Loudspeaker Protector, my version

[zristic]

Hi,

As there are many people interested in loudspeaker protection circuit I made, I decided to place the project here.

Before I start, I want to let you know that there are other solutions such as discussed here:
http://www.diyaudio.com/forums/showt...13#post1304013

What I am writing about is just another implementation of the same thing.

Let's start with the schematics:

[zristic]

Motivation. I respect my loudspeakers since I worked hard to get them and I want to keep them healthy as long as possible. Moreover, a protection circuit is a must in any amplifier. It solves a lot of problems which sooner or later appear in amplifiers. It is much better to burn a couple of cheap components, then to destroy expensive speakers, or even worse, to cause fire.

Theory. The basics of the circuit is DC detector, described here:

http://sound.westhost.com/project33.htm

It is a simple circuit which detects average DC voltage in a period of time. Parameters of the circuit can be altered depending on requirements. Inputs of this detector are signals from both speakers. They are rectified and smoothed into a DC voltage.

The second part of the circuit is an AC detector on the upper right of the schematics. Again, another simple sircuit, which detects a drop in AC voltage. It is used to turn off speaker relays when AC voltage turns off. It helps to avoid noise in speakers when powers turns off.

The third part of the circuit is a soft-start circuit which consists of four 5W resistors in series and a relay. When the power is turned on, the current of the primary coil runs through those resistors thus reducing the effect of core magnetization. Relay bypasses the resistors after 0.7-1 sec of time, leaving the primary of the transformer to operate in usual way.

On the right of the image are output relay drivers which are connecting the speakers to the amplifier.

Finally, the core of the device is PIC16F84 which drives all the logic and determines delays.

Practice. I was critised that I should have put optocouplers in transistors collectors in order to physically separate PIC from analog section of the amplifier. According to the critic, PIC can cause unwanted noise in output signal. Since it was too late for me to change the PCB, I took the risk and wired it all as is. Up to today, I have not heard any unwanted noise cause by PIC. However, I recommend to all to indeed use optocouplers for galvanic separation of the digital and analog sections. This is a practice in modern devices and there are strong reasons for doing it.

Parts used do not need to be expensive. They are all on a side of signal path and really they make insignificant interference to the quality of the output signal.

One detail should be wll thought of though, which is that 10uF capacitors in DC detector must not be bipolar. Otherwise they will explode due to potential inverse voltage which can be applied across their terminals.

Finish. This is how the whole circuit looks:

[zristic]

Documentation. I am providing SCH and PCB files for Protel 99. You do not have to follow them in details, it is enough to take them as a starting point.

Also, in the attachment you will find full source code for mikroPascal for PIC. Again, the code is very simple, and it can be easily ported to C, Basic, Assembly or whatever. It has no more than 60 lines of code and it is very easy to understand.

Please note that I will not take responsibility for any kind of damage you make by misusing these documents. Use them as a guide for your own development, completely at your own risk.

[zristic]

Performance. I am very happy with this circuit. Firstly, it is very simple. It took me one afternoon to assemble it and to test it.
Secondly, program for PIC is very short. It solves a lots of headaches with various delays, logic etc. I wrote it in a moment of inspiration and I will keep it that way.
Finally, the protection is very fast. "AC off" detector is capable of detecting AC noise caused by vacuum cleaner. DC detection circuit is fast too, on ESP site it is said that it detects 5V DC in 20ms.
One improvement is recommended though, as I already said, try to use optocouplers to fully separate digital and analog grounds.

Recently I had a problem with one of the transitors in the amplifier. Actually, I do not know exactly what the problem was, but the outcome was that the insulation below the transistor did not do its job properly. Collector of the ouput transistor got short circuited to heatsink and therefore to chasis. Fortunately, the protection acted so fast that I did not hear any noise on the speakers, while the amp was sparking inside until the fuse went off. I think a better spontaneous test could not be done.

[muzza]

zristic, forgot to mention the AC detect and triggering on vacuum cleaner switchon....
This sounds like a mini brown-out affecting the AC detect cct.

A suggestion would be to increase the value of R? (the resistor between base and ground on Q1)....try 2K2 or 4K7.
Increasing the voltage at the base should implrove riding out the brown-out.

You could put a small capacitor 100nf - 330nf to ground on the base of Q1 as well. This would hold the base voltage during any downward spike.

[gootee]

Originally posted by zristic

<snipped>
One detail should be wll thought of though, which is that 10uF capacitors in DC detector must not be bipolar. Otherwise they will explode due to potential inverse voltage which can be applied across their terminals.

Did you mean that the 10uF capacitors "must be bipolar" (rather than "must not be bipolar")?

- Tom Gootee

[muzza]

Interesting comment on use of bipolar capacitors.

As far as I was aware the bipolars were created mainly for crossover networks. Even Black Gate had a range of these for that purpose.
Many commercial speakers use bipolars in the crossover.
I woud agree that the "poly" and mica types would be better.

Would be interested in the reference to that info .....

[gootee]

Originally posted by muzza
Interesting comment on use of bipolar capacitors.

As far as I was aware the bipolars were created mainly for crossover networks. Even Black Gate had a range of these for that purpose.
Many commercial speakers use bipolars in the crossover.
I woud agree that the "poly" and mica types would be better.

Would be interested in the reference to that info .....

I would be surprised to learn that any basic type of passive electronic component was created specifically for any audio application. Bipolar electrolytic capacitors, and capacitors in general, are used in many orders-of-magnitude more non-audio applications than audio applications. As it is for most and maybe all mainstream passive electronic components, if they weren't broadly applicable in other types of equipment, we probably wouldn't even have them available just for audio, at least not with reasonable availability, pricing, and quality.

At any rate, bipolar electrolytic capacitors are usually just two polarized capacitors packaged together, in series, back to back. I often just use two polarized caps (each with twice the desired "bipolar" capacitance), instead, and sometimes put a large-value resistor in parallel with each one, to keep the voltages across them more equal.

And it is very well-known that significant reverse-biasing of a "regular" polarized aluminum or tantalum electrolytic capacitor is "a bad thing", and can result in an explosion. I don't have any references at hand. But you could probably find warnings on some of the capacitor manufacturers' websites.

[muzza]

Gootee, thanks.
I did read somewhere a long, long time ago about bipolars and their development for use in crossovers... that just stuck in memory ever since.
I will do more investigation just to sate my own interest.

Yes, reverse biasing of normal polarised electrolytics does lead to interesting effects.

But this should not be an issue with bipolars as long as their rated voltage is not exceeded.
I have seen a bunch of examples where bipolar caps were used in crossovers in commercial designs.

I was more interested in zristics comment:-
"10uF capacitors in DC detector must not be bipolar. Otherwise they will explode due to potential inverse voltage which can be applied across their terminals."
That runs contrary to my understanding and experience.
Just wish to gain clarity of understanding.

[zristic]

I was more interested in zristics comment:-
"10uF capacitors in DC detector must not be bipolar. Otherwise they will explode due to potential inverse voltage which can be applied across their terminals."
That runs contrary to my understanding and experience.
Just wish to gain clarity of understanding.

Sorry it is a typo, I meant they should be bipolar. As you can see on the image, I put bipolar MKT 10uF/100V. I apologize once again for the typo, I should not post after midnight...

My Sansui speakers originally had polarized caps in the crossover. I replaced them with bipolar caps. The only problem I've heard of concerning using unipolar caps in crossover was aging.

I, however, witnessed an explosion of a unipolar capacitor which was polarized reversely. In order to produce the effect you need to keep it runing reversely polarized sometime, which is rarely hapening in a crossover... Even if it happens, it is more likely that your speaker driver would die first, rather than a unipolar capacitor.