Project 3A fusing problem

Just finished my amplifier using Mr.Elliott 3A amplifier design. Had no problem at all with it except the other day when one of my rail fuse ruptured. Has I read somewhere on this forum there is big danger in this kind of failure as the output falls all the way to the opposite rail. It resulted in a burned out crossover and speaker element but hopefully nothing to the amp. I was now wondering if you DIYers had any idea to protect the amp from such a problem? Something that could monitor the circuit and see that the output is cut when one fuse fails.

Thankyou for all your replys


2001-06-01 4:53 pm
Uh oh, I was hoping that if the fuses went they would at least protect the speakers. I defended this approach some time ago around here somewhere. The idea is to sacrifice the output transistors because they are cheap in order to save the speakers. Obviously this hasn't worked out. I'm a bit concerned now about my version of this amp.
I think somebody suggested adding overcurrent protection to this amp, you might search around to find this idea. Check Rod's original version of the circuit (the Project 3), and perhaps email him and ask, he has been very responsive to emails.
Did you have the correct fuses in place? What conditions caused the fuses to blow?
I should point out that the speaker was not particularely high priced so Its not such a big deal. The fuse died from over current, normal type 5 amp fuse. I've been pushing the amp for some time before it went numb. I read on previous treads that quick blow fuses were thought to be better suited in this design to pretect the rails, wouldn't it be better with slow blow?

I hope we can find an easy solution to this problem.

thank you for your concern paulb

No Slow Blow

Hello PaulB, standard equipment in commercial gear is " T " type glass fuses.
These often have a tiny blob in the middle of the fusewire link.
These are slower than quick blow, so will handle turn on transient currents, but not so slow as Slo-Blo.
Slo-Blo are not a good idea at all for protection of audio gear.
Hugh Dean mentions output transistor base series resistors to limit fault currents.

Regards, Eric.


2002-04-28 2:11 pm
I use 5A glass "T" type fuses in my P3A, and have a speaker protection circuit, the Velleman K4700. This disconnects the speakers very quickly (via relays) should a DC voltage greater than about a volt or so be present. It serves to protect my precious Dynaudios should the amp develop a fault, or if one of the supply line fuses blows. Its a good kit, and only cost 13GBP. It also protects the speakers from switch-on transients and allows manual muting via a front panel switch or, as in my case, a thermal switch on the heatsink.

See ya,
Velleman K4700

Hi trwh,

I am building a pair of mono amps, I wonder would that be better for me to get the K4701 in line DC Protect kit for each amp. I think the K4700 is for the stereo amp and would be difficult fo rmy set up. Is that correct? If so, do you know is that ok to put the kit inside the amp and connect it between the amp output and speaker binding posts?

Thanks in advance,


2002-04-28 2:11 pm
Chris, a pair of K4701's will work for two mono amps. However, this kit is passive, and does not offer protection from switch-on and switch-off thumps. Also, it draws its operating current from the speaker line and as such may affect sonics (although its effect is probably negligable).

If you are building monobloc amps, you are probably interested in very good performance, so why not use two K4700's? The kit is very cheap, and it can be used in mono or bridged amps. That way you get the switching transient suppression, provision for manual muting (very useful), and I suspect that the active circuit has better (i.e. quicker) performance.
Has anybody tried a procection cricuit that upon sensing DC: the circuit disconects the power supply rails instead of the amplifier outputs?

In other words: put the relay conatcs between the power supply filter caps and the amplifier. Certainly removes the relay contacts from the direct signal path.

The only catch I can see is that you would need to make sure once the circuit has been tripped, it must be reset. Otherwise you could get a on-trip-off-on-trip-off-on-trip-off-on cycle. Getting up to push a reset button is better than replacing drivers or outpout devices.


From another thread that I read and came across Kristijan and Panos post a soft start circuit last week. I made the PCB and bought all the parts then I saw the K4700 and K4701. The PCB that I made does not have the DC protection that is why I thought about using the passive kit in combination. But I think Trwh has a good point about using two K4700, I can save the PCB and parts for other projects in the future.

Actually would someone please enlighten me that why when the fuse blown in the supply rail it causes harm to the speaker. I thought that is what the fuses are for to break the circuit and stop doing any more harm to the rest of the circuit(ie speaker as well).
Well Chris ma

The fuses are there to protect the circuit in case of transistor malfunction...

BuT, if only one of the two rail fuse break and the other stay intact then there will be a rude voltage swing at the output. Normally the speakers see AC voltage coming at them now they see almost the full voltage of your intact fuse rail being thrown at them. In no time your crossover and speaker elements will cook (quite a particular smell believe me!).

Aud_Mot: I really like your idea to cut the rail of the amplifier instead of the amplifier output to the speaker. This could be an even better way to protect the output transistor. Do you think rod protection circuit could be used for that? (ESP project 33)

DC protection kit and fuse?

I have built the Velleman K4700 soft start and DC protect kits for my current project JLH amp mono blocks. My question now is that do I still have to have in line fuses (slow blow 5A) for the mains and in particular the (fast blow 3A) fuse in-line with the speaker output?

If the fuses are still required than any suggestion of the hi quality fuses that I can use and purchase. I would love to not have the fuse in the output there since it is in the signal path and the dc protect circuit may already degrade the sound some what.


My idea is to shut down both rails. This would prevent the output from swinging to the other rail, and you hoping that other fuse blows before your speakers and/or cross overs.

The relay conatcts would be after the PS filter caps, so there would be zero volts into tamp very quickly. I agree with paulb, the PS is in the signal path. I don't know which is better to have in the signel path, relay contacts or a fuse. So I like the design that seems to be the most fail safe.


You are talking about 2 protection devices. 1. DC sensor (the Velleman) and 2. Current limiter (the fuse) One does not replace the other.

Typically current protecting the input and voltage gain stages is not needed. You can concentrate on protecting the output stage.

You are using a very sensible (in my book) approach. Use a soft start to limit turn on current. Then you can use a fast blow fuse on the mains. (power transformer primary side) Use the lowest rated fuse that does not blow under normal listening situations.

IF and this is a big IF..... You are making mono blocks and the output stage is designed modestly, 4:1, 6:1 safety margin, the main fuse will go before the output stages.

Be carefull about that last paragraph. You must look at your design carefully and menatlly short out different devices in a game of "what if...."