Hi!
I'm currently deciding what to do with all the fuse holders installed in the case of an old amp I'm repurposing. How do you best fuse an analog power amp with transformer-based power supplies?
There's two independent power supplies, one for each stereo channel, delivering 2x +-50Vdc @ 5A. Obviously, one slow-blow fuse will be placed on the primary of each transformer. That leaves me with 4 fuse holders 🙂
Option 1: Fuse the amp output, in the signal path
Pro: Best possible protection for speakers
Con: Amplifier not really protected against malfunction
Con: The fuse, being essentially a temperature(current)-dependent resistor, will introduce distortion
Option 2: Fuse each of the positive and negative rails, right after the main power supply capacitors
Pro: Protect amp and speakers against malfunction
Con: Bypass capacitors on amp board can still discharge into speakers during malfunction
Con: When a fuse blows, one power rail will disappear, but the amp might continue with just the other rail. This would make the amp clip either positive or negative half-waves. A surefire way to kill speakers. Am I mistaken or is this a real concern?
Option 3: Fuse GND, right after the main power supply capacitors
Pro: Similar to option 2, but would alleviate the concern regarding disappearing rails. They'd both disappear at once.
Con: No protection should I ever run the amp in bridged mode
Option ?: Something else?
I'd love to hear your thoughts. Thanks!
I'm currently deciding what to do with all the fuse holders installed in the case of an old amp I'm repurposing. How do you best fuse an analog power amp with transformer-based power supplies?
There's two independent power supplies, one for each stereo channel, delivering 2x +-50Vdc @ 5A. Obviously, one slow-blow fuse will be placed on the primary of each transformer. That leaves me with 4 fuse holders 🙂
Option 1: Fuse the amp output, in the signal path
Pro: Best possible protection for speakers
Con: Amplifier not really protected against malfunction
Con: The fuse, being essentially a temperature(current)-dependent resistor, will introduce distortion
Option 2: Fuse each of the positive and negative rails, right after the main power supply capacitors
Pro: Protect amp and speakers against malfunction
Con: Bypass capacitors on amp board can still discharge into speakers during malfunction
Con: When a fuse blows, one power rail will disappear, but the amp might continue with just the other rail. This would make the amp clip either positive or negative half-waves. A surefire way to kill speakers. Am I mistaken or is this a real concern?
Option 3: Fuse GND, right after the main power supply capacitors
Pro: Similar to option 2, but would alleviate the concern regarding disappearing rails. They'd both disappear at once.
Con: No protection should I ever run the amp in bridged mode
Option ?: Something else?
I'd love to hear your thoughts. Thanks!
Last edited:
Ok, option 3 is an awful idea, as a fault might aswell shoot through from the positive all the way to the negative rail.
Option 1 is also awful. As I've feared, they introduce distortion (source: Self's Audio Power Amplifier Design book)
This only leaves option 2. The way of dealing with one rail's fuse blowing is to have additional DC offset protection on the output. Oh well.
Option 1 is also awful. As I've feared, they introduce distortion (source: Self's Audio Power Amplifier Design book)
This only leaves option 2. The way of dealing with one rail's fuse blowing is to have additional DC offset protection on the output. Oh well.
There's no perfect protection scheme, but rail fuses are the preferred method.
And they are also useful for disconnecting one channel for troubleshooting, and
for measuring current.
Some have included the speaker fuse within the nfb loop.
And they are also useful for disconnecting one channel for troubleshooting, and
for measuring current.
Some have included the speaker fuse within the nfb loop.
I never ever fuse rails in split supply amps because one will always open before the other and guarantee full rail voltage at speaker out, worst of all situations.
II much prefer fusing AC windings, before the bridge rectifier.
Now IF single supply and capacitor output, yes, fuse the single +Vb rail.
Never ever ground, of course.
II much prefer fusing AC windings, before the bridge rectifier.
Now IF single supply and capacitor output, yes, fuse the single +Vb rail.
Never ever ground, of course.
In a split supply, that still allows for the loss of one of the rails and thus same problem as fusing the rails themselves. Worse, even if the fuse blows, the tank capacitors can still discharge through the amp unfused, potentially causing quite a lot of damage by themselves.
No, it does not.
If one AC winding blows in a CT power transformer, the other winding will KEEP supplying BOTH rails equally.
Of course, it will blow sometime later, what made one blow will also blow the other.
But both rails will turn off at the same time.
"Unfused" here must be taken literally, meaning the fuse is no more, meaning it blew OPEN so no more energy fed into the main capacitors.
The amp is already damaged/shorted or it would not have made the fuses blow open in the first place, and shorted rectifiers/caps/power transistors (only components "hungry" enough to blow a fuse) will hog any remaining current,for the milliseconds it will take filter caps to discharge.
I already consider the amp dead/damaged and needing service, also blowing speakers at the same time won´t make me happier.
Last edited:
A lot of hype on the evils of fuses I don't buy. Now, old style 3AG style do have steel caps and the fuse holders were often steel if you think that is terrible. Fusing too close may make it more thermally sensitive than desired.
One can use ATO fuses which are more reliable and not steel
One can size a copper wire strand as a fuse soldered on the board for the rails. ( works quite well)
One can consider measurable vs audible and ignore the issue
One must consider the distortion protection circuits cause. It can be orders of magnitude more than a fuse.
One should consider the probability of a fault.
One should consider the cost of a fault.
What is the probability of a fuse fault? Yes, they do.
Are you careful enough never to mess with wiring with power on?
Do you have insane speaker loads and push them too far?
Anyway, think about what fault each fuse is going to protect. Look at a schematic and draw out each probably fault.
Think about why a fuse in that location could cause distortion.
Fuse the output, outside the feedback point
Fuse the output, inside the feedback point
Fuse in the rails, which in a way are inside the feedback loop.
What would an output mute/DC sense relay do instead?
Fast acting (evil) relay on the output
Fast acting relay on the rails
As far as DC servo as protection, unless it has an independent supply, one rail failing might drive the feedback to the other rail instantly.
I have seen folks get carried away with things like a fuse, when they use a single transistor/resistor current source, don't bother with a Darlington VAS, have not analyzed the sonic differences trading IPS LTP degen for global feedback. You get my point. Fix the big things first. If you have perfected an amp to the extent a fuse in the circuit is measurable, let alone audible, then wow, have you done your home work.
If you have an output current sense and DC detection that does not cause more distortion than a fuse, please share it with us. I have never seen one.
Yes Virginia, even a fuse is complicated.
One can use ATO fuses which are more reliable and not steel
One can size a copper wire strand as a fuse soldered on the board for the rails. ( works quite well)
One can consider measurable vs audible and ignore the issue
One must consider the distortion protection circuits cause. It can be orders of magnitude more than a fuse.
One should consider the probability of a fault.
One should consider the cost of a fault.
What is the probability of a fuse fault? Yes, they do.
Are you careful enough never to mess with wiring with power on?
Do you have insane speaker loads and push them too far?
Anyway, think about what fault each fuse is going to protect. Look at a schematic and draw out each probably fault.
Think about why a fuse in that location could cause distortion.
Fuse the output, outside the feedback point
Fuse the output, inside the feedback point
Fuse in the rails, which in a way are inside the feedback loop.
What would an output mute/DC sense relay do instead?
Fast acting (evil) relay on the output
Fast acting relay on the rails
As far as DC servo as protection, unless it has an independent supply, one rail failing might drive the feedback to the other rail instantly.
I have seen folks get carried away with things like a fuse, when they use a single transistor/resistor current source, don't bother with a Darlington VAS, have not analyzed the sonic differences trading IPS LTP degen for global feedback. You get my point. Fix the big things first. If you have perfected an amp to the extent a fuse in the circuit is measurable, let alone audible, then wow, have you done your home work.
If you have an output current sense and DC detection that does not cause more distortion than a fuse, please share it with us. I have never seen one.
Yes Virginia, even a fuse is complicated.