fuse // cap for basic passive speaker protection ?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've read three objections to using a fuse in the output of an amplifier for speaker protection

a) they affect the sound because of non-linear resistance / heating effects
b) they have to be sized large enough to avoid nuisance blowing and as such they react too slowly to protect the speaker
c) they fatigue ( and fail ) over time due to ac heating affects

How about putting capacitance in parallel with the fuse. The capacitance (perhaps a combination of capacitors in parallel) would be chosen to have low ESR and the fuse would be chose to blow at low power levels.

a) the music is bypassed through a good capacitor to avoid non-linear effect of fuse on the music
b) the fuse bypasses the capacitor at very low frequencies to minimize capacitor distortion and phase shift
c) the fuse will experience the average d.c. power into the speaker, it will not get stressed as much thermally and so will last much longer
d) the fuse can be sized for lower power since the a.c. power will bypass through the caps and therefore the fuse will respond faster to protect the speaker from a d.c. fault

Whadayathink ?

The same arrangement could be used in the power rails but would have to be careful about one rail failing and leaving the other rail alive - guarantees dc at the output.
You need 3300 uf at 8 ohms to pass enough bass. More capacitance at lower impedance. The biggest NP (non-polar electrolytic) capacitor I know of is 220 uf, about $4 each. So 15 of those is $60. I haven't tried that, but I have tried 6600 uf and a 10000 uf polar capacitors minus to minus on 8 ohm speakers. The test amp was a Peavey CS800s, which has no capacitors in the path. You could hear the capacitors; they made the top octave of a grand piano and tinkly bells sound a little funny compared with without series capacitor.
Somebody suggested putting a diode and a 100k resistor in to charge the middle (+ to +) of the capacitors up and make sure they don't go through the chemical events around zero volts. I haven't tried that yet.
With the capacitors, the fuse isn't needed, IMHO. I've installed some fuses in the rails to the output transistors, to try to prevent 18 of them going at the next meltdown. $4 MJ15024-5 each, $72. I'm still using test car salvage speakers, not the good ones, so I really don't know how that sounds yet. I don't know if PV-1.3K sounds that good anyway as designed. I I don't use it on my $300 each speakers yet since the DC detect crowbar circuit burns the triac land the PWB instead of protecting the speakers and blowing the main circuit breaker.
Amps that don't cross the speaker through zero volts, can use a one polarity capacitor, and sound great. My djoffe-bias modified ST120, sounds good. The Phillips 15-20-30 watt amp, the Mullard 25 w amp, the Leak Delta 70, the MJR7 lateral fet amp, the G-amp, that is about it on the list of single supply amps I know of that use silicon output transistors. LM3886 can be single supply and has its own current limiter, but that is sort of cheating and limited to 35 W.
I'm trying to something like your setup. I have salvage tungsten contact 22A @ 600 VAC 4 pole relays, that I am trying to install in the power rails. The DC detector will set it to open. The power supply first passing through +-30 VDC will cause it to close to try to weld the contacts slightly together. I'm going to bypass the contacts with 10 uf 200 v film caps to try to stop any possible turn on pop and pass low level signals without any modification.
Last edited:
Why do we need a large cap on the output for bass if we have a fuse that works down to DC ?

My default design (in progress) for my 7 channel amp is also a relay in the power rails. It's dual rail, so a fuse won't cut it, I need to switch off both rails at the same time. And this is where I started, a relay with big cap across it - which lead me to think about simple fuse. Trouble with the power rails is that it's a lot of current.
You are trying to have it both ways. You want the fuse for protection, but are concerned that the miniscule resistancce of the fuse will vary enough with current to degrade the sound. But then you want to add a bunch of capacitance to somehow bypass the fuse. (try calculating how much signal will flow through a capacitor with essentially zero ohms of the fuse in parallel) And we then ignore the signal consequences of capacitors? I don't care how low the ESR is in your cap, it will still be a lot higher than the resistance of the fuse.

Your fuse indeed works down to DC, but then you are back to using it as a fuse, and you objected to that. The low frequencies are the ones that carry the most current

Try it. Set up an amp and a speaker, put a fuse in series with the speaker wire. Now try putting any cap you care to try in parallel with the fuse. Report the difference.
sounds like it is worth investing a bit more time. This could give effective DC offset protection and yet pass the higher AC signals through the capacitors.
Maybe back to back 220uF 63V or 100V in parallel to 4u7F poly something for all the >100Hz signals.
How about a mini air cored inductor in series with the fuse? and then parallel that with the capacitor combination?

This could all be experimented with in the speaker cable and would be relatively quick to swap cables for sound comparisons.
If a fuse is going to create distortion this will only be for low frequencies. Bypassing it with a cap still sends the low frequencies through the fuse, unless the cap is very large. A very large cap will have to be electrolytic. Electrolytics don't like running with no DC bias. I suppose you could use a bipolar, but I suspect that even they benefit from some voltage across them from time to time.

I think this is one of those situations where if the answer was that easy people would have been routinely doing it since the 1960s.
If you aim to DC protection, why not put a fuse series with a 10 Ohm resistor to the ground cord of the transformer output. You will find in normal operation, there is not much current there, even less than 100mA for some amplifier when fully loaded. The idea is to let all the AC current go through filter capacitor in PSU instead of the transformer ground cord. What's more, it won't degrade the sound.

I use this way to protect my 30W amplifier, but I am not sure whether it is suitable for HIGH power application.
Just checked some impedance numbers.
1mF back to back with 1mF electros in parallel with 10uF poly gives an effective capacitance of 510uF.
@ 1kHz the impedance is 0.3ohms.
Most of the signal will still prefer to follow the lower impedance of the fuse+inductor.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.