Hi
Sorry for covering it once again, but if one has designed the power amp so that at disconnected one rail voltage output stayes close to ground potential, say, below a volt or two, does he need any form of DC output protection?
I guess it is now nothing wrong with rail fuses being tightly tailored to maximum expected current.
How about output relay?
regards
Adam
Sorry for covering it once again, but if one has designed the power amp so that at disconnected one rail voltage output stayes close to ground potential, say, below a volt or two, does he need any form of DC output protection?
I guess it is now nothing wrong with rail fuses being tightly tailored to maximum expected current.
How about output relay?
regards
Adam
I'm of the philosophy that you can't foresee or protect against everything, so I build robust amps with no SOA or DC protection. I do use rail fuses, but not of low enough values to prevent anything other than complete destruction of my living room. In 30 years I've never lost a speaker to an amp failure, in fact, I've never had a catastrophic amp failure anyplace other than on the test bench. Had quite a few there though! OTOH, I have had problems with bad relay contacts on commercial amps that used them. I've also seen lots of commercial amps with SOA protection that blew up anyway. That's not to say the theory is bad, just the usual implementation.
Hi Adam,
I have an opposite feeling on the subject.
I'll start by stating that output relays are easily replaced at reasonable cost. Output fuses are a bad idea.
To design an amplifier without some way of disconnecting or protecting the load (speaker) is irresponsible in my view. Some amplifiers can shut down everything quickly (Carver for example - often only one output gone). Speaker relays are useful for muting thumps and protecting the speaker. You need one rated for the energy you are interrupting. You can also short the output to ground with a triac.
Whatever you do, do something. The only time I don't worry about protection is when I am designing the basic amplifier. Protection happens after I get the rest figured out.
-Chris
I have an opposite feeling on the subject.
I'll start by stating that output relays are easily replaced at reasonable cost. Output fuses are a bad idea.
To design an amplifier without some way of disconnecting or protecting the load (speaker) is irresponsible in my view. Some amplifiers can shut down everything quickly (Carver for example - often only one output gone). Speaker relays are useful for muting thumps and protecting the speaker. You need one rated for the energy you are interrupting. You can also short the output to ground with a triac.
Whatever you do, do something. The only time I don't worry about protection is when I am designing the basic amplifier. Protection happens after I get the rest figured out.
Since I worked in audio service, I saw a constant stream of dead things. Some did cause speakers to ignite. So this is a very real possibility. Very few have expired on my bench thankfully.
-Chris
Hi Conrad...
Like anatech, I'm a tech. I see lots of blown gear, and very few bad relays. A speaker output relay is generally one of the most reliable parts of an amplifier in instances where they are utilized.
As far as your statements about relays, contacts can become worn or even pitted (or fused!) when abused or installed in an amp that has been torched a couple of times, but they are easy to source and replace. Thank God most manufacturers have not adopted your viewpoint about thinking their gear is so bulletproof that they are unnecessary.
...not to mention how handy they are for a nice turn-on delay, and immediate disconnection on power-down.
Not being able to foresee or protect against everything is not a very strong argument for making no attempt at all. I wear a helmet when I ride my motorcycle, and a seat belt when in the car...they may or may not save my life if worst comes to worst, but they improve my chances. Same philosophy ought to apply to DC and SOA protection.
Like anatech, I'm a tech. I see lots of blown gear, and very few bad relays. A speaker output relay is generally one of the most reliable parts of an amplifier in instances where they are utilized.
As far as your statements about relays, contacts can become worn or even pitted (or fused!) when abused or installed in an amp that has been torched a couple of times, but they are easy to source and replace. Thank God most manufacturers have not adopted your viewpoint about thinking their gear is so bulletproof that they are unnecessary.
...not to mention how handy they are for a nice turn-on delay, and immediate disconnection on power-down.
Not being able to foresee or protect against everything is not a very strong argument for making no attempt at all. I wear a helmet when I ride my motorcycle, and a seat belt when in the car...they may or may not save my life if worst comes to worst, but they improve my chances. Same philosophy ought to apply to DC and SOA protection.
Thank you for opinions.
Conrad
Anatech
Generally I agree, but I am asking, because a failure did really happen to my amp, positive rail got disconnected atfer some transport (Polish roads
) and only symptoms were a bit of hum and DC below 2V. Screwdriver cured the problem. Topology is fully symetrical. I guess that disconnected rail voltage is kind of worst case failure in terms of DC.
So my question is: in an amplifier designed THIS WAY to stay close to ground even with disconnected one rail voltage do I still need to worry about speaker protection?
Adam
Conrad
Yes, of course
Anatech
Generally I agree, but I am asking, because a failure did really happen to my amp, positive rail got disconnected atfer some transport (Polish roads
) and only symptoms were a bit of hum and DC below 2V. Screwdriver cured the problem. Topology is fully symetrical. I guess that disconnected rail voltage is kind of worst case failure in terms of DC.So my question is: in an amplifier designed THIS WAY to stay close to ground even with disconnected one rail voltage do I still need to worry about speaker protection?
I never used one at output. Only on rails.
Adam
Hi,
I'm agreeing with the majority.
An output relay AND rail fuses seems the safer way to go compared to nothing at all.
Once that output relay has a control circuit, with delayed start and instant off as well as DC detect, it takes just one more relay to incorporate an input mute function from the same control circuit.
After that tiny bit of circuitry is added then consider whether short circuit and/or IV protection is necessary.
BTW,
I have now tested full power from two different power amps on rail fuses that are far too small, they survived.
eg1. 311W into 4r0 on F3.1A, that's 12.5Apk through a F3.1A fuse.
eg2. 453W into 2r0 on F5A, that's 21.3Apk through a F5A fuse.
I wonder just how low these fuses can go to avoid nuisance blowing on full power music?
I'm agreeing with the majority.
An output relay AND rail fuses seems the safer way to go compared to nothing at all.
Once that output relay has a control circuit, with delayed start and instant off as well as DC detect, it takes just one more relay to incorporate an input mute function from the same control circuit.
After that tiny bit of circuitry is added then consider whether short circuit and/or IV protection is necessary.
BTW,
I have now tested full power from two different power amps on rail fuses that are far too small, they survived.
eg1. 311W into 4r0 on F3.1A, that's 12.5Apk through a F3.1A fuse.
eg2. 453W into 2r0 on F5A, that's 21.3Apk through a F5A fuse.
I wonder just how low these fuses can go to avoid nuisance blowing on full power music?
For a commercial product, you guys are absolutely right. My diy motto is "I built it, so I can fix it." That includes the speakers too. Now, if I had very expensive speakers, I might feel differently. The motto has served me well, since I very rarely have to fix anything. More parts usually equals less reliability. As for the benefits of relays, I'll agree, but they tend to change models too often and by the time the contacts become unreliable, an exact replacement might be hard to find. On the higher end amps with more open layout, they're easy to replace, but I've also seen amps where they were inaccessible without deconstructing half the amplifier. Being basically a contrarian, I've also used speaker fuses successfully, but you have to choose wisely.
Let's break things down in a few categories:
Commercial:
SOA protection. Fuses prevent catastophies.
Consumer:
You have to make it so fuses "can't be upgraded". DC protection.
Knowledgeable audiophile:
A simple approach will do. AGX speaker fuses.
On the bench I was testing a Leach amp with a 20 Amp secondary using a CVT transformer. The cap ratings were not up to snuff and one shorted blowing the rail fuse. There was nothing to shut down the AMP.
I would advocate that if one rail fuse blows the speakers disconnect and the amp shuts down. I do think clipping indicators are usefull as well, but DC protection can be, in esscense shutdown due to clipping.
Thump and pop suppression are essential and they can be integrated into the design.
I fixed a commercial disco amp, about 250 W/channel. The speaker leads were shorted by not paying attention to how they were attached. Basically a bus bar across the speaker terminals. Resistors were puddled on the circuit board. 4 other shops would not touch it. I did modify it, such that when the DC protection kicked in the added speaker relay turned off, I think.
Commercial:
SOA protection. Fuses prevent catastophies.
Consumer:
You have to make it so fuses "can't be upgraded". DC protection.
Knowledgeable audiophile:
A simple approach will do. AGX speaker fuses.
On the bench I was testing a Leach amp with a 20 Amp secondary using a CVT transformer. The cap ratings were not up to snuff and one shorted blowing the rail fuse. There was nothing to shut down the AMP.
I would advocate that if one rail fuse blows the speakers disconnect and the amp shuts down. I do think clipping indicators are usefull as well, but DC protection can be, in esscense shutdown due to clipping.
Thump and pop suppression are essential and they can be integrated into the design.
I fixed a commercial disco amp, about 250 W/channel. The speaker leads were shorted by not paying attention to how they were attached. Basically a bus bar across the speaker terminals. Resistors were puddled on the circuit board. 4 other shops would not touch it. I did modify it, such that when the DC protection kicked in the added speaker relay turned off, I think.
Hi Adam,
There is a simple protection for rail reversal if one rail goes open. Place reverse biased fuses from supply (after the fuse) to ground. Protected this way a supply rail will only ever go one diode drop reversed.
I have tested some amplifier designs this way and it works well. Also, if your voltage map stage runs off it's own supply, your output voltage will tend to remain close to 0 VDC if a rail fuse blows. If you run your front end off it's own regulator (good idea for sound quality), this happens naturally.
-Chris
There is a simple protection for rail reversal if one rail goes open. Place reverse biased fuses from supply (after the fuse) to ground. Protected this way a supply rail will only ever go one diode drop reversed.
I have tested some amplifier designs this way and it works well. Also, if your voltage map stage runs off it's own supply, your output voltage will tend to remain close to 0 VDC if a rail fuse blows. If you run your front end off it's own regulator (good idea for sound quality), this happens naturally.
-Chris
looking at rms current and that the two rail fuses are sharing the time, it would appear that I could have done the full power tests on even smaller fuses.Nordic said:
12.5Apk=8.8Arms. F2A passes 4Arms for 1sec and F2.5A passes 5Arms for 2sec.
Similarly the 21.3Apk could be passed by F4A.
I shall need to try one size smaller and see if fatigue becomes a problem with prolonged use.
Nordic,
do you have tables/graphs for standard fuses and for T rated fuses?
My basic protection approach would include:
- Either rail fuses for traditional transformer/rectifier/caps PSU, or PSU current limiting/shutdown in case of SMPS. A switching power supply can detect overcurrent and shut itself down fast enough to leave you with just a single output device shorted, as Anatech pointed out. Rail fuses require diodes to ground after them to prevent reverse biasing of the rails, and those diodes should be rugged enough so that the other fuse can blow before the diode itself blows in case that one pair of output devices becomes shorted. PCB track sizes have also to be considered to prevent them from blowing, particularly the ground path of those diodes.
- Single slope current limiting for linear output stages. This is effective and very easy to implement, usually requires two transistors, two diodes and a few resistors. Current limit becomes inversely proportional to output transistor Vce/Vds.
- Either average-current or peak-current limiting for class D switching output stages. With average current limiting, the overcurrent event will become almost unnoticed as in linear amplifiers (this is my preferred method). On the other hand, with peak limiting you can either terminate the current switching cycle (which usually results in control loop instability and audible noise) or shut down the channel for some time (typical approach), which is also obviously noticed.
-Triac crowbar, triggered in case DC is detected at the output. This does not even require a dedicated supply voltage as it happens with relay coils, the own DC is used to trigger the triac with the help of a low-pass filter. There is a wide selection of low cost high current triacs that are rugged enough to blow rail fuses, furthermore, such triacs will fail shorted if they are abused, unless the I^2*t product of their bonding wires is lower than the I^2*t of the fuse (only designer should care about this). Also, PCB has to be laid out in such a way that in case one trace blows after triac triggering, the speaker becomes disconnected rather than the triac (this would only happen if I^2*t products of PCB tracks are lower than I^2*t products of fuses, which should never happen if a fuse of the right size is employed).
- Either rail fuses for traditional transformer/rectifier/caps PSU, or PSU current limiting/shutdown in case of SMPS. A switching power supply can detect overcurrent and shut itself down fast enough to leave you with just a single output device shorted, as Anatech pointed out. Rail fuses require diodes to ground after them to prevent reverse biasing of the rails, and those diodes should be rugged enough so that the other fuse can blow before the diode itself blows in case that one pair of output devices becomes shorted. PCB track sizes have also to be considered to prevent them from blowing, particularly the ground path of those diodes.
- Single slope current limiting for linear output stages. This is effective and very easy to implement, usually requires two transistors, two diodes and a few resistors. Current limit becomes inversely proportional to output transistor Vce/Vds.
- Either average-current or peak-current limiting for class D switching output stages. With average current limiting, the overcurrent event will become almost unnoticed as in linear amplifiers (this is my preferred method). On the other hand, with peak limiting you can either terminate the current switching cycle (which usually results in control loop instability and audible noise) or shut down the channel for some time (typical approach), which is also obviously noticed.
-Triac crowbar, triggered in case DC is detected at the output. This does not even require a dedicated supply voltage as it happens with relay coils, the own DC is used to trigger the triac with the help of a low-pass filter. There is a wide selection of low cost high current triacs that are rugged enough to blow rail fuses, furthermore, such triacs will fail shorted if they are abused, unless the I^2*t product of their bonding wires is lower than the I^2*t of the fuse (only designer should care about this). Also, PCB has to be laid out in such a way that in case one trace blows after triac triggering, the speaker becomes disconnected rather than the triac (this would only happen if I^2*t products of PCB tracks are lower than I^2*t products of fuses, which should never happen if a fuse of the right size is employed).
Good information, Nordic!
I might add a PDF-catalog for further information:
http://www.qualtekhk.com/pdf/nonfuses_section.pdf
The main thing to be learned: The fuse's rated current is a current that will always flow, it's by no means the trip current (took me years to get that).
Regards, Klaus
I might add a PDF-catalog for further information:
http://www.qualtekhk.com/pdf/nonfuses_section.pdf
The main thing to be learned: The fuse's rated current is a current that will always flow, it's by no means the trip current (took me years to get that).
Regards, Klaus
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.