 1970
0
N said:
I'm just trying to engineer a more respectable equivalent of fuses for trace
rupture.
Eventually one DC rail ruptured as a fuse. I've now decided to do some
calculation to determine what the rupture current for that trace was.
Then as the common return almost fused and is the same dimension but in the
even worse case then carries twice the current , halving this trace rupture
current to use as the fuse rupture current and reduce that for current
carying capacity , by half again ?
If all output trannies fused short circuit all round so loading both DC
rails the combined common return would have ruptured at whatever that
rupture current is for the trace that did burn up.
Anyone know the term for laying double traces of solder along copper traces
to increase current carrying , so I can research it ?
I'm intrigued how you lay quite accurate molten solder runs and parallel
over and with existing traces.

Anyway measurements:
Copper trace 3.5 x .02mm and as 2 half approximated elipses of solder then
area of 1 elipse of tin+lead which is Pi x a x b , a and b minor and major
axes of .15mm and .8mm.
From tables of fusing currents for copper, also tin and lead (NB in the form
of circular wires ) for different diameters.

Copper fusing current of the trace = 12 amps
Lead+Tin elipse then 6 amps (not as much as I would have intuitively
thought)
Total 18 amps so 18/4 = 4.5 amps conventional fuse rating.
Presumably the circular to sheet allowance would up this 4.5 amp figure ,
but by how much ?
What sort of correction factor for thin sheet/non-circular heating then
rupture allowance?

Don't waste time.

The customer's not paying for you to redesign the amp. Fit some wire mesh or
whatever to stop things being dropped inside it in future as per safety regs.

Graham
 
E

Eeyore

Jan 1, 1970
0
Ron(UK) said:
You're never going to be able to protect against stray metal object clattering
around inside,

Quite !
and I`m sure the original designers never gave it a thought.

In which case it shouldn't have a CE sticker.

Graham
 
E

Eeyore

Jan 1, 1970
0
N said:
None of that on this one but I know what you mean. But some extrusion of
solder paste process via a vinyl sign cutter type plotter mechanism in place
of the cutter would be more likely than molten solder. Then fused onto the
board before populating with components but would that take solder bath
soldering of the components afterwards ?

No.

You've gone completely berserk.

Graham
 
E

Eeyore

Jan 1, 1970
0
N said:
I'm just trying to fashion a more respectable equivalent of fuses instead of
trace burning and rupture.

You can't.

Fix it and tell the owner not to drop metal bits inside it.

Graham
 
R

Ron(UK)

Jan 1, 1970
0
Eeyore said:
Quite !


In which case it shouldn't have a CE sticker.

Graham

I would imagine that the stray object originated inside the amp,
probably a panel screw or captive nut never secured in the first place.

How often you you pick up an item of gear and hear something loose
rolling around inside


Ron(UK)
 
R

Ron(UK)

Jan 1, 1970
0
Eeyore said:
No.

You've gone completely berserk.

Graham

As I understand it, in modern pcb production, the majority of the solder
is 'silk screened' as a paste, onto the board as part of the process,
then once the majority of components have been positioned, the whole
assembly is passed between radiant heaters to melt the solder paste and
fuse the whole lot together.

Of course, I could be wrong.

It's a miracle that any electronic stuff works at all these days!


Ron(UK)
 
Top