Then, by the same argument, fuses have the exact same fault ... times two. The push on connectors on the speaker driver, crossover, terminals and all cables.
So everything cancels as there is no escape from contacts. When a relay becomes bad, change it. Just like everything else in this world.
There were other options given as well.
-Chris
So everything cancels as there is no escape from contacts. When a relay becomes bad, change it. Just like everything else in this world.
There were other options given as well.
-Chris
Gentlemen (and Ladies, if we are so blessed),
I don't look here very often, so I missed these.
(a) Ability of MOSFETs to handle the voltages and currents involved in cutting off a shorted-output amplifier:
The IRF IRFB260N is a 200V, 56A n-channel device. The max surge current on this device is 220A.
A power amplifer will likely have one output shorted, and so half the power supply will be funneled through the DC resistance of the speaker. The current will be perhaps as much as 90V through 2 or more ohms of wire. The fault current is then less than the rated max current for the device.
I think it can turn off that fault current with no damage. It's designed to do that. If you're worried that you will have 500A fault currents, you may need four or five of them. Try that with a relay.
(b) Static dissipation of the MOSFET
The same IRFB260N is specified at RDSon of 0.040 ohms. It can't be dissipating more than .04 ohms/ 4 ohms (speaker load) or 1/100 the power dissipated in the speaker. If that's too big, use paralleled MOSFETs. Dissipation for saturated MOSFETs even with big currents should be under 10W.
(c) Failure modes
Bipolar devices fail shorted almost every time, because the failure mode is usually localized hot spots doing localized melting. Power MOSFETs tend to fail open most of the time, based on experience with switching power supplies.
(d) de-powering the output stage only, leaving feedback, etc.
This is a fault amelioration technique. There is no real idea here that the MOSFETs are going to protect non-failing output devices. They are there to avoid the use of failure prone electromechanical devices.
(e) MOSFETs (or semiconductors in general) being delicate and failure prone:
If you want to compare the failure rates of relays to MOSFETs, go get the failure rate numbers. A MOSFET will switch the same current as a relay would, except it will do it without arcing or dying for years - and in a high frequency power supply it will do this hundreds of thousands of times per second for years. Both are adequate when you want them to turn a power amp on or off a few hundred or thousand times. Repair techs are typically good at recalling funny failure modes but bad at guessing failure frequencies, because they only see the failed units. It's a human thing.
(f) cost of MOSFETs
My current favorite is the IRFB260N. It's $3.90 in ones at Digikey, and $2.73 in tens. I can get cheap relays, OK, but what is the cost of a speaker driver replacement if the relay is welded or a tad slow?
It's another technique. It works for me, and very well. Use it or not, as you like.
I don't look here very often, so I missed these.
(a) Ability of MOSFETs to handle the voltages and currents involved in cutting off a shorted-output amplifier:
The IRF IRFB260N is a 200V, 56A n-channel device. The max surge current on this device is 220A.
A power amplifer will likely have one output shorted, and so half the power supply will be funneled through the DC resistance of the speaker. The current will be perhaps as much as 90V through 2 or more ohms of wire. The fault current is then less than the rated max current for the device.
I think it can turn off that fault current with no damage. It's designed to do that. If you're worried that you will have 500A fault currents, you may need four or five of them. Try that with a relay.
(b) Static dissipation of the MOSFET
The same IRFB260N is specified at RDSon of 0.040 ohms. It can't be dissipating more than .04 ohms/ 4 ohms (speaker load) or 1/100 the power dissipated in the speaker. If that's too big, use paralleled MOSFETs. Dissipation for saturated MOSFETs even with big currents should be under 10W.
(c) Failure modes
Bipolar devices fail shorted almost every time, because the failure mode is usually localized hot spots doing localized melting. Power MOSFETs tend to fail open most of the time, based on experience with switching power supplies.
(d) de-powering the output stage only, leaving feedback, etc.
This is a fault amelioration technique. There is no real idea here that the MOSFETs are going to protect non-failing output devices. They are there to avoid the use of failure prone electromechanical devices.
(e) MOSFETs (or semiconductors in general) being delicate and failure prone:
If you want to compare the failure rates of relays to MOSFETs, go get the failure rate numbers. A MOSFET will switch the same current as a relay would, except it will do it without arcing or dying for years - and in a high frequency power supply it will do this hundreds of thousands of times per second for years. Both are adequate when you want them to turn a power amp on or off a few hundred or thousand times. Repair techs are typically good at recalling funny failure modes but bad at guessing failure frequencies, because they only see the failed units. It's a human thing.
(f) cost of MOSFETs
My current favorite is the IRFB260N. It's $3.90 in ones at Digikey, and $2.73 in tens. I can get cheap relays, OK, but what is the cost of a speaker driver replacement if the relay is welded or a tad slow?
It's another technique. It works for me, and very well. Use it or not, as you like.
Hi R.G.,
I have to disagree with you on a few points.
a) Faults occur sometimes very quickly. Both polarities may have shorted device(s). The DC servo action almost ensures this. I've seen it often in service
b) Carver used Mosfets as switches for DC power supplies. Sometimes they go bang in a fault. Fixed a few so it happens ( a shorted device can not interrupt anything)
c) Mosfets tend to short drain - gate when they die, They certianly do not go open. Many repairs to attest to that. Your enviroment is different. Also, when used as a downconverter in the Lightstar, many die at once. Same in a car amp, except those that are open have melted the leads off or blown the package to bits. More experience on my part in service.
d) The entire point is to protect the (often expensive) electromechanical devices. Also to avoid the possible resultant fire(s).
e) There is no nice way of running an AC signal through a Mosfet in the same way. You can't do it in an acceptable manor. Why even comment. There is a drop using the mosfet as a switch. If we are to believe we can hear the difference in rectifiers, we are sure as heck going to hear your fets!! 'nough said.
f) Who cares about the cost? Relays won't kill the pocket book and they are easy to replace. Small cost compared to the speaker going. Besides, the amp is damaged anyway. That's why the protection activated to begin with!
PMA, the contact is moving when the relay is opening in a fault condition. As stated above, the Fet may have already shorted. So it's possible for either method to fail. If there is audio when it closes and it pits - fine. Owner is not operating the equipment properly, or the input hasn't been muted long enough (designers fault). A relay isn't the worst thing you can use. Each method has drawbacks. I do have to comment when a relay is slammed down this hard. They are generally very reliable. They all eventually need replacement. They have saved more speakers than some give them credit for.
-Chris
I have to disagree with you on a few points.
a) Faults occur sometimes very quickly. Both polarities may have shorted device(s). The DC servo action almost ensures this. I've seen it often in service
b) Carver used Mosfets as switches for DC power supplies. Sometimes they go bang in a fault. Fixed a few so it happens ( a shorted device can not interrupt anything)
c) Mosfets tend to short drain - gate when they die, They certianly do not go open. Many repairs to attest to that. Your enviroment is different. Also, when used as a downconverter in the Lightstar, many die at once. Same in a car amp, except those that are open have melted the leads off or blown the package to bits. More experience on my part in service.
d) The entire point is to protect the (often expensive) electromechanical devices. Also to avoid the possible resultant fire(s).
e) There is no nice way of running an AC signal through a Mosfet in the same way. You can't do it in an acceptable manor. Why even comment. There is a drop using the mosfet as a switch. If we are to believe we can hear the difference in rectifiers, we are sure as heck going to hear your fets!! 'nough said.
f) Who cares about the cost? Relays won't kill the pocket book and they are easy to replace. Small cost compared to the speaker going. Besides, the amp is damaged anyway. That's why the protection activated to begin with!
PMA, the contact is moving when the relay is opening in a fault condition. As stated above, the Fet may have already shorted. So it's possible for either method to fail. If there is audio when it closes and it pits - fine. Owner is not operating the equipment properly, or the input hasn't been muted long enough (designers fault). A relay isn't the worst thing you can use. Each method has drawbacks. I do have to comment when a relay is slammed down this hard. They are generally very reliable. They all eventually need replacement. They have saved more speakers than some give them credit for.
-Chris
richie00boy said:
Ordinary EM relay drive circuit can be straightfowardly modified to drive MOSFET.....just include one opto-coupled drive IC per MOSFET...
I think main problem is dissipation in MOSFET due to its on-resistance...
richie00boy said:
This is not as bad as it seems......only the bass unit in 'speaker endangered by DC...typically...This can tolerate caps discharging into it...no problemo....
Actually, it can be argued that main reservoirs discharging into 'speaker through its passive crossover network is no problemo either.....
anatech said:
This is a bad idea! For, as you have observed, DC faults are typically due to an output device failing short-circuit....
Disabling the amps. current sources will not have any effect on this whatever....'speaker still goes up in smoke...
Fuse in output should never be used.....it's a complete waste of time......and affects THD....not 'politically correct'...in all respects....
Hi mikeks,
Agreed. Just trying to put out an example while noting it's drawbacks. Fixed lots of amps, Adcom included. I don't like output fuses either. Found a fair number shorted by foil over the years.
I like relay or total supply shut down (with smaller supply caps to limit energy). I really like "crowbar" at higher powers.
-Chris
Agreed. Just trying to put out an example while noting it's drawbacks. Fixed lots of amps, Adcom included. I don't like output fuses either. Found a fair number shorted by foil over the years.
I like relay or total supply shut down (with smaller supply caps to limit energy). I really like "crowbar" at higher powers.
-Chris
R.G. said:
actually...your MOSFET will NOT have Rds(on) of .04ohms at typical operating currents and junction temps.....its more like 1ohm.... (probably 2ohms or more!
){See figure 4 (pg 3) here:
http://www.irf.com/product-info/datasheets/data/irfb260npbf.pdf }
This gives something like 400W dissipation for 20A!! Not good.
You would require an awfull lot of MOSFETS and substantial heatsinking to accomodate this problem.....
mikeks said:
Thanks what I was doing. Still a little more complex than a relay. Also this method has to use a separate supply, which was something I could do without on space and cost grounds as well.
mikeks said:
That's fine, but I'm trying to build protection suitable for active crossover systems
An output relay here has the double advantage that it can function as a handy startup/shutdown mute as well.
mikeks said:actually...your MOSFET will NOT have Rds(on) of .04ohms at typical operating currents and junction temps.....its more like 1ohm.... (probably 2ohms or more!
{See figure 4 (pg 3) here:
http://www.irf.com/product-info/datasheets/data/irfb260npbf.pdf }
This gives something like 400W dissipation for 20A!! Not good.
You would require an awfull lot of MOSFETS and substantial heatsinking to accomodate this problem.....
I follow the graph. So how do IR claim 0.04 ohms max on resistance then? It doesn't get anywhere near that according to the graph.
richie00boy said:......... I'm trying to build protection suitable for active crossover systems
Always remember to connect a capacitor in series with your tweeter, even with active x-overs....as the relay will still not be fast enough to protect the unit in the event of a DC fault.
I reckon i'll stick with this for the moment:
http://www.diyaudio.com/forums/showthread.php?postid=440786#post440786
http://www.diyaudio.com/forums/showthread.php?postid=440786#post440786
mikeks,
this is not true. At Tj= 120 degrees the normalised On-Resistance of the IRFB260 is twice the reststance at 25 degrees which is 0.04 ohms max. If power dissipation of the MOSFET is a problem the output trannies will glow under normal operation. Of course the MOSFET will dissipate some watts but if it run hot just mount it on the heat sink.
bocka said:
Thanks Bocka....i missed the 'normalized' scale on the y-axis
Thus given Tj=120deg. C, then Rds(on) is roughly 0.1ohms...which gives a dissipation of just 40W for 20A...not bad..
However, for waste-case current-draw, running from say +/-50V rails, the MOSFET may be called upon to pass 50A...requiring 250W dissipation...thermal management accomodating this may have to be provided....Still..this is not as bad as 400W+
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.