This should also be detected by HF Ringing Sensors.
In my experience, woman and kids are Tweeter killers, not amplifiers.
Always if they see a shiny round thingie on a speaker, its unavoidable that they put their finger in it, you tell them, sorry you killed my TW, then its unavoidable they go to second speaker, do the same and say, hey, i just touched like it this one, you stupid
...Groove T, I completely agree on BOTH counts. My son was attracted to the tweeter of my wife's speakers when he was around three. I caught him just before he stuck his finger in the titanium tweeter and gave him the hiding of his life, the only one he ever got. But he never tried to repeat the feat.
Women are far more dangrous. A cleaning lady was explained that she was under no circumsrances to come near to my speakers and especially the titanium tweeter, and sure enough, it was bent all over the very same day. Like waving a red blanket in front of the bull. At first, she denied any knowledge, but later on said "But I only touched it lightly!" Solved that one by making her pay about $150 for a new tweeter. Never again, curiousity is expensive.
Women are far more dangrous. A cleaning lady was explained that she was under no circumsrances to come near to my speakers and especially the titanium tweeter, and sure enough, it was bent all over the very same day. Like waving a red blanket in front of the bull. At first, she denied any knowledge, but later on said "But I only touched it lightly!" Solved that one by making her pay about $150 for a new tweeter. Never again, curiousity is expensive.
I remember the one going around about 10 years ago about travellers to the Far East who were getting drugged and waking up in a bath, covered in blood, with a note saying 'call this number' to the nearest hospital to get urgent medical attention. They then found out one of their kidneys was gone. It appears there was a kidney harvesting industry in full swing to service rich Americans.
I was at that time (before moving out there in 2006) always on flights to China, Taiwan, Malaysia, Thailand etc. Scared the hell out of a lot of folk.
I never did find out who took my one kidney on the last trip though 😀
I was at that time (before moving out there in 2006) always on flights to China, Taiwan, Malaysia, Thailand etc. Scared the hell out of a lot of folk.
I never did find out who took my one kidney on the last trip though 😀
Snopes is always handy for when people spout egregiously stupid stuff (not you, Bonsai).
Kidney Theft : snopes.com
Your legend is actually more venerable than ten years, it was already old when Brunvand collected it in 1991.
Kidney Theft : snopes.com
Your legend is actually more venerable than ten years, it was already old when Brunvand collected it in 1991.
Well, here we go again. No audio information being discussed.
For the record, most serious audio amplifier designers now use protection that either they independently develop, or use specific IC chips that the Japanese made over the decades to protect their power amps. Of course a relatively small amplifier, like 50W into 8 ohms might not need too much protection if done with state of today devices. Perhaps just DC line fuses, to keep it from going into some mode that might wreck the amp, but when you go to 100W/channel or more, then other problems develop and you need some sort of E-I sensor, carefully calibrated to not fire in normal operation, and then we usually use some sort of relay turnoff in order to avoid turning the output stage in common with the loudspeaker into the equivalent of a spark coil ignition circuit. This is why we don't use the 'simple' E-I protection anymore.
These Japanese IC protection chips do other things too! They note if there is excessive DC offset that might damage the speaker, and detect if there is a lot of ultrasonic oscillation that would burn your tweeters. Other things can be detected as well, such as short circuit detection, and output device overheating. All in all it is the best way to go, but LESS IS BEST!
For the record, most serious audio amplifier designers now use protection that either they independently develop, or use specific IC chips that the Japanese made over the decades to protect their power amps. Of course a relatively small amplifier, like 50W into 8 ohms might not need too much protection if done with state of today devices. Perhaps just DC line fuses, to keep it from going into some mode that might wreck the amp, but when you go to 100W/channel or more, then other problems develop and you need some sort of E-I sensor, carefully calibrated to not fire in normal operation, and then we usually use some sort of relay turnoff in order to avoid turning the output stage in common with the loudspeaker into the equivalent of a spark coil ignition circuit. This is why we don't use the 'simple' E-I protection anymore.
These Japanese IC protection chips do other things too! They note if there is excessive DC offset that might damage the speaker, and detect if there is a lot of ultrasonic oscillation that would burn your tweeters. Other things can be detected as well, such as short circuit detection, and output device overheating. All in all it is the best way to go, but LESS IS BEST!
UPC1237 for all the old 20th century lower power stuff.
Ten transistor DC/E-I/Flip-flop ... simple chip.
What other Asian protection IC's are there , and which one would cherry pick "ultrasonic oscillation" as a shutdown condition ?
"Independently developed " would be this -
http://www.diyaudio.com/forums/solid-state/264313-how-build-21-st-century-protection-board.html
I'm running one , takes care of both power supply and amp errata.
I can't destroy either my speakers or amp - even with extended "fiddling".
The new versions even monitor and disconnect rails.
(I still believe in oversized fuses - combined with active E-I )
OS
Ten transistor DC/E-I/Flip-flop ... simple chip.
What other Asian protection IC's are there , and which one would cherry pick "ultrasonic oscillation" as a shutdown condition ?
"Independently developed " would be this -
http://www.diyaudio.com/forums/solid-state/264313-how-build-21-st-century-protection-board.html
I'm running one , takes care of both power supply and amp errata.
I can't destroy either my speakers or amp - even with extended "fiddling".
The new versions even monitor and disconnect rails.
(I still believe in oversized fuses - combined with active E-I )
OS
Even with a large nearly indestructible output stage you almost need some sort of protection if you want to pass certification standards. These testing labs find unusual ways to break things some of which I can't see happening in the real world. Then again my Swiss iron says not to iron clothes well wearing them so people do interesting things.
Amps usually don't get certified by safety testing labs. However, it is EXPENSIVE and dangerous to not have decent protection, especially overheating.
It never used to be required to have safety testing but is getting hard to ship to some places without it. I suspect for low volumes many companies don't test or just put stickers on saying they did especially for CE.
Yet as far as I can tell people only believe the opposite. The more money they spend, the more they believe this to be true: they assume trusted companies test their gear as if they're utterly unsure of the engineering behind it... Customers rarely seem to understand the difference between engineering something and total experiment. They're basically synonyms to them because they don't understand it; and often refuse to. For the same reason they may have total fear of say swapping internal parts of the same value.
How fashionable or unfashionable are microcontroller ICs these days, when implementing protection(s) in > USD10K high end audio products? Undeniably they are digital devices and they run at high frequencies compared to audio. BUT uCs now have 100% internal, on-die clock generator circuits, so there are no pins or nets or PCB traces swinging 5 volts at 1000 volts per microsecond. Also uCs can operate from an LRC isolated and sub-regulated supply, and an LRC isolated ground (if you talk to them via optoisolators), so whatever crud they happen to create on their own power pins, is deeply attenuated before it can get anywhere near 130 dB SNR analog circuits.
Me I've used uCs for soft start timing(s) and for output mute timing(s), simply because the uC solution occupies significantly less board space than the all-analog solution. But those were daughterboards in the AC mains section of the chassis, far removed from the analog payload circuitry on 6 layer teflon boards with silk dielectric capacitors and unobtanium 2SJ74BL PJFETs from someone's private collection.
Me I've used uCs for soft start timing(s) and for output mute timing(s), simply because the uC solution occupies significantly less board space than the all-analog solution. But those were daughterboards in the AC mains section of the chassis, far removed from the analog payload circuitry on 6 layer teflon boards with silk dielectric capacitors and unobtanium 2SJ74BL PJFETs from someone's private collection.
I find many of the MCU based "solutions" to be less reliable than some higher-parts-count logic gate and transistor based circuits, mainly because of firmware issues that haven't been thoroughly thought out.
I recently did a headphone amp design and used a rather brute-force, inefficient and inelegant approach to protecting the output devices of an all-discrete CFA amplifier: I put some resistors in the collectors of the bipolar emitter-follower output parts. We will see how it sounds.
For protecting loudspeakers, one really needs an accurate "plant model" of the transducers. Since the great emphasis is on teeny-tiny ones, a lot of semi houses are devoting considerable effort to this, and apparently getting some successes. There is no reason why this couldn't be applied to larger drivers. The appeal is that you could really stay "out of the way" until some threshold for damage was approached. Individual amps per transducer would make life a little easier.
I am using now in all my amps this kind of PS regulator with all protection needed, overcurrent protection, short circuit protection, DC offset at the amp output. http://www.diyaudio.com/forums/powe...ply-cap-multiplier-electronic-protection.html
This is similar approach JLH used in his 80W MOSFET amp, and in my opinion that is the best way to do amp and loudspeaker protection, no high frequency noise from microcontroller, and addition, there is PS voltage regulation too.
http://www.diyaudio.com/forums/solid-state/243481-200w-mosfet-cfa-amp-92.html#post4580606
cheers Damir
This is similar approach JLH used in his 80W MOSFET amp, and in my opinion that is the best way to do amp and loudspeaker protection, no high frequency noise from microcontroller, and addition, there is PS voltage regulation too.
http://www.diyaudio.com/forums/solid-state/243481-200w-mosfet-cfa-amp-92.html#post4580606
cheers Damir
My take on this protects the amp and speakers from:
1, Overvoltage
2. Overcurrent
3. Excess DC
4. Overheating
5. Overdrive (via a small LED indicator panel)
6. Initial thump on switch-on
I evolved this over the years and the whole is actually split up around the place and is in fact split into a total of three boards.
The reason why this is so is because people have asked to not use this or that part, as the hype from the High End sector has blackened the name of any protection circuit, and people are afraid they might interfere with the sonic performance. And I actually believe classic V/I limiting circuits can be adjusted for full protection function with no audible side effects simpy by using a turn on delay. Perferctly good circuits provided one pays some attention to them, and over the years I saw them activate only a few times, and that only when the amp was severely overdriven with behemoth speakers capable of really accepting bursts exceeding the maximum safe voltage, by which time you already had the whole Syrian war compressed into your room.
The added advantage is that the protection function can be very predictably worked out abd will work guaranteed to fully utilize the capacity of the output stage, using data for 250 mS power bursts, always more than continuous. Also, instead of using standard fuses, one can use self resetting thermal switches.
But I do like Dado's solution.
1, Overvoltage
2. Overcurrent
3. Excess DC
4. Overheating
5. Overdrive (via a small LED indicator panel)
6. Initial thump on switch-on
I evolved this over the years and the whole is actually split up around the place and is in fact split into a total of three boards.
The reason why this is so is because people have asked to not use this or that part, as the hype from the High End sector has blackened the name of any protection circuit, and people are afraid they might interfere with the sonic performance. And I actually believe classic V/I limiting circuits can be adjusted for full protection function with no audible side effects simpy by using a turn on delay. Perferctly good circuits provided one pays some attention to them, and over the years I saw them activate only a few times, and that only when the amp was severely overdriven with behemoth speakers capable of really accepting bursts exceeding the maximum safe voltage, by which time you already had the whole Syrian war compressed into your room.
The added advantage is that the protection function can be very predictably worked out abd will work guaranteed to fully utilize the capacity of the output stage, using data for 250 mS power bursts, always more than continuous. Also, instead of using standard fuses, one can use self resetting thermal switches.
But I do like Dado's solution.
Hear, hear.
Especially in these days of ever shrinking heat sinks in commercial gear.
There are different kinds of 'protection'.
We all agree that major DC offset isolation via some form of implementation is required.
The typical implementation is a series output relay controlled by a sensing circuit (IC), which usually also provides turn on 'thump' delay and immediate isolation in the case of loss of AC power, ie turn off.
One refinement of this approach is incorporated in Accuphase E202 (maybe other models also ?) where upon power up, the load impedance/resistance is sensed before enabling the series output relay.....ie if the load resistance is lower than limits the output relay will not operate....clever and sensible, maybe unique ?.
Other approaches for DC protection include 'crowbar' short protection of the load, causing blowing of AC or DC supply fuses....also perfectly valid.
The next subject is output device dissipation protection.
Rod Elliot has a page - Safe Operating Area (SOA) for semiconductors
Our Bonsai/Andrew has published this investigation - THE SAFE OPERATING AREA (SOA) PROTECTION OF LINEAR AUDIO POWER AMPLIFIERS
Crown (Amcron) take things to another level with ODEP protection/sensing - Crown’s ODEP Circuit
I generally run a pair of tweaked Behringer 2031A 8" 2 way powered speakers (LM3886).
These active loudspeakers run an optocoupler that shunts the input signal in the case of sustained/average 'defined' overload.
In practice this approach works very well and is reasonably sonically transparent and in the case of overdrive rather than objectionable clipping distortion, this causes useful compression which can make some types of music (rock/blues) even more fun.
Crown run an indicator for overload/distortion condition - Input-Output Comparator
The mid-fi consumer world is full of shelf systems/AV systems that run 'soggy/saggy' power supplies....interestingly these systems generally do not get nasty when driven into overload, but become limited/compressed and carry on party after party without dire consequence to amplifier/loudspeaker.
Loudspeaker protection and amplifier protection is a necessary property of any commercially available audio system.
Instantaneous output transistor peak current protection/limiting as has been typically implemented by output transistor emitter resistor voltage detection/sensing is subjectively sub optimal due to alteration of half cycles which causes 'un-natural' distortions and consequent spectral alteration.
Any 'transparent' SOA protection scheme must be symmetrical in nature.
Dan.
We all agree that major DC offset isolation via some form of implementation is required.
The typical implementation is a series output relay controlled by a sensing circuit (IC), which usually also provides turn on 'thump' delay and immediate isolation in the case of loss of AC power, ie turn off.
One refinement of this approach is incorporated in Accuphase E202 (maybe other models also ?) where upon power up, the load impedance/resistance is sensed before enabling the series output relay.....ie if the load resistance is lower than limits the output relay will not operate....clever and sensible, maybe unique ?.
Other approaches for DC protection include 'crowbar' short protection of the load, causing blowing of AC or DC supply fuses....also perfectly valid.
The next subject is output device dissipation protection.
Rod Elliot has a page - Safe Operating Area (SOA) for semiconductors
Our Bonsai/Andrew has published this investigation - THE SAFE OPERATING AREA (SOA) PROTECTION OF LINEAR AUDIO POWER AMPLIFIERS
Crown (Amcron) take things to another level with ODEP protection/sensing - Crown’s ODEP Circuit
I generally run a pair of tweaked Behringer 2031A 8" 2 way powered speakers (LM3886).
These active loudspeakers run an optocoupler that shunts the input signal in the case of sustained/average 'defined' overload.
In practice this approach works very well and is reasonably sonically transparent and in the case of overdrive rather than objectionable clipping distortion, this causes useful compression which can make some types of music (rock/blues) even more fun.
Crown run an indicator for overload/distortion condition - Input-Output Comparator
The mid-fi consumer world is full of shelf systems/AV systems that run 'soggy/saggy' power supplies....interestingly these systems generally do not get nasty when driven into overload, but become limited/compressed and carry on party after party without dire consequence to amplifier/loudspeaker.
Loudspeaker protection and amplifier protection is a necessary property of any commercially available audio system.
Instantaneous output transistor peak current protection/limiting as has been typically implemented by output transistor emitter resistor voltage detection/sensing is subjectively sub optimal due to alteration of half cycles which causes 'un-natural' distortions and consequent spectral alteration.
Any 'transparent' SOA protection scheme must be symmetrical in nature.
Dan.
Last edited:
Dan, that paper was written by Michael Kawanuka - Its a good paper, but I cannot take credit for it - I merely agreed to put it up on my website for him. (I believe he is persona non-grata in these parts nowadays 😉 )
For my sins, I've used a microcontroller to do the protection on my e-amp. It does overtemperature, current (sensed across the emitter degen resistors via an opto), DC offset, output muting using a mosfet SSR and in-rush current limiting for the transformer.
Most uCs can be put into a sleep state with the clock OFF, but still able to wake-up on an interrupt so they generate no clock noise whatsoever during normal operation.
I used analog protection on my nx-Amp and it was a little problemetic on some builds, but solved now.
On a professional amp, I would always recommend going the uC route.
😎
For my sins, I've used a microcontroller to do the protection on my e-amp. It does overtemperature, current (sensed across the emitter degen resistors via an opto), DC offset, output muting using a mosfet SSR and in-rush current limiting for the transformer.
Most uCs can be put into a sleep state with the clock OFF, but still able to wake-up on an interrupt so they generate no clock noise whatsoever during normal operation.
I used analog protection on my nx-Amp and it was a little problemetic on some builds, but solved now.
On a professional amp, I would always recommend going the uC route.
😎
And then for DC protection, there is also the good old, effective but unfashionable output capacitor method.
- Status
- Not open for further replies.