for power rail fusing the fuse is shown in the correct position.
A fuse after the main smoothing has to survive the Music Transients and the first charging of the local decoupling.
A fuse that is upstream of some or all of the smoothing capacitance has to survive those previous current sna d the first charging of the downstream smoothing. This relocated fuse needs to be so big to survive that first charge that it provides virtually no safety benefit.
The amplifier fuses are Primary fuse (preferably close rated requiring a soft start) for the transformer and if desired rail fusing before the decoupling.
for power rail fusing the fuse is shown in the correct position.
A fuse after the main smoothing has to survive the Music Transients and the first charging of the local decoupling.
A fuse that is upstream of some or all of the smoothing capacitance has to survive those previous currents and the first charging of the downstream smoothing. This relocated fuse needs to be so big to survive that first charge pulse that it provides virtually no safety benefit.
The amplifier fuses are Primary fuse (preferably close rated requiring a soft start) for the transformer and if desired rail fusing before the decoupling.
highly undesirable, especially if you got the usual + 0 - type of powersupply
What if we move the fuse to the last R position and have 20,000uF after it to provide a low PSU impedance to the output devices. We limit the inrush current to 5A only via the softstart circuit until the 20,000uF is about 85% charged. Would it work? Can you see any problem with it? I have not got time to simulate this yet. Let me know your thoughts.
This is why one pours planes and uses local decoupling. The common practice of routing traces isn't what it could be but if you do the layout right 100 milliohms at tens of MHz is no problem. Not that this is needed in audio power amps...
Yes, it's trivially easy---see post 1903. Also, most chip amps implement current limiting and shutdown on overheat. Won't protect the driver in the event of catastrophic damage to the chip amp but definitely does protect against your shorted terminals scenario. One can build equivalent protection circuits in a discrete amp, of course, and it's commonly done in better commercial products.
Passively filtering mains ripple off the supply only plays nicely with class A amplifiers. For class B type loads it generally makes the problem worse, not better. You could structure the filter such that the cutoff remains above audio frequencies over the range of the fuse resistance but there's no benefit in doing so.
If there is one I'm not aware of it. Basically it comes down to fuses and overcurrent detection. Tripping a relay on overcurrent and open circuiting the driver is the default output protection scheme; it's left to fuses to blow if an output device has failed and crowbarred the supply.
Last edited:
Why? Are you trying to protect the silicon? Fuses change resistance with changing current amplitude. And they raise the impedance of the rails.
My opinion: Fuse the speakers, to protect them (if you must). Fuse the AC input, to protect people. Done.
The sound is usually better without any fuses anywhere after the reservoir caps, according to the golden-earred opinionistas. For example, defeating (shorting across) speaker fuses is one of the most-recommended tweaks for new Magnepan speaker owners. (I do still use my speaker fuses, so far.) There is even a healthy market for special (and very expensive) "hi fi" fuses. None of that is actual "evidence". But a lot of people swear by it.
My opinion: Fuse the speakers, to protect them (if you must). Fuse the AC input, to protect people. Done.
The sound is usually better without any fuses anywhere after the reservoir caps, according to the golden-earred opinionistas. For example, defeating (shorting across) speaker fuses is one of the most-recommended tweaks for new Magnepan speaker owners. (I do still use my speaker fuses, so far.) There is even a healthy market for special (and very expensive) "hi fi" fuses. None of that is actual "evidence". But a lot of people swear by it.
Last edited:
rail fusing is just as effective as speaker fusing at preventing burning out the speaker driver/s
But a speaker fuse is AFTER the NFB and thus is not usually corrected for the distortion a speaker fuse usually adds.
Whereas a rail fuse is before the decoupling and before the NFB and thus the amplifiers PSRR quite effectively prevents rail fuse distortion effects from being audible.
But there is a problem.
If one rail fuse blows due to slight overload, then most amplifiers send the other rail voltage to the speaker, creating a bigger problem than was first detected. You need to then ensure that the second rail fuse ruptures very soon after the first.
But a speaker fuse is AFTER the NFB and thus is not usually corrected for the distortion a speaker fuse usually adds.
Whereas a rail fuse is before the decoupling and before the NFB and thus the amplifiers PSRR quite effectively prevents rail fuse distortion effects from being audible.
But there is a problem.
If one rail fuse blows due to slight overload, then most amplifiers send the other rail voltage to the speaker, creating a bigger problem than was first detected. You need to then ensure that the second rail fuse ruptures very soon after the first.
Inline fuses makes a difference , fact everything makes a difference ..
I usually up the fuse rating and type for better sonics or bypass with silver wire , does make a difference in my setup and others we have tried, Of course in less revealing systems or running at a pretty benign 8 ohm load may not yield the same results and may not show any sensitive to such changes ..
Last edited:
They are usually of different sizes , 10 with an 8 for eg. My Thresholds are like that ...
If I am fusing a 100W into 8ohms amplifier, I have a choice of using two rail fuses @ F2A and/or one speaker fuse @ F4A.
They would definitely not be the same fuse value in those two quite different locations.
The two Rail fuses would be of the same value.
The mains fuse which I will always fit will be T1A, or maybe T800mA
They would definitely not be the same fuse value in those two quite different locations.
The two Rail fuses would be of the same value.
The mains fuse which I will always fit will be T1A, or maybe T800mA
Smaller transformer, bigger/more output devices. For example, 36va with LM1875, or 300va with Honey Badger. The conservative transformer selection has decreased the likelihood of blowing the outputs/fuses. These BJT's have more capacity for peak power, so it is possible to build the power supply for. . . at some risk of terminology mishap. . . speed. Practically speaking, you'd just build the power board that can get charged on transformer current insufficient/unlikely to blow the output devices. And, it is the concept of: Don't put more in than can come out. This also applies to eating cheese--far too much in could blow the output device or fuses. See? Easy to remember!
this is where the speaker relay comes in handy, the output offset will trip the relay off...
Tony,
I think you have missed the point.
If a rail fuse blows due to an AC overload condition, then the speaker and the amp may be protected from damage due to the short term AC overload, if the fuse was small enough and thus blew fast enough.
The problem then becomes worse.
The amplifier with one live rail sends a very damaging DC to the speaker.
Some or many speaker relays will find it impossible to break the DC fault current flowing to the inductive speaker coil.
The fuse blow due to AC, can become a major DC event.
That is a good reason to NOT fit rail fuses. I have not made up my mind which is better. I have amplifiers of both types. Fused supply rails and unfused supply rails. I have no amplifiers with speaker fuse. I have no amplifiers with secondary fusing before the smoothing capacitors.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.