Select from your junk box a small transformer...something that has an already know
power....... let's say that it can produce 1 ampere when holding the voltage at the terminal at a reasonable level.
Now, after rectification, make a short circuit with an amperimeter that can read 20 Amperes continuous and able to hold 3 times more for half a second (time needed to read digital ones).
You will see that your supply, during short circuit, will be able to produce more than 10 amperes during short circuit....for sure the voltage there will be very small...but 10 amperes will be circulating.
Now let's imagine your amplifier....let's say 50 volts and 10 amperes..... imagine how many amperes it can produce during a short.....maybe 100 or more.
This is the problem with short circuits, accidental shorts, speaker shorts or shorts caused by some types of protections.... for a small period of time 100 amperes or more will circulate.
And they will circulate inside the output transistor or output of Field Effect transistor that is unable to hold that current.
If you have nine transistors...and 10 amperes maximum capacity each one, for instance, 90 amperes will be held for sometime....and this will happens only if you have emitter resistances to equalize current or analog parts to field effect and others.
Having not equalizer resistance, the unit that has bigger gain will hold bigger current than the others...so.... those 90 amperes will not be divided equally between all transistors....one of them may hold 13 amperes and burn (if limit was 10 amperes).
This is the problem you may have during short circuit..... and not having enougth output transistors (twice the maximum short circuit current available by your supply)...and not having equalizer resistances.
regards,
Carlos
power....... let's say that it can produce 1 ampere when holding the voltage at the terminal at a reasonable level.
Now, after rectification, make a short circuit with an amperimeter that can read 20 Amperes continuous and able to hold 3 times more for half a second (time needed to read digital ones).
You will see that your supply, during short circuit, will be able to produce more than 10 amperes during short circuit....for sure the voltage there will be very small...but 10 amperes will be circulating.
Now let's imagine your amplifier....let's say 50 volts and 10 amperes..... imagine how many amperes it can produce during a short.....maybe 100 or more.
This is the problem with short circuits, accidental shorts, speaker shorts or shorts caused by some types of protections.... for a small period of time 100 amperes or more will circulate.
And they will circulate inside the output transistor or output of Field Effect transistor that is unable to hold that current.
If you have nine transistors...and 10 amperes maximum capacity each one, for instance, 90 amperes will be held for sometime....and this will happens only if you have emitter resistances to equalize current or analog parts to field effect and others.
Having not equalizer resistance, the unit that has bigger gain will hold bigger current than the others...so.... those 90 amperes will not be divided equally between all transistors....one of them may hold 13 amperes and burn (if limit was 10 amperes).
This is the problem you may have during short circuit..... and not having enougth output transistors (twice the maximum short circuit current available by your supply)...and not having equalizer resistances.
regards,
Carlos
To be more clear Sakis... even suspecting that you already know that
...But certainly there are folks that do not know those basic things.
Imagine, please, that you have two transistors in some kind of circuit..... the supply is huge enougth to hold 40 volts in both colectors when enormous current is crossing....more than 25 amperes in total are crossing.
Having emitter resistances...when current crossed those ones will develop voltage into those resistance's extremes.... some voltage will be there.
As power, measured over some transistor (or FET transistor) is the result of multiplication of the voltage measure from the colector to the emitter multiplied by the current that is crossing them, you can see the results in the image attached.
Those currents were different because the rigth transistor has much more gain that the left one...so...more current will cross the one that has more gain.... the resistance, beeing there will reduce the the voltage at the emitter....that could be zero (having in that worst sittuation 40 volts from colector to emitter into the left one).
In the left transistor as a result of the resistance beeing there...the emitter voltage raises to 10 volts (ohms law).... colector voltage is 40.... emitter voltage is 10..... over the transistor, measured from colector to emitter..the voltage is 30... and this will be multiplied by the current to know the power over this selected transistor.... will result 300 watts.
But remove those resistances and you will have the red values...very different...much more different and really more dangerous to the higher gain unit.
Because of that, output transistors, beeing mounted in parallel..will need equalizing resistances.
regards,
Carlos
...But certainly there are folks that do not know those basic things.
Imagine, please, that you have two transistors in some kind of circuit..... the supply is huge enougth to hold 40 volts in both colectors when enormous current is crossing....more than 25 amperes in total are crossing.
Having emitter resistances...when current crossed those ones will develop voltage into those resistance's extremes.... some voltage will be there.
As power, measured over some transistor (or FET transistor) is the result of multiplication of the voltage measure from the colector to the emitter multiplied by the current that is crossing them, you can see the results in the image attached.
Those currents were different because the rigth transistor has much more gain that the left one...so...more current will cross the one that has more gain.... the resistance, beeing there will reduce the the voltage at the emitter....that could be zero (having in that worst sittuation 40 volts from colector to emitter into the left one).
In the left transistor as a result of the resistance beeing there...the emitter voltage raises to 10 volts (ohms law).... colector voltage is 40.... emitter voltage is 10..... over the transistor, measured from colector to emitter..the voltage is 30... and this will be multiplied by the current to know the power over this selected transistor.... will result 300 watts.
But remove those resistances and you will have the red values...very different...much more different and really more dangerous to the higher gain unit.
Because of that, output transistors, beeing mounted in parallel..will need equalizing resistances.
regards,
Carlos
Attachments
i dont get it
sorry carlos ...i dont get it ......as far as i know this is only to protect the speakers on running on rail voltage .....
meaning that if something goes wrong with amp and any of transistors is shorted then dc comes in the out and then triac short circuits the out to ground and rail fuses goes of ..... no mater if 7 amps or 100 amps makes them go off ...
the point is that if a circuit like that is quick enough to make shure that works a lot faster than dc will take to go to your speaker ....
i dont think or is expecteted from a thing like that mosfets or bjt to be able to stand a short circuit of as you said 100 amps may be and still working after that ...
it seems to me that this protection is only active if we take as a fact that the amp is allready doomed and one or all of trs is gone to mars ...
correct ????
sorry carlos ...i dont get it ......as far as i know this is only to protect the speakers on running on rail voltage .....
meaning that if something goes wrong with amp and any of transistors is shorted then dc comes in the out and then triac short circuits the out to ground and rail fuses goes of ..... no mater if 7 amps or 100 amps makes them go off ...
the point is that if a circuit like that is quick enough to make shure that works a lot faster than dc will take to go to your speaker ....
i dont think or is expecteted from a thing like that mosfets or bjt to be able to stand a short circuit of as you said 100 amps may be and still working after that ...
it seems to me that this protection is only active if we take as a fact that the amp is allready doomed and one or all of trs is gone to mars ...
correct ????
Sorry... i tried my best...there's a problem on me related English language
Well...one problem is not really true...i have "zilions" of problems with language.
Also you have to translate things..this turns things even worst.
Portuguese to English and than English to Greek.
Also i am not clear enougth i suppose...let's wait someone that can explain this more clearly to you.
Maybe fótios, as he can send you explanations directly in Greek...or..at least someone native in English language that can remove one step into the language versions made...using brain translation or software to translations.
I use my old brain.... has a very slow chemical microprocessor..something alike Zilog Z80.... and the astounding auxiliary memory of 2Kilobytes to speak and write in english..for sure a big mess will result.
Not correct Sakis..... an amplifier with many transistors in the output...many pairs...each pair able to hold 10 amps...may hold shorts of 100 or 200 amperes if your supply and condensers could supply such current.
This was the base of advertizing, the main point into Harman Kardon published texts into electronics magazines...the possibility, because supply and big condensers, to hold, short time short circuits providing 60 to 150 amperes.
The idea was not to provide energy to face real shorts..but to face strange speaker impedances that would need enormous current to face the deep of impedance in some frequencies...this intended to show us that those amplifier were able to hold "difficult loads"...strange speakers that has enormous capacitance and inductances into their internal crossovers.
regards,
Carlos
Well...one problem is not really true...i have "zilions" of problems with language.
Also you have to translate things..this turns things even worst.
Portuguese to English and than English to Greek.
Also i am not clear enougth i suppose...let's wait someone that can explain this more clearly to you.
Maybe fótios, as he can send you explanations directly in Greek...or..at least someone native in English language that can remove one step into the language versions made...using brain translation or software to translations.
I use my old brain.... has a very slow chemical microprocessor..something alike Zilog Z80.... and the astounding auxiliary memory of 2Kilobytes to speak and write in english..for sure a big mess will result.
Not correct Sakis..... an amplifier with many transistors in the output...many pairs...each pair able to hold 10 amps...may hold shorts of 100 or 200 amperes if your supply and condensers could supply such current.
This was the base of advertizing, the main point into Harman Kardon published texts into electronics magazines...the possibility, because supply and big condensers, to hold, short time short circuits providing 60 to 150 amperes.
The idea was not to provide energy to face real shorts..but to face strange speaker impedances that would need enormous current to face the deep of impedance in some frequencies...this intended to show us that those amplifier were able to hold "difficult loads"...strange speakers that has enormous capacitance and inductances into their internal crossovers.
regards,
Carlos
Yes sakis, your question is simple and you are correct in your assumption. The danger I think Carlos points out is that if just one transistor goes short/faulty in a bank of several, you may damage all of them when the crowbar comes into operation.
There is also the danger that somebody uses an incorrect fuse/nail(!) at some time and the crowbar comes into force and sets the amp on fire.
There is also the danger that somebody uses an incorrect fuse/nail(!) at some time and the crowbar comes into force and sets the amp on fire.
Hi,
Richie has hit it on the head.
F4A allows about 250W into 8ohm.
The crowbar can be linked from output to ground or one can fit twin crowbars one from each supply to ground. They can even work differentially. If the output goes +ve then the +ve rail crowbar fires and vice versa for the -ve output offset.
Richie has hit it on the head.
one MUST use a fuse on each supply to allow the crowbar to be effective. I have found that F3.1A rail fuses allow 6Apk to flow to the load without melting. This allows 150W into 8ohm speakers.
F4A allows about 250W into 8ohm.
The crowbar can be linked from output to ground or one can fit twin crowbars one from each supply to ground. They can even work differentially. If the output goes +ve then the +ve rail crowbar fires and vice versa for the -ve output offset.
The output power is irrelevant.
The time it takes the fuse to blow at X*current rating is the part the triac (or other switch) has to survive.
This X can be ten times or a hundred times or even higher.
I just looked back at this post http://www.diyaudio.com/forums/showthread.php?postid=1164234#post1164234
to see that 35Vpk into 4r resulting in 2.8times the fuse rating did not blow the rail fuse. On a short time scale fuses can pass large currents.
The time it takes the fuse to blow at X*current rating is the part the triac (or other switch) has to survive.
This X can be ten times or a hundred times or even higher.
I just looked back at this post http://www.diyaudio.com/forums/showthread.php?postid=1164234#post1164234
to see that 35Vpk into 4r resulting in 2.8times the fuse rating did not blow the rail fuse. On a short time scale fuses can pass large currents.
Here we have difference between practice and theories.
Fuse blows slowly.... triac, beeing solid state reacts faster...it produces short term current near the limit of the supply... output transistors can goes shorted before the fuse blows.
Fuse is something that serves for "half of nothing"...it blows when the transistor is already dead....it will just avoid the short to continue to burn the board and wires melt.
I suggest, strongly, to avoid triac and shorts into the output.
To protect you have to put a lot of transistors in parallel..better other sensors and relays to switch the output off.
Do not trust in fuses Sakis.... they will do a very bad job for you.
During audio reproduction you have bursts of high current...and fuse do not blows....it blows only after the transistor already gone.
regards,
Carlos
Fuse blows slowly.... triac, beeing solid state reacts faster...it produces short term current near the limit of the supply... output transistors can goes shorted before the fuse blows.
Fuse is something that serves for "half of nothing"...it blows when the transistor is already dead....it will just avoid the short to continue to burn the board and wires melt.
I suggest, strongly, to avoid triac and shorts into the output.
To protect you have to put a lot of transistors in parallel..better other sensors and relays to switch the output off.
Do not trust in fuses Sakis.... they will do a very bad job for you.
During audio reproduction you have bursts of high current...and fuse do not blows....it blows only after the transistor already gone.
regards,
Carlos
well said ....
very correct andrew t .... that was my exact point .....
i kept following carlo's latest posts and seems that too much Italian food effected his english also except his weight .....
just a joke my friend Carlos ....
andrew t understood exectly what i ment ..... my point was that in any case that amp fails i consider it lost any way and not really importand if one or all mosfets gone ....or how many fuses ....the point is at least to save the speakers ..... far more expensive than few mosfets
best regards sakis
very correct andrew t .... that was my exact point .....
i kept following carlo's latest posts and seems that too much Italian food effected his english also except his weight .....
just a joke my friend Carlos ....
andrew t understood exectly what i ment ..... my point was that in any case that amp fails i consider it lost any way and not really importand if one or all mosfets gone ....or how many fuses ....the point is at least to save the speakers ..... far more expensive than few mosfets
best regards sakis
I reccon it would be pretty easy to set up some kind of overvoltage protection useing an LM311 chip, for voltages up to about 35V, when triggered it can simply latch a relay over the output and maybe open another one in the PSU section... look at the datasheet.
http://cache.national.com/ds/LM/LM111.pdf
I'm not smart enough to figure out how to build a DC senceing circuit yet though... the circuit above will need a simple bridge to rectify the signals to DC...
I suppose the DC protection side of things can look like the above circuit adjusted for the right voltages, preceeded by a lowpass filter set to say 10hz or so...
Its a nice little chip... it has 2 inputs + and - which you use to provide one with a reference voltage (the one one you use will decide positive or negative peak), using a zener for instance, and then using a resistor divider you can make the other pin sense a voltage.... when the two pins are at about the same voltage the output goes positive and the sensed voltage passes through...
http://cache.national.com/ds/LM/LM111.pdf
I'm not smart enough to figure out how to build a DC senceing circuit yet though... the circuit above will need a simple bridge to rectify the signals to DC...
I suppose the DC protection side of things can look like the above circuit adjusted for the right voltages, preceeded by a lowpass filter set to say 10hz or so...
Its a nice little chip... it has 2 inputs + and - which you use to provide one with a reference voltage (the one one you use will decide positive or negative peak), using a zener for instance, and then using a resistor divider you can make the other pin sense a voltage.... when the two pins are at about the same voltage the output goes positive and the sensed voltage passes through...
I've been thinking about a DC protection system for my website. I've come up with a conceptual circuit. If triggered it disconnects the DC rails from the amp.
I believe this method is better than the crowbar because it works even if the wrong fuses have been used and poses no additional risk to the output devices or wiring, PCB etc. A relay in the output would need to be very rugged as it would be working into an inductive load, it may even fail to protect regardless.
You can sense DC simply by low-pass filtering the amp output signal - music should not have any DC or very low frequency content.
I believe this method is better than the crowbar because it works even if the wrong fuses have been used and poses no additional risk to the output devices or wiring, PCB etc. A relay in the output would need to be very rugged as it would be working into an inductive load, it may even fail to protect regardless.
You can sense DC simply by low-pass filtering the amp output signal - music should not have any DC or very low frequency content.
This technique is used on power supplies for over-voltage protection. If the output voltage exceeds a preset limit, a comparator trips an SCR, crowbaring the supply to ground, probably through some current limiting resistors. An additional pin on the input circuit breaker is also activated, tripping the circuit breaker to remove the mains supply. The supply is designed to survive this with no problem.
I'm pretty sure a similar technique has been used with commercial amplifiers, but the assumption is that the amp has failed, so you just clamp the output, blow the fuse, and that's that. It's probably better than a protection relay because relay contacts can weld under high currents, preventing them from opening. I think I've even seen high current protection schemes where it was expected that the SCR or TRIAC would fail in a shorted condition while protecting the output.
Remember that most speakers can survive very high peak currents, many hundreds of watts, for a number of milliseconds, so if the circuit can dump most of the power supply in that time, the speaker will be protected.
In the end, I just don't know how real of a problem shorted drivers are. Generally when I blow an amp, both top and bottom devices fail shorted, and that takes care of the problem. With more output devices, the chances are greater that the full supply will be applied to the load. To date, my preference is for fuses in the supply lines and no output protection at all- if I had very expensive speakers, my logic might be quite different!
I'm pretty sure a similar technique has been used with commercial amplifiers, but the assumption is that the amp has failed, so you just clamp the output, blow the fuse, and that's that. It's probably better than a protection relay because relay contacts can weld under high currents, preventing them from opening. I think I've even seen high current protection schemes where it was expected that the SCR or TRIAC would fail in a shorted condition while protecting the output.
Remember that most speakers can survive very high peak currents, many hundreds of watts, for a number of milliseconds, so if the circuit can dump most of the power supply in that time, the speaker will be protected.
In the end, I just don't know how real of a problem shorted drivers are. Generally when I blow an amp, both top and bottom devices fail shorted, and that takes care of the problem. With more output devices, the chances are greater that the full supply will be applied to the load. To date, my preference is for fuses in the supply lines and no output protection at all- if I had very expensive speakers, my logic might be quite different!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.