It is more complex and expensive than the circuit it needs to protect, that is a little pointless. My amplifier designs have worked for decades without protection, why add some superficial circuit to it that costs money and real estate while do nothing. All that you have done is creating something with a much higher failure rate than the amplifier itself. Just perform a failure rate analysis on the system, and it shoots up orders of magnitude.
Start with a block diagram, to have folks involved and participating.
Ref #6/7/8/9/11/12/21 (not to ignore #16 either).
Ref #6/7/8/9/11/12/21 (not to ignore #16 either).
Thanks Nico and welcome back seems this thread will be well warmed by your charm and jokes.🙂 We did spend a lot of quality time here https://www.diyaudio.com/community/threads/symef-amplifier.198500/ over 12 years ago. I guess I spent over 20,000 hours designing that amplifier alone, maybe that was too much😉.
Using protection has its advantages, the phrase eyes only don't touch and don't use my equipment goes up in vapor. Your right Nico well designed class A amps don't need protection from the grid, the grid needs protection from them. Since they don't break you can directly couple them to speaker with no protection. The circuit breaker in the house suffices. But they are big heavy and expensive. Yes protection has a failure rate but many times it works. Imagine life without something as simple as a fuse! Class AB amplifiers are fragile. If one is spending $2500 plus to buy or build an amplifier, even this protection may seem insufficient. Does protection add distortion and noise and take away the essence of the music? Not really.
Last edited:
A protection circuit is usually cheaper than the speakers it protects. Separate protection is usually there to react to failure of the amp, not prevent it. It functions to prevent the failure cascading - but protection usually also serves as a de-thumping circuit and inrush-limiter, so its complexity is paid back several ways.
Protection of the amp circuit itself is typically built-in to the amplifier in the form of I or IV limiting and fuses.
Protection of the amp circuit itself is typically built-in to the amplifier in the form of I or IV limiting and fuses.
)n Audio (John if I remember correctly) in my opinion there is only one protection needed in an amplifier and that is when the offset after years drifts above a set value and can lead to damage to woofers and cause long term overheating and eventual damage. Turn-on plop is usually just anoying. The circuit you present, and I do not call your integrity nor design capabiity into question, in my opinion is too much, It should only protect and do nothing else. It should also not contain many components because easch component adds to the mean time between failure and can sometimes become far more problematic than the equipment it is going to protect. From what I recall you like showing your superior skils with very complex designs. I commend you for you superior circuit development, but in my opinion unnecessary.
Puzzled by above response, reads like you've never worked on a unit with failed output or driver transistor. An output c-e short dumping 60Vdc on the speakers makes a mess real quickly. Not uncommon for spkrs to be valued at USD $$$$, I will continue to explore protection options.
Touch wood, in the past 50 years I have never had a design go bad, why don't you design protection for battery cars to not burn out, that would be quite novel. Do you tie a Rotweiler toeach of your speakers incase your kid pokes his finger into the cone?
Last edited:
Show me an electronic component suddenly fail catatstrophically without your fingers in it somewhere.
There are many protection solutions, some in house some off the shelf, some simple, some more complex. Some work good, some work bad and others no work at all.Current management can be by a precise selected fast blow 3A-10A fuse depending on how big the amplifier is to amplifier over current shut down, not forgetting compressors and VI limiting. Heres an example of current detection that even upc1237 can use. Detection is usually on the positive going waveform as these devices have higher gain.
Makes me wonder why "crowbar" was used in old computer power supplies, while "open the relay contact" is used in Audio? Seems slow.
Although, using a Chroma programmable AC source, I was able to put several 1/2 cycles of full (US) AC line voltage through a generic computer speaker woofer, without destroying it. That'd be on the order of 10 ms. Unsure how many ms pass from when a relay coil energize signal deasserts, until the mechanics start moving.
Would it be faster to use a SSR to clamp the amplifier output to ground in response to some detected event; let the fuses on the +/- power supply lines blow?
Although, using a Chroma programmable AC source, I was able to put several 1/2 cycles of full (US) AC line voltage through a generic computer speaker woofer, without destroying it. That'd be on the order of 10 ms. Unsure how many ms pass from when a relay coil energize signal deasserts, until the mechanics start moving.
Would it be faster to use a SSR to clamp the amplifier output to ground in response to some detected event; let the fuses on the +/- power supply lines blow?
Power supplies are usually more robust than amplifiers with respect to shorted outputs. If the amp had failed and that’s what triggers the DC protect I suppose there is nothing to lose. But on a recoverable DC offset or sub-Hz output, shorting it may be asking the amp to fail. One could design the VI limiters for absolute short circuit protection, but that usually means they become audible under some legitimate load conditions. Or you need a ridiculous quantity of output transistors to drive a low Z load without nuisance activation.
Once you disconnect the signal, were assuming it was a short to ground and ground is ground and amplifier output rests at ground. For DC detection a solid state relay is even too fast. If it is heat and over current at the same time ,once the reaction starts even the SSR is insufficient as at that time the die might be too far gone. For events that require early detection, electro magnetic mechanical relays are ok. SSRs are great and can be used, but they cant do magic outside a certain threshold.
All of this kind-of reminds me of the Crown ODEP circuitry > highly thoughtful, but complex.
The use of opto-coupling in protection circuits is probably under used but very good.
HOWEVER
When it comes to protecting speakers from DC there has never been a 'LAW calved in stone' stating that -
the output of amplifiers must be DC coupled. Provided you have a capacitance high enough to not introduce
phase shift or current limiting, with AC coupling you really do have fail-safe protection, and achieved in a very simple way.
A capacitor bank to achieve this can be made cost effectively by combining:
a quantity of high value/voltage electrolytics back to back + some bipolars + some quality film types. (all in parallel)
The use of opto-coupling in protection circuits is probably under used but very good.
HOWEVER
When it comes to protecting speakers from DC there has never been a 'LAW calved in stone' stating that -
the output of amplifiers must be DC coupled. Provided you have a capacitance high enough to not introduce
phase shift or current limiting, with AC coupling you really do have fail-safe protection, and achieved in a very simple way.
A capacitor bank to achieve this can be made cost effectively by combining:
a quantity of high value/voltage electrolytics back to back + some bipolars + some quality film types. (all in parallel)
33,000 uF output capacitor….
Could still put out a dangerous enough transient into a cute little 5” woofer.
Could still put out a dangerous enough transient into a cute little 5” woofer.
For beginners dont believe everything you see or hear on forums, use your common sense and lecture notes, although right now forums might be more helpful on some matters than chatgpt, next gpt update might be able to sort through forums. Sometimes veterans are just having a little fun horsing around. Hope no next generation engineers have had they minds bungled up. Even class A amplifiers burn and their control systems are sometimes as complex as a small computer. Heres an example of a burned out beloved class A amplifier, what a beauty.
The “magic” of SSRs is that they won’t weld their contacts shut in the event of trying to open in the presence of sufficient stored energy in your speaker’s VC. Relays big enough to handle the energy must be used, and they can get big and slow. And expensive. When you can find them for $2.99 from a surplus house thats one thing. But that stock is dwindling all over the world, leaving you to buy new at prices that make a pair of $10 hexfets and a couple assorted small parts quite attractive.
Those ice cube relays were more than fast enough, if the speaker is robust enough to tolerate the amp being run to hard clipping when playing music. But if the speakers are undersized and the amp only used for “headroom” better protection may be needed. And when you start looking at multi kW amps, no reasonable relay is big enough to safely open under all fault conditions. Well, the ones they use for a 5 ton AC compressor are, but you won’t find an amp with one of those in it. Hell, they probably use SSRs in the modern ones anyway.
In post 32 one can gun for 60A performance if the power supply allows especially for those bass bursts in home theatre listening however, the amp will be gone in a flash if anything goes wrong. Since our output sensing is digital in nature one can replace the thyristor with a capacitor further stretching burst performance to 60A, its quite a popular approach, but it gets quiet suddenly😉. The amplifier must also have pleasant clipping nature where the sound gets fuller. The amp should not be like walking on glass when driven hard. That's why some manufacturers stick to house brew EI cores and tube like sounding amplifiers.