Hi,
I was unsurprised.
Non I care to to use.
I work on the principle that the speaker protection is there to protect the speaker from a defective amplifier. I therefore use crowbar circuits on the principle that when they actually engage the amp is already broken.
Just in case this is actually not the case and the protection circuit was triggered by DC on the CD (there are some such) I first mute the Amp which has a relay that pulls much faster than the big output crowbar.
So far I have not had any amp fail that had this arrangement and take a speaker out and I am not aware of any Amp that was taken out by the protection circuit engaging in error, from multi-killerwatt 20" borg cube Pro-Amp's I build in the 80's (they used Fets for mutes and industrial triacs for crowbar) to more modest domestic Amp's.
On the other hand in my Pro-Day's I had multiple times speakers taken out by amp's blowing up when the relay did not release cleanly and instead arced and welded shut...
Ciao T
I was unsurprised.
Non I care to to use.
I work on the principle that the speaker protection is there to protect the speaker from a defective amplifier. I therefore use crowbar circuits on the principle that when they actually engage the amp is already broken.
Just in case this is actually not the case and the protection circuit was triggered by DC on the CD (there are some such) I first mute the Amp which has a relay that pulls much faster than the big output crowbar.
So far I have not had any amp fail that had this arrangement and take a speaker out and I am not aware of any Amp that was taken out by the protection circuit engaging in error, from multi-killerwatt 20" borg cube Pro-Amp's I build in the 80's (they used Fets for mutes and industrial triacs for crowbar) to more modest domestic Amp's.
On the other hand in my Pro-Day's I had multiple times speakers taken out by amp's blowing up when the relay did not release cleanly and instead arced and welded shut...
Ciao T
I am confused already on questions how acceleration of air movement is related to electrical power applied to the speaker, how SPL is related to loudness, and audibility of reproduction errors to SPL.
My point was, subconscious competence to kill that flying beasts without paying attention. Just by sound localized in 3-D around head and body. Now, calculate precision that subconscious mind uses to localize sound sources when they are dangerous, or annoying...
Hi,
Sure. Sounds like fun.
Would you mind if I decline any actual empirical research on the subject and leave that part to you...
I already did my stint in west africa.
Ciao T
Sure. Sounds like fun.
Would you mind if I decline any actual empirical research on the subject and leave that part to you...
I already did my stint in west africa.
Ciao T
Hmmm, If I sense DC, then I really should put the relay in the AC and that way it stays clear of the audio. It would let me put in a remote trigger to. Then rely on the fuses to protect the amp. It does not solve a turn on delay mute though.
That leads me to another question. I have seen the speaker fuses put inside the feedback loop. The DH-200 does so. If the fuse blows, does that not remove the feedback and do potentially not very good things to the amp, or do things just bounce off the rails a bit and do no damage as the output current is zip?
That leads me to another question. I have seen the speaker fuses put inside the feedback loop. The DH-200 does so. If the fuse blows, does that not remove the feedback and do potentially not very good things to the amp, or do things just bounce off the rails a bit and do no damage as the output current is zip?
Err, aren't DC protection circuits in amplifiers to protect the speakers not the amp ?
Once there is a fault that is putting DC on the speaker output something major has usually already blown, (shorted output transistor etc) and its too late for the amp. Let the fuse blow to prevent a fire or complete meltdown of the amp...
The important thing is to get the DC disconnected from the speakers ASAP, (within milliseconds) and that can only be done by disconnecting the output of the amp from the speakers.
If you just switch off the AC (before or after the mains transformer) there will be enough storage in the filter capacitors to maintain DC output long enough to potentially do serious damage to the woofers, and it may prevent the fuse from blowing as well - thus subjecting the speakers to second and subsequent blasts of DC if the owner switches the amp off and on again a few times.
With a conventional DC protection system with the relay on the output the speakers are disconnected quickly enough to avoid damage and the fault current is allowed to continue long enough to blow the fuse thus preventing any further DC spikes if the equipment is turned off and on. (Albeit at the expense of probably blowing up more components in the output stage)
Good amps are relatively cheap, good speakers can be very expensive. (Not to mention the sentimental attachment if they're DIY) I know which of the two I'd rather be protecting in the event of catastrophic failure
Last edited:
Appreciating this is yet another very complex problem, and neither easy nor cheap to do well. Going on the assumption a DC fault is inductive of blowing half the outputs or a driver already, then the crowbar approach makes sense. Thinking out loud here, I could then assume a triac would be sufficient to protect the speakers and eventually blow the rail fuses if an output was shorted. To keep it from stewing in it's own juices, I still should take down the AC. Fifty percent solution maybe. An external fault like a shorted cable would still likely take the amp out. I can vouch for a fuse not protecting the amp anyway. (Waiting on my mosfets from England).
Then I address the turn on delay with just a decent relay on the output. But now I need to protect it from arcing, so I need to mute the input until it has closed. Round and round it goes. I can see the advantage in just buying an amp and assuming the designer did everything perfectly. Hard to worry about what you don't know to worry about.
Going by my rule that I am too cheap to follow, speakers should cost several times the amp or the next best investment would still be speakers. Well, I buy my amps cheap used, and build my speakers, so maybe not "several times".
Then I address the turn on delay with just a decent relay on the output. But now I need to protect it from arcing, so I need to mute the input until it has closed. Round and round it goes. I can see the advantage in just buying an amp and assuming the designer did everything perfectly. Hard to worry about what you don't know to worry about.
Going by my rule that I am too cheap to follow, speakers should cost several times the amp or the next best investment would still be speakers. Well, I buy my amps cheap used, and build my speakers, so maybe not "several times".
Yes. A fellow I used to know in Salt Lake City developed a triac crowbar circuit which he used on his modified SWTP Tiger amps. He removed all other protection circuitry, and if the circuit sensed DC or an excursion outside the output devices' SOA, the triac would short the power rails and blow a fuse. It was a remarkably good solution; he used to demonstrate it by running the output at full power, then shorting it. The circuit worked so fast, he didn't even get a spark or a pop, the amp would just go black.
John,
Exactly what I am attempting to do. As I am not exactly 20 any more and intend to retire before I could quit and go back to finish my EE, it's all self study. Textbooks don't answer questions, so I have to rely on forums with real people for where the text leaves off. Believe it or not, this kind of study is stress relief compared to trying to put RDBMS problems into Mongo or RDF as they are too big for an RDBMS and too relational for any thing else.
Device protection seems far from "sound quality vs. measurement" discussion but if I twist the argument sufficiently I could say an amplifier or speaker that has failed definitely has very bad sound quality.
I was about to write more about this when I recalled writing about it before:
http://www.diyaudio.com/forums/digital-source/170966-microcontroller-based-protection-system.html
And for what it's worth I stand by what I wrote then.* If one is going to have protection, a well-designed and programmed system involving a microcontroller is better (as in it can have an envelope in which it can detect more faults than a simple analog circuit can, including things such as whether an output transistor is operating within its SOA) than a few analog components.
And yes, the design of such complicated things goes "round and round." But that's true of most everything. A "straightforward" standard audio power amplifier design has a dozen or more semiconductors when half that many would provide the same gain and power output (if not the same low distortion!).
Getting to the Moon is a lot more complicated than just having a rocket with enough thrust to get you there.
* Some people seem to be allergic to microcontrollers and computers in general when it comes to "important" applications. I've sometimes shared this apprehension. Over 30 years ago when I was in college for electrical engineering a classmate and I discussed things such as a car with a gas pedal that controls a potentiometer rather than a carburetor throttle linkage. He said he wouldn't drive a car in which a gas pedal controls a pot that gets read by a microprocessor for acceleration, and I agreed. But these kinds of devices (and cars overall nowadays) are much more reliable than what we saw and experienced back then in the Apple ][ days.
ETA: a couple more thoughts on this. Here is a mission-critical computer, which if you look at "Applications outside Apollo" it led to other, more complex mission-critical computers:
http://en.wikipedia.org/wiki/Apollo_Guidance_Computer
Also, writing reliable protection code may be beyond the abilities of many DIY'ers, but I have no doubt there are several here who COULD do it. It may not be worth it for a DIY design, but it could pay for itself in reducing repair numbers and costs in a commercial design, especially a powered, multi-amped speaker system, as I mention in the other thread.
I was about to write more about this when I recalled writing about it before:
http://www.diyaudio.com/forums/digital-source/170966-microcontroller-based-protection-system.html
And for what it's worth I stand by what I wrote then.* If one is going to have protection, a well-designed and programmed system involving a microcontroller is better (as in it can have an envelope in which it can detect more faults than a simple analog circuit can, including things such as whether an output transistor is operating within its SOA) than a few analog components.
And yes, the design of such complicated things goes "round and round." But that's true of most everything. A "straightforward" standard audio power amplifier design has a dozen or more semiconductors when half that many would provide the same gain and power output (if not the same low distortion!).
Getting to the Moon is a lot more complicated than just having a rocket with enough thrust to get you there.
* Some people seem to be allergic to microcontrollers and computers in general when it comes to "important" applications. I've sometimes shared this apprehension. Over 30 years ago when I was in college for electrical engineering a classmate and I discussed things such as a car with a gas pedal that controls a potentiometer rather than a carburetor throttle linkage. He said he wouldn't drive a car in which a gas pedal controls a pot that gets read by a microprocessor for acceleration, and I agreed. But these kinds of devices (and cars overall nowadays) are much more reliable than what we saw and experienced back then in the Apple ][ days.
ETA: a couple more thoughts on this. Here is a mission-critical computer, which if you look at "Applications outside Apollo" it led to other, more complex mission-critical computers:
http://en.wikipedia.org/wiki/Apollo_Guidance_Computer
Also, writing reliable protection code may be beyond the abilities of many DIY'ers, but I have no doubt there are several here who COULD do it. It may not be worth it for a DIY design, but it could pay for itself in reducing repair numbers and costs in a commercial design, especially a powered, multi-amped speaker system, as I mention in the other thread.
Last edited:
Tvr, you are right, about solid state protection, but in a 'sophomoric' sort of way. Now that I have your attention let's discuss the problem, in general.
First, when you stated 'problems into mongo', I had to look it up on Google. I presume you mean mongodb.
You see, my understanding of 'mongo' was that it was a planet in a serial film in the 1930's where Ming ruled.
Or, Mongo Santa Maria, a drummer in Cal Tjader's (sp) jazz group.
Or Mongo, the giant character in 'Blazing Saddles'.
However, if I had a question about your use and interest in 'mongo' I would not think any of these three definitions gave me any hint to what you were talking about.
Now, what would you think, if I started to ask you, in detail, WHY you use 'mongo' and why not Fortran or something else?
This is what I am attempting to point out.
Now for a general answer to your general question about solid state amplifier protection.
We have been looking at the problem for about 50 years.
First, was probably a fuse placed in the output of the amp, OPEN LOOP.
This protected the speaker, but fuses are nonlinear, even if it is difficult to measure their nonlinearity with continuous tones. So what to do?
Second, some put the fuse within the feedback loop. Same nonlinearity, but reduced in amplitude. To keep the amp from 'freaking out' with the opening of the fuse, often, a resistor is put in parallel with the fuse, to keep some sort of continuity.
About 45 years ago, effective E-I sensing protection circuits were developed. You might look at the Crown 300 power amp to learn more. This is still the big 'mid-fi' favorite. Its problem is that it is often audible, because it turns the output into a current source, when activated, and with the loudspeaker attached, a SPARK COIL. That is the major reason (at least as far as I know) why we have REVERSE DIODES across the supplies to the amp output, in many designs.
Parasound uses output relays, a compromise, but a necessary one, with such high single output power. Relays are problematic, some of my colleagues even burn them out in my designs, BUT they DO protect the loudspeaker, and they can remove turn-on pops, and provide overtemp protection as well.
Finally, triacs have been used over the decades. Their big problem is where to dump the energy, if you attempt to remove the energy in the power supply to protect the amp and loudspeakers.
I hope that this helps.
First, when you stated 'problems into mongo', I had to look it up on Google. I presume you mean mongodb.
You see, my understanding of 'mongo' was that it was a planet in a serial film in the 1930's where Ming ruled.
Or, Mongo Santa Maria, a drummer in Cal Tjader's (sp) jazz group.
Or Mongo, the giant character in 'Blazing Saddles'.
However, if I had a question about your use and interest in 'mongo' I would not think any of these three definitions gave me any hint to what you were talking about.
Now, what would you think, if I started to ask you, in detail, WHY you use 'mongo' and why not Fortran or something else?
This is what I am attempting to point out.
Now for a general answer to your general question about solid state amplifier protection.
We have been looking at the problem for about 50 years.
First, was probably a fuse placed in the output of the amp, OPEN LOOP.
This protected the speaker, but fuses are nonlinear, even if it is difficult to measure their nonlinearity with continuous tones. So what to do?
Second, some put the fuse within the feedback loop. Same nonlinearity, but reduced in amplitude. To keep the amp from 'freaking out' with the opening of the fuse, often, a resistor is put in parallel with the fuse, to keep some sort of continuity.
About 45 years ago, effective E-I sensing protection circuits were developed. You might look at the Crown 300 power amp to learn more. This is still the big 'mid-fi' favorite. Its problem is that it is often audible, because it turns the output into a current source, when activated, and with the loudspeaker attached, a SPARK COIL. That is the major reason (at least as far as I know) why we have REVERSE DIODES across the supplies to the amp output, in many designs.
Parasound uses output relays, a compromise, but a necessary one, with such high single output power. Relays are problematic, some of my colleagues even burn them out in my designs, BUT they DO protect the loudspeaker, and they can remove turn-on pops, and provide overtemp protection as well.
Finally, triacs have been used over the decades. Their big problem is where to dump the energy, if you attempt to remove the energy in the power supply to protect the amp and loudspeakers.
I hope that this helps.
Yea, MongoDb, Hbase, etc. It's what yo do when you have a petabyte of data. The guys up on top of the hill from you can tell you all about it. Really sharp folks up at LBL. Some rather interesting places to eat along the road there too.
Anyway. Protection. This is home DIY scale.
If the triac could handle at least as much power until both the rail fuses blew,
and the output transistors could handle more than the rail fuses,
and an output relay needed for muting anyway would disconnect the speaker,
and the sense circuit also shut down the AC, as that relay needs to be there for remote control anyway,
then traditional bleeder resistors across the filter bank should take care of the rest. Trip the triac on excess or sustained DC. 50%, 20 year old solution, but reasonable at the DIY scale? I need to add the remote relay and turn on delay anyway.
Off to search for old Crown schematics.
I proved the simple rail and output fuses in a DH-120 are not sufficient.
Do I need to do anything to protect the output relay from arcing, like mute the input until the output picks, and reverse on power down so it never opens with substantial voltage across it? Low resistance contacts if I remember are not happy about arcing. They may have been improved.
Anyway. Protection. This is home DIY scale.
If the triac could handle at least as much power until both the rail fuses blew,
and the output transistors could handle more than the rail fuses,
and an output relay needed for muting anyway would disconnect the speaker,
and the sense circuit also shut down the AC, as that relay needs to be there for remote control anyway,
then traditional bleeder resistors across the filter bank should take care of the rest. Trip the triac on excess or sustained DC. 50%, 20 year old solution, but reasonable at the DIY scale? I need to add the remote relay and turn on delay anyway.
Off to search for old Crown schematics.
I proved the simple rail and output fuses in a DH-120 are not sufficient.
Do I need to do anything to protect the output relay from arcing, like mute the input until the output picks, and reverse on power down so it never opens with substantial voltage across it? Low resistance contacts if I remember are not happy about arcing. They may have been improved.
- Status
- Not open for further replies.